key: cord-0252685-puwg1fjq
authors: nan
title: 23rd Annual Meeting
date: 2000-04-03
journal: J Gen Intern Med
DOI: nan
sha: d33e9f84139ebc7bf4c5e9d9be18e89e650d7513
doc_id: 252685
cord_uid: puwg1fjq

nan

PURPOSE: To delineate the optimal role for attending physicians (AP) and optimal site for AP-houseofficer (HO) interactions in a HO-based primary care clinic. METHODS: Initially, 227 patient encounters were randomized to either exclusive HO-AP conference room discussion or to HO-AP conference room discussion plus AP seeing the patient (randomized controlled trial, . Subsequently, 186 patient encounters were randomized to either exclusive HO-AP conference room discussion or to exclusive HO-AP exam room discussion with the patient present . Post-visit, questionnaires were obtained from patients (related to visit satisfaction) and from HO and AP (related to AP teaching, AP diagnostic and therapeutic contributions and HO comfort and autonomy). RESULTS: Patient demographics and self-rated health were comparable in all study groups in both RCTs. Patient overall visit satisfaction, satisfaction with medical care and feeling of reassurance with seeing their doctor(s) as well as HO and AP assessment of AP contributions to teaching, diagnosis and therapy did not differ in RCT-1. In RCT-2, patients involved in exam room encounters significantly more often felt more comfortable with the discussion (p Ͻ 0.001) and would prefer to listen to the discussion in the future (p Ͻ 0.001) than those in the conference room group. In RCT-2, patient overall visit satisfaction and satisfaction with their medical care were comparable with exam room and conference room presentations. HO and AP assessment of AP contributions to teaching, diagnosis and treatment were nearly identical when conference room and exam room presentations were compared in RCT-2. APs felt that 83% of exam room encounters were beneficial to patient care. A small but significant (p Ͻ 0.001) per cent of HOs (10%) but not APs (0%) were made uncomfortable by exam room presentations and some HOs (11%) felt exam room presentations diminished their autonomy. CONCLUSION: Patients and APs perceive that exam room presentations are beneficial and exam room presentations do not decrease patient rating of overall visit satisfaction. Also, AP contributions to teaching, diagnosis and therapy are maintained in the setting of exam room presentations. Some HOs perceive that exam room presentations cause discomfort and diminish autonomy. student, and compared using paired t-tests and Pearson correlations; 38 pairs of ratings were analyzed for year one, 36 for year two. RESULTS: Preceptors tended to rate students higher than students rated themselves. For year one, the mean of student self-ratings of skills was approximately three quarters of a standard deviation less than the mean of preceptors' ratings of the same skills; for year two students, student ratings were a full standard deviation less than preceptors' ratings. Both differences were statistically significant (p Ͻ .000). With respect to confidence/comfort in skills, student ratings were also lower than preceptors' ratings of apparent confidence/comfort. For year one, student ratings were approximately one half of a standard deviation lower, for year two, student ratings were more than one half a standard deviation lower. Again, mean differences for both years were statistically significant (p Ͻ .01). Correlations between student and preceptor ratings varied: for year one, the correlation between students' and preceptors skills ratings was .33 (p Ͻ .05), while for year two, the correlation was .07 (not significant). The correlation between students' confidence ratings and preceptors' ratings of students' confidence/comfort was .41 (p ϭ .01) for year one, and .33 (p ϭ .05) for year two. CONCLUSION: The finding that the students in this study tended to rate themselves lower than their preceptors has several possible explanations. Students' experience with self-assessment may be limited; their understanding of performance criteria may be incomplete; medical students may be critical of their own performance, preceptors' may tend to be lenient. In addition, preceptors may rate students based on a very small number of observations or no observations. While the current study does not allow us to determine the reason for the the results found here, it does highlight the fact that it is important to look carefully at such ratings. Training of students and preceptors in the specific items, explicit definitions of performance criteria, and repeated opportunities for evaluation are recommended. In addition, frequent direct observations by preceptors are advised. PURPOSE: Although most patients with diabetes are managed in primary care sites, we have little understanding of the attitudes of their providers. Such information is particularly important to guide the training of resident physicians who will be the primary care practitioners of the future. METHODS: To meet this need, we developed a 34-item questionnaire to assess 1) barriers to delivering care as perceived by internal medicine residents in an outpatient primary care clinic at Grady Hospital, that primarily serves urban African-American patients, and 2) thresholds for glucose control at which residents would start or intensify diabetes medications. RESULTS: A total of 153 ⁄ 185 residents completed the survey (83% response rate); their mean age was 28 years; 29% were women; 35%, 37%, and 28% were in their first, second, or third year of training; and 26% indicated plans to practice in a primary care setting. There was no consensus that time or staffing were inadequate, or that patient psychosocial problems were not typical of other patient populations. However, the residents saw barriers in several areas: 1) System: 82% had problems because charts and data were unavailable; 2) Patients: although 77% felt that diabetes risk factors were similar to those in other populations, 62% said that patient literacy levels limited care, and 40% felt that patients did not view diabetes as a serious problem; 70% said that diabetes control could not be improved unless patients made lifestyle changes; 3) Disease: 87% felt that diabetes is not too complex to be managed in primary care sites, and 71% said their diabetes training was adequate; 4) Physician: 75% said they were not reluctant to be aggressive in diabetes management despite the presence of other medical problems. However, 88% were not sure that glucose levels measured during clinic visits could be used to guide therapy, and their stated thresholds for intensification of therapy varied widely: fasting glucose 139 ϩ / Ϫ 26 mg/dl and HbA1c 7.4 ϩ / Ϫ 0.7% (M ϩ / Ϫ SD); 80% said they would appreciate feedback on their diabetes performance. CONCLUSION: Barriers to outpatient diabetes management reported by internal medicine residents include both typical municipal hospital problems and probable sociocultural misperceptions. Although training was perceived as adequate, many residents do not understand how to use glucose levels to inform clinical decisionmaking. New approaches to clinical education will be needed to improve diabetes management. Early clinical experiences are becoming a common method to expose students to medical practice at the beginning of their cirriculum. What is unclear is what students find most or least rewarding in these early clinical experiences. Our early clinical experience was a primary care externship in which ninety first year medical students spent one week observing a volunteer community based primary care physician. After the externship each student completed a survey which included open-ended questions regarding their experiences. Responses were coded for prevalent themes. The most common themes in response to "my most rewarding experience" were interactions with patients (41.6%), observing the physician interact with patients (36.0%), interactions with physician (10.1%), applying knowledge (10.1%), and helping patients (4.5%). In response to "my least rewarding ex-perience" the common themes included, downtime when not interacting with physician or patients (22.5%), difficult patient interactions (16.9%), only being able to observe the physician (14.6%), "nothing" (11.2%), poor interaction with physician (5.6%). Other responses included writing assignments, long distance commute, and aching feet. There were no correlations between student responses and their demographic data. We examined the associations between these responses and students' ratings of their level of understanding of primary care and time spent in various activities. Students that felt that observing the physician interact with patients was the most rewarding experience were more likely to feel that the externship improved their understanding of what a physician does (r ϭ 0.22, p Ͻ 0.05). These students also tended to work with preceptors of a younger age (r ϭ 0.21, p Ͻ 0.05). On the other hand, students that felt that the most rewarding experience was patient interactions were less likely to feel that they had a good understanding of what being a physician is all about (p Ͻ 0.05). Students who felt that the most rewarding aspect of the externship was helping patients were more likely to have spent a significant amount of time observing the physician at other sites other than the office and hospital i.e., free outreach clinics (r ϭ 0.39, p Ͻ 0.0001). These students were also more likely to want to go back to the preceptors' office during clinical years (r ϭ 0.20, p Ͻ 0.05). The preceptors of these students tended to be older and women (r ϭ 0.30, p Ͻ 0.05). Understanding what students view as most and least rewarding experiences will help structure externships so that future experiences are positive. Positive interactions with physicians and patients may give students a better understanding of primary care. If the goal of a preclinical externship is to provide a better understanding of what a physician does, students may benefit more by spending most of their time observing the physician rather than interacting with the patients on their own. went off, at random, an average of 3.2 times an hour. Whenever the pager went off, the intern made note of the person with whom he or she was in contact. We also obtained DRG weights and information on these intern factors: census size, number of admissions during the 24-hour admitting period, previous postgraduate training, faculty ratings of humanistic qualities, and in-training examination scores. RESULTS: Of the 25 interns, 64% were men; 32% were in the categorical program, 16% in medicine-pediatrics, and 52% in a preliminary year. Four reported previous postgraduate training. Preliminary interns spent less time with patients (5.3% vs. 16.0%, p ϭ 0.009). The fraction of time spent with patients was higher with more experience before starting the internship ( r ϭ 0.67, p Ͻ 0.001) and better examination scores ( r ϭ 0.42, p ϭ 0.036). Among the 90 patients with PSQ data whose interns carried the workload pager the day they were admitted, the mean PSQ was 1.80 (s.d. 0.67) . Their mean LOS was 7.63 days (s.d. 10.0). Interns caring for them had a mean of 2.5 admissions (s.d. 1.5 ) and a census of 5. 1 (s.d. 2.1) . The mean proportion of contacts with patients was 11.6%. LOS was related to DRG weight (Pearson r ϭ 0.22, p ϭ 0.37) but there was no significant correlation with admissions, census, or proportion of contacts with patients. Neither was there a significant correlation of PSQ with those three variables. CONCLUSION: As in previous studies, internal medicine interns spent less than 15% of their time with patients. Preliminary year interns spent even less time with patients. Time spent with patients was not correlated with patient satisfaction or length of stay. Nor were these outcomes related to intern census or admissions.

ENVIRONMENTS. JL Bowen, SS Desai, RA Harrison, Medicine, OHSU, Portland, OR PURPOSE: Research in the area of resident continuity clinic experiences has focused on resident satisfaction. Satisfaction instruments are often constructed without resident input and may not ask questions that explore the residents' underlying needs to enhance learning. Needs assessment instruments ask questions of residents that probe for key underlying issues important to residents. This study is designed to assess residents' needs for optimal learning in continuity clinic settings using a critical incident technique. METHODS: We convened five focus groups of internal medicine residents. All 29 OHSU continuity residents were invited to participate in this study on a voluntary basis. Consent was obtained before audio-recording each confidential session. Each session, facilitated by the investigators, consisted of two to six resident physician volunteers who were asked to recall and describe a memorable "good" day and "bad" day in clinic. The audiotapes were transcribed, and independently analyzed and coded by each investigator. The coded transcripts were then compared and discussed, generating a thematic coding schema. Disagreements were discussed and resolved by consensus. OHSU IRB approved this study. RESULTS: Sixty-eight percent of eligible participants completed the study. Seven major themes were identified. Although residents addressed teaching issues (e.g., preceptor style, availability, role modeling) and learning issues (e.g., autonomy, patient-centered learning), the themes of clinic efficiency, workload, and physical space often dominated the narratives. Problems with teamwork and communication were emphasized, especially the importance of teamwork and consistent pairing with support staff. Residents also described the impact of both positive (e.g., self-confidence, control, predictability, empowerment) and negative (e.g., intimidation, isolation, vulnerability, guilt) feelings on their learning experiences. CONCLUSION: The critical incident technique is an effective method for uncovering the conscious and unconscious needs of residents in continuity clinic. Clinic inefficiency distracts the residents' ability to focus on their education. In order to optimize residents' learning experiences in their clinics, clinic process issues and the affective domain of resident education needs as much attention as typical internal medicine curricular issues. Information gathered from our qualitative study can be used to create quantitative assessment tools that better reflect the needs of residents in continuity clinics.

Brandenburg, L Adams, CT Lin, M Blake, M Lemenger, Division of General Internal Medicine, University of Colorado Health Sciences Center, Denver, CO PURPOSE: Determine the impact of the Electronic Residency Application Service (ERAS) on Internal Medicine Residency programs, specifically on number of applicants, screening, interviews, ranking, and quality of final match. METHODS: A Web-based survey was developed based on consultation with Association of Program Directors in Internal Medicine (APDIM) and ERAS officials and a review of the literature. The survey was designed as a web page to allow ease of response and data collection. The survey consisted of descriptive questions about the program (geographic region, size of program, and type of program-university or community), questions that could be answered on a Likert scale, and open fields that invited individualized comments. It was then sent to program directors via the APDIM list serve. In order to increase the response rate, this was followed up with personalized e-mails, letters, and faxes that were sent to non-responders. The quantitative data was collected in an Access database and is being analyzed using univariate and multivariate statistics. RESULTS: The final response rate was 81 percent. ERAS was well received with 86 percent reporting they found overall screening of applicants easier. The results show that the overall number of applicants varied greatly (49 percent reported more, 33 percent reported less). There was a wide distribution as well regarding the number of IMG applicants, but there was a tendency to have fewer IMG applicants (48 percent reported less). Most programs interviewed the same number of applicants, but 35 percent interviewed more. Most programs also ranked the same number of applicants, with 33 percent ranking more, and 13 percent ranking less. The quality of final match was rated the same as previous years by 47 percent, better by 38 percent, and worse by 15 percent. Qualitative comments reveal a high level of satisfaction with ERAS, specifically the efficiency of the process and the ability to communicate with applicants by e-mail. Specific recommendations for improvement included: more versatile filters, improved readability of scanned data, and increased use of shortcut keys to facilitate review of applications. CONCLUSION: The transition to ERAS was successfully accomplished in internal medicine. However, lessons learned will be useful for both continuously improving the ERAS process and for guiding new specialties as they make the transition to ERAS. PURPOSE: Teaching to learners of varying levels in traditional ward attending rounds is a challenging task that runs the risk of satisfying no one. We hypothesized that adding teaching rounds dedicated to students would improve (1) the perceived quality of the educational experience for students and housestaff, and (2) teacher and learner satisfaction, without negatively impacting ward team dynamics. METHODS: Student Teaching Attending (STA) sessions were added to traditional Ward Attending (WA) rounds during the third year Internal Medicine clerkship for the 1998-1999 academic year. STA rounds replaced a didactic lecture series and were similar in organization to traditional ward rounds except that only students were present. Students, housestaff, and WA's were surveyed prior to the introduction of STA sessions to document baseline attitudes. These same groups were surveyed again, along with STA's, during the 1998-1999 academic year. RESULTS: Baseline survey response rates were: students 25 (18%), housestaff 11 (10%), WA's 33 (48%). For the 1998-1999 academic year, response rates were: students 121 (88%), housestaff 81 (56%), WA's 45 (46%), STA's 18 (82%). After the addition of STA rounds, students felt more strongly that they achieved their educational goals during the rotation (38% strongly agree in 1998-1999 vs 16% baseline, p ϭ .033), while housestaff opinions did not change significantly (p ϭ .47). Students who experienced STA rounds felt more strongly that the rotation increased their interest in a career in Internal Medicine (50% strongly agree vs 24%, p ϭ .016). We had anticipated a shift in the focus of ward rounds towards housestaff and away from students. Contrary to expectation, students, housestaff, and WA's perceived the success of ward rounds in meeting individual learner needs as similar between the two years. WA's felt more strongly that they enjoyed teaching rounds after STA sessions were added (89% strongly agree vs 67%, p ϭ .016), while enjoyment was less improved for students (p ϭ .071) and stable for housestaff (p ϭ .50). There was no measured impact on perceived team dynamics. CONCLUSION: The addition of a supplementary teaching attending for students during the Internal Medicine Clerkship improved student and teacher satisfaction with the clerkship and enhanced career interest in Medicine without impacting the quality of ward attending rounds or team dynamics.

Sciences, Tulane University School of Public Health; Medicine, Tulane University of School of Medicine, New Orleans, LA PURPOSE: This study was designed to ascertain the feasibility of introducing second year medical students to clinical practice guidelines and to determine if they could use these guidelines to manage a simulated patient (SP) with newly diagnosed Type 2 diabetes mellitus. METHODS: Half the second year class was divided into 18 groups of 4 students each and assigned an SP with hyperglycemia. The other half of the class served as controls. Both groups attended lectures and small group on diabetes in their pathophysiology course. In addition to the hyperglycemia, the SPs gave a history of smoking, were overweight, had elevated BP's, hypercholesterolemia and proteinuria. The students were given ADA guidelines as well as several well-done clinical studies on the management of diabetes. The students' task was to manage the SPs appropriately. BP's and laboratory parameters changed depending on therapeutic interventions. At the end of the semester, the SP's charts were audited to determine what interventions were instituted. All second year students were examined using an OSCE, which included a case of a young type 1 diabetic who required counseling. The performance on this station was compared between the control and experimental groups. RESULTS: All groups obtained glycated Hgb, fasting lipids, urine for albumin and ophthalmology consults. Podiatric exams were documented on all charts. All groups initiated ACE inhibitors. Hyperglycemia was treated with a sulfonylurea by 54%, metformin (24%), both (12%) or acarbose (6%). ASA was prescribed by 72% of the students. All groups counseled the SPs on smoking cessation and an appropriate diet. On the diabetes OSCE station the students who participated in the study scored significantly better than the students who did not participate (66% vs 59% p Ͻ 0.001). CONCLUSION: : Second year students with minimal faculty input can successfully use practice guidelines to manage patients with diabetes. The knowledge ob-tained from this innovative teaching methodology can be transferred to similar but non-identical situations This problem based approach may be an ideal way to teach disease management. In addition, it is hoped that introducing practice guidelines to unbiased preclinical students will allow future clinicians to be more accepting of clinical guidelines.

MM Budev, DG Litaker, HL Thacker, M Hewson, General Internal Medicine; Education Division Office, The Cleveland Clinic Foundation, Cleveland, OH PURPOSE: As Odysseus entrusted his son to his friend/teacher Mentor, modern day academic medicine informally entrusts the guidance of young physicians to the tutelage of experienced educators. In the past, efforts at describing factors of mentorships have been poorly defined and lacking empirical justification. This needs assessment is aimed to identify :1) the perceptions of internal medicine residents regarding mentoring, 2) their overall mentoring experience 3) possible impacting factors, and 4) their willingness to participate in a formal mentoring program. METHODS: A cross sectional survey was conducted of all internal medicine residents at a tertiary care institution. The respondents voluntarily completed a 29item questionnaire asking for perceptions on various aspects of mentoring. The responses were scored using a 5 point Likert scale (1 ϭ not important to 5 ϭ important). Descriptive analyses, using frequency and means were performed. RESULTS: 96 of 133 surveys were returned (72%). Respondents consisted of 74% (n ϭ 71) males with 47% PGY-1, 21% PGY-2, and 30% PGY-3, and 2% PGY-4 (med/peds). A majority (75%) of residents felt a mentoring relationship had been important to their careers. 60% residents indicated never having a mentoring relationship, although 57% of these individuals were only into the first 4 months of training. The most important limiting factor cited was the absence of a formal mentoring program (X ϭ 4.8). Paucity of staff interest also contributed to the perception of scarce mentor availability (X ϭ 4.17). The majority of respondents (91%) expressed interest in having a formal mentoring program. A majority of respondents believed that mentoring impacts careers, and that an individual has to be mentored to become a mentor (88% respectively). The most important criteria for mentor selection were; being a role model (X ϭ 4.25); high professional achievements (X ϭ 3.48); serving as a networking source (X ϭ 3.40); and high academic rank (X ϭ 3.29). The least important criterion was mentor gender (X ϭ 1.95). In fact, the data further supported this showing equal numbers of male and female residents had selected female mentors. CONCLUSION: The current mentoring environment has changed with the addition of fellows who serve as liaisons between experienced attendings and residents, and an emphasis on service rather than teaching. Our data reveals a strong resident desire for a formalized mentoring program. This could overcome some of the constraining factors in initiating and nurturing mentoring relationships. The data also indicates that gender is not a limiting factor. We believe formal mentoring will not only benefit the mentee and the mentor, but academic medical institutions as well.

RESIDENT SURVEY ON DOMESTIC VIOLENCE. MM Budev, D Abood, C Henry, General Internal Medicine, The Cleveland Clinic Foundation, Cleveland, OH PURPOSE: To determine the knowledge base, attitudes toward, and experience with domestic violence among Internal Medicine (IM) and Medicine/Pediatric (Med/Peds) residents at the Cleveland Clinic Foundation. Other secondary objectives included determining if educational predictors, personal experience contribute to making residents more likely to suspect domestic violence (DV) in their longitudinal clinics. METHODS: A survey consisting of twenty-eight questions was distributed over a four-week period to members of the Cleveland Clinic Foundation's IM and Med/ Peds residency program. The Chi-Square test, Fisher's Exact test, Mantel-Haenszel Chi-Square test, and Wilcoxon rank sum test were used where appropriate, using version 6.12 of SAS. RESULTS: Approximately 68 out of 121 residents completed the survey, yielding a response rate of 56%. Distribution of residents responding to the survey included 43 male (63%) and 25 female (37%) residents. Of the respondents, 32% were PGY-1, 24% were PGY-2, 41% were PGY-3, and 3% were PGY-4 Med/Peds residents. Ninety-one respondents felt that DV was an important issue to address in residency training. Only 27% of respondents recalled formal DV training in their residency. In fact, 63% of respondents felt that routine DV screening should be implemented in their outpatient history and physical exams. Yet, only 35% of residents suspected abuse in their longitudinal clinic patient population. Of the residents who suspected abuse, 88% questioned the patient, 48% felt comfortable with questioning the patient, 30% felt their training had prepared them for the discussion, and 34.8% were able to make the appropriate referrals. When those residents who reported having received DV training during their residency were looked at independently, they were more likely to be able to offer the appropriate referrals to victims of DV (p ϭ 0.02). They may also have been more likely to suspect DV in their clinics, although this association did not reach statistical significance (p ϭ 0.054). There was no significant association between gender, age, level of training, country of origin, medical school location, or having a friend or family member who was a DV victim, and suspecting DV in the longitudinal clinic. Even those residents who had previous training in DV issues were not more likely to answer more than three questions correctly in the general knowledge section.

CONCLUSION: Despite the heightened awareness about DV in the healthcare community, surprisingly few residents recalled formal education on the topic during their training program. Those who did recall any formal training may have been more likely to suspect DV in their longitudinal clinic patients and were more likely to be able to offer victims of DV the proper referrals. Further research is needed to determine the most effective clinical training strategy in this area. PURPOSE: To assess and characterize the perceptions of internal medicine residents regarding their most memorable "finds" and "misses" on physical examination of their patients. METHODS: In May 1999, forty-one first year residents in an internal medicine training program at the Mayo Clinic participated in an objective structured clinical examination (OSCE). Immediately afterward, the residents were asked to complete a brief, anonymous questionnaire which surveyed their perceptions regarding the relative importance of clinical exam skills and asked them to record their most memorable physical examination "find" and "miss", how they felt about it, and what effect it had on the overall care of the patient. RESULTS: All residents completed at least a portion of the survey. When asked to rate their physical examination skills, 56% felt that they needed some degree of improvement. Thirteen residents (32%) believed that it was possible to complete the residency program with major deficiencies in physical exam skills and yet still maintain a reasonable degree of overall clinical competence. Twenty-four residents recorded their impressions of a memorable exam "find", while only eleven of this group could recall an example of a notable "miss". While the discovery of a positive finding was judged to have helped expedite the diagnosis and management of the case 87.5% (21/24) of the time, a corresponding delay in diagnosis/ appropriate therapy with a missed finding could be recalled in only 18% (2/11)of the cases. CONCLUSION: Most internal medicine residents at the end of their first year of training seem to view the physical examination as a significant contributor to their overall clinical competence and recognize a need for self-improvement in this area, but a significant minority do not seem to share this perception. Although a considerable number of residents can easily recall a case in which they discovered an important physical exam finding which seemed to expedite patient care, only half as many could remember a single case of a missed finding, and these were generally judged to have had no detrimental impact on the management of the case. While the phenomenon of psychologic denial may play a role in these perceptions, it is also quite possible that they simply reflect the recognition of multiple diagnostic "cues" in the majority of cases seen by the internist (such that missing one of them will not usually cause harm). At the intern level, there is also possibly an awareness that missed findings will be shortly thereafter detected by others in the "supervisory safety net". These potential explanations remain to be explored, as does the intriguing possibility that perhaps these perceptions may be countered by an educational emphasis upon critical findings and patient scenarios which are not subject to these criteria. RESULTS: The response rate was 90% (n ϭ 110). Of those responding: 72% were men, 45 was the mean age and 58% had completed a fellowship. CD's received: (mean) 22% salary support for while spending 28% of their time on the clerkship. They spent 3 half days weekly in clinic, 3 months on inpatient services and had been clerkship directors 6.5 years. The mean number of articles was 2 (range 0-20) and grants was 0.7 (range 0-4). Only 38% had any grants, so they were considered as a dichotomous variable in the comparative analysis. In univariate analysis, the factors associated with publishing articles (p Ͻ .05) were: male gender (2.7 v .9) , having 3 or less outpatient clinic sessions weekly (2.8 v 1.3) , fellowship training (3.0 v 1.2), institutional faculty development program (3.4 v 1.8) , teaching in other courses (F ϭ 5.9) and using the "expectations article" (3.7 v 1.8) . The variables associated with grants were identical. Clerkship characteristics and inpatient clinical activity were not associated with scholarship. In multivariate analysis, a model using the teaching, fellowship training, outpatient clinic and the "expectations article" had R 2 ϭ 22 for articles published. Regarding grants, a model using teaching, outpatient clinic, faculty development, gender and "expectations article" had R 2 ϭ .35. CONCLUSION: Several factors were noted to be associated with increased scolarship. The explanation for the gender differences was unclear. A threshold effect for outpatient clinical activity was noted, which was not seen for inpatient work. Faculty development, either through a fellowship or a formal program was beneficial. This national sample of internists gives insight to factors influencing scholarship which should be considered by those responsible for the career development of clerkship directors.

MJ Fagan, RA Griffith, LCW Obbard, CJ O'Connor, Division of General Internal Medicine, Brown University School of Medicine, Providence, RI PURPOSE: To determine if the introduction of a curriculum in evidence-based physical diagnosis (PD) into a third year medical student clerkship in internal medicine could improve student knowledge, skill, and confidence in PD. METHODS: We developed an evidence-based PD curriculum based on 8 articles selected from JAMA's Rational Clinical Examination series. The topics included central venous pressure, splenomegaly, low back pain, ascites, carotid bruit, systolic murmur, deep venous thrombosis, and hypovolemia. Each weekly, one-hour session included a discussion of the article emphasizing the sensitivity and specificity of PD findings and a practice session with an inpatient. Students at one hospital site were the intervention group, and students at 3 other clerkship sites served as controls. At the end of the clerkship, all students completed a PD quiz based on the articles, a skills assessment using standardized patients, and a survey about confidence in their own PD skills. Confidence scores were calculated by summing the responses to the 7 point Likert scale survey questions, with 7 indicating "very confident" and 1 indicating "not confident". The intervention group also received a pre-clerkship assessment of PD skill using standardized patients. RESULTS: For the first academic quarter of 1999-2000, 12 intervention students and 19 control students completed the quiz, skills assessment, and survey. Intervention students scored significantly higher on the end-of-clerkship skills assessment compared to control students, mean correct score (Ϯ SD) 92.8% (Ϯ6.8%) v. 51.4% (Ϯ12.5%), p Ͻ 0.001. Intervention students also scored higher on the endof-clerkship skills assessment than they did on their pre-clerkship skills assessment, mean correct score (ϮSD) 92.8% (Ϯ6.8%) v. 40.5% (Ϯ12.3%), p Ͻ 0.001. Compared to controls, intervention students reported greater confidence in their PD skills related to the 8 topics in the curriculum, mean confidence score (ϮSD) 41.6 (Ϯ4.6) v. 36.4 (Ϯ4.8), p ϭ 0.005. In contrast, there was no significant difference in the confidence scores of the intervention and control groups for 4 PD topics not covered in the curriculum, mean score (ϮSD) 18.9 (Ϯ2.4) v. 18.2 (Ϯ3.2), p ϭ 0.52. There was a non-significant trend toward higher mean quiz scores in the intervention group compared to controls, mean correct score (ϮSD) 69.5% (Ϯ9.3%) v. 65.5% (Ϯ7.3%), p ϭ 0.18. CONCLUSION: We conclude that our evidence-based PD curriculum was successful in improving student skill and confidence in specific PD domains. Medical educators attempting to improve physical diagnosis skill should consider adopting an evidence-based approach.

Farber, A Friedland, BM Aboff, DB Ehrenthal, T Bianchetta, Dept of Medicine, Christiana Care Health System, Wilmington, DE PURPOSE: Communication skills, including the communication of bad news to patients, have been taught to residents using a variety of methods. Standardized patients have become increasingly used; however, the use of actual patients with life-threatening illnesses has not been studied in resident education. We therefore conducted a workshop with residents using such patients. METHODS: Four patients with cancer from the Wellness Community of Delaware were recruited to participate in a 2 hour resident workshop on giving bad news to patients. The curriculum was based on work by Urban and Rabatin (1) , with patients using their own diagnoses and histories as the role play scenarios. Residents were instructed to give bad news as they usually do, then feedback from the patients, discussion about the role play, and about the method as described by Buckman (2) ensued. A second role play using the learned concepts was conducted, and a final debrief then occurred. Prior to the workshop, residents com-pleted an 11 item questionnaire about actions to be taken during the communication of bad news, with a similar post-test being administered at the end of the workshop. The residents answered the items using a 4 point Likert scale (strongly agree to strongly disagree), and demographic questions were included. Pre and post tests were compared using paired student's T tests, while the impact of demographic data on responses was analyzed via analysis of variance (ANOVA). RESULTS: Twenty-five residents participated in the workshop, with 15 (60%) completing both pre and post tests. Attitudes toward ensuring hope is conveyed to patients (p Ͻ 0.05), starting the discussion by ascertaining the patients' understanding of the condition (p Ͻ 0.01), and encouraging the the patient to express his/her feelings (p Ͻ 0.01) all significantly improved with the use of the workshop. Most of the improvement occurred in residents with previous training. CONCLUSION: A role play workshop using untrained patients with cancer can significantly improve residents' attitudes about giving bad news. Volunteers are without cost to the institution, and have a positive influence in such workshops. Residents who are previously trained in giving bad news may hold counter-productive attitudes, and therefore all residents should participate in workshops using role play by actual patients about giving bad news. PURPOSE: Although depression is one of the most frequent psychiatric disorders seen in the primary care setting, research suggests that patients with depression are seriously underdiagnosed and undertreated. Since approximately 40% of depressed adults receive mental health care in primary care settings, we wanted to explore the prevalence and treatment of this disorder in our residents' continuity clinic. METHODS: After review of the literature, variables were identified and used to develop an instrument that served as the basis for data extraction from patients' charts. One hundred twenty-two (122) charts were selected for review and data collection. Two PGY3 medical residents conducted the chart reviews. Comparisons were made between those identified as depressed and those not identified using either a t-test or Chisquare test. RESULTS: Of the 122 charts reviewed, 23 (19%) had the diagnosis of depression noted on the patients' active problem list. Twenty of the patients with depression (87%) were receiving some form of treatment, either medication or medication with counseling, and 3 (13%) had no documented treatment. Only 5 (22%) of the diagnoses of depression were substantiated by the use of an instrument or appropriate DSM4 criteria. CONCLUSION: Although the prevalence of depressive disorders in the primary care setting is purportedly in the 5% to 9% range, a higher prevalence rate was observed in our residents' continuity clinic (19%) . The accuracy of this prevalence rate is questionable given the limited use of instruments to substantiate the diagnosis of a depressive disorder. However, the residents predominantly followed recommended treatment guidelines for depression. This study raises concerns regarding the accuracy of diagnoses and corresponding documentation issues. Additionally these results have broader implications regarding medical error.

Outpatient Clinic, University of Lausanne, Switzerland PURPOSE: Proficiency in physical diagnostic skill seems to decline. Few data have examined this question and generally by using recorded cardiac events from patients instead of testing directly the bedside diagnostic skills. Objectives: 1. To compare the cardiac auscultatory skills of physicians in training in comparison with skilled cardiologists by using real patients to test bedside diagnostic skills 2.

To evaluate a teaching program at bedside in order to improve auscultatory diagnostic skills. METHODS: In an academic primary care center, 10 physicians (internal medicine and family practice trainees) and two skilled academic cardiologists, blinded to the patients' characteristics, examined 13 patients and documented their findings on a questionnaire. They should find 33 cardiac events and 13 diagnosis. Cardiac sound, murmurs and diagnosis were previously determined by an independent skilled cardiologist using echocardiography. The physicians participate in a 45 minute-course, once a week for 5 months. After the course, they listened again consecutively the same patients. The percentage of correct diagnosis was the main outcome of interest. RESULTS: The experts were the most skillful with 62% of correct diagnosis and heard 100% percent of diastolic murmur. The physicians gave only 21% of correct diagnosis and detected 62.5% of diastolic murmurs. 2) After the course the mean percentage of correct diagnosis was 35% (p Ͻ 0.05, paired t-test), a relative increase of 54%. CONCLUSION: The bedside diagnostic skills of internal medicine and family practice trainees are effectively low but could be improved by a course focusing on bedside teaching. A randomized clinical trial is needed to confirm these results.

PURPOSE: Direct-to-consumer pharmaceutical advertising has increased pressure on physicians to accommodate marketing-inspired patient expectations for drug therapy. The purpose was to assess resident confidence in prescribing appropriate therapy while satisfying a patient in response to a patient treatment request. METHODS: A cross-sectional written survey was administered to 50 internal medicine resident's in their 1st-3rd year. Resident characteristics studied were gender, year of training, and USMLE admission score. Residents self-rated their confidence in prescribing appropriate therapy in response to patient specific requests and in achieving a satisfactory patient visit from the patients point of view on 2 separate 5-point scales (1 ϭ not very confident; 5 ϭ very confident) for each of 6 scenarios. The 6 scenarios represented consumer requests based upon information from a newspaper article, an internet resource, a television ad, a magazine ad, prior experience, and testimonial from a relative. Resident's self-rated attitudes (perceived accuracy, accessibility, applicability of information, frequency of resource use in practice) toward medical knowledge resources using a 5-point scale. Medical knowledge resources evaluated were the internet, drug reference manuals, electronic medical text, peer reviewed research, and pharmaceutical sales representatives. RESULTS: No differences in responses were found in their confidence with either treatment or satisfaction based on gender, year in training, or USMLE score. No relationship was found between confidence in the accuracy of all 5 medical knowledge resources studied and confidence in treating or satisfying a patient. A significant correlation was found between ready availability of the internet with confidence in prescribing for 5 of the 6 scenarios (newspaper article, internet resource, television ad, magazine ad and prior experience). A significant correlation was also found between availability of all 5 medical knowledge resources and resident confidence in prescribing for the newspaper article scenario. A negative correlation was found between frequency of use of pharmaceutical representatives and confidence in treatment success and patient satisfaction for several of the scenarios. CONCLUSION: Ready availability of the internet is associated with resident confidence in responding to a specific patient request for treatment. Greater interaction with pharmaceutical sales representatives may be associated with lower resident confidence in ability to manage patients with specific treatment requests.

R Germany, L Cation, Medicine, University of Illinois College of Medicine, Peoria, IL PURPOSE: Internal medicine residency programs have traditionally used costly paper residency brochures to provide information on their programs to applicants. The advent of the Internet allows residency programs to post program information on their websites with ease and minimal expense. Now that the National Residency Matching Program requires that all internal medicine residency programs use the Electronic Residency Application Service (ERAS) for applications, we examined how many programs have Internet websites and how many are still distributing paper residency brochures. METHODS: A list of the United States internal medicine residency programs was obtained from the AMA Directory of Graduate Medical Education 1999-2000 edition. Each residency program was contacted by phone or email between September and December 1999 and a survey was completed. The survey determined if the program had a paper brochure and/or an Internet website. Websites were then examined for their content. RESULTS: Of the 380 internal medicine residency programs listed, we were able to contact 347 (91%). Of these 347 programs, 293 (84%) had a program website, 159 (50%) had a paper brochure, 134 (39%) had both, and 9 (3%) had neither. Regarding website content, 225 (77%) of websites contained the program phone number, 179 (61%) had an email address listed, and 168 (57%) had an email direct link. Website program information included the specific yearly rotation schedule (77%), conference schedule (64%), and resident clinical research opportunities (57%). Other common features included vacation (62%), salary (56%), and community links (44%). Sixty-eight percent of websites had a picture of the residency program director and 53% had a picture of the department chairperson. Seventy percent of websites had a reference to ERAS in their application section but only 35% had a direct link to ERAS. Only 19% of websites had a reference to the program's ABIM Certification Exam first-taker pass rate. CONCLUSION: Our study reveals that of the internal medicine residency programs we were able to contact, most have an Internet website with information on their program. The reasons for this are many and include cost and ease of updating information. The content of program websites varied widely but most had information on their curriculum, benefits, and a way to contact the program. With the boon of the Internet and the move to the ERAS system, most internal medicine Overall Overall teaching effectiveness Ϫ0.14 Ϫ0.01 0.32 PURPOSE: Teaching physical diagnosis skills lends itself particularly well to multimedia and virtual reality (VR) presentation, but the efficacy and efficiency of this approach has not been tested. We compared the achievement of students who used a web-based multimedia tutorial (consisting of interactivity, pictures, and virtual reality simulating the examination of the eye and ear) with that of students who used a printed tutorial presenting the same information written by the same authors. METHODS: Student scores on a pretest covering information about the eye and ear were used to divide the first-year medical student class at the University of Nebraska into two equal groups. Group A (n ϭ 60) used the written tutorial for the ear material and the multimedia tutorial for the eye material. Group B (n ϭ 61) did the reverse. We measured knowledge gained with a multiple-choice post-test that included both paper (40 questions) and web-based (10 questions) sections. We recorded time by recording check-out and check-in times for the written tutorial and log-in times for the computer program. RESULTS: We found that students who used the multimedia version scored higher on the post-test than those using the written version. Scores for the eye test showed a larger gain (15.9 vs. 13.4 of 25, p Ͻ 0.001), than for the ear. (16.1 vs. 14.5 of 25, p Ͻ 0.03). These effects persisted after controlling for the time spent and the pre-test score: (p Ͻ 0.001, ANCOVA). Students using the multimedia versions spent more time than those using the written versions (mean time ϭ 44.1 min vs. 33.4 min, p Ͻ 0.01). The correlation between the time spent and the post-test score, however, was higher for those using the mulimedia tutorial than for those using the written tutorial. Regression analysis found three independent predictors of the post-test score: use of the multimedia version (rather than the written tutorial), time spent on the computer section, and pre-test score. 117 of 121 students (97%) returned post-surveys: 78% preferred the multimedia version; 63% rated the multimedia learning more effective. CONCLUSION: In conclusion, this controlled trial with 121 students at one institution found that computer learning incorporating VR and interactivity resulted in increased learning as well as increased time on task. The increase in learning exceeded what would be expected from the increased time alone. Differences in effect size suggest computer learning benefits some areas more than others. These results are promising and suggest that computer learning contributes uniquely to learning physical diagnosis.

INTRODUCTION: Few studies have evaluated the effectiveness of subspecialty consult rotations as an educational method for preparing general internist for practice. The primary objective of this study is to determine to what extent subspecialty education during residency training prepares graduates for general internal medicine practice. The secondary objective is to assess the impact of subspecialty consult rotations on graduates' sense of being prepared for practice. METHODS: Surveys were sent to all 92 OHSU graduates (1995) (1996) (1997) (1998) . Demographic questions assessed type and location of practice. Graduates were asked to indicate if they had taken the rotation or not. Using a 1-5 Likert scale (unprepared to prepared) each subspecialty was assessed in three areas; 1) general knowledge, 2) diagnosing common disorders, and 3) management of disease. Graduates who responded with a 4 or 5 were considered prepared. They were asked to describe their exposure to each subspecialty and to identify other areas that they wished they had received more training during residency. RESULTS: A total of 61 ⁄ 92 (66%) graduates responded to the survey. Thirty five percent were associated with academics. Over half spent greater than 50% of their time in an outpatient practice. Of graduates who rotated through cardiology, rheumatology, gastroenterology, infectious disease, 83% or higher reported being prepared Nephrology scored higher in general knowledge (90%) compared with diagnosis (73%) and management (73%). Geriatrics, pulmonary, endocrinology, and hematology/oncology had a comparatively smaller proportion of graduates reporting they felt prepared (47%-79%). Graduates reported being least prepared in Neurology (24%-45%). Graduates who did not take a specific rotation reported feeling less prepared in nearly all subspecialties compared with those who completed rotations. Graduates reported that they had too little exposure to endocrinology, pulmonary, geriatrics and neurology. Orthopedics, neurology, womens' health and dermatology were most frequently cited as needing more exposure during training. CONCLUSION: The majority of recent graduates believe their subspecialty training experiences has prepared them for their current practices yet there are several areas needing improvement. Graduates who did not do a consult rotation were less likely to feel prepared. This study demonstrates that the subspecialty consult rotation is an important method for assuring that graduates feel prepared for practice. Ongoing challenges include deciding how subspecialty material will be taught to all residents and how predominantly hospital-based medicine programs will be able to meet the needs of graduates in primary care.

PURPOSE: To evaluate the first in a series of national conferences designed to improve faculty development in ambulatory teaching on a national level, and to compare the experience of hospital versus community-based attendees. METHODS: The General Internal Medicine Faculty Development Project: Generalist Faculty Teaching in Community-Based Ambulatory Settings, funded by HRSA, is a collaborative effort of the major organizations in internal medicine. (Bowen et al. AJM: 107, (193) (194) (195) (196) (197) Teams from internal medicine teaching programs are selected to attend one of three national conferences based upon a submitted plan for ambulatory faculty development. Teams included an administrative team leader, a community-based teacher, and other key faculty members. Participants attend facilitated team meetings, plenary sessions, and 4 of 17 workshops related to educational skills and program implementation. RESULTS: Sixty-two teams (259 persons) were selected to attend the first conference in December, 1999. Response rate for the evaluation was 91%. Overall, 1% would not recommend the conference to a colleague, 9% would recommend as a satisfactory experience, 42% as a good experience, and 48% as an outstanding ex-perience. Mean ratings for individual workshops ranged from satisfactory to outstanding. Workshops with the highest attendance were "Teaching in the outpatient setting: precepting skills" and "How Doctors Learn: Tips for Teaching Faculty to Teach." The "precepting skills" workshop received the highest rating (70% outstanding). Participants who identified themselves as being primarily communitybased (CB) (N ϭ 85) rated the conference similarly to hospital-based teachers (HB) (N ϭ 130) (p ϭ 0.4). On a scale listing prior training in 24 teaching skills, on average, CB teachers were trained in 4.5 fewer teaching skill areas, compared with the HB teachers (p Ͻ 0.001). CB teachers were less likely to have protected time for their faculty development project compared with HB teachers (odds ratio 0.49, 95%CI 0.24-0.98), but supervisor support and salary support rates were similar. Participants rated the likelihood that their group will accomplish the objectives of their plan with a mean ϭ 4.2 (SD 0.73) on a likert scale from 1 ϭ Not Likely to 5 ϭ Extremely Likely. CONCLUSION: This collaboratively implemented national faculty development project on ambulatory teaching succeeded in soliciting acceptable faculty development plans from teams composed of hospital and community-based teachers from 62 teaching institutions. Despite disparities in previous training and protected time, both CB and HB teachers rated the conference highly and were confident in the potential accomplishments of their teams. We will assess the accomplishments of individual teams through prospective follow-up.

A CRITICAL APPRAISAL OF THE CRITICALLY APPRAISED TOPIC. MR Huber, VM Montori, Internal Medicine, Mayo Clinic, Rochester, MN PURPOSE: The purpose of this study is to describe the Critically Appraised Topics (CATs) currently available on the Internet. METHODS: We searched the World Wide Web using the string "critically appraised topic" in the meta-search engine Metacrawler (www.metacrawler.comusing Altavista, Infoseek, Webcrawler, Thunderstone, About, Excite, Lycos, Looksmart, GoTo, and DirectHit search engines). The last search date was November 28, 1999. We also used personal and institutional websites that list evidencebased medicine web pages (Netting the Evidence, www.ceres.uwcm.ac.uk, members.tripod.com/jepling/EBM.htm) and institutional and commercial medical websites (www.hon.ch/MedHunt/, www.medweb.emory.edu/MedWeb/, www.healthatoz.com, healthweb.org). We obtained 105 hits of which only 6 websites featured CATs of interest to general internists. One of us (MRH) collected data from each CAT including number of authors, each author's academic degree or training level, the existence of a clinical scenario, the existence of a focused clinical question and the number of stated components-if all components were present we considered them "well-built", type of question (i.e., therapy, diagnosis, prognosis, harm), the existence of a search strategy description, date of searching, list of databases searched, number of articles included in the appraisal, existence of a statement describing the article's validity, results and applicability, an expiration date and if the CAT had expired based on it, and if the CAT had been peer-reviewed before publication. RESULTS: A total of 236 CATs were identified. Most CATs are single-authored (94.5%) by housestaff or junior faculty. Only 18% of CATs described the search strategy followed to find the appraised evidence. Sixty-four percent of CATs found described a single article. Of all CATs with expiration date, 88% had expired based on it. Most CATs did not fulfill the stringent validity criteria for systematic reviews of the literature or the less rigorous validity criteria set forth by the Health on the Net Foundation or by the Discern project. CONCLUSION: CATs are readily available but potentially biased, incomplete and outdated sources of medical information on the Internet. Evidence-based medicine educational activities that include publishing CATs on an institution's Internet website need to be mindful of the high standard expected of medical information on the World Wide Web by patients and healthcare providers. At the time of this evaluation, CATs do not meet those standards and should only be shared within the setting in which they were created. PURPOSE: To assess inpatient faculty teaching effectiveness in the setting of the hospitalist and traditional academic models of inpatient practice. METHODS: At the end of each month for the first half of 1999, a 5-point Likerttype survey was administered to residents on general medical services. This questionnaire was broken into categories assessing the effectiveness of teaching, feedback, collegiality, allowance of autonomy, and efficiency of daily rounds. Residents were asked to rank-order the importance of feedback and collegiality, quality of teaching, autonomy, and efficiency of daily rounds in the overall scheme of their perception of satisfaction on general medical services. RESULTS: There was a 54% response rate (104/180 surveys). Residents prioritized quality of educator (34%) in ranking contributing factors to their satisfaction with a hospital rotation. This was followed by efficiency of morning teaching rounds (18%), type of patient population (15%), team dynamics (13%), collegiality (11%) and finally autonomy (8%). In areas demonstrating quality of education, faculty with increased presence on the wards were perceived to demonstrate physical exam (p Ͻ 0.05), cite literature (p Ͻ 0.01), discuss differential diagnosis (p Ͻ 0.01), and provide more formal teaching (p Ͻ 0.001) than those in the more traditional model of inpatient teaching. Perceptions were similar for areas of collegiality: giving regular feedback (p Ͻ 0.01) and comfort with asking faculty questions PURPOSE: As of 1998 International Medical Graduates (IMGs) constituted 27% of all residents and 40% of internal medicine residents in the United States. Pre-residency evaluation has shown that IMGs perform as well as US medical graduates (USMGs) on tests of medical knowledge and are proficient in history taking and physical examination. However it is not known whether IMGs have had similar educational experiences as USMGs in all areas important for their preparation for internship in the United States.

METHODS: To answer this question, we surveyed a convenience sample of 60 first-year internal medicine residents at an academic hospital and a public hospital in Chicago, IL. The 54-question survey assessed resident's previous educational experiences in medical school and residency, attitudes towards certain patient problems, and confidence in their abilities to address those problems. RESULTS: Compared to the 28 USMGs, the 32 IMGs were significantly less likely to have had formal coursework in how to take a sexual history (41 vs. 89%; p Ͻ 0.01), how to recognize domestic violence (19 vs. 68%; p Ͻ 0.01), the impact of culture on health (45 vs. 79%; Ͻ 0.01), how to discuss the process of dying with patients (19 vs. 82%; p Ͻ 0.01), and how to assess patients for substance and alcohol abuse (41 vs. 94%; p Ͻ 0.01). Most IMGs and USMGs had training in interviewing skills (84 vs. 96%) and psychiatry (91 vs. 100%). Few in either group had formal coursework in how to manage patients' pain (34 Vs 39%). Despite these differences, IMGs were as confident in their abilities to address these issues as USMGs. They also endorsed certain attitudes more often than USMGs reflecting discomfort in discussing personal issues with patients, lack of awareness of the signs and treatment of domestic abuse, and cultural differences in the care of dying patients. CONCLUSION: There were important differences in the educational experience of international and US medical school graduates prior to internship in this study. IMGs were significantly less likely than USMGS to have had formal skill training in several areas essential to the practice of internal medicine in the United States. These differences should be addressed in residency curriculum given the large and important role IMGs play in US Graduate Medical Education.

PURPOSE: Many medicine residency programs have started to use communitybased practices as ambulatory care training sites for their residents. However, there are not many studies evaluating the influence of the use of community base practices on resident education. The purpose of this study is to determine whether the setting for ambulatory care training influences resident education. METHODS: We conducted a survey of all internal medicine residents at a tertiary care hospital. The survey included questions regarding training year, ambulatory care training site, average number of patient seen per session, satisfaction with the ambulatory experience, diversity of diagnosis and socio-economic background, independence managing their patients, satisfaction with teaching, evidence based medicine, alternative medicine, cost effective medicine, preventative medicine, exposure to the business aspects of medicine, and having adequate time with their preceptor. A preliminary survey was piloted on interns for reliability. Those questions that reached reliability score (alpha) above 0.75 were used in the main survey. The main survey results were analyzed by using Kruskal-Wallis Test. RESULTS: Three questions (cost effective medicine, preventative medicine, and exposure to the business aspects of medicine) were eliminated during the pilot phase because they did not reach a reliability score of 0.75. The main survey was sent to 33 residents and 28 responded (85%). Twenty-one percent of residents practiced in medical clinic, 64% in community based practices and 14% in fulltime faculty practice. Most residents (83%) saw 5 to 6 patients per session. Using a Likert scale of 1 to 5, the average score for satisfaction with the ambulatory experience was 4.5, diversity of diagnosis 4.1, diversity of socio-economic background 3.1, independence managing their patients 4.2, satisfaction with teaching 4.3, evidence based medicine 3.7, alternative medicine 2.1, and having adequate time with their preceptor 4.3. Statistically significant differences were seen in questions regarding independence managing their own patients and diversity of diagnosis between residents from medical clinic and other training sites.

PURPOSE: To explore the patient's perception of primary care received from a resident working with assigned faculty versus a student-faculty pair or faculty alone. METHODS: The study design was a cross sectional survey of patients attending two academic hospital based Internal Medicine clinics, one a predominantly faculty practice with resident and student participation (FWest) and the second being predominantly resident practices with faculty supervision (FEast). Two focus groups, one from each clinic, were recruited to identify issues relating to the teaching encounter. A survey was constructed using 7 questions derived from the focus groups; 5 questions from the Medical Outcomes Study (MOS)-9 item visit rating form; the MOS-6 item general health survey; and 22 questions from the Components of Primary Care Index (CPCI) which includes scales for continuity, interpersonal communication, coordination of care, knowledge of the patient and advocacy. All patients attending the clinics during a twenty day study period were eligible for the study. Surveys were given to the patients at the time of arrival in the waiting room and collected upon departure. RESULTS: Surveys were completed by 361 patients. Overall satisfaction was high, 4.5 on a 5 point Likert scale. However, among patients seen by residents, satisfaction and CPCI scores were significantly worse than either the student/faculty pair or the faculty alone. In contrast, among patients seen by a student, ratings were comparable or better than that of faculty alone. Resident patients perceived themselves as sicker (general health ϭ 3.1) vs. faculty patients (3.65 ) and a greater number were indigent (general assistance/T19 ϭ 33%) vs. those of students (12%) or faculty (7%). However no correlation between ratings and health status, insurance status or site of care was noted. CONCLUSION: These results suggest that in this institution, resident participation in care is perceived more negatively than that of students or faculty care alone. Clearly the patient populations differ even though there was no apparent correlation with health or insurance status. Other hypotheses might involve resident attitudes, competing agendas, or level of faculty involvement. PURPOSE: Although medical students may initially learn how to perform pelvic exams on paid volunteers, they often achieve proficiency by practicing on anesthetized women. Practicing pelvic exams on women without consent is a violation of patients' right to self-determination. We sought to determine the prevalence of medical students' performing pelvic exams under anesthesia for practice and to assess students' attitudes toward the need for explicit informed consent. METHODS: We surveyed 1500 randomly selected, fourth year, US medical students using a 44 item self-administered questionnaire. Descriptive statistics were used to characterize students' experiences and attitudes about performing pelvic exams under anesthesia. Logistic regression was used to identify variables associated with beliefs that consent was implied or that explicit consent should be obtained before students perform pelvic exams under anesthesia. RESULTS: After one month, 387 completed surveys have been returned and an additional 142 had invalid addresses (preliminary response rate 28%). Among respondents 50% were women and 74% were ages 25-29. Among the first 387 respondents, 90% had performed a pelvic exam on an anesthetized patient. Among those who performed exams under anesthesia, the main reasons for the exam were: to palpate an abnormality (64%) and to practice pelvic exams (29%). Over half (58%) of students believed consent for pelvic exams under anesthesia was implied by patients being in a teaching hospital. However, only 9% believed that women in teaching hospitals for surgery understand that medical students perform pelvic exams under anesthesia. 66% of students believed written or verbal consent for students to perform pelvic exams under anesthesia should be ob- Since then it has acquired the prestige of the best formal medical education but also the tradition of an arduous and stressful activity. Shortness of resources and an authoritarian style has prevailed in the country's recent history and consequently in the hospital institutions. In this milieu mistreatment should be considered an issue of concern. The purpose of this study is to explore mistreatment experiences reported by residents during their training. METHODS: Design: Focus group interviews. Setting: Ambulatory care center fom Hospital de Clinicas, Buenos Aires University. Population: Internal Medicine and General Medicine residentes (PGY2 and PGY3) volunteered to participate in two 90 minutes sessions coordinated by two of the authors. Anonymity was guaranteed. Sessions were tape recorded, transcript and analyzed by two of the investigators. Participants were invited to define by there own mistreatment and openly worked out personal episodes. Measurements: Qualitative, consensus review of the content of session transcripts. RESULTS: Personal mistreatment has been perceived by almost all the participants. Mistreatment often took the form of public humiliation and belitled. Although task for punishment were commonly infringed. High concern was elicited by the obligation to work in impaired personal conditions as a result of sleep or food deprivation. They also expresed as with high frequency been obliged to work in non medical tasks replacing nurses or as ancillary services as stretch-bearers. Threats of physical harm and of being slapped, pushed or hit were rare. Respondents believe that the origin of the mistreatment is in the relative powerlessness of residents, who are on the lowest rung of the medical hierarchy. In their opinion this situation is well known and tolerated by authorities. Senior residents also exert mistreatment to their lower degree companions. CONCLUSION: This study reveals that this group of residents perceive mistreatment and hostility in their training environment. Cultural determinants and shortness of resources may be responsible for this situation. Further research as well as educational interventions designed to prevent and address these issues can have positive impact on the experience of medical residents.

PURPOSE: Academic general internists are increasingly expected to teach EBM to students and residents. However, it is not known how they actually use evidence in their own patient care. METHODS: We conducted a national, web-based survey of academic internists as part of a randomized trial of an electronic journal club, to assess use of evidence in practice, attitudes toward EBM, and ability in critical appraisal. Thus far we have 123 (9%) responses to an e-mail invitation recently sent to 1375 academic internists. RESULTS: Mean age was 40.1 years and 40% were female. Most have teaching responsibility for residents (95%) and students (84%). Half, 50%, teach EBM to medical students or residents. When faced with clinical questions, they reported finding an answer 73% of the time. They were most likely to seek answers from colleagues (93%), Medline (84%), paper textbooks (79%), and research summaries like ACP Journal Club (57%). In a typical month of practice, they reported using published evidence in 65% of patient encounters. They actually read published evidence to guide clinical decisions for 30% of patients. They were confident in critical appraisal skills (mean ϭ 3.2 on 4-point, 14-item scale) and had positive attitudes about the role of evidence in patient care. Those who teach EBM reported using evidence in their clinical practice more often (69% vs. 61%, p ϭ .08), were more confident in their critical appraisal skills (3.4 vs. 3 .0, p Ͻ .001), and read more journals/week (4.5 vs. 3 .8, p ϭ .007) compared with those who do not teach EBM. However, they reported similar success rates in finding answers to questions arising in practice. CONCLUSION: Academic general internists, especially teachers of EBM, appear to use evidence in their clinical practice most of the time, and are successful finding answers to questions using a variety of resources. It is reassuring that teachers of EBM seem to practice what they preach. METHODS: Prospective interinstitutional study in which 30 Internal Medicine Residency programs 1) completed profiles of their end-of-life (EOL) care teaching practices and 2) administered EOL knowledge and self-assessed competency assessment instruments to their residents and faculty. Programs were recruited by the American Board of Internal Medicine. Residency program directors identified the EOL education offerings within their institutions, and the means used to evaluate this teaching. Performance on the knowledge examination is represented by the percentage correct score. Ratings on the self-assessed competency assessment ranged from "1" (need further basic instruction") to "4" (competent to perform independently"). The alpha value, representing a measure of the reliability of the instruments, was .62 for the knowledge test and .95 for the self-assessed competency instrument. Distribution of training levels among the 1,139 participants was 29% PGY1, 44% PGY2-4, 27% faculty. RESULTS: Ethics and pain assessment and treatment constituted the EOL topics most often represented as required instruction, in just over half (54%) of programs. Non-pain symptom treatment and education about hospice care were the EOL topics least often presented, in either required or elective form. The most consistent source for programs to assess their residents' competencies in EOL care were faculty ratings. Performance on the knowledge examination revealed small but statistically significant increases over training levels (PGY1 mean ϭ 50.5%, PGY2-4 ϭ 55.2%, to faculty mean ϭ 60.7%) Self-assessed confidence moved from the PGY1's mean rating (2.6) indicating they could perform EOL practices with supervision to the PGY2-4 (3.2) and faculty (3.5) mean ratings corresponding to judgments that they could perform practices independently (F ϭ 272.9, p Ͻ .000.) CONCLUSION: The study's findings speak to the willingness of Internal Medicine Residency programs to critically assess their EOL educational offerings. Although the study did find statistically significant increases across training levels in EOL knowledge, the relatively small magnitude of change and the overall knowledge mean score (55.2%) indicate considerable gaps remain in the EOL knowledge base. Despite these gaps, the mean value of the ratings of self-assessed competencies of the residents and faculty reflect their judgment that they are prepared to carry out EOL care responsibilities independently.

PURPOSE: With increasing technology, medicine is becoming more invasive in its practice. With this comes the mandated responsibility for residency programs to prepare graduates to perform a core set of procedural skills. We describe a 1.5hour workshop, which provides standardized, pedagogic training in one of these core procedural skills-insertion of the central venous line (CVL). METHODS: Nineteen interns (76%) of the University of Hawaii Internal Medicine Residency Program were surveyed using a self-reported questionnaire regarding their attitudes and knowledge of CVL insertion. They identified discomfort with procedural technique and knowledge. Instruction was described as haphazard with no standardized, uniform training in existence. In collaboration with faculty of the Department of Anatomy, we executed a multi-disciplinary workshop with the goals of improving resident attitudes about CVL insertion and improving clinical knowledge. Curricular content included informed consent, indications and contraindications for CVL insertion, technique, complications, and documentation. Instruction was via interactive discussions, practical skills demonstrations, and a cadaver demonstration. RESULTS: Comparison between pre-and post-questionnaires revealed improvement in intern comfort with informed consent, insertion technique, recognition of proper placement, knowledge of complications, and writing a proper procedure note (p Ͻ 0.05). Comparison between pre-and post-objective assessment showed significant knowledge gain (p Ͻ 0.05).

CONCLUSION: All participating interns agreed that they enjoyed the workshop and recommended it to be continued for future interns. Our data demonstrates that a systematic approach to CVL insertion has a positive impact on intern attitudes and knowledge. PURPOSE: To ascertain whether alcohol restriction lowers blood pressure in hypertensive and normotensive individuals. METHODS: An English-language MEDLINE search from 1966 to January 1999 was conducted. Bibliographies of review articles were also searched. Trials that had randomized allocation to control and alcohol reduction groups, monitored by timed alcohol consumption, with outcome measures of systolic and/or diastolic blood pressure were selected by blinded review of the methods section. Two investigators extracted data independently, using purpose-designed forms, and discrepancies, if any, were resolved by discussion. Q statistics was used for a test of homogeneity and fixed-effect-model was used for weighted-pooled estimates of net blood pressure change. RESULTS: Seventeen trials that met our inclusion criteria did not show heterogeneity. The mean baseline alcohol consumption was 49.7 g/d in 13 trials with 1365 hypertensive subjects and 52.8 g/d in 4 trials with 929 normotensive subjects. The mean net reduction in alcohol consumption was 30.0 g/d in hypertensive trials and 22.3 g/d in normotensive trials. The net decrease (95% confidence interval) in blood pressure was 4.0 mm Hg ( PURPOSE: Many medicine residency programs have begun to use communitybased practices as ambulatory care training sites for their residents. However, there is a paucity of studies evaluating the influence of community base practices on resident education. The purpose of this study is to determine whether residents practicing in different ambulatory training sites care for patients with similar complaints. METHODS: We conducted a survey of all internal medicine residents (PGY-2 and 3s) at a tertiary care hospital. The survey included questions regarding training year, ambulatory care training site, average number of patients seen per session and how frequently they encountered the 20 most common presenting symptoms from the National Ambulatory Medical Care Survey (NAMCS), a probability sample survey conducted by the Centers for Disease Control on patients' office visits since 1973. The 1996 NAMCS dataset was used to generate a list of the most common presenting symptoms seen by internists in outpatient setting. Sampling weights were applied to generate national estimates of the total number of visits. Resident survey results were analyzed by using Kruskal-Wallis Test. RESULTS: The survey was sent to 33 residents and 28 responded (85%). From these 50% were PGY-2s. Twenty-one percent of residents practiced in medical clinic, 64% in community based practices and 14% in full-time faculty practice. Eighty-three percent of residents saw 5 to 6 patients per session. The most common presenting symptoms in the NAMCS were: general medical exam 8.4%, progress visit 6.7%, cough 3.9%, hypertension 3.8%, chest pain 2.4%, back symptoms 2.3%, stomach/abdominal pain 2.3%, diabetes mellitus 2.2%, blood pressure test 2.1%, throat symptoms 2.0%, shortness of breath 1.7%, headache 1.7%, tiredness 1.5%, vertigo/dizziness 1.4%, upper respiratory infection 1.4%, rash 1.4%, medication 1.2%, nasal congestion 1.0%, low back symptoms 1.0%, and neck symptoms 1.0%. Residents in all settings perceived seeing the top 20 presenting symptoms more frequently than the national estimates. Once adjusted for multiple analysis there were no statistically significant differences in the frequencies of the most common presenting symptoms between all three training sites. CONCLUSION: This study suggests that residents see the most common presenting symptoms in similar or greater frequency than the national estimates, but no differences were seen among the training sites. Further studies are needed to determine if these symptons correlate to similar diagnoses and the educational impact of training in different ambulatory sites.

H Onishi, Y Oda, S Emura, S Yamashiro, S Koizumi, General Medicine, Saga Medical School, Saga, Japan PURPOSE: To formulate and investigate feasibility of the objective structured clinical examination for basic neurological examination (Neuro-OSCE) as a formative assessment tool for undergraduates. METHODS: We formulated a 31-item Neuro-OSCE with 3 point scale (0, 1 or 2) based on the four-minute neurologic exam (Goldberg). In 1998, we administered it to 104 medical students in senior year. Responses to our questionnaire from 103 students provided the information on their self-learning behavior and self-evaluation. For statistical analysis, Mann-Whitney's U-test and Spearman's correlation test were used. RESULTS: Cronbach's coefficient alpha was 0.71 indicating fair internal consistency. Mean score for the entire students was 53.6 ϩϪ 6.1 (SD). Female students marked significantly higher score than male students (P Ͻ 0.001). Higher score was associated with better self-evaluation (p ϭ 0.013) and longer self-practicing time (p ϭ 0.011) but not with knowledge-based-self-learning time (P ϭ 0.84). CONCLUSION: We designed Neuro-OSCE for the assessment of skills in psychomotor domain. As expected, higher score predicted longer self-practicing rather than knowledge-based-self-learning time. Extracting essential items from complex neurological examinations enabled undergraduates to self-evaluate and to master the skill proportional to the length of practice. To verify the external validity of our Neuro-OSCE neurology specialists should be involved. The didactic sessions start with a brief case presentation by each resident, this is followed by an in-depth discussion of the topics of the day with expected resident participation. All necessary reading materials are handed out at the beginning of the month. Each three hour didactic session ends with role-plays with the instructor taking the role of the patient and one resident the doctor. This enables the residents to put into immediate practice what they have just learned. There is also a half-day visit to a domestic violence prevention center. Clinical practice of women's health is expected to occur when residents see female patients in their continuity practices. To expand the scope of this curriculum beyond these limited topics both residents and faculty have been encouraged to incorporate into everyday teaching and learning how various conditions differ in women. During the last session of the month each resident is responsible for presentation of a topic in women's health not covered in the curriculum. This well researched topic is presented to the instructor and other resident(s). Learners are evaluated by the completion of a standard evaluation form used in the program. Evaluations are based upon residents' performance on a pre-and posttest, participation during the didactic sessions and the quality of the end of the month presentation. Learners evaluate the curriculum by filling out a survey. RESULTS: The curriculum was well received by all residents. All residents surveyed either agreed or strongly agreed that because of the knowledge gained during the month they would be able to evaluate, manage or appropriately refer female patients who presented to them with the conditions discussed. The mean posttest scores were 80%, an improvement over the mean pretest score of 62%. Residents stated particularly that the role-plays helped improve their understanding. A survey of these residents in their third year regarding their perceptions on the quality of women's health education in this program is planned. CONCLUSION: The curriculum has improved residents' knowledge in selected women's health topics. Using this same curricular model, topics could be adapted to meet the learning needs of other residency programs.

RESIDENT BEHAVIOR? YL Oppenheim, JH Oh, AM Eichorn, AB Jotkowitz, Division of General Internal Medicine, Long Island Jewish Medical Center, New Hyde Park, NY PURPOSE: The increased focus on ambulatory care in Internal Medicine residencies has necessitated the use of community based practices as training sites for residents. However, there is a paucity of data regarding the influence of community versus hospital based training on resident practice. The purpose of this study is to determine whether setting of ambulatory training influences resident practice patterns when screening for colorectal cancer. METHODS: We conducted a survey of all internal medicine residents at a tertiary care hospital. The survey included questions about resident training level, gender, ambulatory care setting and practice patterns regarding colorectal cancer screening. Endpoints on interest were based on current screening recommendations and consisted of appropriate age of screening (50 years old) and methods of screening of asymptomatic individuals. Appropriate methods of screening were defined as the use of home based fecal occult blood testing, flexible sigmoidoscopy, colonoscopy or barium enema but not digital rectal exam or office based guaiac testing after digital rectal exam. We also surveyed resident response to a positive homebased fecal occult blood test. Appropriate response to a positive home based fecal occult blood test was defined as performing a colonoscopy only. Data was analyzed by logistic regression to determine independent predictors of the endpoints. RESULTS: Surveys were sent to 55 residents. Responses were received from 31 residents, of whom 29% were first year, 32% were second year and 39% were third year residents. A total of 55% practiced in community based practices while 45% practiced in hospital based practices. Seventy-four percent of residents started screening at the appropriate age and 58% did not have an upper age limit for screening. Eighty-seven percent of residents utilized inappropriate initial screening tests in addition to or in place of the recommended screening tests. Seventyseven percent of residents did not perform colonoscopy in response to a positive home based fecal occult blood test. There was a strong trend toward hospital based setting being a predictor for appropriate age of screening (OR ϭ 9.03, CI ϭ 0.95 to 86.04) and of performing colonoscopy after a positive home fecal occult blood test (OR ϭ 9.19, CI ϭ 0.88 to 95.61). CONCLUSION: In our study, residents who practiced in hospital based settings more closely followed the guidelines for colorectal cancer screening. If confirmed in further studies, these results are concerning given the current trend toward ambulatory training in community based offices. Further studies are needed to determine the influence of ambulatory care training setting on the quality of resident education. Each examination consists of multiple choice questions that tests skills in data recall, data interpretation, diagnosis and/or treatment, and clinical decision making, as it pertains to the particular subspecialty being tested for that month.

Results of ITE and IHE examinations for academic years 1997-98 and 1998-99 were reviewed and residents' scores in each post-graduate (PGY) class were compared amongst themselves and a numerical rank was assigned to each resident (i.e., the highest scorer received a rank of one; the second highest scorer a rank of two, etc.). Each resident was assigned two ranks: one based on the score achieved on the ITE, and the other based on the yearly average score for the monthly IHE. A total of six ranking groups were, therefore, obtained for each academic year, corresponding to the PGY-I, PGY-II, and PGY-III IHE and ITE scores.

In each academic year, the rank on the IHE for each resident was compared to his or her ITE rank, in each PGY class. RESULTS: Using the Spearman Rank Test, a high concordance () was found between the numerical rank the residents achieved on the IHE compared to that of the ITE. In both academic years, this concordance was highest in the PGY-III classes, with the ϭ 0.92 in 1997-98, and the ϭ 0.70 in 1998-99, and was statistically significant (p Ͼ 0.0001 and p Ͼ 0.002, respectively).

Concordance for the PGY-I and PGY-II classes in 1997-98 was ϭ 0.52 and ϭ 0.84 (p Ͼ 0.054 and p Ͼ 0.0001), respectively, and in 1998-99, was ϭ 0.67 and ϭ 0.64 (p Ͼ 0.013 and p Ͼ 0.027), respectively. CONCLUSION: The IHE is a fairly reasonable tool to predict which residents will score highest and which will score lowest on the ITE, especially in the PGY-III classes.

SATISFACTION? R Pirkle, DR Campa, C Grudzen, G Gildengorin, M Rabow, EJ Perez-Stable, Medicine, University of California, San Francisco School of Medicine, San Francisco, CA PURPOSE: Managed care organizations often exclude residents from providing care to their enrollees in part because of the perception that residents provide care that is less than satisfying. We set out to test whether there is a difference in satisfaction between patients in a predominantly managed care practice receiving primary care through resident-faculty teams versus full-time faculty alone. METHODS: A 15-item patient satisfaction questionnaire (PSQ) was administered to patients at a community-based academic practice via telephone by 2 medical students blinded to the purpose of the study. The questions were derived from 4 previously validated instruments and presented in a 5-point Likert scale format. They addressed the accessibility and convenience of care as well as the physicians' thoroughness and competence, and communication and interpersonal skills. Two full-time faculty preceptors worked with six residents to form management teams consisting of one resident and one faculty preceptor per patient (cases). Randomly selected comparison patients (controls) were matched by primary faculty physician, age, and gender. RESULTS: 66 cases and controls were interviewed over an 18-month period. De-mographic characteristics of the two groups were similar: 74% in capitated managed care plans, 37% women, 34% White, 28% Latino, 25% Asian, and 5% African American. There were no significant differences in mean scores between patients seen by resident-faculty teams vs. faculty alone for all areas evaluated (see Table) .

CONCLUSION: Our results imply that patients are as satisfied receiving primary medical care through resident-faculty teams as they are from faculty alone. They also suggest that managed care organizations could allow residents, under faculty supervision, to be primary-care clinicians for their members without sacrificing patient satisfaction.

Pursley, JF Wilson, CH Griffith, DS Kwolek, Division of General Internal Medicine, University of Kentucky, Lexington, KY PURPOSE: The American Board of Internal Medicine has stated that women's health and gender-based medicine should be included in internal medicine residency curricula. METHODS: Two instruments were used in order to evaluate perceived curricular adequacies and core knowledge in women's health. We designed and administered to 53 internal medicine residents a 17-item questionnaire using a 5 point Likert scale. This survey rated the extent to which medical residents felt that their residency training addressed gender differences in subject matter such as presentation of coronary artery disease, cardiac risk factors, and signs of physical abuse. Also administered was a short answer questionnaire to evaluate actual knowledge of gender differences in risks, presentation, and prognosis of coronary artery disease; and factors related to the detection, triage and treatment of domestic violence. The results of the Likert scale and the short answer questionnaire were then compared using chi-square analysis. RESULTS: Only 5 of the 53 residents felt there were no inadequacies in women's health curriculum. Ninety-four percent and 96%, respectively, of residents felt that gender differences in presentation of coronary artery disease and cardiac risk factors were covered adequately. Seventy-four percent of residents rated their curriculum adequate in the area of signs of physical abuse. A marginal association existed with gender (p Ͻ 0.09), in that females related more inadequacies in this aspect of the curriculum (physical abuse) than males. Of those residents who had rated the curriculum adequate in regards to cardiac disease, 26% were unable to describe two gender differences in risk factors, presentation, and prognosis of coronary disease. Of those residents who rated adequacies in the signs of physical abuse curriculum, 26% were unable to describe even two factors important in the area of detection, triage and treatment of domestic violence cases. However, 87% of residents did know whether or not to report domestic violence to authorities. CONCLUSION: Residents' perceptions of adequacies in internal medicine women's health curriculum include cardiac disease (p Ͻ 0.05), and less so signs of physical abuse. However, there exists a gap to be filled between perceptions of curriculum adequacies in these areas and actual knowledge in these domains. PURPOSE: Successful mentoring has been shown to promote career satisfaction in non-medical fields, but occurs less frequently among women and minority populations. Our aim was to identify factors associated with having a mentor and being satisfied with mentorship in medicine. METHODS: We distributed 8,093 questionnaires (response rate 47%) to full-time assistant professors, instructors, fellows and house officers of Harvard Medical School (HMS). We collected data on demographic factors (sex, ethnicity, academic rank and commitment to a career in academic medicine) and characteristics of the mentoring relationship (personal communication, professional development, academic guidance, research and skill development). We used logistic regression to identify factors significantly associated with having a current mentor and being satisfied with the relationship. RESULTS: Of 3913 respondents, 38% were women, 72% caucasian, 16% Asian American, and 5% were underrepresented minorities (URM). Most respondents (66%) reported a current or past mentor; 55% of respondents identified a current mentor. Multivariable analysis showed factors associated with having a current mentor were being female [ Scoring scale: 1 ϭ excellent; 2 ϭ very good; 3 ϭ good; 4 ϭ fair; 5 ϭ poor with current mentoring [3.0 (2.5,3.5) ]. Satisfaction was independent of sex or URM status. Characteristics of mentoring which were associated with satisfaction included keeping in touch regarding progress [3.8 (2.8, 5.1) ]; advice on professional decisions [2.9 (2.0,4.2) ], research [2.7 (2.0,3.8) ] and career plans [1.6 (1.0,2.7)]; not abusing power [2.0 (1.5,2.6) ]; helping with funding [2.0 (1.4,2.7) ]; and taking a respectful attitude toward work [1.8 (1.3,2.5) ]. For women, additional factors associated with satisfaction were: accounting for cultural issues [2.5 (1.2,5.6) ], discussing pitfalls in academic growth [2.3 (1.1,4.9) ] and developing communication skills [2.1 (1.1,3.9) ]. No additional factors were identified among URM. CONCLUSION: Most junior faculty and trainees have had a mentoring relationship while at HMS, and 63% report satisfaction with mentorship. Women and URM had equal levels of satisfaction as their male and majority counterparts. Fellows were most likely to be satisfied with mentoring. Keeping in close touch with mentors and receiving useful advice were key attributes of satisfaction. Factors identified in this study may help guide the development of mentorship programs. PURPOSE: Peers may provide valuable feedback to medical students learning interviewing skills. However, even in non-grade situations, students may be hesitant to verbally critique their colleagues. We assessed the use of a written anonymous peer-and self-evaluation of videotaped interviews during a first-year interviewing course. METHODS: Videos of student's interviews with standardized patients were viewed by groups of 5 to 10 students along with faculty preceptors. Students were asked to evaluate their peers and themselves in the areas of knowledge of structure, techniques, and style of the interview using a 15 point scale as well as providing written comments. Faculty completed an identical form. Written comments were classified as either positive or negative. The ratios of positive to negative comments were also assessed. RESULTS: Because there was little variation in individual responses across the three categories (reliability: self 0.86, faculty 0.79, peer 0.91), these results were averaged and the composite scores analyzed. The written comments were quantified as either positive or negative. * Statistically different between the 3 groups. CONCLUSION: Peers generally rated their colleagues higher than the students did themselves or did the faculty, and provided very few negative comments, even in a non-graded situation. Peers did give their colleagues a "feedback sandwich" of 4 positive to 1 negative comments, which is what is often suggested as appropriate in the literature. Students were more critical on their own performances and were more likely to give themselves negative rather than positive comments. Peer evaluations using an open-ended format for feedback may be more beneficial than quantitative "grades." Students' self-perceived areas of weakness may be useful for instructors to develop further educational interventions.

MD Schwartz, AL Kalet, K Mukohara, Div. of Primary Care, NYU Medical Center, New York, NY PURPOSE: Can clinical epidemiology training turn Primary Care (PC) residents off to EBM as they learn the limits of published evidence, turning them into clever nihilists? METHODS: We surveyed 2nd year PC residents (PGY2) before and after a 60-hour, 6-week course in EBM using an instrument designed to measure "clever nihilism," self-reported critical appraisal (% of maximum score on 13 items) and electronic searching skills (% of maximum score on 17 items). The 6-item, 4-point nihilism scale, confirmed by factor analysis (with Cronbach alpha ϭ 0.7), addressed distrust of research methods and of applicability of published studies, and difficulties of practicing EBM. RESULTS: Over the last 2 years, 14 residents and 38 PC faculty were surveyed. Compared with the pre-course survey, after the course residents had a decrease in nihilism while improving their self-reported critical appraisal and electronic searching skills. Compared with PC faculty, residents after the course had higher scores for critical appraisal (85% vs. 71%, p ϭ 0.002) and electronic searching skills (83% vs. 68%, p ϭ 0.001), but had similar levels of nihilism. CONCLUSION: PGY2 PC residents completing an intensive course in clinical epidemiology improved their perceived critical appraisal and electronic searching skills without becoming more nihilistic regarding EBM. In fact, nihilism decreased after the course. It remains possible, however, that less intensive interventions could breed clever nihilism among our trainees and long-term impact on use of evidence in practice is unknown.

COST AWARENESS AMONG MEDICAL RESIDENTS. SI Shaikh, PH Mehta, NB Patel, Internal Medicine, New York Methodist Hospital, Brooklyn, NY PURPOSE: Managed care has revolutionized the practice of medicine. Learning cost-effective medicine without compromising the quality of care is indispensable for successful medical practice especially in contemporory managed care era. In teaching hospitals, residents often order many laboratory and radiological tests, prescribe medications, decide length of hospital stay and frequency of follow-ups in clinics, which requires rational, scientific and managed approach for patient care. So we decided to survey cost-awareness of this modalities among residents in our teaching hospital. METHODS: We conducted a cross-sectional survey of sixty postgraduate (PGY-1, PGY-2 and PGY-3) medical residents at our medical clinic. They were given a multiple choice questionnaire listing cost of twenty four different laboratory and radiological tests, cost of hospital beds, cost of clinic visits and cost of common medications used. The data were analyzed statisticaly and used to find out their projection and comparing intergroup variability. RESULTS: Correct costs for any test were chosen only by 1.6% to 35.5% of respondents for twenty four given items. Most of the cost were underestimated except cost for regular bed in hospital per day was overestimated by 12% of its correct cost and cost for aspirin by 600%. Less than 5% of respondents knew the correct cost for common tests like echocardiography (1.6% of residents) and cost of first clinical visit (3.4%) . Cost of items like urine analysis, echocardiography and first clinic visit were underestimated by Ͼ90%, Ͼ75% and Ͼ90% of responders respectively. Highest rate of correct responses was for cost of regular bed (35.5% of responders), chest x-ray (35.5%), MRI of head with contrast(33.8%) and cost of ICU bed per day in hospital (30.5%). There was no statistically significant difference or variation for correct responses between PGY-1 (least experience), PGY-2 and PGY-3(most experience) residents. CONCLUSION: This results bring out several serious issues about cost awareness. First, costs estimated by most of the residents in this study were highly inaccurate as compared with the actual costs indicating great lack of knowledge. Second, there was the consistent tendency among residents to underestimate the costs, with few exceptions. Finally, noteworthy conclusion was the absence improvement in cost awareness at higher levels of training (PGY3 or PGY2) where residents are preparing to enter the world of practice. Degree of illiteracy was equivocal but not equal in all areas tested, there was variable and stastically significant difference among extent of correct response among given group. While we do not advocate a financial basis for clinical decisions, it is clear that major educational efforts to rectify this lack of awareness should be seriously considered. To determine how many internal medicine residency training programs have hospitalists on staff, describe the role of hospitalists in teaching activities, and define attitudes regarding hospitalists. METHODS: All program directors of accredited internal medicine training programs were surveyed by mail starting in April 1999 about the presence of hospitalists, attitudes about hospitalists, and participation in teaching activities by hospitalists. The response rate was 57.6%. RESULTS: There was no apparent response bias. Overall, 50.5% of the programs employed hospitalists. General attitudes about hospitalists were positive. For example, the majority agreed that hospitalists are more familiar with practical aspects of inpatient care (58.4%), patients are satisfied with the inpatient care that they receive from hospitalists (54.0%), and they expect the use of hospitalists to increase over the next few years (85.2%). Most disagreed that hospitalists needed more training beyond IM residency (53.1%). Most also thought that use of hospitalists disrupted continuity of patient care (54.1%). Attitudes were significantly more favorable for program directors that had hospitalists. For the 109 programs with hospitalists, nearly all reported that hospitalists participated in teaching of medical students (80.2%) and residents (84.5%). Other educational activities in which they participated include attending rounds (74.7%); management conferences (53.5%); journal club (48.5%); resident report (58.6%); and curriculum development (55.6%). The majority agreed that hospitalists are viewed as good educators (78.5%), have led to improved housestaff supervision (52.2%), are more accessible to housestaff (62.2%), and are respected (72.3%). They are less certain that hospitalists had an impact on housestaff's consideration of length of stay (37.8%) or costs of tests and procedures (37.4%), or will impact on their future ordering behaviors (24.7%). CONCLUSION: Half of the medicine programs have hospitalists and nearly all participate in teaching. Attitudes regarding hospitalists are generally favorable though there is not consensus regarding their impact on residents' future behaviors. Given trainees' substantial exposure to hospitalists, there may be a role for faculty development targeted at improved hospitalist teaching and feedback methods. Future studies should examine what, and how, students are taught by hospitalists and whether this teaching differs from that provided by other attendings. Studies should also look in-depth at hospitalists' relative contribution to residents' education and the impact on their behaviors. OSCE Format: Sixteen 6-minute stations that tested students' abilities in historytaking, physical examination, and differential diagnosis. Each station included 2 minutes for faculty-to-student feedback on performance. Participants: 130 second-year medical students and 156 faculty members Measurements: 8 Likert-type scaled items for students and 6 items for faculty members, assessing the OSCE overall, components of the OSCE, and the relevance of the OSCE to the year-long course. RESULTS: All students and faculty members completed and returned the survey forms; responses were missing for 3% of items. Ninety percent of students agreed that the OSCE was a fair evaluation of the skills that were taught in the course, and 98% agreed that the OSCE helped them to identify skills that needed improvement. Ninety-nine percent of students agreed that the feedback they received from faculty examiners during the OSCE was helpful. Overall, 97% of students agreed that the OSCE was a worthwhile experience, but only 9% thought the OSCE should contribute to their grade in the course. All faculty members agreed that their participation in the OSCE was worthwhile; 70% strongly agreed and 30% agreed. Ninety-one percent of faculty members agreed that the station they supervised was a fair evaluation of students' skills and 56% agreed that their stations were difficult for students. Fifty-seven percent agreed that the OSCE should contribute to students' grade in the course. CONCLUSION: Faculty members and second-year medical students perceived a feedback-laden OSCE to be a valuable component of a year-long physical diagnosis course. However, most students were resistant to having the OSCE contribute to the grading process of the course.

Western Reserve University at MetroHealth Medical Center, Cleveland, OH; University of Nebraska, Omaha, NE BACKGROUND AND OBJECTIVES: Proficiency in procedural skills is increasingly important in primary care, yet little is known about how best to teach these skills. We examined the comparative effect of 3 pedagogical approaches on new interns': 1) performance rates of ABIM-required procedures over their first 10 months; and 2) changes in their scores on an objective test of knowledge. DESIGN, SETTING, SUBJECTS: Controlled (Firm System) trial among all interns in 3 parallel groups at an urban teaching hospital, from 7/98-6/99. In prior work, we have documented equivalent distributions of patients and conditions across firms. INTERVENTIONS: In their first 6 months, CDROM interns (N ϭ 8) completed a CD-ROM text and video guide on 7 procedures: arthrocentesis (Arth), paracentesis (Par), thoracentesis (Thor), lumbar puncture (LP), n-g intubation, and ABG and central line placement. CDROMϩLEC interns (N ϭ 8) also received lectures on each procedure. CONT interns (N ϭ 10) underwent traditional training and served as the control group. MAIN MEASURES: Interns completed an 89-item knowledge test during orientation and again in 6/99. Their performance of 4 procedures (Arth, Par, Thor, and LP) through April, 1999 was determined by administrative (i.e.; lab) data and chart abstraction (to reliably link the MD to the procedure). Cross-group differences in performance rates were adjusted for in-patient months of opportunity to perform these hospital-focused procedures. We also examined cross-group changes (post-pre) in test scores. MAIN FINDINGS: At baseline, interns across the 3 groups were similar in demographics, previous training and procedural experience, track within the residency, and knowledge. All interns completed both the pre-and post-tests. All groups improved in knowledge from pre-to post-test (mean % correct: pre ϭ 58.0%, post ϭ 65.6%; p Ͻ 0.001), although there were no cross-group differences in improvement (p ϭ 0.38). Of 396 procedures performed during the 10 months, interns (vs. others) performed 72, over 140.5 intern in-patient months (0.51 procedures/MDmonth). Intensity of intervention was associated with performance of procedures: rates were 0.61, 0.60, and 0.44 in the CDROMϩLEC, CDROM, and CONT groups, respectively. The CD-ROM interns performed procedures at a higher rate than the control group (p ϭ 0.04), although the CDROMϩLEC interns did not perform more than the CDROM interns (p ϭ 0.84). CONCLUSIONS: The CD-ROM guide is a promising, standardized approach for teaching procedural skills to medical interns. Further investigations should examine timing of the teaching, longer term effects, the incremental value of other teaching, and the impact of its use on appropriateness and outcomes of procedures. To determine whether a structured program of bedside teaching can improve the procedural and perceptual skills of medical residents in performing the cardiovascular examination. METHODS: A firm-based controlled trial was performed comparing three groups. Participants consisted of residents rotating on the inpatient medicine services over a two-month period. The two intervention groups received bedside teaching sessions (three 2-hour sessions over 4 weeks) with identical content but different teaching styles: (1) a traditional bedside teaching method, "demonstration and practice" (n ϭ 26); (2) an innovative method, "collaborative discovery" (n ϭ 24); and (3) a control group, which received usual teaching (n ϭ 25). Outcome measures were scores on procedural and perceptual components of an objective structured clinical examination (OSCE) using two patients. Participants were tested before and after the intervention. RESULTS: There were no significant differences between groups at baseline. For procedural skills, adjusted mean differences on a scale of 35 were: collaborative discovery vs. control, ϩ3.3 points (95% CI, 0.7 to 5.9; p ϭ 0.01) and demonstration and practice vs. control, ϩ3.9 points (95% CI, 1.3 to 6.4; p ϭ 0.004). For perceptual skills, adjusted mean differences on a scale of 65 were: collaborative discovery vs. control, ϩ4.2 points (95% CI, 0.6 to 7.8; p ϭ 0.03) and demonstration and practice vs. control, ϩ1.2 points (95% CI, Ϫ2.3 to ϩ4.7; p ϭ 0.5). CONCLUSION: Both educational interventions improved procedural examination skills for the cardiovascular examination. One intervention, collaborative discovery, also improved perceptual skills. Design: Prospective randomized trial of students to an identical 90 minute casebased lecture on screening in primary care "online" (audio of the class instructor with power point slides) or "live" from the same instructor in a conference room.

Measures: Survey data of baseline items including career choice, computer skills, access to computers, and attitudes about the use of computers for education were obtained. Improvement in knowledge about and application of screening principles measured by a pre-and a post-lecture exam. Both exams contained a 12point scale of different multiple-choice and discussion (clinical case scenarios) questions that have been used in the rotation for two years. The post-lecture exam was taken on average 2 weeks after the lecture. A priori, a 25% improvement in pre-to post-exam scores was considered a "clinically significant" improvement in knowledge. Survey data was used to monitor students' time to complete the lecture, satisfaction, and narrative comments about the experience. RESULTS: Sixty students (30 in each group, mean age 26, 39% female, 98% response rate) completed the study. There were no differences between the groups on all baseline items and on the pre-lecture exam (6.5 on the 12-point scale in both groups). Forty-four percent of the online group and 54% of the live group showed a 25% improvement on the post-lecture exam (p ϭ 0.4 for the betweengroup improvements). Students in the online group saved 53 minutes in completing the lecture (less driving time to access the lecture and less time listening to the lecture). Seventy-eight percent of the online students verses 93% of the live students were satisfied or very satisfied with their experience (p ϭ 0.1). Satisfaction was independent of baseline student computer skills and attitudes. Many students offered creative suggestions and some offered a few warnings about integrating online teaching into primary care clerkships. CONCLUSION: This study demonstrates that senior student clerks improved their knowledge about screening equally by an online and a live lecture of similar material. Students gave various reasons for preferring one format to the other, but overall, students were satisfied with either venue, even considering that the traditional live lecture required almost an additional hour to complete. Further studies of the effectiveness and efficacy of the online lecture will elucidate its role in clinical clerkships. Educators have reported that a faculty development course increased their ability to effectively address unprofessional learner behavior (ULB). However, instruments to objectively evaluate faculty response to ULB have not been described. METHODS: We developed a 22 item instrument of behaviors that educators potentially employ in addressing ULB. 4 standardized students (SS ϭ 4th year medical students) & 7 clinical observers (CO ϭ 2MD, 2RN, 3MSW) trained Ͼ10 hours to recognize, categorize & record these behaviors as present or absent. SS trained 6 hrs further to portray consistently provocative, unprofessional behaviors. 20 live objective structured teaching evaluations (OSTE), in which volunteer educators interacted with SS who behaved unprofessionally, were used to assess inter-rater reliability of the CO & SS, using % agreement with the majority response & kappa scores. 10 prerecorded OSTEs were used to assess intra-rater reliability at 0 & 1 month to determine the % answers scored consistently per rater. Questionnaire items were classified as low [1-Ͻ2] , moderate [2-Ͻ4] , or high [4] [5] inference behaviors by the developers of the ULB faculty development workshop. RESULTS: In each 5 minute interval of the 20 live OSTEs, raters observed 6 Ϯ 2 educator behaviors in reaction to ULB. Raw agreement was high among all 11 raters (86 Ϯ 7%), while kappas varied with the inference level of the behavior. For instance, 5 of 6 low inference behaviors had kappas Ͼ.4; specifically, clarification of words or actions [.66 ], acknowledging learner's emotions [.63 ], interrupting learner [.53 ], repeating disrespectful words [.46 ], rescheduling discussion time [.43 ]. Only 1 of 12 moderate inference behaviors (assessing learner's motivating behavior [0.51]) & none of 3 high inference behaviors had kappas Ͼ.4. Agreement rates within CO & SS were similar. Observed frequency & inter-rater kappa were not statistically associated (r ϭ .16, p ϭ .50). Based on 10 pre-recorded OSTEs, 7 CO & 2SS rated each behavior identically 86 Ϯ 0.1% of the time at 1 month. CONCLUSION: Accurate feedback on faculty interactions in addressing ULB is essential for improved faculty development. Raters are able to rate low inference faculty behaviors more reliably than higher inference behaviors. Overall test-retest agreement using our instrument is high. Traditionally, educators have felt that faculty behaviors in response to ULB were difficult to observe and categorize. Our preliminary evidence suggests that, with extensive training and inclusion of items worded as low inference behaviors, educators' strategies for handling unprofessional learner behavior can be reliabliably evaluated. PURPOSE: To identify characteristics of primary care research fellows and their training that are associated with productivity in publication early in their careers. METHODS: Retrospective cohort study of 1988-97 graduates of 25 National Research Service Award primary care research fellowships funded by the Health Resources and Services Administration. A 40-minute, self-administered survey was sent to participants to assess the characteristics of their training program, current position, and publications (defined as original research reports, editorials, commentaries, and reviews) since fellowship. A current curriculum vitae (CV) was requested to corroborate self-report of publications. Standard bivariate analyses and multiple logistic regression models identified predictors of publishing an average of у1 paper/year (a common standard for academic promotion) since the end of fellowship training. RESULTS: 146 ⁄ 215 program participants (67%) completed the survey. Of them, 111 (76%) returned a CV. Respondents were 39 Ϯ 5 years of age; 51% were male and 84% were white. Clinical disciplines were GIM (35%), family medicine (24%), general pediatrics (31%) and other (10%). Respondents had completed fellowship 4 Ϯ 2 years prior to the survey. They reported publishing a total of 3.2 Ϯ 4.5 first-authored papers and 2.9 Ϯ 4.7 co-authored papers since fellowship; 45 (31%) published an average of у1 paper/year. The correlation between self report and CV was r ϭ 0.91. Predictors of publishing at this rate are shown in the Table. The Cindex for the multivariate model was 0.81. Characteristics unrelated to publication productivity included: age, race, years since training, an advanced degree (e.g., MPH), research training model (apprenticeship vs. independent project), and writing a grant during fellowship.

CONCLUSION: The time allocated to research and the presence of a significant mentor during fellowship were associated with publication productivity early in the career of these primary care fellows. These associations may be due to structural characteristics of the training programs or to the choices made by fellows during and after their training. In either case, evaluation of fellowship outcomes may assist in the design of more effective training programs. PURPOSE: The involvement of the pharmaceutical industry in residency training programs has engendered much debate, yet little is known about the perceptions and behaviors of housestaff. We studied the attitudes and practices of internal medicine residents toward gifts from the pharmaceutical industry. METHODS: Confidential surveys were mailed to all first-and second-year medicine residents at a large university-affiliated residency program. Attitudes about nine types of industry promotions (e.g., inexpensive trinkets, dinner lectures) were scored from 1 to 4 on a Likert scale from "very inappropriate" to "very appropriate." Promotions were then grouped by cost and educational value, and summary scores calculated as the mean of the numerical responses within each category. Residents also reported their practices, perceptions of influence, and attitudes about institutional policies.

RESULTS: Ninety percent of residents (105 of 117) completed surveys. We observed a wide variability of attitudes about different types of gifts, ranging from 92% who considered it appropriate to accept a pocket antibiotic guide to 15% who stated it was appropriate to receive luggage. More than half of residents made a distinction (a difference between category summary scores у1 standard deviation of the overall sample mean) between appropriate and inappropriate promotions based on their cost. On the contrary, only 15% distinguished the appropriateness of promotions by their educational value. Behaviors were often incongruent with attitudes: every resident who considered free pens and conference lunches inappropriate (n ϭ 18 and 13, respectively) had accepted these gifts, and half of those who felt industry-sponsored recreational events were inappropriate (n ϭ 36) had either participated in such events or intended to do so. Most respondents (61%) stated that their prescribing practices were not affected by industry promotions or contacts, whereas only 16% believed other physicians were free of influence. Nonetheless, more than two-thirds of residents agreed that it is appropriate for a medical institution to have rules on industry interactions with residents and faculty. CONCLUSION: Residents distinguish the appropriateness of industry promotions primarily on their financial, rather than educational, value. While viewing themselves as less susceptible to influence than other physicians, more than two thirds of housestaff support the existence of policies that govern their interactions with pharmaceutical industry representatives. Residency program leaders should address these attitudes in creating educational and policy initiatives about housestaff-industry interactions. Are there differences between medical schools in the degree to which their graduates will behave professionally once in practice? If so, are there differences due to selection (admission variables) or due to socialization (medical school variables)? METHODS: The National Practitioner Data Bank (NPDB) is a federally-mandated collection of all malpractice payments and certain adverse actions against licensed physicians in the US since September, 1990 (156,369 events) . For the purpose of this study, only the 1200 adverse actions for unprofessional conduct were included. Malpractice payments and all other adverse actions (e.g., substance abuse, felony) were excluded. Each action was linked with the medical school from which the physician graduated. Medical school characteristics were obtained from Association of American Medical Colleges resources. We used Poisson regression with estimation of an extra Poisson component of variation to model the variance in the school rates of reporting. An offset representing the number of graduates in practice from each medical school, during the 10 years the NPDB has been in existence, was included in the model. RESULTS: The rate of unprofessional conduct identified in graduates is approximately two actions per thousand graduates. There is one outlier school, with a rate of nearly one action per fifty graduates (excluded in subsequent analyses). The mean number of reports per medical school is 7.9. The number of reports varies significantly across schools, with a standard deviation of 1.3 counts (95% CI [1.2, 1.5] ). The percent of instate admissions per instate applicant explains about 12% of the variance at the school level. Stated another way, a 10% increase in the instate acceptance rate is associated with a 15% increase in the future rate of actions for unprofessional conduct. Other measures (public/private, US News rank, overall selectivity) did not contribute significantly to the model. CONCLUSION: Medical schools graduate students with varying degrees of future professional conduct. Some schools have no graduates with reported adverse actions for unprofessional conduct, while others have as many as one action per fifty graduates. Some of this effect appears to be related to the admissions process, but much of the variability remains unexplained. Further study is needed to uncover the institutional variables that predict future unprofessional conduct. 3.5) or both. Demographic information about the residency programs was also collected and analyzed. The chi-quare test was used for analysis of nominal variables. RESULTS: Of the 397 programs surveyed, 312 responded, yielding a response rate of 78.6%. Sixty-eight percent of programs included home care instruction in their residency curriculum, consisting of house call(s), lecture(s), or both. Half of the programs provided opportunities for residents to make house calls with physicians or nurses, but only 25% of all residency programs had a mandatory house call training experience. Less than half (45.5%) of programs included lectures about home care in their curriculum. Of those that did offer lectures, the mean duration was 3.2 hours/year. Residency programs with a primary care track were more likely to include house calls and/or lectures in their curriculum (X 2 ϭ 7.44; p ϭ 0.006). CONCLUSION: Nationwide, internal medicine residents receive limited training in home care. In fact, one in three internal medicine residency programs offer no instruction whatsoever in home care. Absence of this curricular piece may leave future internists inadequately prepared to coordinate and supervise their patients' home health care and may leave them out of step with the changing needs of their patients. METHODS: A review was performed of all morning reports over three years at a tertiary teaching hospital. Morning report topics were categorized by clinical condition, organ and system competency, and inpatient or outpatient venue. The educational content of morning report was analyzed and compared to the curriculum guidelines created by the FCIM. RESULTS: Of 583 consecutive morning reports at the end of the three-year study period, 331 (57%) focused on inpatient cases and 252 (43%) focused on outpatient cases. The FCIM Curriculum Task Force organ and system competencies discussed most often were infectious disease (14%) and cardiovascular illness (14%). The organ and system competencies most likely to be discussed in the context of inpatients were infectious disease, cardiovascular illness, and pulmonary medicine. The competencies most likely to be discussed in the context of outpatients were rheumatology, general internal medicine, and endocrinology and metabolism. With the exception of general internal medicine topics, between 60-86% of all organ and system competency focus areas were the subjects of a morning report by the end of the study period. CONCLUSION: Depending on format and educational content, morning report can be a valuable educational conference and a key component of an internal medicine residencys' overall curriculum. The FCIM report can serve as a tool in creating, implementing, and assessing a program's curriculum.

G Talente, D Rudy, D Barnett, C Griffith, J Wilson, Division of General Internal Medicine, University Of Kentucky, Lexington, KY BACKGROUND: As the new millenium begins, there has been discussion about the competencies new physicians will need to be successful. Another question is how do we integrate new competencies with the traditional physician skills and qualities being taught in medical school? Are we accomplishing this now? We looked at the competencies that first year students reported as important following an externship with a primary care physician early in the first year and again upon completion of the year in order to evaluate the effect of their early training on their views. METHODS: Ninety seven first year students were asked to answer the question, "The competencies needed to practice medicine in the 21 st century are?" The question was first presented to them following a weeklong observation at a community practice in October of the first year. The question was repeated upon completion of the first year. At this time they were also asked if their views of important 21 st century competencies had changed. Responses were analyzed by content domain, correlated with demographic data and compared over time. RESULTS: Similar competencies were reported on initial and follow up questioning: knowledge 48% and 44%, compassion 47% and 50%, communication skills 46% and 48%, life long learning 20% and 13%, application of new technology 28% and 27%, computer skills 13% and 18%, business skills 18% and 7%, and understanding of insurance systems 11% and 7%. Twenty-one students believed their view of important 21 st century competencies changed during their first year of school. For these students, communication skill was recognized more frequently after completion of the first year (29% post extemship vs. 57% post first year, p Ͻ .018). Business skills were mentioned less frequently (43% vs. 10%, p ϭ .005).

There was no correlation between responses and demographic data except for a tendency for students from small towns to find business skills important. CONCLUSION: Early medical school training and clinical experiences appear to reinforce the traditional competencies of knowledge, compassion and interpersonal skills with much less effect on students' appreciation of less traditional skills they will need when they enter practice, and even a negative effect on their appreciation of business skills. The autopsy has been recognized as a valuable educational tool that helps residents understand the link between clinical findings and pathology. Although recommended for accreditation of teaching hospitals, the autopsy rate has fallen over time. We analyzed the impact of autopsies on resident clinical education. METHODS: The data were collected at a 520-bed private teaching hospital in Tokyo from 1993 through 1997. The performance of autopsies at this institution is actively promoted by teaching staff and administration. Participants included 89 1st post-graduate year (PGY) and 88 2nd PGY residents from 4 departments: internal medicine, surgery, obstetrics & gynecology, and pediatrics. The number of autopsies attributable to each resident and results of standardized resident performance evaluations by teaching staff in the same academic year were compared using linear regression analysis. Admission score (AS) is the result of the exam taken by medical students to qualify for residency positions. RESULTS: Average numbers of autopsies and deaths per resident were 3.6 and 8.2 for the 1st PGY and 5.5 and 12.5 for the 2nd PGY, respectively.

CONCLUSION: A negative association existed between the number of autopsies obtained by residents and performance evaluations. It is possible that a higher autopsy rate led to a higher discovery rate of missed diagnoses or incorrect premortem management decisions, thereby leading to lower performance ratings for residents. Although implausible, another interpretation of these data is that higher clinical competence was related to a lower autopsy rate. PURPOSE: Effective curriculum development requires extensive time and manpower utilization. The purpose of this study was to develop teaching strategies for our inpatient teaching service using time-efficient qualitative and quantitative group process techniques. METHODS: The residency review committee for internal medicine requires that 4.5 hours per week of inpatient rounding be dedicated to explicit teaching apart from patient management rounds. Fifteen internal medicine faculty who comprise 85% of the total inpatient rounding schedule responded to an electronic mail (email) inquiry: "What specific teaching strategies would you employ during these 4.5 hours of inpatient teaching time?" Using the Delphi technique the list of ideas were reported in a second email in which the faculty were asked to clarify their ideas and list the strengths and weaknesses. A list of 13 teaching strategies was developed from this iteration. The faculty were then asked to use the Nominal Group Technique to score the strategies by impact and feasibility using a 10 point scale (1-least, 10 most) . RESULTS: Thirteen teaching strategies were compared by impact (I), feasibility (F), and total score (S) of I and F. There is a significant difference between the teaching strategies in terms of their mean scores on I (p Ͻ .001), F (p Ͻ .001) and S (p Ͻ .001). Case based teaching strategies had the highest I F and S score. Five teaching strategies were significantly different from case based teaching using the Mean and Standard Deviations of the total score. Methods with the highest impact also had the highest feasibility with six strategies demonstrating a significant correlation using the Pearson Correlation Coefficient. Five strategies were distinguished by significantly higher total scores and correlation coefficients including: case based teaching, team teaching, primary care case discussions, ABIM board review questions, radiology rounds and feedback sessions. CONCLUSION The generalist disciplines of family medicine, internal medicine, and pediatrics have increasing numbers of academic fellowship positions and programs. We examined the association between fellowship training and academic career outcomes among primary care faculty. METHODS: In 1995, we conducted a cross-sectional study of 500 full-time academic primary care physicians from a national sample of 24 randomly selected US medical schools. The self-administered questionnaire included items regarding fellowship training. Primary outcomes were grant submissions and funding during the most recent 2 years, career refereed publications, rank (being a full or associate professor), and salary. Secondary outcomes were preparation for and commitment to research, self-perceived research and teaching skills, career aspirations, and job satisfaction. All results were adjusted for age, race, gender, years since first faculty appointment, and institution. RESULTS: Among study respondents, 234 physicians had completed a fellowship, while 266 had not. Fellowship-trained physicians devoted almost triple the amount of time at their current job to research (17% vs. 6%, p ϭ 0.0001), and in the past 2 years, were 4.8 times more likely to have submitted a grant (p ϭ 0.0001) and 4.2 times more likely to have had a grant funded (p ϭ 0.0001). They were also more likely to have had any refereed publications (OR ϭ 3.8, p ϭ 0.0001) and to have a higher rank (OR ϭ 1.9, p ϭ 0.02). Fellowship-trained physicians felt more prepared for research (p ϭ 0.0003), more committed to research (p ϭ 0.0001), had better self-reported research skills (p ϭ 0.0001), and were more likely to aspire to full professorship (p ϭ 0.02). The groups did not differ with respect to salary, self-perceived teaching skills, or job satisfaction. Among faculty with fellowship experience, those with at least 1 year of research training during fellowship were even more academically productive with grant submissions (OR ϭ 1.9, p ϭ 0.02), grant funding (OR ϭ 2.9, p ϭ 0.0003), and publications (OR ϭ 2.4, p ϭ 0.02). In addition, those with extensive research training were 2.3 times more likely to have a higher rank (p ϭ 0.03), although salary remained comparable. All secondary outcomes were significantly better among those with more research training, except for job satisfaction which was equivalent. CONCLUSION: In our study of academic primary care physicians, fellowshiptrained faculty submitted more grants, were funded more often, and published more than their non-fellowship trained peers. They also held higher academic ranks, although salaries were similar. Among faculty with fellowship experience, additional research training as a fellow was associated with higher productivity and rank.

PA Thomas, JH Shatzer, Dept. of Medicine, Johns Hopkins University School of Medicine, Baltimore, MD PURPOSE: Medical students in ambulatory settings are encouraged to develop focused history and physical examinations of common patient problems. A potential drawback of teaching the focused history and physical to early learners of medicine is that attention to checklists may result in the curtailment of diagnostic decision-making. The purpose of this study was to evaluate the impact of an evidence-based clinical practice guideline approach to low back pain on ambulatory clerks' diagnostic decisions. METHODS: All 108 students taking a required 4 week ambulatory clerkship in one academic year were included. 80 students assigned to one regional site by residential zip code received the intervention of a 2-hr. small group instruction, using the AHCPR Clinical Practice Guideline for Acute Low Back Pain. 28 students assigned to a second regional site did not receive the intervention and were used as controls. Evaluation consisted of student performance in a standardized patient examination at completion of the clerkship. A low back pain case, developed as a possible pathologic vertebral fracture, contained "red flags" in the duration and quality of the back pain, and the finding of vertebral point tenderness. Data-gathering was assessed during the 15 minute encounter by per cent checklist items done. The student post encounter note was scored for: documentation of 5 red flag items in history, documentation of spinal point tenderness in physical, extending the differential diagnosis beyond regional back pain, and inclusion of further diagnostic testing. Measurement of differences in interventions were analyzed by effect size calculations,d ϭ effect size. RESULTS: Students taught the clinical practice guideline scored higher on SP checlist performance than controls (mean .64 vs. 60, d ϭ 0.4, with no difference in interpersonal scoring between the two groups. Students taught the guideline recorded the same number of red flags in their progress notes, 3.0 vs. 2.7, and 83% reported the finding of point tenderness on examination. Students taught the guideline were less likely however to extend the differential diagnosis of the patients' problem beyond regional back pain (30% vs. 36%), and less likely to consider diagnostic testing at the time of the visit (32% vs. 39%). 34 ⁄ 108 students appropriately extended the differential diagnosis. These students had similar SP checklist scores as those who did not (.63 in both groups), but recorded more red flags into their post-encounter notes (3.2 vs. 2.7, d ϭ 0.4. CONCLUSION: The use of an evidence-based clinical practice guideline for low back pain did not improve student identification of a potential high risk low back pain patient. Students using abbreviated history and physical approaches need to be impressed with the significance of positive findings in their assessments.

COST. P Turner, K Galt, E Rich, B Houghton, Creighton Center for Practice Improvement and Outcomes Research; Department of Medicine, Creighton University, Omaha, NE PURPOSE: Direct-to-consumer and -physician pharmaceutical advertising has increased pressure on physicians to make pharmacotherapy treatment decisions upon marketing rather than the best cost-effective medical treatment option available. The purpose of this study was to assess the association between medical residents' attitudes and characteristics with the cost of prescribing. METHODS: We surveyed medical residents at Creighton University regarding their attitudes toward pharmaceutical information resource use, knowledge of professional position statements on physician interactions with industry, attitudes towards pharmaceutical representatives, and resident demographics (i.e., gender, ethnicity, USMLE scores, and residency year). Attitudes towards pharmaceutical information resources were assessed by residents' perceived accuracy and use of the Internet/Worldwide Web medical resources, clinical/drug reference manuals, electronic medical text CD-ROMs, peer reviewed research/review journals, and pharmaceutical sales representatives. Knowledge of professional position statements was evaluated with items developed from the literature (i.e., AMA, ACP). Attitudes towards psr's related to the usefulness/accuracy of information and residents' interactions with psr's. The cost of prescribing was measured with three clinical case scenarios of common ambulatory problems (supraspinatus tendinitis, mild hypertension, and acute cystitis). Academic internists and pharmacists with extensive experience in treating each problem served as consultants in developing each case. Four choices of drug treatment for each scenario were offered based on current recommendations of the literature, with widely varying costs but equal efficacy. A relative value index was used to calculate the cost of prescribing for each resident. Bivariate and multivariate (ANOVA with Tukey HSD post hoc comparisons) statistical analyses were use to evaluate the relationship between resident attitudes/characteristics and the cost of pharmacotherapy treatment. RESULTS: Fifty of the 59 residents completed the survey. Year in residency was significantly associated with prescribing costs (p ϭ .025), with 3rd year residents having lower prescribing costs than 1st year residents (p ϭ .019). Residents more confident in the accuracy of information in clinical/medical reference manuals were associated with higher prescribing costs (p ϭ .037). There were no other significant relations between residents' resource use, knowledge of professional position statements, or attitudes towards psr's associated with the cost of prescribing. CONCLUSION: The year in residency training and type of pharmaceutical information resource utilized may be associated with prescribing costs.

JA Underwood, KA McGarry, Department of Medicine and Psychiatry, Brown University School of Medicine, Providence, RI PURPOSE: To investigate the self-administration of psychotropic medications among resident physicians for the symptoms of depression, anxiety and sleep disorders. METHODS: An anonymous survey was distributed to all medical residents at the internal medicine residency program at Rhode Island Hospital. The survey had three separate sections, one each for depression, anxiety and sleep disorders. The residents were asked if they had self-prescribed for the above diagnoses with any prescription or non-prescription medications. The duration of treatment, the acquisition of the medications and whether they were under the care of a clinician was also asked. RESULTS: Seventy-nine (63%) of the 125 medical residents responded. Of the three disorders, residents self-treated for the symptoms of a sleep disorder most often. Twenty-five residents (32%) used at least one medication for the symptoms of a sleep disorder. Only one resident was prescribed a medication. The most commonly used medication was diphenhydramine, which was most frequently used during night float and emergency room rotations. The majority of the residents used a sleep agent for less than or equal to one month. Thirteen residents (16%) used at least one medication for the symptoms of depression. The most commonly used medications were selective serotonin reuptake inhibitors and St. John's Wort. Only two residents (2.5%) were under the care of a clinician. Three (23%) of those who self-treated for depression took a medication for greater than or equal to one year. The majority of the residents obtained the medications from a sample closet. Five residents (6%) used at least one medication for the symptoms of anxiety. Only one resident (1%) was under the care of a clinician. The most commonly used medications for anxiety were beta-blockers and alcohol, both self-prescribed. CONCLUSION: This study suggests that medical residents often self-prescribe for sleep disorders, depression and anxiety. The ethical implications of this practice warrant further consideration. The lack of clinician input is concerning considering the possible impairment of a residents' judgment while treating patients.

self-perceived skill and comfort in various aspects of end-of-life care. A self-administered questionnaire was distributed during orientation in 1998 and 1999 (n ϭ 115, response rate 96.7%). Measures included: 1) the role of the physician in 9 aspects of end of life care; 2) amount of formal education, observation of attending physicians and clinical experience with pain management, giving bad news, and family, prognosis and advance directive discussions; 3) comfort in giving bad news, managing pain, and having discussions about prognosis and advance directives; 4) skill in discussing prognosis and caring for dying patients; and 5) demographic and personal experience with death and dying. A large majority of interns rated pain management (96.5%) and giving bad news (93.9%) as very important to the physician's role. About three quarters rated helping patients with emotional concerns related to illness (79.1%), fears of death (75.7%) and maintaining hope (73.9%) as very important, and fewer thought responding to emotional needs of family members (62.6%) or patient concerns about family well-being (60.0%) was very important. Family bereavement (44.3%) and religious concerns of patients (29.6%) were least likely to be rated as very important. The importance of the types of physician roles was not associated with skill, comfort, educational experiences (formal education or observation), demographics or personal experiences with dying, with the exception of:

These findings indicate that physician involvement in emotional concerns of dying patients is not seen as very important by one quarter of young doctors, and that even fewer view attention to family issues as very important. Skill in caring for dying patients was associated with medical (giving bad news and pain management), but not with emotional, spiritual or family concerns that are viewed with great importance by patients. Role modeling (through observation) may be an important factor in how young physicians perceive the physician's role in end-of-life care. Further research is needed to better understand what factors determine a physicians' perception of his/her role in end-of-life care. To assess whether a case-based palliative care curriculum improved the opioid prescribing practices of medical residents, we performed a prospective review of the charts of medical residents' as well as a comparison group of neurology and rehabilitative medicine (neuro/rehab) residents. METHODS: The curriculum consisted of 10 one-hour case-based seminars that addressed attitudes, skills and clinical practice issues. It included two seminars on pain management that emphasized the side effects of meperidine, the need to prevent constipation and the additive effect of nonsteriodal anti-inflammatory drugs (NSAIDS). Consecutive billing records of patients who received an opioid during hospitalization on the medical (n ϭ 600) and neuro/rehab (n ϭ 300) services from eight month periods before (1/1/97-4/30/97) and after (1/1/99-4/ 30/99 ) the implementation of the curriculum on the medical service were reviewed. The data abstraction protocol was validated by standard chart review of a random subsample of 45 charts (concordance ϭ 89.9%). Three outcomes were measured: 1) percent opioid orders for Meperidine in non-gastrointestinal (non-GI) patients (excluded because of a longstanding belief that meperidine is the opioid of choice in biliary disease); 2) percent opioid orders accompanied by a bowel regimen; and 3) percent opioid orders accompanied by adjuvant nonsteroidal anti-inflammatory drugs. RESULTS: The results represent a preliminary analysis of 80.3% (n ϭ 723) of the total number (n ϭ 900) of charts calculated by Fisher's exact method (power of 80% to detect a 10% difference at ϭ ␣ ϭ .05, two-tailed). CONCLUSION: Although preliminary, this data suggests that a palliative care curriculum can change the opioid prescribing practices of medical residents, over and above secular trends among all house officers. Larger samples and further research are needed to develop a better understanding of these findings and to learn how palliative care education can improve patient care.

PURPOSE: To evaluate the implementation of an evidence-based medicine (EBM) curriculum for internal medicine housestaff in a large, urban, university-based residency program. METHODS: An eight-week EBM curriculum was developed, and implemented over one year with second-year residents (N ϭ 42) in their two four-week ambulatory care blocks. Topics covered in the curriculum included developing clinical questions from patient encounters, using the best quality information sources, searching MEDLINE and the Internet for information, and critical appraisal of various types of journal articles. Residents also gave presentations in the last week of the curriculum. The curriculum was evaluated with written pretests and posttests for each of the four week blocks, pretest and posttest MEDLINE and Internet searches based on clinical scenarios, as well as an independent MEDLINE search at the end of the curriculum. Also, a survey of EBM attitudes and abilities was administered to the participants at the end of the academic year, and responses were compared to a baseline survey that was completed at the beginning of the year. Third year residents who were one year ahead of the curriculum participants also took the survey, and served as a quasi-control group. RESULTS: For each of the four-week blocks, residents' mean written test scores increased significantly (pretest 49%, to posttest 86% for block one, 42% to 57% for block two, p Ͻ 0.001 for each), and residents' computer searching skills improved markedly. Residents felt the curriculum was useful for them, and were generally enthusiastic about participating. At the end of the year, participants reported increased use of research journals and the Internet, compared to use before the course and to the quasi-control group. Participants also had more confidence in an internist's ability to use the literature to answer patient care questions instead of relying on consultants, compared to baseline. They reported increased skills in assessing the relevance and validity of journal articles. There was no improvement in their skills in choosing which journal articles are important to read. Residents demonstrated only a slight improvement in their ability to detect design flaws in a simulated journal article. CONCLUSION: A block curriculum in EBM for internal medicine residents resulted in improved short-term knowledge, MEDLINE and Internet searching skills, and overall self-reported attitudes toward the use of research literature in patient care. Although residents reported improved skills in critical appraisal at the end of the year, there was less improvement in objective measures in this area. The curriculum content and design can potentially be adapted for other programs that are considering formally teaching EBM to their residents. PURPOSE: We sought to identify patterns of participation in a faculty development workshop (FDW) and to elicit attitudes that might explain why faculty chose to participate. METHODS: All 93 members of the department of medicine at a northeastern urban medical school were invited to attend a FDW. Demographic data were compiled for participants (P) and compared to data for those who did not participate (NP). Teaching reputation (TR), strong, average, or weak, was estimated by a formula using 3 years of survey data from medicine residents. A self administered questionnaire asked faculty to assess the relative importance of various work responsibilities and was used to compare P versus NP. Finally, a second survey was sent to NP asking them to rate various explanations for their decision. The FDW consisted of a 2 hour noon session held in a convenient location on 6 consecutive Wednesdays so that faculty could choose a preferred date. The workshop was led by a senior faculty member experienced in FD. RESULTS: Forty-eight (52%) of 93 faculty attended the workshop; all completed their survey. Of the 45 NP, 19 (42%) returned surveys. Demographic data showed no difference between NP who returned surveys and those who did not. P were more likely to be women (p Ͻ .05), more likely to be assistant professor or lower in rank (x2 ϭ 9.313, p Ͻ .01), and more likely to be generalists than sub-specialists (p Ͻ .01). There was no relationship to TR. There was no difference in how priorities were assigned to teaching, patient care, and administrative responsibilities, but NP ranked research a more important responsibility than P (p Ͻ .05). NP ranked "too busy" as the most important reason for their decision (mean ϭ 4.75, 5 ϭ extremely important, 1 ϭ not at all important), versus "inconvenient time" (3.81) , "not enough advance notice (2.81) , "unaware of opportunity" (2.67), "would not have improved my skills" (2.00) . CONCLUSION: This study suggests that faculty who choose to avail themselves of FDW opportunities are more likely to be women, generalists, of lower academic rank, and less involved in research. Limitations of this study include the imprecision in measuring teaching skills and teaching reputation, and the single institution sample source and single FDW/FDW leader offered. These observations, if confirmed, should be considered in both the design of future FDW programs and in efforts to analyze the impact of FDW on outcome measures such as ratings of teacher effectiveness or ability of FDW to modify learner behavior or enhance learner knowledge. PURPOSE: This study objectively evaluates a curriculum designed to improve the spoken case presentation. METHODS: In a prospective, blinded study, we evaluated the effects of a curriculum to improve presentation skills. Sixty-four third-year medical students at three different hospitals were videotaped presenting a standardized case before and after the medical clerkship. Thirty-two students received a four-week curriculum, while the other thirty-two received the standard ward instruction. Two independent evaluators blinded to the case and control assignments reviewed the before and after videotapes. Each tape was scored on a 1 to 30 scale for quality, time, and accuracy. Students and their respective residents and attendings were surveyed as to the students' communication, history taking, and physical examination skills. RESULTS: The average time of the spoken case presentation at the beginning of the clerkship was similar in both groups (intervention group: 7 min. 59 sec.; control group: 8 min. 10 sec.; p Ͼ 0.50). When compared to the control group, students who received the curriculum reduced their presentation time by 45 percent (intervention: 5 min. 12 sec.; control: 9 min. 29 sec.; p Ͻ 0.001). The quality of presentations improved in both groups, but was significant only in the intervention group (intervention quality score: 17.1 to 26.2 p Ͻ 0.001; control 20.3 to 22.7; p ϭ 0.06). Students taught the curriculum, but not those who received standard instruction, reported increased abilities in communication, history taking, and physical examination skill (performance composite index; intervention: 17.5 to 22.0 p ϭ 0.0001; control: 18.4 to 19.0; p ϭ 0.21). Attending physicians and residents reported improvement in both groups, although the greater improvement was seen in the students given the curriculum. The reviewersassessment's had excellent agreement (r ϭ 0.87). CONCLUSION: A curriculum based upon early diagnosis generation and the iterative thought process improves the efficiency and quality of presentations, and augments overall ward performance. PURPOSE: The College of Graduate Medical Education has stated that many women receive incomplete and poorly coordinated care for their routine and comprehensive health concerns, in part due to deficiencies in physician education. This study reports on a survey designed to determine the perceptions of medical students on the extent to which gender differences in medical topics are addressed by the curriculum. METHODS: The survey, a 17-item Likert scale with 4 ϭ excellent, 3 ϭ good, 2 ϭ adequate, 1 ϭ poor, and 0 ϭ not at all, was given to 180 first-, second-, and thirdyear students, with a response rate of 85%. Data were analyzed with multiple regression using the general linear model, with variables measuring the adequacy of the curricula as dependent variables and with the level of medical education and gender as independent variables. RESULTS: As the students' level of education increased, students rated the coverage of sex differences in given health issues more favorably (year 2 and 3 were significantly different from year 1, p Ͻ .0006, .0001, respectively). For example, year 1 students gave an average rating of 1.59 for "female and male differences in preventive health behavior" whereas year 2 and year 3 students gave ratings of 2.00 and 2.68, respectively (p Ͻ .0001). In addition, male students rated the coverage more favorably than female students, on average (p Ͻ .0048). For example, the average male rating for coverage of "female and male differences in non-reproductive physiology" was 2.28 whereas female students gave an average rating of 1.72 (p Ͻ .001). Likewise, men rated coverage of "signs of physical abuse" as 2.24 whereas women rated it as 1.55 (p Ͻ .0028). CONCLUSION: Although the average rating for coverage of health issues increased as the students' level of education increased, the overall average rating is less than adequate, and the coverage of no health issues is rated as good or excellent. Changes in the medical school curriculum at our institution are needed to better educate future physicians in the gender-based care of patients. PURPOSE: We explored the on-line availability of full-text review articles suitable for accompanying the SGIM-CDIM third year clerkship curriculum in ambulatory Internal Medicine. METHODS: We performed a Medline search of review articles (limited to English language, publication type-review and years-1998-99) for 7 syndrome and 3 disease SGIM-CDIM curriculum topics. Two faculty members determined citation relevance to third year clerkship students and practicing generalists based on the citation title, journal, and number of references. RESULTS: 336 relevant citations were identified from 175 journals (131 for medical students) with only 11 journals with Ͼ4 relevant citations. For the 3 disease curricular topics the number of relevant citations were 108 (diabetes), 40 (depression), 88 (COPD) compared with a mean number of citations of 13.9 (range 6-34) for the 7 syndrome curricular topics. 244 were judged relevant to ambulatory In-ternal Medicine students and 92 to practicing generalists but not to students. 25.0% (61/244) of citations relevant to medical students were syndrome-oriented compared with 42.4% (39/92) of citations relevant to practicing generalists. 28.9% (97/336) of all relevant citations were available in electronic, full-text version (free -5.1%, 17/336); through Ovid TM -20.2%, 68336%; or through MDConsult TM -6.0%, 20/336%. Only 29.5% (72244) of citations relevant to students were electronically available as full-text articles. CONCLUSION: Disease-oriented review articles are more numerous than syndrome-oriented review articles. Electronically available or not, the small number of syndrome-oriented review articles may make it difficult to avoid the use of textbooks for an ambulatory internal medicine clerkship. The contrast in availability of full-text reviews for clinical syndromes versus those for diseases likely reflects the specialization within academic medical centers. PURPOSE: Many medical journals are currently offering physicians the option to subscribe electronically, allowing the reader to access publications via the Internet. In addition to assessing the current status of electronic publications, a study was conducted to better understand the opinions and attitudes of physicians regarding electronic journals. METHODS: Data collection from books and websites was performed to determine which biomedical journals offer a full-text version electronically and the specific characteristics of this format. Also, a three-page questionnaire was delivered to all active physicians at the Johns Hopkins Bayview Medical Center that explored features of electronic journals that make them distinct from printed journals. RESULTS: Hundreds of biomedical journals offer full-text versions of their printed journal electronically. Rarely, access is offered free of charge. Most journals with an electronic format offer access either: a) as a bonus to those who subscribe to the print version, b) at an additional cost to those who subscribe to the print version, or c) as a stand-alone subscription. An increasing number of electronic journals offer the electronic format in a form that is identical to the printed version (as a down-loadable pdf file). Of the 314 physicians surveyed, 255 (81.2%) returned a completed questionnaire. The mean age of the respondents was 40.7 years, 65% were male, 20% were housestaff, and 63% were from the Department of Medicine. Twenty-six percent of respondents thought that electronic journals would lower the quality of the medical literature and 25% believed that the prestige of authorship would be lessened. Seventy to eighty percent of physicians responded that electronic journals would: i. decrease clutter in their offices and homes, ii. be more environmentally friendly than the current system, iii. make it easier to locate research reports that they had read, iv. offer the benefit of linkage to other related articles. Seventy-four percent of physicians were concerned about losing the convenience of being able to read their printed journal anywhere. In multivariate analyses, the independent variables that were associated with positive attitudes towards various aspects of electronic journals included: female gender, faculty member (vs. house officer), fewer publications, better computer skills, and more frequent internet use, [all p Ͻ 0.05]. CONCLUSION: The availability of full-text online versions of biomedical journals is steadily increasing. Physicians responded favorably to the many potential values and applications of electronic publications but are most concerned with the loss of convenience that printed journals offer.

Medicine, Johns Hopkins Bayview Medical Center, Baltimore, MD PURPOSE: Limited data are available on " problem resident", which is defined by the American Board of Internal Medicine as "someone who demonstrates a significant enough problem to require intervention." A questionnaire was developed to survey internal medicine program directors to gain an increased understanding of the issues related to "problem residents", including their identification, contribution of underlying causes, and management. METHODS: We recently mailed a confidential survey to all 406 accredited internal medicine residency programs in the United States, and recruitment is ongoing. RESULTS: Preliminary results from the 247 (61%) surveys already returned show that the prevalence, during academic year 1998 to 1999, of "problem residents" in internal medicine residency programs ranged from 0% to 39%, with the mean of 7% (SD ϭ 5.8). Only 6% of program directors reported an absence of "problem residents"" in their training program. Program directors stated that "problem residents" often (half of the time or more frequently) have insufficient medical knowledge (49%), poor clinical judgment (47%), and inefficient use of time (44%). Stress and depression are the most frequent underlying issues for "problem residents" (41% & 25%); cognitive dysfunction and substance abuse are felt to be infrequent. Sixty percent of program directors agree or strongly agree that it has been difficult to convince "problem residents" of their deficiencies because of lack of honest and accurate written evaluations. Furthermore, while 76% of program directors reported that attending physicians made verbal comments that helped them recognize "problem residents", only 41% included these comments in their written evaluations (p Ͻ 0.01). Chief residents and program directors most frequently discover the "problem residents" (83% and 77%); "problem residents" rarely come forward and identify themselves. In managing "problem residents", program direc-tors feel that more frequent feedback sessions (63%) and an assigned mentor for structured supervision (55%) are the two most helpful interventions. Measures such as regular meetings with the program director (94%) and timely evaluations from attending-physicians (85%) are helpful to prevent deterioration in the performance of the "problem residents". When asked about their comfort level in the identification of "problem residents", diagnosing underlying causes, and management of the issues involved, program directors had decreasing levels of confidence with each progressive step in the process (89%, 76% and 58%, respectively). CONCLUSION: The vast majority of internal medicine residency programs have "problem residents". The presenting characteristics and underlying issues are diverse, suggesting that an individualized approach is needed for each case. BACKGROUND: Medical providers are ideally situated to identify and intervene on behalf of women who are experiencing domestic violence. Although, awareness of the frequency and importance of domestic violence (DV) has been increasing in the community, few recent studies have examined how this increased awareness has impacted domestic violence screening by physicians. METHODS: Copies of all consults performed by 50 different third year internal medicine residents participating in an ambulatory and inpatient general medicine consult service at a large woman's hospital between 1995 and 1999 were obtained. 6-16 consults were randomly selected for each resident. These consults were reviewed for the presence of somatic complaints that are believed to be markers of potential DV. These consults were then reviewed for documentation of inquiries regarding whether the woman was or had been a victim of DV. RESULTS: In total 674 consults were reviewed. 92.7% met our criteria for being appropriate for domestic violence screening. 50% of these patients were pregnant. In only 9 of the 674 consults reviewed was the presence of absence of DV noted. In none of these cases was DV present. CONCLUSION: Third year residents on consultative services at our center are not documenting DV screening as part of their evaluation of women with somatic complaints that have been associated with DV. Determining why this is so and how best to address it should be a priority for individuals interested in improving the health of women.

A META-ANALYSIS OF EFFECTS OF GARLIC ON SERUM LIPID PROFILES. RT Ackermann, CD Mulrow, G Ramirez, University of Texas Health Science Center, San Antonio, TX PURPOSE: To assess whether oral garlic supplementation improves serum lipid profiles in human subjects.

METHODS: We searched 1800 English and non-English citations through December 1999 using 11 electronic databases (e.g., MEDLINE, EMBASE, the Cochrane Library, NAPRALERT), article bibliographies, manufacturers, and experts. Search terms included: aglio, allium sativum, garlic, garlic extract, garlic oil, knoblauch, kwai, kyolic, and 17 additional related chemical terms. Selection criteria were randomized trials of at least 4 weeks duration that measured lipids and compared garlic with placebo, no garlic, or another active agent. Two independent physicians abstracted data from trials. Lipid outcomes were pooled using standardized mean differences. RESULTS: Of 43 trials meeting selection criteria, 1 had clearly adequate randomization processes; 33 used double-blind designs; and 4 had Ͼ20% drop-outs. Trial interventions included standardized dehydrated garlic preparations (n ϭ 22), aged-garlic extract (n ϭ 4), garlic oils (n ϭ 6), raw garlic (n ϭ 1), various non-standardized dried preparations (n ϭ 8), and combination products (n ϭ 2). Pooled analyses of placebo controlled trials showed that garlic preparations significantly reduced total cholesterol by 9.9 mg/dl (CI 1.7 to 18.2) PURPOSE: Antibiotic prophylaxis for bacterial endocarditis (BE) is recommended by the American Heart Association for patients undergoing invasive dental procedures. The cost-effectiveness of this strategy is not clear. METHODS: We compare antibiotic prophylaxis versus no-prophylaxis for a hypothetical cohort of patients (age 42 years) with native heart valves undergoing an invasive dental procedure. A standard cost-effectiveness analysis from a societal perspective is conducted. Effectiveness is measured as quality-adjusted life years (QALYs) saved by the prophylaxis program. The model included direct and indirect medical care costs and cost of productivity losses. All future costs and benefits were discounted at 3%. Values for model parameters were extracted from published literature. Estimates for antibiotic efficacy were derived by conducting a meta-analysis of available data. All parameters were subjected to extensive sensitivity analysis. RESULTS: Under base-case assumptions, antibiotic prophylaxis (amoxicillin 2 gm po) for 10 million patients undergoing a dental procedure would prevent 33 cases of BE and 5 deaths, while 9 lives would be lost due to fatal antibiotic side effect. As a result there would be net loss of 4 lives but a net saving of 289 QALY due to cases of BE prevented. The incremental cost-effectiveness (CE) ratio for prophylaxis was $245,549/QALY. The analysis was most sensitive to changes in estimates of antibiotic efficacy. If antibiotic efficacy is raised from 30% (base-case) to 64% efficacy (best-case), the CE ratio is $20,323/QALY. The percentage of BE that is attributable to dental procedures also changed the results. If this percentage is raised from 11% (base-case) to 29% (best-case), the CE ratio is $50,418/QALY. Similarly, the population incidence of BE affected the results. If incidence is raised from 38 cases/million (base-case) to 99 cases/million (best-case), the CE ratio is $49,943/QALY. Prophylaxis with clarithromycin 0.5 gm po had similar results to amoxicillin, but ampicillin iv was not cost effective and led to a net loss of life & QALY due to higher incidence of fatal side effects. CONCLUSION: Antibiotic prophylaxis is not cost-effective as a public health policy. In our analysis if antibiotic efficacy is assumed to be 53% or higher, then prophylaxis can be considered as cost-effective, i.e., CE ratio less than $50,000/ QALY. Due to the lack of good human data on antibiotic efficacy we recommend that decision for prophylaxis be individualized. Patients that are highest risk for endocarditis, such as patients with prosthetic valves, cyanotic congenital heart defect, and prior endocarditis should be candidates for prophylaxis. Patient risk for antibiotic side effect should also be evaluated. Recruitment and retention of minority participants into clinical trials is an important component of research and often the rate-limiting step towards successful completion. Participants who are eligible for, and express interest in, enrolling into a trial will sometimes not return for their randomization visit. Factors associated with failure to return for randomization are largely unknown. We compared 287 eligible African American (AA) smokers who enrolled, but did not return for randomization (NR), to 500 AA smokers who returned and were randomized (R) to participate in a study comparing culturally sensitive educational materials to usual care materials for smoking cessation AAs. The 500 participants who returned for randomization received brief counseling and 8 weeks of transdermal nicotine patches free of charge. An analysis was conducted on variables potentially associated with not returning for randomization. As shown below, those not returning for randomization (n ϭ 287) were significantly younger, were more likely to be male, smoked fewer cigarettes, were less likely to have been told to quit in the past year, and less likely on planning to quit in the next 30 days. In addition, they were less educated, had greater heavy drinking, were more likely to have been proactively recruited (participant was approached), and were less likely to have their own transportation.

Potential participants who were eligible for randomization, but did not return, differed in a number of ways from those who did return for randomization. Better understanding of these factors may allow researchers to target recruitment efforts, potentially resulting in enhanced accrual and retention, and therefore, generalizability. with moderate to severe knee and/or hip pain of more than six months duration were surveyed. To assess patients' perceptions of the role of prayer in their own management of arthritis, we used a component of Efficacy of Specific Treatments Scale (Bill-Harvey) "Do you consider prayer to be very helpful; somewhat helpful; or not at all helpful "and whether they have actually tried prayer as a form of therapy for arthritis. An assessment of religion included religious denomination as well as 4 items to assess religiosity (i.e., frequency of praying, religious service attendance, importance of religion, and self-perception of religiosity). Other measures we examined included severity of their arthritis (Lequesne Scale); Arthritis-Specific Functional Health Status (WOMAC); Comorbidity (Charlson); and Fatalism (Powe). RESULTS: AA and white patients in this cohort were comparable with respect to age (67 Ϯ 9 vs. 68 Ϯ 9, p ϭ 0.53), Lequesne score(mean 11 vs. 12, p ϭ 0.09) and WOMAC score (mean 17 vs. 18, P ϭ 0.25) and Fatalism (mean score, 1.98 vs. 2.04; P ϭ 0.37). AA reported higher Religiosity (Mean score 76.5 vs. 70.21, P ϭ 0.005) and were more likely to perceive prayer as "helpful" in the management of arthritis compared to whites (crude odds ratio 3.70, 95% CI 2.32 to 5.90). This difference persisted after controlling for severity of arthritis, Fatalism, religious denomination and other demographic, clinical and psychosocial covariates (adjusted odds ratio 2.44, 95% CI 1.27 to 4.70). Furthermore, we found that compared to whites, AAs were more likely to have actually tried prayer as a form of therapy for their arthritis (74% vs 55%, P ϭ 0.002) and were more likely to be Baptist (60% vs 11% P ϭ 0.000). CONCLUSION: In this cohort of AA and white male veterans with moderate to severe chronic knee and/or hip pain, AA patients were more likely to perceive prayer as "helpful" in management of arthritis and to have tried prayer as a form of therapy compared to whites. These ethnic differences in patients' perceptions of the role of prayer in the management of arthritis may explain disparities in health care utilization. PURPOSE: Patients who require mechanical ventilation (MV) after hematopoeitic stem cell transplantation (HSCT) have a high probability of death. We conducted a study to determine the accuracy of previously published characteristics said to predict mortality of these patients, and to estimate the long term survival of patients who recover from this condition. METHODS: Prospective evaluation of predictive characteristics on a multi-center inception cohort of 226 HSCT patients requiring mechanical ventilation. Estimation of predictive power of characteristics using (Bayesian) combination of prior probability of mortality generated from a systematic review of the literature and likelihood ratios of predictors as determined from the prospective multi-center cohort. RESULTS: Expected mortality of patients requiring mechanical ventilation after HSCT is in the range of 82% to 96% based on our review of 14 published studies. We assessed ten predictive models that were statistically significantly associated with death in the original studies. Of these, only the presence of combined hepatic and renal dysfunction was strongly associated with death, with a post-test probability ranging from 99% to 100% (Table contains sample analyses) . Other reported predictive characteristics were not useful. Of patients who recover, six month survival ranged from 25% to 88%.

CONCLUSION: Expected mortality of patients requiring mechanical ventilation after HSCT is high, ranging from 82% to 96%. However, survival for those who recover is substantial. Of models available for identifying which of these patients are likely to die, the presence of combined hepatic and renal dysfunction is the only one associated with a very high probability of death (99% to 100% Between 1992-96, 19 .4% of people with у1 episode without insurance had a major health decline compared to 8.8% of the continuously insured (p Ͻ 0.001). The risk of a major decline was similar regardless of the number of episodes without insurance (1, 2, or 3) or when the episode occurred (i.e., 1992, 1994, or 1996) . After adjusting for the covariates listed above, the odds ratio for a major health decline was 1.64 (95% CI 1.39 to 1.95) for people with у1 episode without health insurance compared to those who were continuously insured. CONCLUSION: Periods without health insurance independently increase the risk of a major decline in self-reported health for adults age 45-61. Transitions between private insurance and being uninsured are common in this group; crosssectional studies therefore underestimate the number of older Americans who are at risk from being uninsured. Changes in health or social policies that affect the number of uninsured adults age 45-61 may greatly affect the number of elderly in poor health and lead to higher Medicare costs as this cohort reaches retirement age. Asian and 5% other. 33% reported a family history of breast cancer. Women's estimates of their lifetime risk of breast cancer ranged from 0 to 90%; the most frequent response was 10%, and 1 ⁄ 2 estimated their risk between 0 and 20%. 1 ⁄ 10 th of the women estimated their lifetime risk of breast cancer at 45% or higher. Health status was significantly associated with perceived risk (see table, p ϭ 0.05).

Neither psychological status nor age was related to perceived breast cancer risk. Twice as many non-white as Caucasian women reported their risk at у45% (20% versus 10%, p ϭ 0.03). Women with a history of breast cancer in a first-degree relative, and those with lower educational attainment estimated their risk of breast cancer higher than did other women. CONCLUSION: Women participating in mammography screening in a managed care setting estimated their own risk of breast cancer to be between 0 and 90%: the most common estimate was near the population average risk. In contrast with previous studies, age was not significantly associated with perceived risk. Nonwhite women and those with poor health, relatives with breast cancer, and lower education gave higher estimates than others.

BACKGROUND: Cigarette smoking is associated with lower education and socioeconomic status and smokers are more likely to have a history of major depression than are non-smokers. The purpose of this study was to examine differences in health related quality of life among smokers, former smokers, and non-smokers. METHODS: The data are drawn from the 1997 Kentucky Health Survey administered by the UK Survey Research Center. The survey included 677 Kentuckians who were 18 years of age or older. Participants were contacted during October 1997 by telephone using a random digit dialing method. The response rate was 56% of all eligible respondents. The margin of error Ϯ 4% at the 95% confidence level. 92% of those surveyed were white and 5% were African-American; 56% of respondents were female; 45% had some education beyond high school and 18% did not complete high school. By self-report, 30% resided in a rural community and the median income was $30,000. A mental health score as well as role functioning, social functioning, physical functioning, pain perception, and health outlook were calculated using the MOS-20 format and compared to self reported smoking status. The effects of self-reported smoking status on these MOS-20 scales were examined using multiple regression analysis controlling for age, sex, education, and rural status. RESULTS: 29.8% were current smokers; 24.1% were former smokers; and 45.8% were non-smokers. 30% of current smokers and 29% of former smokers began smoking before the age of 15. In comparison to those who never smoked, current smokers were younger, less well-educated, and predominantly urban dwellers. year old Asian Indian man presented with low grade fever, general weakness, myalgias, sore throat and dry cough which was worse on extension of the neck for 3 months. He was PPD ϩve 2 yrs ago but did not take prophylaxis. He was aferbrile without lymphadenopathy and the examination of heart, lungs, abdomen, extremities, joints was normal. RESULTS: Lab data : WBC 6600/ul; Neutrophils 67%, Lymphocytes 17%, Monocytes 10%, Hb/Hct 12.6/36.8; Platelets 222000/ul; ESR ϭ 74 mm/hr. Chemistry results and chest X-ray were normal. Gallium scan showed increased uptake in mediastinum. CT chest revealed multiple mediastinal adenopathy, 2 ϫ 2 cm, in retrocaval, precarinal, and subcarinal spaces with a 5 mm nodular density of the left upper lobe of the lung without hilar adenopathy. Biopsy of the mediastinal lymph node showed confluent necrotizing and non-necrotizing granulomas. Repeat smear from the lymph node was ϩve for acid fast bacillus, although culture was Ϫve. The diagnosis of tuberculosis was made by DNA probe. He responded to standard anti-tuberculous treatment. CONCLUSION: Mediastinal lymphadenopathy is an unusual presentation of reactivation tuberculosis. Even though PPD was ϩve, Tuberculosis was not suspected because of lack of loss of weight or pulmonary symptoms, with a normal physical exam and chest X-ray. C.T. chest and gallium scan are sensitive and indicated when clinical and chest X-ray findings are Ϫve and when reactivation is suspected. CT scan differentiates active and inactive healed lesions. Increased uptake on Gallium scan indicates active disease and serial scans are used for monitoring disease activity and response to treatment. PCR: Rapid confirmation of tuberculosis with species identification and MDR strains can be done in few hours by PCR enabling early treatment and control measures. It is comparable in sensitivity and specificity to culture requiring less than 10 organisms for detection compared to the 10,000 necessary for AFB smear. PCR does not differentiate between live and killed organisms and patients on therapy may remain PCR ϩve for an unknown period despite mycobacterial sterilization. DNA probes directed against RNA may be capable of detecting only live bacilli. Rapid diagnosis of TB meningitis is made by PCR of CSF and detection of tuberculostearic acid in CSF. Rapid culture methods within one to two weeks is possible by BACTEC and SEPTI-CHECK methods. Twenty patients (42%) had a Charlson comorbidity score of 2 or greater, but 12 of these 20 (60%) patients were still eligible for at least one trial. There was a modest positive correlation (r ϭ 0.27) between eligibility for any RCT and treatment with warfarin and a modest negative correlation (r ϭ Ϫ.30) with age and warfarin use. There was no association between warfarin use and Charlson comobidity score (r ϭ Ϫ0.04). CONCLUSION: These preliminary data found that the majority of patients in a teaching clinic would have been eligible, based on published criteria, for at least one RCT studying the efficacy of warfarin therapy to prevent stroke in patients with atrial fibrillation, and that advancing age, but not comorbidity, was associated with less warfarin use. PURPOSE: Despite convincing evidence linking poor social support with poor health outcomes, questions about social support are not routinely asked as part of the medical interview. The purpose of this study was to determine the prevalence of poor social support among hospitalized patients and to determine the impact of poor social support on length of stay and cost of hospitalization. METHODS: During a six week period, patients admitted to the internal medicine service were asked to complete a questionnaire designed to assess the psychosocial needs of hospitalized patients. Social support was determined by asking patients whether or not they had someone who they could talk to about important things in their lives. Patients were also asked to rate their self-perceived health status. The burden of comorbid illness was determined by the Charlson comorbidity index. Diagnosis related groups (DRG) were abstracted from patient charts. Variation in the length of stay was determined by calculating the difference between estimated length of stay for that DRG and the patients' actual length of stay. Total cost of hospitalization was obtained from the hospital's computerized data base. RESULTS: Of the 317 patients interviewed, 12% reported that they did not have a source of social support. Their mean age was 63 years, 54% were female, and 30% were African-American or Latino-American. Patients who did not have a source of support were more likely to describe their health as poor (P ϭ .01). The most common reasons for admission among patients with poor support were for cardiovascular, pulmonary, and HIV related illnesses. The mean length of stay for patients who were admitted for these conditions and who lacked a source of support was 13 days as compared to 7.5 days for those patients who had a source of support.

(P ϭ 001). Patients with poor social support stayed in the hospital an average of 5 days longer then expected for their DRG. The total hospital cost for these patients was on average $3,000 higher. On multivariate analysis, after adjusting for age, race, sex, functional status and Charlson comorbidity score, poor social support remained as a significant predictor of length of stay for patients admitted with cardiovascular, pulmonary, and HIV related illnesses. (P ϭ .02) CONCLUSION: Having someone to talk to is an important factor which physicians may often over look. Approximately one out of ten patients admitted did not have a source of social support. These patients were more likely to view their health as poor and stayed in the hospital longer than expected. Innovative methods of identifying patients with poor social support and interventions which enhance social support may improve patient outcomes as well as decrease health care cost. , employment and alcohol-use outcomes is essential for counselling pre-and post-transplantation patients. Our purpose was to evaluate patterns of QOL, employment, and alcohol use among liver transplant recipients with ALD and other etiologies of liver failure (non-ALD). METHODS: We identified 5473 potentially relevant articles using structured MEDLINE and Embase searches and 32 additional references from articles' bibliographies. We included studies if: they reported an assessment of quality of life (QOL), employment or alcohol consumption; reported either pre-and post-transplantation data or had a comparison group; and were written in English. We combined studies to calculate summary proportions, odds ratios, and performed a sign-test to evaluate the direction (positive or negative) of the effect of transplantation on QOL. RESULTS: Of the 5505 identified articles, 89 studies reporting data on 6099 transplant recipients met our inclusion criteria. We found significant post-transplantation QOL improvements (p Ͻ 0.05) in Karnofsky, Sickness Impact Profile, and Nottingham Health Profile scores; physical health, sexual functioning, daily activities, general QOL, and social functioning; but not psychological health. Employment among ALD vs. non-ALD recipients was 29% vs. 59% pre-transplantation and 33% vs. 85% at three years post-transplantation (p Ͻ 0.00001 for differences between groups at each interval). Non-ALD recipients using alcohol had lower rates of employment than those who abstained; however there was no such association for ALD recipients. Although, there was no difference in the proportion of ALD and non-ALD recipients reporting alcohol use post-transplantation (4% vs. 5% at 6 months, and 17% vs 16% at 12 months), the non-ALD recipients were more likely to drink moderately (p Ͻ 0.0001), whereas ALD recipients were more likely to drink excessively (p Ͻ 0.05). The odds ratio for alcohol use among those patients with less than 6 months of pre-transplantation abstinence was 7.8 (95% confidence interval: 4.0-15.3). CONCLUSION: Liver transplantation provides clinically important increases in functional status in both ALD and non-ALD recipients. Prior to transplantation and at long-term followup, substantially fewer ALD recipients are employed than are non-ALD recipients. Less than 6 months of abstinence from alcohol pre-transplantation is a strong predictor of post-transplantation alcohol use. 

Use of evidence-based therapies did not differ by race/ethnicity. Use of cholesterol medication or beta-blockers for secondary prevention did not differ by sociodemographic characteristics. CONCLUSION: In this cohort of Medicare MC patients with diabetes, all of whom had the same pharmacy benefit, there were several sociodemographic differences in the use of evidence-based therapies. While women more often received cholesterol-lowering medications, older, poorer persons and those with less education were less likely to receive several recommended therapies. PURPOSE: Given the increased risk of prostate cancer among AA men and the uncertain benefit and potential harms of screening, we sought to develop an educational brochure which advocated neither for nor against testing but rather promoted shared decision-making by providing information and encouraging discussion with one's doctor. METHODS: The target population was AA, predominately Medicaid-eligible men enrolled in two health care organizations. The 12-page brochure was tailored to the needs and preferences of these men based upon results from a telephone interview and a series of focus groups. To then evaluate the brochure we telephoned 200 eligible practice visitors 40 to 75 years of age, of whom 104 completed a baseline interview. The brochure was then mailed to these men, 58 of whom also completed a post-brochure interview. Those completing versus not completing the second interview did not differ in measured factors. RESULTS: Half are over the age of 60 and have fewer than 12 years of formal education. Knowledge of prostate testing was limited. At baseline 37% identified either rectal exam or PSA test as included in testing and 73% were uncertain about the possibility of a false positive test result. Attitudes toward testing were highly favorable. Over 85% believed that "testing saves lives" and that "all AA men should be tested". At baseline, 35% had made a testing decision: 30% to be tested and 5% to talk to their doctor. Factors independently associated with having made a decision include age Ͻ60, recognizing possibility of false test results, and belief that one's doctor recommends testing (a factor closely correlated with favorable test attitudes). Among the 58 men completing pre and post interviews, 97% agreed the brochure was valuable and easy to understand and 74% believed it provided the right amount of information about pros and cons. The knowledge score increased from 2.5 to 3.2 on a 5-point scale (p Ͻ 0.001) but there was no change in the favorable pre-brochure testing attitudes. Testing decisions did change. The proportion of men choosing to be tested increased by 20%. No man decided to not be tested.

CONCLUSION: We conclude that a carefully developed and targeted "non-advocacy" brochure may increase knowledge but does not consistently reduce (and may increase) enthusiasm for testing among men with pre-existing and strong beliefs about its value. PURPOSE: Acute uncomplicated cystitis is one of the most common community acquired infections in women. Empiric treatment with trimethoprim-sulfamethoxazole for uncomplicated cystitis has been a widely utilized strategy. However, there has been increasing resistance to this antibiotic in Escherichia coli, the most common cause of these infections. The purpose of our study was to determine the trends in this resistance among E. coli uropathogens and to ascertain if there are risk factors, which predict this resistance in acute cystitis. METHODS: There were two parts to the study: Part A was a retrospective study of antibiotic resistance in 786 urine cultures from the microbiology computer system of all patients with E. coli (у10 5 /ml). Part B was a prospective observational study performed at multiple different clinic sites at Denver Health. All patients with symptoms typical of uncomplicated cystitis seen between July 1998 through September 1998 had a questionnaire completed. Treatment details were at the discretion of the care provider. All patients with an isolate resistant to trimethoprim-sulfamethoxazole were phoned and asked to return to clinic where a repeat urine culture was obtained and a second questionnaire was completed. RESULTS: The rate of resistance increased from 2% in 1981 to 26% in 1998. Factors associated with a resistant strain were a urinary tract infection in the past six months (OR ϭ 2.5 [1.0-6.1]), and travel outside the United States in the past six months (OR ϭ 6.0 [1.3-28.6] ). Hispanic ethnicity, compared to persons of other racial/ethnic backgrounds and age less than 3 years (P Ͻ 0.01) also predicted a higher rate of resistance in the retrospective study. Rates of resistance were still greater than 15% in patients without any of these risk factors. The rate of trimethoprim-sulfamethoxazole resistance was similar between inpatient versus outpatient isolates (P ϭ 0.35). Resistance to trimethoprim-sulfamethoxazole predicted resistance to at least one other antibiotic in these urine cultures. CONCLUSION: There is increasing resistance of E. coli in cystitis to a commonly used antibiotic. This is not strictly a nosocomial problem. Prior use of antibiotics is one of the significant risk factors for this resistance, but rates were high even in patients without risk factors. This will likely require closer attention to local antibiotic susceptibility data and a change in empiric treatment strategies. PURPOSE: Spirituality and prayer are important to many people. Two-thirds of the American public believe physicians should talk to their patients about spiritual issues and pray with those who request it, yet only 10% of patients recall their physician doing so. Interested physicians might be more willing to incorporate spirituality as an aspect of health care if they can discreetly identify receptive patients. Our objective was to determine in different settings how often patients desire prayer with physicians and to identify the characteristics of these patients. METHODS: General Internal Medicine patients were surveyed equally from a University, Veterans, and County Hospital (50% outpatients and 50% non-ICU inpatients). 735 patients were invited to participate and 600 agreed to do so (18% refusal rate). The survey instrument and medical record review included common patient demographics, measures of physical health, functional status, social support, quality of life, medical prognosis, and various religious and spirituality assessments. Patients also indicated whether or not they desired prayer with their physicians, and if so, whether they desired this "now" (during the immediate outpatient clinic visit or inpatient hospitalization) or in the "future" (sometime later). RESULTS: Seventy-one percent of patients desired prayer at some time with their physicians. Surprisingly, the following variables were NOT predictive of patients' desire for prayer with their physician: patients' perceived quality of life and state of health, functional status, degree of social support, Charlson's comorbidity score, number of medications, hospital type, or inpatient vs outpatient status. After multivariate analysis, only gender (female Ͼ male, 83% vs. 63%, p ϭ 0.023) and some spirituality variables were statistically significant predictors for patients' desire for prayer with their physician. The most significant predictor (p ϭ 0.003) was a simple, single item 1-9 self-rated spirituality scale (see Table) . PURPOSE: Candidates for implantable cardioverter-defibrillator (ICD) insertion are usually characterized by the presence of severe heart disease, which would make them prone to develop atrial fibrillation (AF) following surgery. A high incidence of post-operative AF (up to one third of cases) was reported in the era of epicardial ICDs. On the other hand, the incidence and risk factors for AF in the recipients of endocardial ICDs are less well characterized. AF is an important cause of inappropriate ICD discharges, since it may be associated with fast ventricular response and misdiagnosed as a ventricular tachyarrhythmia by single chamber ICDs. The identification of patients at especially high risk for AF after ICD implantation may justify the prophylactic use of a dual chamber system, capable of discriminating between supraventricular and ventricular tachyarrhythmias. The purpose of this study was to evaluate the incidence and risk factors of AF in a cohort of patients undergoing ICD implantation at our institution.

[cdot] METHODS: 8 consecutive patients who had transvenous ICD implantation between 4/94 and 7/99 were included in this retrospective review. These patients were followed at 3-month intervals, at which time ICD was interrogated. The characteristics of patients with or without AF during the follow-up were compared. RESULTS: Mean follow-up was 21 Ϯ 15 months. Fifteen patients (17%) developed AF during follow-up, while 73 did not. AF was associated with inappropriate ICD therapy in 13 out of 15 patients. The two groups with or without AF didnot differ significantly regarding: duration of follow-up (24 Ϯ 16 vs 20 Ϯ 15 months), history of previous AF (13% vs 11%), age (67 Ϯ 12 vs 64 Ϯ 10 years), gender, type of underlying heart disease, history of antiarrhythmic therapy (mostly indicated by the presence of frequent ICD shocks during follow-up), left atrial size, presence and severity of mitral/aortic regurgitation. Left ventricular ejection fraction was depressed in both groups, but to a lesser degree in patients with AF, compared to those without AF (39 Ϯ 10 vs 29 Ϯ 13, p Ͻ 0.01). CONCLUSION: 1) AF is a frequent phenomenon in first 2 years following endocardial ICD implantation, even though its incidence is less than that reported in earlier studies utilizing epicardial systems. 2) This arrhythmia is a common cause of inappropriate ICD shocks due to fast heart rate, suggesting the use of prophylactic beta-blocker therapy may be indicated in this population. 3) In our study, we were unable to identify significant differences between ICD patients with or without AF. In a compromised population, such as that of ICD recipients, conventional clinical or echocardiographic risk factors of AF appear to be frequently present, and may have a limited usefulness for stratification into subgroups at higher or lower risk of AF. Regression models adjusting for age, gender, and race, and subsequent transition matrices confirmed these general trends, and suggested threshold levels of health status beyond which prospects of recovery were significantly less. For example, 70 year old women who rated their health perception as "very good" in 1991 had a 53% chance of being "very good" or "excellent" in 1992, compared with 24% for those who rated their health as "good" in 1991. The decision to take HRT has become increasingly complex with the introduction of selective estrogen receptor modulators. Therefore, it is even more important to understand how women, including those with a history of breast cancer, view outcomes associated with HRT. Our aim was to measure the utilities for HRT-related health states in peri-and postmenopausal women with and without breast cancer. METHODS: We interviewed peri-and postmenopausal women from general medicine (GIM) clinics and those in breast clinics with a history of early stage breast cancer at 2 tertiary care centers. We excluded women with high levels of depression or anxiety. Women used a computerized program (U-titer) with a standard gamble technique to assign utilities to their current health and HRT-related health states including: coronary artery disease-acute myocardial infarction and chronic angina; hip fracture-acute and long-term; breast cancer-newly diagnosed and long-term follow-up; HRT use-withdrawal and unexpected bleeding; and menopause-acute symptoms and chronic effects. RESULTS: The first 54 women interviewed had a mean age of 63.6 years (SD 8.8), 56% were married, 62% white, 32% African American and 40% had a high school degree or less. There were no significant differences between the women from the 2 clinics. 62% of the women have ever taken and 27% were currently taking HRT. Utilities for current health and the chronic health states are shown below.

Although we have not yet reached the projected sample size of 300 that will give 80% power to detect a difference of 0.1 in utilities between groups, the utilities for all acute and chronic states were similar for women in both clinics (p Ͼ .10 LS-spine BMD was somewhat higher in blacks than in whites (P ϭ .06). In hispanics aged 60-79 years neither hip nor LS-spine BMD were significantly different from whites. (See Table) CONCLUSION: While common wisdom holds that hispanics are relatively protected from osteoporosis, these data suggest that older hispanic women in this largely Caribbean Latina population had BMD closer to that of whites than that of blacks. These results support the need for more research on the epidemiology, prevention, and treatment of osteoporosis in diverse Latina populations.

SL Cohn, S Chhabra, S Chhabra, Medicine, SUNY Downstate Medical Center-Kings County Hospital, Brooklyn, NY PURPOSE: The optimal antihypertensive medication and level of blood pressure for surgery are unknown. Beta-blockers are felt to be potentially beneficial in this setting whereas dihydropyridine type calcium channel blockers have been implicated in adverse effects. We reviewed data from a series of 314 consecutive hypertensive patients seen in our Preoperative Medical Consultation Clinic to examine the safety and efficacy of aggressive preoperative blood pressure reduction. METHODS: Using an arbitrary preoperative BP goal of 160 ⁄ 100 , we previously reported lowering BP by 14% in the subgroup not meeting this criteria. 1 The mean BP's initially and prior to surgery for the entire group were 157 ⁄ 92 and 146 ⁄ 86 . We randomly selected half (157) We developed decision support software based on published risk models that provides women with personalized information about their menopausal symptoms, risks for coronary heart disease, breast cancer, osteoporosis, and pulmonary embolism, the impact of HRT on these risks, other treatment options, and strategies for risk reduction. Peri-and postmenopausal women between the ages of 45 and 70 without breast cancer or any terminal disease were recruited through hospital fliers, advertisements, and clinician referral. Participants completed a brief questionnaire about their lifestyle, family, and medical history. This information was entered into a software program that generated a 35-45 page printed personalized health report (the DSA) that was mailed to the patient. Patients later completed a written questionnaire that assessed the impact of the DSA on decision making, attitudes, and intended health behaviors. RESULTS: Of the 81 women who have been enrolled in this ongoing pilot study, 13 have completed responses to date. Over half of the respondents reported that the DSA changed the way they thought about HRT and 42% reported that it helped them make a decision. The DSA changed the way many women thought about their chances of getting breast cancer (54%), osteoporosis (46%), and heart disease (31%). The majority of respondents (75%) reported that they planned to make changes in their lifestyle based on the DSA, and 25% reported already having implemented substantial lifestyle changes. The DSA increased both motivation and confidence in their ability to improve their diet, lose weight, exercise more, and quit smoking. Seventy-five percent of the women who were not already receiving regular screening for colorectal cancer reported being more likely to get screening sigmoidoscopy. This pilot study is limited by small sample size, lack of a control group, and a focus on process measures. CONCLUSION: These pilot data suggest that a personally tailored DSA may help women make informed decisions about HRT, change their risk perceptions, and motivate behavioral change. This pilot study is being followed by a randomized, controlled trial to determine the long-term impact of this DSA. Of the perimenopausal and postmenopausal women who completed baseline and follow up surveys, those exposed to the risk management program were significantly more likely to report at the conclusion of the study satisfaction with overall menopause care (68% vs. 55% ( 2 ϭ 11.8, p ϭ .001). They were more likely to be satisfied with discussions they had had with health care providers about menopause management ( 2 ϭ 5.9, p ϭ .02) and about the risks and benefits of HRT ( 2 ϭ 150, p ϭ .001).

On most dimensions of self-reported knowledge, such as knowledge about ways to lower the risk for heart disease and methods for managing menopausal symptoms, the risk management participants were significantly more likely to describe themselves as having a good or great deal of knowledge compared with those in the usual care arm. A difference was not demonstrated when both study arms reported a high level of knowledge about a topic, e.g., the importance of not smoking(98% vs. 96%, 2 ϭ 3.7, p ϭ .06). No differences were found between intervention and usual care in their level of conflict about the decision regarding the use of HRT. CONCLUSION: Participants in a year-long risk management program designed to support women in the menopause transition were more satisfied with their menopause care and reported that they were more knowledgeable than the usual care arm about menopause-related subjects. Conflict regarding the HRT decision was not diminished by this intervention. These data support the benefit of such care models for this population. We assessed psychiatric severity using the Brief Psychiatric Rating Scale (BPRS) which rates psychiatric severity from 0 (none) to 6 (extremely severe) in 18 symptom areas. Oral health was assessed using the Oral Health Impact Profile-screening version, which allows respondents to report the frequency of 14 oral health problems on a 5-point scale from 'never' to 'very often'. RESULTS: 51% of our sample were men, 55% were Caucasian, and 54% diagnosed with schizophrenia. The mean age was 42 and the mean BPRS score was 22.9, reflecting moderate psychiatric impairment. 89% of respondents reported having at least one oral health problem. 58% reported having at least 5 oral health problems. The most common complaints were self-consciousness (64.3%), embarrassment (57.9%), and discomfort eating foods (57.1%) because of problems with teeth. In bivariate analysis, Hispanic ethnicity was associated with worse self-reported overall oral health (p ϭ . 02). In addition, the total number of dental problems reported by the participants was inversely related to their scores in all eight domains of the SF-36 (r ϭ Ϫ.23 to Ϫ.42; all p Ͻ ϭ . 001). After controlling for sociodemographic factors, psychiatric illness, and medical conditions, the total number of dental limitations remained significantly associated with lower patient ratings of perceived health in 6 of the 8 SF-36 domains: bodily pain (p ϭ .0004), general health status (p ϭ .0004), vitality (p ϭ .004), role limits-emotional (p ϭ .004), social functioning (p ϭ .05) and mental health status (p ϭ .009). CONCLUSION: Persons with severe and persistent mental illness have markedly elevated rates of oral health limitations compared to the general population. These problems are associated with worsened self-perceived health status. Addressing oral health may be an important way to improve the overall health of this population. We assessed psychiatric severity with the Brief Psychiatric Rating Scale (BPRS) which rates psychiatric severity from 0 (none) to 6 (extremely severe) in 18 symptom constructs. Podiatric health was assessed using 10 items from the National Health Interview Survey and an additional item addressing foot pain. RESULTS: Fifty-one percent of our sample was male, 55% were Caucasian, and 54% diagnosed with schizophrenia. The mean age was 42 and the mean BPRS score was 22.9, reflecting moderate severity of psychiatric illness. Eighty-four percent of patients surveyed reported at least one podiatric problem and 56% reported at least 3 problems. The most common problems were foot pain (47.9%), nail disorders (33.6%), and foot infections (30%). These conditions were much more common than 1990 NHIS general population reports of nail disorders (5.7%) and foot infections (5.0%). In bivariate analysis, the total number of podiatric problems reported by the participants was inversely related to their scores in seven of the eight domains of the SF-36 (all p Ͻ ϭ .01). After controlling for sociodemographic factors, psychiatric illness, and medical conditions, the total number of podiatric limitations remained significantly associated with lower patient ratings in 3 of the 8 SF-36 domains: bodily pain (p ϭ .02), vitality (p ϭ .02), and general health status (p ϭ .02). CONCLUSION: Persons with severe and persistent mental illness have markedly elevated rates of podiatric problems when compared to the general population. These problems are associated with worsened self-perceived health status. Addressing podiatric health may be a successful way to improve the overall health of this population. Although blacks and drug users have been found to underuse highly active antiretroviral therapy (HAART), it is not known whether this results in poorer health outcomes. To examine this issue we followed from 1996 to 1998 a cohort of 2267 adults in the HIV Costs and Services Utilization Study (HCSUS), a nationally representative study of persons under care for HIV, and examined predictors of survival. The population represented was 34% aged 18-34, 22% female, 51% white, 31% black, 15% Hispanic, 50% male-to-male sexually exposed, 23% drug use exposed, 19% heterosexually exposed, 8% blood product exposed, 24% less than high school educated, 19% with annual income less than $5,000, 21% with CD4 Ͻ50, and 20% uninsured. By January 1998, 261 persons (12%) had died, and in bivariate analysis certain groups had a greater relative risk of death: blacks (RR ϭ 1.73, p Ͻ .001) and Hispanics (RR ϭ 1.48, p Ͻ .04) compared to whites; those exposed to HIV through blood product transfusion compared to those male-to-male sexually exposed (RR ϭ 3.17, p Ͻ .0001); those with less than a high school education compared to college grads ( In this first population-based study of lesbian health, we found that lesbians were more likely than heterosexual women to have poor health behaviors, and less likely to receive preventive health care services. These findings support our hypothesis that sexual orientation/behavior is an independent predictor for health behaviors and receipt of care, and supports the need for the increased systematic study of lesbian health issues. We constructed a discrete-event simulation model of the natural history of colorectal neoplasia in women and men. Probability estimates for the risk of CRC, the incidence of colorectal adenomas and cancers, the mortality from CRC, the diagnostic performance of CS, the complication rates of CS and surgery, and the direct costs of CRC diagnosis and treatment were summarized from the literature. Quality adjusted life years (QALYs) were computed as the measure of effectiveness using utilities as the weights. We used published utility estimates for the outcome states of CRC we previously determined using an instrument specifically designed to measure utilities for outcome states of CRC. We created separate hypothetical cohorts of 100,000 persons who survived to the target age without prior colonoscopy or previously diagnosed colorectal adenomas or cancer for each gender and age group (60, 65, 70, 75, 80, 85 , and 90 years of age). We modeled one-time colonoscopic screening for each gender-age group to examine the costutility compared to no screening. The screening strategy included the removal of all identified polyps, resection of all identified CRCs and CS surveillance for all patients with adenomas or CRCs. We discounted costs and QALYs at 3% per annum. RESULTS: Compared to no screening, screening CS among 60, 70, and 80 year old women:men was associated with an absolute reduction in CRC incidence of 3.77%:3.98%, 3.14%:3.14%, and 1.48%:1.18%, respectively. In addition, screening CS was associated with an absolute reduction in overall deaths related directly to CRC or its diagnosis and treatment among 60, 70, and 80 year old women:men of 1 , and false positive rates associated with the 5 screening programs endorsed by the 1997 guidelines published by the American Gastroenterological Association: annual fecal occult blood tests (FOBT), flexible sigmoidoscopy (FS) every 5 years, annual occult blood tests and flexible sigmoidoscopy (F&S), double contrast barium enema (BE) every 5-10 years, and colonoscopy (CS) every 10 years (Gastroenterology 1997:112:594) RESULTS: Age significantly affected the programs' relative performance on all outcomes except mortality; these results are summarized in the table below, which shows the estimated probabilities (p) of these outcomes associated with the best and worst program in each category. BE is the most effective program in reducing cancer risk for patients between 50 and 65. At 65, F&S becomes associated with the lowest cancer rates & becomes progressively more effective than the other programs as patients grow older. At all ages, FS is the least likely to cause major side effects and CS the most likely; as patients age, the differences between these 2 programs, and among all the programs, become progressively smaller. At all ages, FS is least likely to result in false positive results while F&S is the most likely. The differences between these 2 programs, and among all the programs, become progressively smaller as patients age.

CONCLUSION: Patient age significantly impacts the results of colorectal cancer screening. Because these age-related differences in expected outcomes could affect the choice of optimal screening program for many patients, this effect should be taken into account when making screening recommendations for individual patients. Screening guidelines should also adjust for these differences by including sequences of age-appropriate screening tests in future recommendations. likely to indicate readiness to to control weight than normal weight women and white women respectively. Women who had never smoked were more likely to report actively trying to control weight than current smokers (OR ϭ 2.57, p ϭ 0.01). There was no difference in readiness to control weight among women of different life phases once adjustments were made for BMI. Although 102 (76%) of obese women claimed to be actively trying to control weight, only 34 (25%) had been engaged in physical activity for at least 1 month and only 24 (18%) reported consuming a low fat diet. CONCLUSION: Black women and women smokers were significantly less likely to report actively trying to control weight than white women and non-smokers respectively. This may present unique interventional and/or research opportunities. There is a large discordance in women's reported readiness to control weight and their reported physical activity and dietary fat consumption. Elucidation of the reasons for this discordance would be helpful in the fight against obesity with its attendant morbidity and mortality. have been limited to participants of clinical trials. We sought to examine the quality of life of survivors of AMI in 5 hospitals in 2 mid-Michigan communities, 3 and 6 months after AMI, and to examine potential racial and gender-based differences. METHODS: We performed 3-and 6-month interviews in 241 out of 393 consenting patients prospectively identified during hospitalization for AMI between January 1994 and April 1995. Demographic variables and quality of life after myocardial infarction (QLMI) were measured with items from standardized instruments. The mean (and standard deviation) at 3 months and 6 months for each of 5 domains of QLMI (emotions, restrictions, symptoms, confidence and self-esteem) were calculated for different groups. Best quality of life score ϭ 7 and worst score ϭ 1. Appropriate t-tests were performed to compare between and within group differences. RESULTS: With the exception of symptoms at 3 months,means for all domains of QLMI were Ͼ5. The means improved from 3 months to 6 months for restrictions (5.47 to 5.65, p ϭ 0.006), symptoms (4.8 to 5.63, p Ͻ 0.001), and confidence (5.11 to 5.3, p ϭ 0.001). There was no difference between the overall mean 3-month and 6-month scores for emotions and self esteem. In subgroup analyses, blacks reported slightly worse emotions, and whites reported slightly improved emotions (see Table) . Blacks had a significantly lower score than whites at 3 months in restrictions (4.03 vs 5.42, p ϭ 0.001) and confidence (4.53 vs 5.11, p ϭ 0.04) as well as in emotions, confidence (4.00 vs 4.9, p ϭ 0.001) and symptoms at 6 months. Females had scores similar to men in restrictions, confidence and self-esteem both at 3 months and at 6 month. Other gender and racial differences are presented in the Table. CONCLUSION: The quality of life for this community cohort of patients with AMI was good and improved from 3 months to 6 months. However, there were significant racial and gender differences that warrant further elucidation. and asked if they had diabetes. All respondents who did not report having diabetes were eligible for the screening study. We enrolled subjects at the time of an outpatient visit to the Medical Center. We obtained demographics, a structured medical history, height and weight for all subjects. We screened patients by using an initial random Hemoglobin A1c (HbA1c) measurement, and then obtaining follow-up fasting plasma glucose (FPG) for all subjects with HbA1c Ͼ ϭ 6.0%. We defined a case of unrecognized diabetes as either HbA1c Ͼ ϭ 7.0% or FPG Ͼ ϭ 126. Logistic regression was performed to determine independent factors associated with undiagnosed diabetes and adjusted odds ratios. RESULTS: We enrolled 1253 subjects. The prevalence of unrecognized diabetes was 4.5% (95% CI, 3.4, 5.7). Independent risk factors for diabetes identified by logistic regression were: self-reported hypertension (adjusted OR ϭ 2.5, p ϭ 0.004); obesity as defined by weight more than 20% above ideal body weight (adjusted OR ϭ 2.2, p ϭ 0.02); and self-reported family history of diabetes (adjusted OR ϭ 1.7, p ϭ 0.06). Non-caucasian race (a known risk factor for diabetes) was not associated with the new diagnosis of diabetes in our sample. Considering obesity, family history, and hypertension as 3 risk factors for unrecognized diabetes, the table shows the prevalence of unrecognized diabetes in patients with each number of risk factors (p ϭ 0.001 for trend):

CONCLUSION: There is an opportunity to find patients with unrecognized diabetes in a medical center setting. Assessing risk factors prior to blood testing will improve the specificity, with little cost in sensitivity, of screening for diabetes in a medical center setting. The target population which optimizes any potential value of diabetes screening is patients with at least 1 of 3 risk factors for diabetes. Hypertension is strongly associated with unrecognized diabetes. (56%) with high blood pressure followed-up with a physician after discharge from the emergency shelter. CONCLUSION: Among homeless persons seeking medical care in an emergency shelter, high blood pressure readings were common, especially in those over 30 years. Most persons with high blood pressure did not report a history of hypertension and were not taking antihypertensive medications. Although many patients were lost to follow-up after discharge from the shelter, a majority of homeless persons actually kept their scheduled appointments. Clinicians working in homeless shelter clinics should consider checking blood pressure in all adults, especially those over age 30. Because many homeless patients with high blood pressure may not follow-up after discharge from emergency shelters, clinicians should consider alternative ways to care for these patients. . We defined patients as having diabetes if they fulfilled any of the following: International Classification of Disease code 250 (ICD-9-DM), glucose Ͼ200 mg/dl, glycated hemoglobin Ͼ6.5%, ketones in blood or urine, or use of insulin or oral hypoglycemic agents. We defined poor glycemic control as: glucose Ͻ60 or Ͼ250 mg/dl at any time during hospitalization. RESULTS: The mean age was 64 (SD 10.1), 57% were male, and 27% were nonwhite. Poor glycemic control occurred in 52%. Among patients with poor glycemic control the length of stay was 2.5 days higher (95% CI 1.4, 3.6), charges were $7,330 higher (95% CI $4,700, $9,970), and mortality was 4% higher (95% CI 0.5%, 7.7%).

CONCLUSION: Poor perioperative glycemic control in patients with diabetes undergoing cardiothoracic surgery is an indicator of higher mortality and increased resource utilization. An opportunity may exist to improve the quality of care in the perioperative period. Only 17% of the patients are current users of alternative medicine and about twothirds had never tried any form of alternative medicine. The most popular was herbal medicine (23.5%) distantly followed by acupuncture (3.5%).The only other modalities reported were nutrition (2.4%) and chiropractic (1.2%). The typical user is middle aged and female. More than half of current users expressed satisfaction. The reasons why some patients discontinued alternative medicine use was not explored.

CONCLUSION: The use of alternative medicine is not common in this predominantly African-American population. This lends strength to earlier observations that the use of alternative medicine is more popular among white, middle class populations. The extent of the functional disabilities associated with breast cancer is critical to policy-makers and to those making therapeutic decisions. However, the available information has been based on relatively short follow-up times and coarse measures of functional status. METHODS: In a retrospective cohort study, we interviewed enrollees in the Fallon Community Health Plan, including both breast cancer survivors who were disease-free 5 to 12 years after initial diagnosis and treatment and randomly selected women of the same age with no history of breast cancer. The standardized interview included components of the Nagi disability scale, the Upper Extremity Function Scale, the MOS-SF12, work history, exercise experience, recent symptoms, comorbid conditions, and demographics. RESULTS: 271 survivors of breast cancer (73% response rate) and 208 comparison women (59% response rate) were interviewed. The two cohorts were highly similar in terms of age, race, education level, and comorbidity. Women who had survived breast cancer did not differ from comparison women on responses to the Nagi disability scale [the odds ratio (OR) for any response of "very difficult" or "can't do this" was .95, 95% confidence interval (CI) .59-1.5, controlling for age, comorbidity, and education level]. Breast cancer survivors had slightly, but not significantly, higher scores on the MOS-SF12 physical and mental health scales and they did not report higher rates of upper body symptoms. However, women who had survived breast cancer were somewhat more likely to score in the lowest decile on the Upper Extremity Function Scale (OR 1.8, CI .92-3.6) and they were significantly less likely to report improved ability to carry out work, either at or outside of home (OR .43, CI .28-.67). CONCLUSION: Breast cancer survivors do not report higher levels of disability than women of the same age. However, they are less likely to report improvements in ability to work over time than women of their age and there is a trend toward reports of worse ability to carry out activities that require use of their upper extremities. (1) were performed in primary care (2) were published in English in peer reviewed journals between 1966-1998 (3) employed a gold standard, and (4) reported the performance characteristics (e.g., sensitivity and specificity) for at least one screening method for alcohol problems. Two reviewers appraised all eligible articles for the accuracy of the screening strategy and pertinent methodological content using pre-specified coding criteria. The methodological standards were: a description of patient spectrum (demographics and comorbidity), use of eligibility criteria and reporting of participation rate, avoidance of workup bias, avoidance of review bias, and analysis in pertinent clinical subgroups. 1.43) . Age of first use of marijuana, depth of inhalation of marijuana and use of pipe vs joint for marijuana was not related to these cancers. Although power was limited, marijuana use was not associated with cancer for those who never used tobacco. Adjusting for all sociodemographic factors, family history of these cancers, lifetime tobacco use and lifetime alcohol use did not change the relationship between marijuana and cancer. CONCLUSION: The balance of evidence from this, the largest case-control study addressing marijuana use and cancer to date, does not favor the idea that marijuana as commonly used in the community is a major causal factor for head, neck or lung cancer in young adults.

PURPOSE: Patients increasingly seek more active involvement in health care decisions, but little is known about how to communicate complex risk information to patients. Our objective was to elicit patient preferences for the presentation and framing of complex risk information. METHODS: Peri-and postmenopausal women were recruited through hospital fliers and a community health fair. Eight focus group discussions and 15 one-onone interviews were conducted, where women were presented with risk data in a variety of different graphical formats, metrics, and time horizons. Risk data were based on a hypothetical womans' risk for coronary heart disease, hip fracture, and breast cancer, with and without hormone replacement therapy. Participants' preferences were assessed using likert scales, ranking, and abstractions of focus group discussions. RESULTS: The 40 participants' mean age was 51 years, 50% were non-Caucasian, and all had completed high school. 83% of women preferred bar graphs over line graphs, 100 representative faces, thermometer graphs, and survival curves. Lifetime risk estimates were preferred over 10 and 20-year horizons, and absolute vs. relative risk estimates were the preferred metric. Few participants liked the concept of number needed to treat.

CONCLUSION: Although there are many different formats for presenting and framing risk information, simple bar charts depicting absolute lifetime risk estimates appear to be the preferred format for communicating health risks to patients. (2049) completed the interview. We found that those in the lower income groups were more likely to be female, nonwhite and have fewer years of education (p Ͻ .05 for each). In addition, lower income was also associated with worse reported health status, more medical conditions, and less health insurance coverage (p Ͻ .05 for each). AM use varied by income quartiles (lowest 43%, 37%, 44%, highest 48%) (p Ͻ .05). The average annual out-of-pocket expenditure by users of AM increased with income quartile ($265, 440, 321, 505 p Ͻ .05). Use of specific modalities was similar across all income groups except that the highest income group reported higher rates of use of chiropractic and massage. The odds ratio for use of AM in the highest income group compared with the lowest was 1.2 (95% CI .93-1.6) and increased to 2.0 (95% CI 1.5-4.8) after adjusting for other demographic variables and medical conditions associated with AM use. When examining chiropractic and massage, a similar pattern emerged. For chiropractic, the odds ratios for use in the highest group increased from 1.4 (95% CI .89-2.1) to 2.0 (95% CI 1.0-3.8) when adjusting for medical conditions and demographics. For massage, the odds ratios for use in the highest group increased from 1.6 (95% CI 1.0-2.4) to 3.4 (95% CI 2.0-8.3). Whether one had personal medical insurance did not affect rates of use of alternative medicine across the income groups. CONCLUSION: Higher income is associated with increased use of AM overall. After adjusting for other demographic variables and medical conditions, these differences become more dramatic. While income is a barrier to AM use, still, among the lowest income group (Ͻ$20,000), 43% of respondents used AM and spent on average over $250 annually. Patients' reports about their care were used to compute a problem score (0-100), with higher scores indicating more problems. Patients were then categorized into either a low problem group (score in the bottom 3 quartiles) or a high problem group (top quartile score) for their hospital and subsequent outpatient care, respectively. Clinical data were obtained from discharge abstracts. Outcome analyses adjusted for chest pain one month post-MI, sociodemographic factors, comorbidities, prior hospitalizations, treatment type, hospital size, and transfers. RESULTS: Most patients had relatively few problems with either their hospitalization (mean score ϭ 15) or the first 3 months of post-MI outpatient care (mean score ϭ 12) but the ranges were large (0-85 and 0-75, respectively). In adjusted analyses, patients in the high problem hospital care group had more problems with subsequent ambulatory care, such as not receiving sufficient information from outpatient providers (OR ϭ 0.3, 95% CI 0.2-0.5) or not having complete confidence or trust in these providers (OR ϭ 0.4, 95% CI 0.2-0.6) than patients in the low problem hospital care group. However, both groups were equally likely to have seen a primary care physician or cardiologist and enroll in cardiac rehabilitation. Patients in the high problem group for both their hospital and subsequent outpatient care were twice as likely (OR ϭ 2.2 95% CI 1. 1-4.5) to report chest pain one year post-MI as patients in the low problem group for both stages of care, after adjustment for potential confounder variables. In contrast, patients in the high problem hospital care group and low problem ambulatory care group were equally likely (OR ϭ 1.0, 95% CI 0. 4-2.4) to have chest pain at one year as those who reported few problems with both stages of care. CONCLUSION: Although negative hospital experiences predispose AMI patients to worse long-term outcomes, these effects can be offset by more positive experiences with subsequent ambulatory care. The purpose of the current analysis is to examine the role of exercise and eating behaviors of Mexican-American adolescents. DESIGN: These results are from a larger study examining the impact of a behavioral counseling intervention on adherence to preventive tuberculosis treatment in Hispanic adolescents. RESULTS: Preliminary data were available from a sample of 242 male (55%) and female (45%) Hispanic adolescents in San Diego County. Subjects were questioned about their level of acculturation as measured by the Bi-Dimensional Short Acculturation Scale for Adolescents (BAS), place of birth, dietary and exercise behaviors and perceived body image as part of the baseline interview for the clinical trial intervention. The average age of the study participants was 16 years with 64% born outside the U.S., primarily in Mexico. On average, 0.4% of the subjects reported eating five or more servings of fruits and vegetables per week; 23% reported exercising at least 3 days/week. In addition, while only 26.9% of the sample perceived themselves as overweight, 49.6% were trying to lose weight. Exercise in this sample of Hispanics is also far below the norm (63.8%) as compared with the Youth Risk Behavior Surveillance of 1997. CONCLUSION: These results indicate that Hispanic adolescents may be at higher risk for obesity and diseases such as type II diabetes that are influenced by sedentary lifestyle and poor eating habits. Further analysis will relate diet, exercise, and other sedentary activities, such as TV-watching to degree of acculturation, at which time the relationship between acculturation and exercise behaviors will be examined. PURPOSE: There are no definitive guidelines for an optimal immunization schedule after primary immunization. We used HBs-antibody level to estimate the appropriate timing of next HBs-antibody measurement for booster immunization. METHODS: 136 HBs-antibody-negative student nurses who receive immunization for hepatitis B virus. 129 with 3 100mIU/ml HBs-antibody were evaluated at 7th month. 92, 44, and 16 were evaluable at 23rd, 35th, and 47th month, respectively. HBs-antibody was measured at 7th, 23rd, 35th, and 47th month from the first immunization of primary immunization. We settled booster immunization value of HBs-antibody at Ͻ100mIU/ml. To evaluate the further necessity of HBsantibody measurement, likelihood ratio of 3 100mIU/ml was calculated in each hextile stratum of 100-300, 300-600, 600-1200, 1200-2000, 2000-4000, and 3 4000mIU/ml of HBs-antibody for predicting 23rd month from 7th month, 35th month from 7th and 23rd month, and 47th month from 35th month HBs-antibody measurement. RESULTS: Likelihood ratio was infinity at 2000-4000 stratum and over in predicting 23rd month from 7th month, 3 4000 stratum in predicting 35th month from 7th month, 600-1200 stratum and over in predicting 35th month from 23rd month, and 300-600 stratum and over in predicting 47th month from 35th month, respectively.

CONCLUSION: We could schedule next HBs-antibody measurement for deciding booster immunization as below: There was no significant differences in gender distribution, mean age, mean ejection fraction, % of patients of age Ͼ70, % with EF Ͻ40, diabetes, unstable angina, multivessel, CAD, prior MI, or acute MI PTCA between the black and white patient groups. However, a greater % of black than white (91% vs 58%) was hypertensive. Stents were part of a similar % of single vs multivessel procedures in both groups, and a similar number of lesions were intervened upon in both groups. Vessel site (LAD vs RCA vs CIRC vs graft) distribution was similar in both groups, as were procedural success rates (96% for whites, 100% for blacks).

In-hospital complications were infrequent in both groups, with no significant difference seen in non fatal MI, CABG, death, need for repeat PTCA, stroke, or vascular complications. Total in-hospital charges were the same for both groups. CONCLUSIONS: Although data is limited, it appears that black patients benefit from stent use for PTCA similarly, as do whites. Our data would suggest that continued study of the use of these devices is indicated in various patient subgroups. The high incidence of hypertension in black patients coming to stent implantation deserves further investigation. PURPOSE: Many clinicians believe that there is a relationship between perioperative anemia and cardiac events. Nevertheless, there is a paucity of literature that supports this assertion. Given the ongoing concern about the use of blood products, further evaluation of this association is warranted. The purpose of this study was to compare post-operative hemoglobin in patients who suffered cardiac complications to that of patients without cardiac complications, with the hypothesis that hemoglobin would be lower in those patients who had a cardiac event. METHODS: We identified 32 patients who underwent orthopedic surgery at one of two London, Ontario hospitals, and who had a perioperative cardiac event, defined as unstable angina, myocardial infarction, acute pulmonary edema or death. For each of these cases we attempted to identify three controls, matched for age, sex, and pre-operative Canadian Cardiovascular Society (CCS) angina grade. The lowest post-operative hemoglobin value obtained for these two groups of patients was compared using a paired t-test. RESULTS: The lowest mean post-operative hemoglobin was 88.7 g/L in the cases and 94.9 g/L in the controls, a difference of 6.2 g/L (95% CI:12.1-0.27, p ϭ .041). The cases and controls did not differ significantly with respect to age (p ϭ .94), lowest intra-operative systolic blood pressure (p ϭ .24), highest intra-operative heart rate (p ϭ .21), pre-operative hemoglobin (p ϭ .33), or CCS angina grade (identical), indicating good matching. CONCLUSION: The results suggest that there is an association between perioperative anemia and cardiac events in patients having orthopedic surgery. Nevertheless, it is not known whether increasing the hemoglobin by transfusion will have an impact on cardiac event frequency. We look forward to future research that will examine this issue. In the postgame evaluation, the vast majority of study participants found CHAT enjoyable (85.7%), easy to understand (93.4%), easy to do (93.4%), and informative (90.8%). The majority of participants indicated that they had learned a lot from playing the game (64.5%) and were motivated to learn more about health insurance (71.8%). More than 85% of participants stated that they would be willing to abide by the group's chosen healthplan. CONCLUSION: CHAT is an enjoyable, easily understood, informative and motivating process which relatively disadvantaged, low income groups of persons without healthcare expertise can use to design healthplans that are cost-conscious and acceptable to them. In this non-randomized study, the use of coronary stents in pts undergoing PTCA for AMI appears to result in a decrease in early mortality (p ϭ .07), especially in pts with 3-vessel disease (p ϭ .04).

Clinic, Northshore, Peabody, MA PURPOSE: Erectile dysfunction (ED) is a common medical condition affecting an estimated 30 million men in the United States. The objective of this study was to determine the efficacy and safety of sildenafil when either medication changes or lifestyle modifications were initiated in patients before treatment with sildenafil in a sexual dysfunction clinic. METHODS: The study comprised 714 men diagnosed with ED of various etiologies with a broad range of concomitant medical conditions treated either by medication changes or adjustment of social risk factors before treatment with sildenafil for ED. Patients were instructed to take a 50-mg dose of sildenafil approximately 1 hour before sexual activity, not more than once in a 24-hour period and at least twice a week. Dosage could be increased to 100 mg or decreased to 25 mg based on tolerability and efficacy. Patients recorded the number of successful intercourse encounters in a patient diary for 6-8 weeks. RESULTS: The primary etiology of ED was determined to be organic in 68% of men, psychogenic in 7%, or mixed in 24%. The incidence of an organic etiology increased with age. The predominant associated risk factors were hypertension (39%) and patients taking multiple medications (34%). The percent of patients taking multiple medications increased with age. Patients that were hypogonadal (36%) had concomitant medical conditions plus ED. Overall, there was an 82% successful intercourse rate with sildenafil treatment in the patient population. A 50-mg dose was effective for 66% of the patients with successful intercourse. There were no significant differences between patients given testosterone replacement and sildenafil or sildenafil alone. Common adverse events-flushing (20%) and headache (17%)-due to sildenafil treatment were mild to moderate in nature and resulted in Ͻ2% discontinuation.

CONCLUSION: The rate of successful treatment with sildenafil in men with ED and concomitant medical risk factors is higher when medication changes and lifestyle modifications are initiated before treatment with sildenafil. METHODS: A conceptual model of patient trust was developed and items to measure trust were generated, reviewed and pilot-tested. Items were grouped into four dimensions of trust: loyalty, honesty, competence, and confidentiality. Twentyfour candidate items were then pilot tested in a national random-digit-dial survey of 989 adults. Data collected included: relationship length and frequency of contact, how physician was selected, satisfaction with care, medical history and health status, demographics of the subjects and their physicians, desire to switch doctors or obtain a second option, preferences for seeking professional help and making medical decisions, and related patient, physician, and situational factors. Factor structure, internal consistency, and construct and predictive validity were assessed. RESULTS: Factor analyses showed a single-factor solution for 12 items. Internal consistency (reliability) was .92 (Chronbach's alpha). Trust in a physician is a concept that can be reliably measured, however, trust does not consist of multiple dimensions, as we hypothesized. Physician trust was correlated with satisfaction with care (.50). The following factors were correlated with physician trust, demonstrating predictive and discriminant validity: length of relationship, number of contacts, choice in selecting physician, physician chosen based on recommendation, seriousness of treatment, having sought a second opinion (Ϫ), past dispute with this physician (Ϫ), and desire/intention to switch doctors or obtain second opinion (Ϫ). The following factors were not significantly related to physician trust: patient or physician demographics, patient education or income, having a chronic or serious condition, being worried about health, willingness to seek professional care or rely on expert judgment, or preferred role in making medical decisions. CONCLUSION: Patient's trust in their primary care physicians can be reliably measured. Physician trust is distinct from satisfaction. The factors that predict physician trust relate to physician characteristics, the nature and origin of the relationship, and the seriousness of treatment, rather than to patient characteris- PURPOSE: Administrative databases, created primarily for fiscal purposes to track health care utlization of enrollees in health insurance plans, are increasingly being used for epidemiologic and health services research on large populations. This investigation was conducted to compare health care utilization in a population identified solely through an administrative osteoarthritis diagnosis, with a group of patients in whom this diagnosis had been validated. METHODS: We identified all enrollees in a group-model HMO with documentation of at least one health care encounter associated with an osteoarthritis diagnosis during the period 1994-1996, and who continued to be enrolled in the health plan for a one-year period following the health care encounter date. This population was comprised of 10,740 individuals. From this population, we randomly selected 700 individuals, whose medical records were abstracted by trained nurse reviewers utilizing a structured data collection instrument to ascertain information relevant to the diagnosis of osteoarthritis. Pairs of physician reviewers evaluated the abstracted information and rated the evidence for the presence of osteoarthritis according to three levels (definite, possible, and unlikely). All persons rated as having definite osteoarthritis were included in the validated group (n ϭ 443). Health care utilization was assessed in both the administrative osteoarthritis diagnosis group and the validated group across the following domains: 1. non-inpatient care associated with an osteoarthritis diagnosis; 2. relevant radiographic studies (e.g., spine, shoulder, hand, hip, knee, and foot x-rays); 3. relevant surgical procedures (e.g., arthroscopic procedures of the knee or hip, or hip/knee replacements); 4. relevant medication dispensings (e.g., NSAIDs, non-acetylated salicylates, opioid analgesics, and intra-articular steroids The potential for HIV testing to lead to improved health outcomes has never been greater, yet little is known about utilization of this service among populations at highest risk for HIV. The aim of this study is to determine the prevalence and predictors of HIV testing in a population of urban, poor, predominately minority homeless women in Los Angeles County. METHODS: Women were interviewed in shelters and meal programs in Los Angeles County to create a probability cluster sample (n ϭ 974). Bivariate analysis of patient characteristics, contact with health and social services, and indications for HIV testing identified variables with p Ͻ 0.1, and these were entered into a multiple regression model. Additionally, stepping algorithms with random subsampling for cross-validation were used to develop a more parsimonious multiple regression model. Samples were weighted and cluster analysis was performed on final models. RESULTS: Approximately 70% of the sample reported receiving an HIV test in the past year. After adjustment for cluster effects, only a history of drug abuse or dependence and having a regular source or care were significant predictors of HIV testing in the past year. Sixty-five percent of respondents reported currently having a regular source of care. Neither income, public benefits, health insurance status, nor having a case manager predicted receipt of HIV testing. Similarly, none of the indications for HIV testing were statistically significant predictors. Of note, 15% of respondents reporting a pregnancy in the last year also reported that they did not receive an HIV test. There were no significant negative predictors of HIV testing in the past year. CONCLUSION: The rates of HIV testing in this probability sample of women are among the highest reported for any population studied and are consistent with prior research indicating that individuals at highest risk are more likely to be tested. However, there is a significant minority of at-risk women with specific indicationsfor testing, including pregnancy, who are not being reached. Our study indicates that policies focused on providing a regular source of care are likely to have the greatest impact on sertesting prevalence. However, the fact that our analyses revealed only two statistically significant predictors demonstrates that our current understanding of homeless womens' use of this critical service is incomplete. Additional research is needed to better understand access to and utilization of HIV testing in this high-risk population. PURPOSE: Black patients on dialysis have lower mortality rates and withdrawal rates for failure to thrive than age-matched White patients. However, there has been little examination of racial differences in health related quality of life (HRQL) or self-reported overall health (OH) among these patients. METHODS: To assess OH and HRQL, we surveyed 1,392 patients (response rate ϭ 82.9%) age 18-54 approximately 9 months after they initiated dialysis for endstage renal disease (ESRD) in five states and the District of Columbia. Respondents rated their overall health for the most recent month on dialysis. Questions adapted from the Kidney Disease Quality of Life (KDQOL) Instrument assessed patients' emotional health, physical activity, energy level, social activity, and burden of ESRD on daily life. All scales ranged from zero (poorest health/greatest burden possible) to 100 (best health/least burden possible). Available medical records (84% of each group) were reviewed for comorbid illness. Linear regression was used compare adjusted OH and HRQL by race. RESULTS: Among both the 721 Black patients and the 671 White patients, 53% were female and the mean age was 42 years. Black patients reported better OH (mean score 65.0 vs. 60.0) and higher energy levels (mean score 47.2 vs. 41.3) (both P Ͻ 0.001). There were no differences between groups in the other four KDQOL domains. Black race remained significantly associated with OH and energy level after adjusting for comorbid illness, body mass index, perceived quality of care, amount of psychological and social support available, income and employment status. After adjustment, black men reported a lower burden of ESRD on daily life.

CONCLUSION: Black race is associated with a better OH and energy levels among dialysis patients, even after controlling for numerous potential confounders. Future research into determinants of overall health and HRQL for each racial group are needed to understand these racial differences. PURPOSE: Many women with breast cancer are menopausal and experience severe menopausal symptoms. Because breast cancer is more common among older women, many are at elevated risk for coronary heart disease and osteoporosis. To better counsel menopausal breast cancer survivors, we explored their decisions, attitudes, and expectations concerning the use of hormone replacement therapy (HRT). METHODS: Women who had been treated for breast cancer at a tertiary hospital were mailed a questionnaire about their decisions, attitudes, and expectations concerning HRT. Subjective data were collected using a 5-point likert scale and the decisional conflict scale. RESULTS: Of the 71 contacted, 51 returned a completed survey (72%). Their mean age was 52 years (range: 32-73), mean time since initial diagnosis was 6 years (range: 0-35), and 80% were postmenopausal. A majority of women were undecided about taking HRT (70%), 4% were already taking or definitely would take HRT, and 26% would definitely not take HRT. Nearly half felt the decision was hard to make. Many reported not knowing the benefits (26%) or the risks and side effects (41%) of HRT. Over half felt they needed more advice and information about the choice. Reasons cited for why they would take HRT included lowering their chances of heart disease (52%) and osteoporosis (48%), looking and feeling younger (36%), relieving menopausal symptoms (34%), and improving their sex life (32%). Reasons reported against taking HRT included the chances of breast cancer recurrence (80%), weight gain (20%), menstruation returning (20%), and vaginal bleeding (15%). Their doctors' recommendation, either for or against, weighed heavily in the decision. Eighty-four percent of women expected that HRT would lower their chances of heart disease and osteoporosis, and over a third were moderately to extremely worried about not receiving HRT's protection against these chronic diseases. CONCLUSION: Many breast cancer survivors are undecided about the difficult decision to take HRT and need more advice and information from their clinicians about its risks, benefits, and side effects. In making this decision, women with breast cancer appear to place the most importance on their risk for breast cancer recurrence, but also seriously consider their risks for heart disease and osteoporosis, as well as their doctors' recommendation. Menopausal symptoms do not play as large a role in decision-making as might be expected.

DIAGNOSTIC PERFORMANCE AND OUTCOMES. RM Hoffman, M Adams-Cameron, WC Hunt, CR Key, FD Gilliland, Albuquerque VA Medical Center; University of New Mexico Cancer Center, Albuquerque, NM; University of Southern California, Los Angeles, CA PURPOSE: We created a prostate-cancer screening surveillance system to determine the diagnostic performance and outcomes of prostate-specific antigen (PSA) testing in a community-based population. METHODS: We obtained all PSA testing data, including results, patient demographics, and date of testing from the major laboratories in the Albuquerque metropolitan area from 1995 through 1997. Medical records abstractors from the New Mexico Tumor Registry (NMTR) obtained data on all benign prostate biopsies at these laboratories. Data for men у40 years were linked with the NMTR to exclude prevalent cases of prostate cancer and to identify incident cases of prostate cancer. The NMTR also provided data on date of cancer diagnosis, staging, and treatment. The diagnostic performance of PSA testing was determined by calculating sensitivity, specificity, likelihood ratios, and the area under the ROC curve (AU-ROC). RESULTS: We identified a cohort of 41,261 subjects without previously diagnosed prostate cancer who underwent PSA testing. The median age was 61 (range 40-107) and 62% were white, 29% Hispanic. Overall, 2,574 subjects (6.2%) were biopsied, including 33% of the 5,298 men with PSA values у4.0. Biopsies were obtained from 44% of men 50-59 with elevated PSA levels, but only 23% of men у70. Cancer was detected in 944 subjects (2.3%), 758 (80%) had localized and 49 (5.2%) had distant stage disease. The cancer detection rate ranged from 0.2% for men in their 40s to 3.8% in men у70. PSA levels were significantly higher in cancer cases (median ϭ 6.6) than non-cases (1.1), P Ͻ 0.0001. Among subjects undergoing biopsy, the sensitivity for PSA у4.0 was 78% and the specificity was 40%, the AUROC was 0.64 (SE ϭ 0.01). Likelihood ratios (95% CI) for PSA levels Ͻ4 ϭ 0.56 (0.49, 0.64), 4-10 ϭ 1.0 (0.9, 1.1), Ͼ10-20 ϭ 1.6 (1.3, 1.9), and Ͼ20 ϭ 5.4 (3.9, 7.5). Sensitivity increased with age, from 68% in men in their 40s to 83% in men у70. Conversely, specificity decreased from 48% to 32%. Seventy-four percent of the patients with localized cancers received aggressive treatment (surgery and/or radiation). However, the proportion of patients receiving aggressive treatment decreased from 82% in men 40-59 to 56% in men у70. CONCLUSION: In a community-based population, PSA testing provided only fair discrimination between men with and without prostate cancer. Very high PSA levels (Ͼ20) were needed to substantially increase the likelihood of cancer. PSA testing appeared to have less clinical utility for men у70 because they were less likely to undergo biopsy or have aggressive treatment for localized cancers.

Rabow, Medicine, UCSF, San Francisco, CA PURPOSE: To compare the views of seriously-ill out-patients with those of their primary care providers (PCPs) regarding the place of spiritual care in the primary care medical encounter. METHODS: Two self-administered surveys on spiritual care were developed and pre-tested. One survey was distributed to all currently active PCPs in the General Medicine Practice (GMP) at UCSF. The other was distributed to GMP patients enrolled in a study of end-stage CHF, COPD, and cancer. Non-responders received two subsequent survey distributions. RESULTS: Of the 38 patient surveys, 26 were returned complete (68%). The average age of the patients was 64. 73% were female and 73% were White. 65% considered themselves "Quite a bit" or "Very much" spiritual. While there was no gender difference between patients' desire for spiritual care, significantly more men reported their providers had asked if they would like a referral to a chaplain (p ϭ 0.01). Of the 84 PCP surveys distributed, 67 were returned complete (80%). The average age of the PCPs was 33. 55% were female and 61% were White. 61% were residents; the remainder were fellows or faculty. Only 30% of PCPs considered themselves "Quite a bit" or "Very much" spiritual. Compared to their male colleagues, female providers considered themselves to be more spiritual (p ϭ 0.05) and believed it more important to attend to patients' spiritual concerns (p ϭ 0.03). Female PCPs more frequently asked whether patients were part of a spiritual community (p ϭ 0.05) and prayed more often for patients (p ϭ 0.001). Faculty and fellows were more likely than residents to ask patients if they were part of a spiritual community (p ϭ 0.0001), ask about the importance of spirituality (p ϭ 0.001), offer to discuss spiritual concerns (p ϭ 0.001), and offer a chaplaincy referral (p ϭ 0.01). This difference persisted even when controlled for age and number of weekly clinical sessions practiced. Similar percentages of patients (60%) and PCPs (68%) believed that it is important for PCPs to attend to patients' spiritual concerns. However, only 19% of patients had ever been asked by their PCP if they would like to discuss spiritual concerns and 19% had been asked if they would like to speak with a chaplain. 68% of patients wanted their PCP to pray for them, but only 23% reported that their PCP had ever offered. When asked what hinders them from addressing spirituality, 82% of providers cite a lack of time, 37% do not feel competent, and 21% do not feel it is part of their job. CONCLUSION: Seriously-ill patients and PCPs in an urban, academic, general internal medicine clinic agree that spiritual care has an important role in the medical encounter. However, discrepancies exist between this stated belief and spiritual care practices, especially among male providers and residents. . Age-gender-specific indices were created for individuals over the age of 50 involving 8 preventive services for women and 5 preventive services for men. Each index was normed to reflect the percent of the age-gender specific recommended preventive services utilized by an individual (i.e., percent compliance). RESULTS: Women reported receiving a mean of 57% (SD ϭ 16%, Median ϭ 63%) of the recommended preventive services compared to 46% for men (SD ϭ 29%, Median ϭ 40%; p Ͻ 0.001). Clinically unimportant differences in the percent of services utilized were found between White Non-Hispanic (53%), Black Non-Hispanic (51%), Hispanic (48%) and Other (51%) individuals (p Ͻ 0.001). Gender differences were consistent across racial-ethnic groups. Both income and years of education are positively related to the percent of services received (p Ͻ 0.001) for both men and women and show a range from 37% for men with less than an elementary education as well as men with incomes of less than $10K per year to only 52% for men who graduated from college and men with annual incomes greater than $75K. The mean percent of recommended preventive services received by women was about 10% higher at each level of annual income and educational attainment (range: 47% to 63%). Nevertheless, gender, income and educational attainment were only weak predictors explaining less than 10% of the variation in the utilization of recommended preventive services. However, of particular interest is the relationship between smoking status and percent of recommended preventive services received (p Ͻ 0.001). While current smokers were least compliant (men 35%, women 50%), self-reported former smokers utilized more of the recommended age-gender-specific preventive services (men 49%, women 61%) than individuals who claimed they never smoked (men 46%, women 56%). CONCLUSION: These findings confirm that most people, particularly men, those with little education and low incomes receive only about half of the preventive services recommended for their age and gender. However, individuals who report they have quit smoking utilize a greater proportion of preventive services recommended for their age and gender than do current smokers and those who never smoked. The health message about the hazards of smoking apparently convinced former smokers of the benefits of healthy behaviors and these 'converts' have become the most compliant in the of utilization of recommended preventive services. PURPOSE: Recent studies have suggested that patients have difficulty comprehending medical risk information. One explanation is that these patients have difficulty interpreting numbers. Most persons, however, use numbers in other uncertain contexts, e.g., when deciding on whether to carry an umbrella to work in the morning or on purchasing services that might vary in price. This study seeks to determine whether subjects are able to use imprecise numerical information when making decisions in non-medical and in medical contexts. METHODS: 145 prospective jurors received a written, self-administered questionnaire involving scenarios with imprecise probability information on the chance of rain, on the cost of a car repair, and on the chance of developing a hypothetical disease. Within each scenario, subjects received a probability estimate with three levels of precision, e.g., a 30% probability of rain Ϯ5%, Ϯ10%, and Ϯ20%. In both the weather and the medical test scenarios, we varied the probability estimate (from 50% to 30%); and in the weather scenario, we also varied the ranges of precision (from Ϯ1, 5, 10% to Ϯ5, 10, 20%) around the probability of rain. The questionnaires were otherwise identical. The study assessed subject's ability to compare the precision provided by the three options within each scenario. RESULTS: Within this convenience sample of potential jurors, the mean age was 40 years; 70% were female; 66% received greater than 12 years of education; 44% were Caucasian; and 76% were employed. Subjects correctly identified the choice with the most precise information in 81%, 88%, and 69% of the responses in the weather forecast, in the car repair, and in the medical test scenarios respectively (p Ͻ 0.001). Furthermore, when the risk estimate varied from 30% to 50%, there was no significant difference in patients' identification of the most precise weather forecast (89% v. 84%, p Ͼ 0.5) or in their identification of the most precise medical test (67% v. 70%, p Ͼ 0.7). When the width of the confidence intervals varied between Ϯ10% to Ϯ20%, there also was no statistically significant difference: 69% v. 84% (p Ͼ 0.09). CONCLUSION: In conclusion, the subjects' ability to distinguish choices on the basis of information precision varied by context, but not by base rate or degree of precision around that base rate. These data suggest that contextual factors may play a larger role in subjects' ability to comprehend imprecise risk information than issues of numeracy, and that presenting uncertainty in medical settings faces particular challenges even to patients who can understand uncertainty in non medical settings. Generalists face the challenge of managing multiple risk factors when caring for patients with type 2 diabetes. We carried out a meta-analysis of hyperglycemia, hypercholesterolemia, and hypertension treatment trials (RCTs) to quantify the magnitude and timing of effect of these strategies. METHODS: Using Medline and review articles, we systematically searched for RCTs focusing on type 2 diabetes or diabetic subgroups. We included RCTs measuring pre-specified macrovascular and microvascular outcomes and abstracted characteristics of the studies, patients, and outcomes. We calculated treatment effect by odds ratio(OR) and patient-years of treatment needed to prevent one event(PYT). ORs were pooled using the Dersimonian-Laird model (RevMan 4.0). We estimated the time to treatment effect (TTE) from the separation of Kaplan-Meier curves. RESULTS: Four glycemic, 7 cholesterol, and 13 hypertension studies were used. For macrovascular events(cardiovascular mortality, myocardial infarction (MI), and stroke), hypercholesterolemia and hypertension control had larger treatment effects than glycemic control (see In small studies, parents of children with these conditions express concern about how to provide adequate care as their children grow into adulthood. We identifed care providers for adults with CP in a national sample and assessed the impact of providing such care on employment of family members. METHODS: We used the 1994-95 National Health Interview Survey (NHIS, a nationally representative sampling of non-institutionalized, civilian U.S. residents) and NHIS Disability Supplement (NHIS-D, a detailed follow-up survey for persons with selected conditions including CP). We used SUDAAN to calculate national population estimates and descriptive statistics. RESULTS: 129 persons age 18 or older reported cerebral palsy (representing an estimated 174,000 civilian, community-dwelling adults), of whom 76% completed the NHIS-D. The mean age of the sample completing both surveys was 36.4y (SE 1.6) . 67% needed at least one regular helper for activities of daily living (ADLs, such as bathing) and/or instrumental ADLs (such as telephoning). A parent was the primary helper for 38%, spouse for 15%, paid employee for 13%, and other relative or friend for 34%. 34% of the total population either lived alone or were married. Of those who lived with someone else but were not married, 62% (or 41% of the total population) said that they did so because of their health or physical problem. 73% of these people lived with parent(s), 12% with a sibling, and 15% with another relative or non-relative; 47% of this group were 40 years old or older. 16% of all persons with CP (and 26% of those who lived with someone else because of health problems) said that their medical problem caused a family member to quit a job, reduce working hours, or change jobs. Family members affected included parents (64%), spouses (19%), and siblings (17%). CONCLUSION: Parents, spouses, and other relatives provide the majority of personal assistance for non-institutionalized adults with CP living in the U.S. Approximately 1 ⁄ 3 of this population continue to live with parents well into middle age because of their medical condition. A substantial minority report adverse impact of their medical condition on family members' employment. These results suggest that greater financial and/or personnel assistance for the medical care of adults with CP may be warranted. To determine the relationship between income levels and alcohol-related hospital admission rates within a large urban area. METHODS: We examined hospitalizations among the approximately 635,000 adult residents of the downtown core of Toronto, Ontario, a socioeconomically diverse city with geographic clusters of high, middle, and low income neighborhoods. We used a comprehensive database of hospital admissions that obtains mandatory reports from every acute care facility in Ontario. Alcohol-related admissions during the period April 1996 through March 1997 were identified by ICD-9 codes for the following conditions: alcohol psychoses, alcohol dependence syndrome or alcohol abuse, alcoholic fatty liver, acute alcoholic hepatitis, alcoholic cirrhosis or other alcohol-related liver damage, toxic effects of alcohol, or accidental poisoning by alcohol. If a patient was admitted more than once, only the first admission was counted. Each patient's place of residence within Toronto was established using postal codes. The study area consisted of 1,408 census enumeration areas. We obtained 1996 census data on population composition and average household income in each enumeration area. The enumeration areas were grouped into 5 income quintiles, each with approximately equal population. In each quintile, age-and sex-adjusted admission rates for alcohol-related conditions were calculated using direct standardization. RESULTS: Alcohol-Related Hospital Admission Rates per 10,000 Population in Downtown Toronto CONCLUSION: Men and women living in the lowest income quintile areas of Toronto have particularly high admission rates for alcohol-related conditions. Among men, admission rates decrease steadily with increasing income, but admission rates among women are similar in all income quintiles except the lowest one. These findings suggest significant gender differences in the correlation between socioeconomic status and health-related consequences of alcohol use. We studied a cohort of 8769 men 18 to 64 years old who used homeless shelters in Toronto, Ontario, at least once during 1995. The cohort was followed until the end of 1997. All shelter admissions were tracked using a citywide database. Deaths were ascertained using the Ontario death registry. Survival analyses were performed using Cox regression. For each month of follow-up, a time-dependent covariate modelled the hazard associated with homeless shelter use during that month compared to no homeless shelter use. In a multivariate model, we adjusted for age and for pattern of shelter use during the one year preceding entry into the cohort. Entry into the cohort was defined as the first shelter admission in 1995. Pattern of shelter use was classified as transitional, episodic, or chronic using a previously described cluster analysis technique. RESULTS:

The hazard associated with months during which shelters were used was consistent across age groups, and persisted when the analysis was restricted to men with little or no history of homelessness prior to 1995. For death due to homicide or suicide, the adjusted relative risk during shelter use months compared to months in which shelters were not used was 5.5 (95% CI, 1.8 to 16.5). An episodic pattern of homelessness prior to cohort entry was associated with a significantly elevated risk of death in univariate but not multivariate analyses. CONCLUSION: Among men who use homeless shelters in Toronto, the risk of death almost doubles during months of shelter use. This association could be due to factors such as increased substance abuse, exposure to violence or the elements, or homelessness-related exacerbation of pre-existing medical problems. Alternatively, some persons may enter the shelter system because of deteriorating health. PURPOSE: Although ethnic disparities in the utilization of joint replacement for osteoarthritis have been documented, the reasons behind these disparities are not clear. Since differences in patient attitudes about joint relacement may play a role in their utilization of joint relacement, older male AA and white veterans with chronic knee and/or hip pain were surveyed with respect to their attitudes toward joint replacement. METHODS: A cohort of 285 patients aged 50 or older (45% AA and 55%white) with moderate to severe knee and/or hip pain of more than six months duration were surveyed. Patients who had had joint replacement were excluded. Patients were asked whether they would consider surgery if their knee or hip pain became severe and the procedure was recommended by their physician. We also assessed their perception of the efficacy of joint replacement using a component from the Efficacy of Specific Treatments for arthritis (Bill-Harvey). Other measures we assessed included the severity of their arthritis (Lequesne Scale), arthritis-specific health status (WOMAC), comorbidity (Charlson), Fatalism (Powe) and educational level. RESULTS: AA and white patients were comparable with respect to age (67 Ϯ 9 vs. 68 Ϯ 9, p ϭ 0.53); severity of arthritis (mean score 11 vs. 12, p ϭ 0.09); arthritisspecific health status (mean score 17 vs. 18, P ϭ 0.25); comorbidity (mean score 2.6 Ϯ 2.1 vs. 2.6 Ϯ 2.1, p ϭ 0.99); Fatalism (mean score 2.1 Ϯ 0.5 vs. 2.0 Ϯ 0.6, p ϭ 0.63). AAs were less likely than whites to be high school graduates (54% vs. 66%, respectively, p ϭ 0.03). AAs were significantly less likely than whites to consider joint replacement as an option (crude odds ratio ϭ 0.42, 95% CI 0.21 to 0.83). This difference persisted after adjusting for age, severity, health status, comorbity, fatalism, educational level and other demographic, clinical, and psychosocial covariates (adjusted odds ratio ϭ 0.48 95% CI 0.23 to 0.99). In addition, AAs were more likely than whites to perceive joint replacement as " Not at all helpful" (14.6% vs 7%, P ϭ 0.006). CONCLUSION: In this cohort of AA and white male veterans with moderate or severe chronic knee and/or hip pain, AA patients were less likely than whites to consider joint replacement as an option for future therapy and they were also less likely to perceive joint replacement as being helpful for arthritis. These ethnic differences in patient attitudes may help to explain differences in the utilization of joint replacement for arthritis. , serious illness worry and mental disorders (PRIME-MD). Surveys immediately after the visit and at 2 weeks assessed patient satisfaction, residual serious illness worry and unmet expectations. Two-week assessment also included symptom outcome and functional status. Health utilization for 3 months was assessed using the DOD computerized health data system. RESULTS: Nearly all patients endorsed at least one pre-visit expectation. Eightyone percent hoped for a diagnosis, 63% prognostic information, 60% a prescription, 54% a diagnostic test, 45% a subspecialty referral and 7% an excuse from work. Immediately after the visit, 33% had an unmet expectation for an explanation of the symptom's cause and 51% an unfilled desire for an estimate of how long to expect the symptom to last. Eleven percent had a variety of other unmet expectations, including desires for diagnostic tests, subspecialty referrals, prescriptions, or sick slips. Patients who reported receiving diagnostic or prognostic information were more likely to be fully satisfied with their care (p Ͻ 0.001) were less likely to be worried their symptom was due to something serious (p Ͻ 0.001) experienced greater symptom (p Ͻ 0.01) and functional status improvement (p Ͻ 0.003) at 2 weeks and had lower 3 month utilization rates. (p ϭ 0.002) Residual expectations for a subspecialty referral or diagnostic testing were also associated with lower levels of satisfaction (p Ͻ 0.00001) and with continued serious illness worry (p Ͻ 0.001). CONCLUSION: Receiving diagnostic and prognostic information from physicians during encounters with symptomatic patients is associated with greater satisfaction, reduced serious illness worry, and greater symptom and functional status improvement by two weeks. PURPOSE: Domestic violence is a serious public health issue but has only recently become part of the curriculum in residency. Several studies have shown high incidence of violence against women among acute and primary care patients. The purpose of this study is to determine current indicience of violence among female clinic patients and gather information on related health problems in order to improve training in violence intervention for residents. METHODS: A self-administered 30-question survey was given to consecutive female patients at the internal medicine residents' clinic during a five day period. The survey was titled "Women's Health Survey" and questions on violence followed several demographic and general health questions. The women completed the survey in the waiting room or in the exam room. RESULTS: Of 264 patients approached, 169 complete surveys were returned (64%). When women were asked if they had ever been physically abused as a child, 28% responded "yes". When asked if they had ever been sexually abused as a child, 35% responded "yes". Physical or sexual abuse in adulthood was present in 47% of women. Physical abuse by a partner within the last year was present in 13% of women. Women with a history of abuse as an adult were significantly more likely to have had a miscarriage than those without an abuse history (55% vs. 32%). They were also more likely to have attempted suicide and visited an emergency room recently. There were no significant differences in age, education or income between those with a history of abuse and those without. CONCLUSION: The rate of current and past abuse among these primary care female patients is alarmingly high. The abuse may be contributing to increased use of health care services. Recognition of this problem is the first step to decreasing the incidence of abuse. Primary care residents should have ample opportunity to become well trained in recognition and initial intervention in violence against women. sistency by a health services researcher and a clinical social worker and further adjudicated as necessary. RESULTS: Four hundred thirty one distinct comments on spirituality were generated with 95% falling into 8 major categories (see Hebert et al). The category of physician-patient communication was particularly interesting to us (n ϭ 104). All group participants joined in describing a need for empathic communication from their physicians. The most common expression was the need for comfort. Patients mentioned prayer with their physicians, physical contact, eye contact, body language, and words that gave messages of comfort. Personality, trustworthiness, ability to give hope, genuine caring, and interest in patients' personal and family life were also used as descriptors. Patients with a strong sense of spirituality spoke of physicians who were able to give them comfort as spiritual. Patients demonstrated a yearning for this level of communication with their physicians. Several patients indicated willingness to break off care and move to a physician who could meet this standard, even in the middle of a serious illness, thereby disrupting their care. When physicians met their standard of spirituality, patients reported improved ability to cope with the burden of illness (medicines, tests, etc.) and to accept the outcome of their illness with peace. CONCLUSION: Our study further supports the therapeutic effects of empathic care giving by physicians. Patients describe empathic physicians as spiritual, indicating that this skill set is sacred to them. Furthermore, this particular aspect of interpersonal care defined through patients' eyes is important to patients' ability to cope with illness and remain in a therapeutic relationship with their physician. (1.83-11.29 )), mediastinal lymphadenopathy on chest radiograph (OR 3.56 (1.12-11.48)), expectoration (OR 0.29 (0.13-0.65)), atypical infiltrate (OR 0.38 (0.17-0.86)) and foreign-born status )) were the most significant predictors of TB. A score was developed which summed the risk factors for each patient by adding ϩ1 for each positive predictor and Ϫ1 for each negative predictor. Having a score of zero or higher had a sensitivity of 85% for predicting TB. CONCLUSION: A positive PPD, mediastinal lymphadenopathy, and foreign-born status are positively associated with TB in smear-negative patients while the presence of expectoration or an atypical chest xray infiltrate are negatively associated. The score of zero or more helps identify patients who are more likely to have TB and those who should be treated while awaiting culture results. Other predictor variables including tobacco use, vitamin use, and personal or family history of breast and/or gynecologic cancer were not significant. In addition, breast cancer or cervical cancer screening did not significantly influence the likelihood of having either sigmoidoscopy or colonoscopy. 43% of the women learned about colon cancer from the media and 18% from health care professionals. None of the patients attributed insurance restrictions as a reason for not having the screening tests done. Only 2 patients could identify a female celebrity who had colon cancer. CONCLUSION: In our study, women who agree to breast and cervical cancer screening are no more likely to have colon cancer screening than those who refuse. Despite an increased incidence of colon cancer in women with a personal history of breast and some gynecological cancers, women with these risk factors are no more likely to have colon cancer screening tests. Women with a family history of colon cancer and women who believe that colon cancer is an important health issue, however, are more likely to agree to colon cancer screening. We suggest that the media and health care professionals further stress the importance of colon cancer screening in women over the age of 50. The increased use of mammography screening has led to an enormous increase in the incidence of non-invasive breast cancer, ductal carcinoma in-situ (DCIS). Because modified radical mastectomy remains a frequent surgical treatment for DCIS there is widespread concern among some patients, advocacy groups and clinicians that breast conserving surgery is heavily underutilized among women with non-invasive disease. However, there have been no population-based studies to assess determinants of surgical treatment type, the process of decision-making or outcomes such as patient satisfaction for patients with DCIS. METHODS: We performed a population-based pilot survey of 183 women from the metropolitan Detroit SEER cancer registry catchment area diagnosed with DCIS and invasive breast cancer during a 5 month period in 1998. RESULTS: The response rate was 71.2% with 80% of women completing the survey within 4 months of diagnosis. Half the women had DCIS. Overall, 51.9% of women received a mastectomy (54.3% of women with DCIS vs 49.4% of women with invasive disease, p ϭ 0.34). Attending surgeons appeared to have played a dominant role in decision-making: One third of women did not perceive that they were given a choice between treatment options; an additional 38% of women received a surgeon recommendation (92% of whom received the recommended treatment). Patient attitudinal factors such as concerns about the clinical benefits and risk of surgical treatment options were important determinants of treatment choice and favored receipt of mastectomy. Knowledge about differences in clinical benefits and risks between surgery options was low: less than one third of women knew that survival was the same for the two treatments. Women who did not perceive choice would have "liked to have been more involved" (38% vs 9% of those who perceived choice, p Ͻ 0.1) and "did not have as much say as I wanted" (35% vs 11% of those who perceived choice, p Ͻ 0.1). These results were the same across stage of disease. CONCLUSION: Determinants of surgical treatment type appeared to be the same for women with DCIS and invasive breast cancer, with surgeons playing a dominant role in the decision-making process. Low patient knowledge of differences in CONCLUSION: These results suggest that women may be particularly vulnerable to the consequences of disability even if they live with a spouse. Programs providing home care support to the elderly need to consider these enormous gender disparities when developing targeting strategies in the community. , to select patients with acute myocardial infarction (AMI) most likely to benefit from tissue-plasminogen activator (t-PA) over streptokinase (SK). Our model suggested that, when patients are stratified, 85% of the lives saved with population-wide use of t-PA could be obtained by treating only 50% of patients. Also, when the incremental risk of intracranial hemorrhage (ICH) was included, our model suggested that patients in the lowest benefit quartile would actually be more likely to be harmed by t-PA relative to SK, so that treating half of patients with t-PA would be essentially "quality-neutral" to treating all, if patients are properly selected. We now attempt to validate this model directly on the population in the GUSTO trial, the landmark study that demonstrated that t-PA was, on average, superior to SK in reducing mortality in AMI. METHODS: Because the GUSTO database did not contain all the variables required by the TPI, we first derived a "Modified (GUSTO-compatible) TPI" on the original TPI database of 4,911 patients, and ensured that the new model's predictions had a very high correlation to the predictions of the original TPI (pearson correlation ϭ 0.98). "Modified-TPI" predictions were then obtained on 24,146 patients in the GUSTO trial database. Patients were categorized into quartiles of predicted mortality benefit and quartiles of predicted composite (mortality and intracranial hemorrhage [ICH]) benefit. RESULTS: Mortality and composite outcomes for patients treated with t-PA and SK are shown in the Table below , divided, respectively, into quartiles of expected mortality benefit (mort) and quartiles of expected composite benefit (comp). In terms of lives saved, 85% of the benefit of t-PA seen in the GUSTO trial was captured in the first two quartiles of mortality benefit, exactly as predicted. Also, when the risk of ICH was considered, a trend toward harm was observed in the lowest benefit quartile and, as predicted, treatment of all patients in the GUSTO clinical benefits and risks between treatment suggest that there is need to improve communication between surgeons and patients. Lower satisfaction among women who perceived less control suggests that the process of decision-making may be more important than the actual treatment chosen. The demand for home care for the disabled elderly has increased markedly in industrialized nations because age-specific disability has increased, populations are aging and more disabled elderly are living alone. This is of particular concern to women who represent a large proportion of the elderly disabled in these countries and may be vulnerable to unmet need because they live alone or they may be caregivers themselves. Yet, we have little information about gender differences in receipt of informal home care in the United States. METHODS: To illuminate this issue we used the first wave of a large nationally representative longitudinal survey of the elderly (AHEAD, 1993, N ϭ 7,443) to assess gender differences in receipt of unpaid (informal) home care to the disabled aged 70 and older. We used a two-part regression model to quantify the number of hours per week of informal home care after adjusting for age, level of disability and living arrangments. RESULTS: Forty two percent reported one or more current ADL or IADL impairments (N ϭ 3109, mean number of impairments ϭ 2.9 out of 11) of whom 65.6% were women. The Table shows the number of informal care hours per week by gender and living arrangement adjusted for age and disability. Disabled women received much less informal home care than their male counterparts because they were much less likely to be married and married women received many fewer hours of care than married men. Indeed, disabled married women received only slightly more care than disabled women living alone; while disabled married men received three times more care than disabled men living alone.

trial would be essentially no better than limiting treatment to the 50% of patients with the highest predicted benefit. CONCLUSION: Using multivariate risk/benefit models within research protocols can help uncover patient subgroups likely or unlikely to benefit. If such models are made clinically available, clinicians could better target therapies with significant costs or risks. CONCLUSION: It appears that heparin is the major contributing factor for vascular complications after PTCA. Addition of IIb/IIIa inhibitors does not seem to increase the risk.In fact it will allow use of lower dose heparin, ultimately decreasing the incidence of vascular complications.The cost of IIb/IIIa inhibitors remains a limiting factor for their routine use. we recorded a total of 432 symptoms. Patients were randomly selected from 4 different months to avoid seasonal bias. We used explicit criteria to assign symptom etiology (physical, psychiatric, or idiopathic) and a 30% sample of the charts were co-audited to minimize investigator bias. Interrater agreement was high (kappa ϭ 0.75 for etiologic category, and 0.84 for outcome). Symptom type, visit type, demographics, provider relationship (new vs established), treatments, tests, referrals, and final diagnosis were examined. Follow-up notes for one year after the index visit were reviewed to determine chart-documented outcomes. RESULTS: Patients were 73% women, 59% African-American, with a mean age of 56. Various joint and URI complaints were the most common followed by the following specific symptoms: back pain (n ϭ 42), headache (27), dyspnea (26), abdominal pain (19), and chest pain (19). In two-thirds of the cases, symptoms noted in the chart were the principal reason for the visit. Prescribing a new (39%) or continuing an old (32%) medication was the predominant symptom-specific physician action, while ordering tests (19%), referrals (11%), or nonpharmacological treatments (10%) were less common. Using explicit criteria, 53% of the symptoms were classifed as physical in etiology, 10% as psychiatric, and 37% as idiopathic. Back pain and headache were the most likely to be idiopathic (71% and 48%). A psychiatric cause was somewhat more common in younger, white patients. Investigators had the highest diagnostic certainty for physical causes and least with idiopathic at both initial and final visits. At the index visit, nearly half (45%) of the symptoms were chronic, recurrent, or persistent since a prior visit. In follow-up notes, symptom outcome was documented as improved in 19%, the same or worse in 25%, and not mentioned in 56%. Poor outcome (i.e., same/worse) was associated with longer symptom duration at presentation, younger age, and psychiatric etiology. Outcome was unrelated to specific provider actions. Serious diagnoses (e.g., cancer) not suspected at the index visit did not emerge during follow-up. CONCLUSION: Nearly half of physical symptoms documented in primary care lack a clear-cut physical etiology. While serious physical causes not suspected initially seldom become manifest at follow-up, symptom persistence is an important concern in at least one out of four patients. Better management strategies for such patients is a priority. ) was calculated for each participant using both the LDL and total cholesterol (TC) scoring tables. FRSs were categorized separately for men and women so that high risk (HR) equaled an equivalent annual CHD risk (Ͼ1%/yr). Two-stage (probit/OLS) regression was used to predict SCA (any CAC) and extent of clinically significant SCA (CACϾ80). ROC areas (reported as the c-statistic from logistic regression) were used to determine the ability of EBCT to discriminate HR determined by calibrated FRS, and the ability of FRS to identify clinically significant SCA. RESULTS: Similar groups were classified at HR by FRS (22%) and EBCT (19.25%). Among men, 10% of those at low risk (LR; Ͻ.5%/yr, n ϭ 240) by FRS had clinically significant SCA by EBCT; 29.7%% of HR men (n ϭ 202) had no SCA as did 15% of HR men aged Ͼ60 (n ϭ 64). Among FRS HR women (n ϭ 92), 47% were without SCA, as were 20% aged Ͼ65 (n ϭ 20). Over 15% of both men and women with CHD risk Ͼ2%/yr (n ϭ 60) were without SCA. Using probit regression, LDL, HDL, age, gender, diabetes, and hypertension predicted the presence of SCA, smoking did not (pseudo R 2 ϭ .13). Using OLS for those with any SCA, age, gender, diabetes, smoking and hypertension predicted the extent of SCA; lipids did not (r 2 ϭ .145). To identify CHD risk Ͼ10%/10-yrs, EBCT had moderate discrimination (ROC areas ϭ .70 men, .66 women PURPOSE: Most arthritis patients are unaware of the benefits of exercise, and chronic arthritis is commonly known to cause pain and limit daily activity. Exercise demonstrations and interactive participation will heighten patient awareness of the benefits of participating in an exercise program. This session will provide a practical way for the patient to learn methods to alleviate pain and discomfort caused by arthritis. Hence, improving their quality of life through life-style modifications. METHODS: In a period of two years (1997) (1998) (1999) , 94 male veteran patients with a variety of arthritis conditions (46 osteoarthritis, 23 rheumatoid arthritis, 7 gout, 4 spondylolisthesis, 2 hemoarthrosis, 3 osteopenia, 9 low back pain) were asked to voluntarily attend a 20-minute interactive exercise demonstration conducted by a licensed physical therapist in a clinic waiting area. Subjects, age range 40-80 years, were recruited from the arthritis clinic while waiting to be seen by a physician. The importance of movement and flexibility were explained to the participants. The exercise sessions consisted of thera-band resistive strengthening exercises, range of motion exercises combined with Tai Chi movements, other general conditioning and energy conservation techniques. After each demonstration, the subjects filled out a questionnaire regarding the effectiveness of the demonstration. One-week later phone surveys were conducted to assess the frequency of use and benefits of the demonstration. RESULTS: The results are as follows:

CONCLUSION: These interactive sessions demonstrated a convenient and practical form of exercise to the subjects. This could boost subjects' self-confidence and knowledge to exercise on their own. Telephone surveys indicated that there is dissemination of knowledge about exercise and arthritis among the participants and their peers. This interactive, user friendly, economical exercise technique is an effective and worthwhile approach for educating individuals about arthritis and pain management.

A NEW MEASURE OF DEPRESSION SEVERITY: THE PHQ-9. K Kroenke, RL Spitzer, JB Williams, Regenstrief Institute, Indianapolis, IN; Columbia University, New York, NY PURPOSE: While considerable attention has focused on diagnosing depression in primary care, it increasingly has been noted that monitoring treatment response and adjusting therapy is equally important. Therefore, we have developed and validated a simple depression outcome measure in two large prospective studies. METHODS: The Patient Health Questionnaire is a recently validated, self-administered version of the PRIME-MD diagnostic instrument for mental disorders in primary care. Its depression measure includes the nine DSM-IV criteria for mood disorders, each item scored from 0 ("not at all") to 3 ("nearly every day"). The PHQ was administered to 6000 patients in 7 general internal medicine and family practice clinics (n ϭ 3000), and 7 obstetrics-gynecology clinics (n ϭ 3000). Internal reliability was good in both samples (Cronbach's alpha ϭ .90 and .86). Functional status was assessed with the SF-20, and sick days and clinic visits during the preceding 3 months by patient self-report. A sample of 580 patients had a telephone-administered structured psychiatric interview by a mental health professional who was blinded to the results of the PHQ. RESULTS: Data for the 3000 general internal medicine/family practice patients is shown in the table. As PHQ-9 depression severity increased, there was a substantial, stepwise decrement in functional status for all six SF-20 subscales. Also, sick days and health care utilization increased. Using the MHP reinterview as the criterion standard, a PHQ-9 score of 10 or greater had a sensitivity of 88% and specificity of 88% for major depression. Analysis by type of depressive disorder-major (10% of sample), other (6%), and none (84%)-revealed that PHQ-9 thresholds of 5, 10, 15, and 20 represent minor, moderate, major, and severe depression cutpoints. Findings were essentially the same for the Ob-Gyn sample, thus providing external validity.

CONCLUSION: A validated diagnostic tool, the PHQ-9 also has considerable promise as an outcome measure in the treatment of depression. It is currently being used in several large treatment trials which will be important to demonstrate its sensitivity to change. regarding pharmacologic stress tests and their listed references were retrieved using the search strategy "coronary disease/diagnosis AND (pharmacologic stress);" experts were consulted to further identify studies. Study Selection: Studies were included if all subjects underwent at least one echo or SPECT pharmacologic stress test with adenosine, dipyridamole, or dobutamine for diagnosis of CAD; all subjects underwent coronary angiography; and data was presented in a manner that allowed the calculation of sensitivity and specificity of the tests. Data Extraction: Two independent reviewers abstracted population characteristics, technical factors, methodological factors, and results. Disagreements were resolved by conference. RESULTS: Ninety-seven studies met inclusion criteria. Adenosine echo and dipyridamole echo had similar sensitivities, 72% (95% confidence interval [CI], 68%-77%) and 71% (95% CI, 68%-73%); and specificities, 92% (95% CI, 86%-95%) and 93% (95% CI, 90%-95%). Likewise, adenosine SPECT and dipyridamole SPECT studies had similar sensitivities, 90% (95% CI, 88%-92%) and 87% (95% CI, 85%-89%); and specificities, 76% (95% CI, 71%-82%) and 70% (95% CI, 66%-73%). The dobutamine echo and SPECT studies had similar sensitivities, 80% (95% CI, 79%-82%) and 82% (95% CI, 79%-85%); but dobutamine echo had a higher specificity, 83% (95% CI, 81%-85%), than dobutamine SPECT, 74% (95% CI, 70%-79%). CONCLUSION: Adenosine and dipyridamole have similar sensitivities and specificities as pharmacologic stressors for the diagnosis of CAD. The use of SPECT with adenosine or dipyridamole offers higher sensitivity but lower specificity than the use of echo. In dobutamine studies, SPECT and echo offer similar sensitivities, but the specificity of echo is higher. To determine the long term smoking cessation rate and the predictors of cessation after outpatient group counseling in an urban public health care organization. METHODS: Smokers (n ϭ 156) who attended the first of a series of four free outpatient stop smoking group classes at an urban public hospital or at either of two associated community health clinics between August 1996 and July 1997 were surveyed using a written questionnaire. A follow-up survey by phone was attempted for all individuals ( 64% response rate, average length of follow-up ϭ 27 months). Results from these surveys along with class attendance and information from a pharmacy database were used to determine the self-reported abstinence rate and to identify factors predictive of abstinence. RESULTS: Features of the initial cohort include 62% women, 40% disabled, 50% minority, average age 48 years, and an average habit of 25 cigarettes per day. Participants attended an average of 3.4 of the 4 classes. 77% filled a prescription for transdermal nicotine at an agency pharmacy. However, only 22% returned for all the patch refills. Of the 101 patients who were contacted, 17 reported no current smoking. Factors associated with self reported abstinence at 2 years using multiple logisitic regression include high self efficacy to quit at 6 months (OR 10.7, p ϭ .05) and age greater than 65 years (OR 8.2, p ϭ .06). None of the following factors correlated with abstinence: stage of change, nicotine dependence, gender, race, use of nicotine patches, class location, cigarettes per day, employment status, marital status, or insurance status. CONCLUSION: The long-term self-reported abstinence rate among our low-income urban participants in group counseling compares favorably with results from the published literature. Similar to our inpatient follow-up program, a simple 10point linear scale of self-confidence to quit at 6 months was highly predictive of abstinence. , and level of social support (unmarried living alone, unmarried living with others, married). RESULTS: About 10% of respondents showed evidence of cognitive impairment consistent with dementia. There was a strong positive association between dementia severity and weekly hours of family caregiving (p Ͻ .01) (Table) .

Using the average hourly wage of a home health aide in 1997 ($8.30) as the value of a family caregiver's time, the expected yearly caregiving cost per case ranged from $5,800 for Mild Dementia to $17,700 for Severe Dementia. This represents a total yearly cost in the US of about $17 billion for dementia-related informal caregiving. CONCLUSION: Family caregiving time and costs for those with dementia are substantial and increase sharply with the severity of cognitive impairment. Accurate estimates of caregiving costs are, therefore, essential to the validity of future evaluations of interventions that decrease the incidence or slow the progression of dementia. We present frequency distributions for each outcome over time and report tests for trend. Outcomes included: estimates of six-month survival; functional status; occurrence of severe physical and emotional symptoms; patients' preferences for care; and the financial impact of patients' illnesses on their families. RESULTS: As death approached, patients' prognosis became poorer and illness more severe. Median Acute Physiology Scores for hospitalized patients rose from 33 in the interval 6 months to 3 months prior to death, to 44 within 3 days of death. However, the median model-based estimate of six-month survival was 54% even within 3 days of death. Functional impairment, measures of depression and the percent of patients reporting severe pain or dyspnea increased as death approached. Perceived quality of life did not change appreciably, with 29% to 58% of patients reporting good to excellent quality of life in all intervals prior to death. As death approached, the percent of patients preferring not to be resuscitated (DNR) rose from 33% 6 months to 3 months prior to death, to 47% 1 month to 3 days before death (p Ͻ .05), and the frequency with which DNR orders were written for hospitalized patients also increased. The patients' illnesses had marked financial impact on their families, with 23% of patients' families reporting the loss of most or all of family savings at the time of the patient's death. CONCLUSION: During the last six months of life in CHF, illness becomes more severe, disability and severe symptoms more prevalent, and patient preference not to be resuscitated more common. However, there is no significant decrement in quality of life as death approaches. The unpredictable course of CHF is reflected in the finding that, during their last 30 days of life, many patients retain good median model-based six-month prognoses and enjoy good to excellent quality of life. The advent of non-surgical means for placement of gastric feeding tubes resulted in a significant reduction in the risk of this procedure and a steady increase in utilization; however, mortality for patients receiving these feeding tubes remains high. The purpose of this study is to determine the expected benefits of gastric feeding tube placement from the patient and surrogate perspectives. METHODS: We interviewed patients and/or surrogates who had an initial gastric feeding tube placed to determine their perceptions of the benefits of feeding tube placement and of the decision-making process surrounding the placement. Patients older than 21 were identified by daily procedural logs of interventional radiology and the gastroenterology suites at a major teaching hospital and a community hospital. Patients were interviewed by telephone or in person within a month of the feeding tube placement. Surrogates were interviewed if patients were unable to be interviewed because of cognitive impairment or intubation. Patients with diagnoses of trauma or non-head and neck malignancy were excluded. RESULTS: To date 159 patients were eligible and 24 patients and 68 surrogates have been interviewed (57%). The mean age was 66 years, 66% were white, and 49% were women. Major diagnoses were: 28% stroke, 19% neurodegenerative disorders, 16% head and neck cancer, and 37% other serious medical illness. Most patients/surrogates thought that the gastric feeding tube would improve their nutrition (99%), and make them more comfortable (72%). A majority also thought that it would allow them not to feel hungry or thirsty (68%) and to have less pain (60%). Many thought the gastrostomy would provide greater independence (54%), and almost all thought it would prolong life (99%). Although most reported that it would improve quality of life (91%), there were no differences when respondents rated the quality of life on a 1-10 scale before the gastric feeding tube was placed (4.9) and after it was placed (5.0). Respondents felt that they had gotten adequate information to make the decision whether to get the feeding tube (98%) and had no regrets about the decision (98%). CONCLUSION: Respondents reported that they had been adequately informed about the feeding tube. Patients and surrogates had similarly high expectations regarding the benefits to be derived from gastric feeding tubes, many of which may not be realized. Follow up interviews are in progress to determine if these expectations change with time and experience with the gastric feeding tube. PURPOSE: Ninety percent of community physicians in the United States report prescribing an antibiotic to smokers with acute bronchitis. We performed a systematic review of the literature to determine the efficacy of antibiotic use in smokers with acute bronchitis without underlying chronic lung disease. METHODS: We searched MEDLINE using the subject headings "bronchitis," cough," and "antibiotics" to identify English language trials, reviews, letters, and editorials published between January 1966 and October 1999. Titles and abstracts were screened for suitability and those deemed appropriate were retrieved and reviewed. Reference lists of articles were used to identify additional appropriate sources. RESULTS: One hundred and six articles were initially retrieved and reviewed, yielding 10 randomized placebo-controlled trials and 19 randomized active-controlled trials. None of the studies specifically addressed antibiotic use in smokers with acute bronchitis. The 10 randomized, placebo-controlled trials tested various antibiotics-demethylchlortetracycline (1 trial), trimethoprim/sulfamethoxazole (1 trial), doxycycline (4 trials), and erythromycin (4 trials)-in both smokers and non-smokers with acute bronchitis. Six of the ten trials, representing over 310 smokers, reported no benefit to antibiotics versus placebo in smokers. In one of these six trials, a benefit was found for non-smokers. Three trials, representing 160 smokers, reported decreased duration of cough, feeling of illness, congestion, or days off work for antibiotic-treated patients. One trial gave no information about smoking status. Meta-analysis of these results was not possible due to (1) data not uniformly reported or available by smoking status, and (2) different outcomes among trials, including duration of cough, sputum production, fever, days off work, general feeling of well-being, and resumption of normal activities. Of the 19 trials comparing two or more antibiotics only three reported baseline information about smoking habits, and none reported outcomes stratified by smoking status. CONCLUSION: Although antibiotic prescription for smokers with acute bronchitis is common, this practice has not been rigorously evaluated and the evidence from placebo-controlled trials suggests it may be ineffective. CONCLUSION: This study reveals that patients prefer physicians to assume more of the decision making power for colorectal cancer screening though patients want some role in this process. Differences exist between certain patient groups for which physicians should be specifically sensitive. Physicians should assess each of their patients' desire to participate in this decision making process and act accordingly. This will allow for the accommodation of each individual patients' preferences and needs. were surveyed in late 1997 about religious involvement and practices. Those questions were used to create a 5-item measure of religiosity (alpha ϭ 0.84) and a 4-item measure of spirituality (alpha ϭ 0.84). We examined associations of religiosity and spirituality with demographics, acute stressors, and circumstances of HIV testing and care. Significant demographic variables were entered with religiosity and spirituality in multivariate models of recent viral load, CD4 count, symptom and disease incidence. RESULTS: Many patients indicated that religion (42%) and spirituality (59%) were very important in their lives. In bivariate analyses, we noted significant positive relationships between religiosity (REL), spirituality (SPI) and need for income assistance (pREL ϭ 0.002, pSPI ϭ 0.004), housing (pREL ϭ 0.002, pSPI ϭ 0.001), financial aid with rent (pREL ϭ 0.008, pSPI ϭ 0.005), and need for home delivered meals or food bank assistance (pREL ϭ 0.036, pSPI ϭ 0.005). Circumstances of initial HIV testing and time to care did not differ by religiosity or spirituality, but individuals most anxious about revealing their HIV diagnosis to family (pREL ϭ 0.002, pSPI ϭ 0.001) or friends (pREL ϭ 0.000, pSPI ϭ 0.038) demonstrated higher religiosity and spirituality as did nonwhite race (pREL ϭ 0.000, pSPI ϭ 0.000), residents of regions other than the Western United States (pREL ϭ 0.000, pSPI ϭ 0.000), and non-homosexuals (pREL ϭ 0.000, pSPI ϭ 0.000). Age, gender, race/ethnicity, region, sexual orientation, and exposure to HAART were entered into multivariate models regressing religiosity and spirituality on clinical variables. A small significant relationship between higher spirituality and increased symptom incidence (beta ϭ 0.0067, p ϭ 0.009) and disease incidence (beta ϭ 0.002, p ϭ 0.000) was noted. CONCLUSION: Religious belief and practices are common among patients with HIV infection. Cross sectional comparison demonstrated no consistent relationships between religiosity, spirituality, and clinical status. A small significant relationship was noted between spirituality and increased symptom and disease incidence; relationships between acute stressors, religiosity and spirituality also suggest that religiosity and spirituality are important resources for people facing HIV infection.

M Louis-Simonet, A Rajeswaran, D Carballo, PF Unger, AF Junod, FP Sarasin, Dpt of Internal Medicine, Geneva University Hospital, Geneva, Switzerland PURPOSE: To measure prospectively in primary-care patients the diagnostic yield of a standardized sequential workup of patients with syncope. METHODS: All consecutive patients presenting in the emergency department with syncope as a chief complaint were included in the study. Stepwise workup included: 1) a complete history, physical examination and 12-lead ECG in all patients; 2) targeted test(s) when a specific entity was suspected by suggesting signs and/or symptoms; and 3) extensive testing (i.e., 24 hrs Holter, ambulatory loop recorder, echocardiography, upright tilt test and electrophysiologic studies in selected cases) for patients in whom the cause of syncope remained undetermined after the initial steps. Diagnostic criteria were based on existing literature. RESULTS: Among all patients (n ϭ 650) included over a two-year period, the cause of syncope was established after the initial evaluation in 452 (69.5%): vasovagal 37% (n ϭ 240), hypotension (drug or hypovolemia-related, postprandial and idiopathic) 24% (n ϭ 158), arrythmias 3.5% (n ϭ 23), other cardiac 1% (n ϭ 7) and others 3.5% (n ϭ 24). In 47 patients (7%) out of 66, targeted test(s) confirmed symptoms-based suspected diagnosis: aortic stenosis (n ϭ 7), pulmonary embolism (n ϭ 7), seizures (n ϭ 22), stroke (n ϭ 8), and others (n ϭ 3). Extensive workup was completed in 110 of the 151 remaining patients. A cause for syncope was established only in 27 of them: cardiac 2.5% (n ϭ 16), neurocardiogenic 1.5% (n ϭ 11), the remaining patients (n ϭ 83, 14%) being considered as having syncope of unknown etiology. CONCLUSION: In unselected patients, the diagnostic yield of a noninvasive workup for syncope, including targeted testing for patients with suggesting signs and/or symptoms, was higher (77%) than currently reported. PURPOSE: Over 500,000 elective joint replacements are performed annually, making it one of the most common surgeries in the United States. Hip and knee replacement improves mobility, decreases pain and improves quality of life for patients with severe large-joint arthritis. Previous studies demonstrate that a patients' tangible and intangible supports ("social support") play an important role in rehabilitation and recovery. The relationship of social support to other healthrelated outcomes is poorly described. The purpose of this study was to examine the association between social support and hospital length of stay (LOS), cost of care, and health-related functional status. METHODS: We surveyed 883 consecutive patients undergoing elective joint replacement at two suburban Indianapolis hospitals between January 1, 1998 and March 31, 1999 . Patients were interviewed by telephone prior to surgery and 6 months post-surgery. The survey included reliable and valid measures of sociodemographic factors, comorbid conditions, total social support at time of surgery (MOS-20), general physical health and emotional function (SF-12 PCS and MCS), and lower extremity joint-specific pain and function (WOMAC). Hospital LOS and actual costs were obtained through the hospital's administrative database. RESULTS: Mean age of the cohort was 67.4 years (range 21-100), 62.8% were women, and 90.8% were Caucasian. Elective knee replacement was performed on 62.4% and hip replacement on 37.6%. The average total social support was 88.7% (range 0-100). Subjects with higher social support were more likely to be undergoing knee replacement, a non-smoker, reside in a non-urban area, and discharged to home following surgery. Median hospital LOS was 4 days, and median hospital costs were $6,418. Using linear regression, controlling for demographic and behavioral variables, procedure, comorbidities and baseline function, increasing social support was associated with a shorter LOS and lower hospital costs (both p Ͻ 0.05), and greater improvement in PCS score (p ϭ 0.07). Social support was not associated with joint-specific outcomes. CONCLUSION: Patients undergoing elective joint replacement have high levels of social support, and greater social support is associated with improved outcomes and lower hospital costs. Whether interventions targeted to patients with lower pre-and post-operative social support will improve outcomes requires further study.

College of Medicine, Bronx, NY PURPOSE: Colon cancer screenings are lower among ethnic minorities contributing to inequalities in cancer mortality. Barriers to cancer screenings need to be explored to reduce health disparities among ethnic groups. METHODS: Participants (n ϭ 172) in an urban, ethnic diverse community were recruited after local church and community events and invited to participate in telephone interviews. The questionnaire focused on knowledge of cancer screenings and various health beliefs (misconceptions, risk perceptions, cancer-related anxiety) based on medical literature and beliefs elicited from the community. RESULTS: Participants mean age was 54.7, with 58% (n ϭ 110) above the age of 50 (mean age: 63, range: 50-84), of which 48% were African-American, 76% were female, 50% had less than a college education, 49% were employed (43% retired, 2% disabled), and 97% had medical insurance. The reported sigmoidoscopy screening rate (37%) was higher when compared to the national average, although only 27% had the appropriate colon cancer screening within the last 5 years. Univariate analyses showed that gender, age and education was not associated with sigmoidoscopy use, however African-Americans and employed participants were less likely to have had a sigmoidoscopy when compared to Whites (71% vs. 51%, p ϭ 0.034) and non-employed participants (72% vs. 50%, p ϭ 0.023) respectively. Knowledge about cancer screenings and anxiety of cancer showed no association with reported sigmoidoscopy use. However, participants who believed that their personal risk of cancer was lower than that of others (81% vs. 54%, p ϭ 0.014) and those who held the misconception that surgery spreads cancer were less likely to have had a sigmoidoscopy (71% vs. 49%, p ϭ 0.034) compared to those who did not hold these beliefs. Multivariate logistic regression analysis of all significant univariate variables revealed that ethnicity (OR ϭ 2.5, CI:1.03-6.18), personal risk perception (OR ϭ 4.6, CI: 1.48-14.4 ) and misconception about surgery (OR ϭ 2.5, CI:1.01-6.11) were significant predictors of self-reported prior sigmoidoscopy use. CONCLUSION: This research suggests that educational interventions could benefit from focusing on cancer risk perceptions and misconceptions in an ethnically appropriate manner to increase sigmoidoscopy screenings. and corroborated by an independent investigator. Patients again completed the AQLQ during in-person interviews during interval follow-ups. Asthma resource utilization was measured by patient report every three months by telephone. RESULTS: A total of 212 patients was enrolled, mean age 41 Ϯ 11, 83% women, mean follow-up 10.0 months. Surprisingly, 35% of patients expected to be cured of asthma. In multivariate analyses, patients expecting a cure were more likely to be currently using inhaled corticosteroids (p ϭ .004), to have never required oral corticosteroids (p ϭ .02), and to have lower asthma self-efficacy (p ϭ .008).

Outcomes: In follow-up analyses of within-patient change in AQLQ scores, patients expecting a cure did not improve by a clinically important difference in any domain, while patients not expecting a cure improved by a clinically important difference in all domains. Also, more patients expecting a cure declined in various AQLQ domains and in the overall score (p ϭ .03), compared to patients not expecting a cure. In multivariate analyses, only older age (p ϭ .02) and expecting a cure (p ϭ .03) remained predictors of worsened function. Patients expecting a cure also had more hospitalizations, emergency room visits and non-routine office visits for asthma compared to patients not expecting a cure (1.8 Ϯ 5.0 versus 1.3 Ϯ 2.8). In addition, patients expecting a cure had a stepwise increase in requiring at least some urgent resource utilization for asthma (52% versus 39%) compared to patients not expecting a cure. CONCLUSION: Surprisingly, 35% of patients had the unrealistic expectation of being cured from asthma treatment. Expecting a cure predicted functional decline and more frequent asthma resource utilization. Addressing unrealistic expectations that are potentially modifiable may help to improve short-term asthma outcomes. We assessed glycemic control in Russian-speaking only patients from the same health care system before and after the employment of a bilingual, Russian-born physician who was medically trained in both countries. METHODS: We conducted a retrospective chart review (1997) (1998) (1999) to compare the values of hemoglobin A1c, as an indicator of glycemic control, in our monolingual, Russian-speaking patients who, prior to October 1997 were seen by Englishspeaking physicians at Denver Health, with the values after they began to receive treatment from a bilingual, Russian physician. All of the patients were treated on the main campus of Denver Health Medical Center, and the chart reviews were performed by the physician primary and co-primary investigators. RESULTS: Twenty-two Russian-speaking diabetics were evaluated for glycemic control. Their average age was 67.0 (11 males and 11 females). Initial mean A1c measurements were 9.23 Ϯ 1.44 with a minimum value of 6.7 and a maximum of 12.5. The second A1c measurements, obtained approximately eight to twenty-four months after the initial values, were a mean of 7.9 Ϯ 0.86 with a minimum value of 6.9 and maximum of 10.0. The mean decline in A1c was Ϫ1.3 Ϯ 1.0 with the greatest decline in A1c values of Ϫ3.2 (Ϫ3.2 to ϩ0.7). A p-value for the t-test of the mean change was significant at p Ͻ 0.0001. CONCLUSION: The significant decline in HgbA1c from the first to second measurement indicates improvement in glycemic control of the Russian-speaking population likely due, in part, to language and cultural concordance between these new emigres and their primary care provider. Further evaluation will be necessary to define the optimal means of caring for new immigrant patients. In multivariable regression analyses among PAD patients only, ABI (regression coefficient ϭ 608 kcals/1 unit ABI, p Ͻ 0.01) and walking velocity over four meters (regression coefficient ϭ 996 kcals/1 m/sec, p Ͻ 0.01) were associated independently with PA, adjusting for age, sex, race, body mass index, exertional leg symptoms, and comorbid diseases. CONCLUSION: Higher physical activity levels are associated with a higher prevalence of exertional leg pain in PAD. Both PAD severity, as measured by ABI, and walking velocity over four meters are associated independently with objectively measured PA in PAD patients. Further study is needed to determine whether slower walking velocity precedes decline in PA or whether slower walking is a consequence of reduced PA. proportion surviving at 5 years was .63 for patients scoring ϭ Ͻ50 on the physical scale compared to .79 for those scoring Ͼ50 (p ϭ .000). For the frequency scale, the cumulative proportion surviving was .63 for patients scoring ϭ Ͻ50 and .70 for those scoring Ͼ50 (p ϭ .005). These differences remained significant when stratified by age and co-morbid conditions. Differences in survival were not significant for the disease perception and stability scales. In a multivariate model including physical function, age, and comorbid conditions, the relative risk of death was 2.0 (95% CI ϭ 1.62-2.46) for patients scoring ϭ Ͻ50 compared to those scoring Ͼ50 ( p ϭ .000). In a model including angina frequency, age, and comorbid conditions, the relative risk of death risk was 1.4 (95% CI ϭ 1.13-1.76) for those scoring ϭ Ͻ50 (p ϭ .003).

CONCLUSION: Lower quality of life, as measured by the SAQ physical function and frequency scales, may be a significant risk factor for all-cause mortality in veterans with CAD. QOL as measured by the disease perception and symptom stability scales did not significantly relate to risk of death. We report on the results of recruitment to a two-armed randomized trial to test an innovative intervention to increase rates of breast and cervical cancer screening among older African-American and Latina women. The intervention relied on younger women, who typically use preventive services more often than do older minority women, to encourage the older women to seek screening. METHODS: Both study recruitment and the intervention were conducted by outreach workers (OW) familiar with study neighborhoods. Younger women (ages 30-45) were asked to nominate and work with at least one older African-American or Latina woman (50 years and older). Recruitment was conducted at sessions randomized in advance to treatment or control conditions. OW were blinded to the randomization until just prior to a session. In treatment sessions, OW explained the study and taught the younger women how to intervene with the older women.

In control sessions, the OW explained the study. One week after recruitment, the nominators (younger women) in both study conditions were contacted to determine the older woman's willingness to participate. Also, in the intervention condition, nominators were contacted six weeks later to reinforce their work encouraging the older nominee to get an exam. RESULTS: A total of 151 recruitment sessions were held over a 22-month period. They were conducted at health fairs, in employment training, housing or energy assistance offices, at churches, hair salons, and health clinic waiting rooms and door-to-door. A total of 549 nominators were willing to contact an older woman to participate in the study, but only half were able to obtain agreement from their nominees. Fifty-eight percent of the older women nominated by the younger women in the control condition agreed to participate, compared to 48% in the intervention condition. After 6 weeks, 75% of the intervention group nominators reported talking with the older women about the screening exams. CONCLUSION: This unique intervention represents an effort to create a community-based approach to contact women who can be difficult to reach for preventive screening exams. Recruitment to the study required considerable effort with outreach in a number of different recruitment sites. Recruitment rates were slightly higher in the control condition which demanded less of the participants.

PS Mehler, S Biggerstaff, Denver Health Medical Center, Denver, CO

INTRODUCTION: There is a markedly increased risk of cardiovascular mortality in type 2 diabetic patients. Inadequate achievement of published guidelines for the control of hyperlipidemia may be contributing to this result. PURPOSE: The purpose of this study is to assess the degree of lipid control, longitudinally over five years, in a cohort of type 2 diabetic patients. METHODS: Nine hundred fifty (950) type 2 diabetic patients enrolled in the Appropriate Blood Pressure Control in Diabetes (ABCD) trial had fasting blood samples obtained to determine low-density lipoprotein (LDL) cholesterol, total cholesterol, high-density lipoprotein (HDL) and triglycerides levels. Temporal trends in the degree of lipid control were analyzed from five years of the study period. Subgroup analysis in patients with concurrent coronary disease was also performed to determine the degree of lipid control in this population. Multivariate logistic regression analysis was used to evaluate for potential factors associated with different degrees of lipid control. RESULTS: The mean baseline total cholesterol levels in these type 2 diabetic patients at the initiation of the study was 218 mg/dL and the LDL level was 130 mg/ dL. Only 18.15% of these patients had a LDL level less than 100 mg/dL at baseline. After five years of being enrolled in the study this number remained unchanged (18.44%). The percentage of patients achieving a LDL level less than 130 mg/dL also did not improve from baseline through the end of the study ( RESULTS: Twenty two of 57 subjects (39%) had a feeding tube placed during their hospitalization. Risk factors for feeding tube placement by logistic regression included poor oral intake while hospitalized (odds ratio of 8.92, 95% confidence interval 2.2-40.0) and being African American or Latino (odds ratio of 5.34, 95% confidence interval 2.1 to 37.0). Previous hospitalizations for pneumonia, a history of weight loss, dementia stage, presence of a pressure ulcer, absence of an advance directive, pre-hospitalization residence (nursing home vs home), or dementia stage did not predict tube feeding. Median survival following hospital admission in subjects receiving a feeding tube was 195 days (range 21 to 1405 days) as compared to 189 days in subject who did not receive a feeding tube (range 4 to 1502 days). Tube feeding was not associated with survival in a Cox proportional hazard model (p ϭ 0.9) CONCLUSION: End stage dementia patients hospitalized with an acute medical illness are at high risk for placement of a gastrostomy tube. Risk factors for tube placement include poor oral intake while in the hospital and being non white. End stage dementia is associated with a median survival of 6 months following hospitalization for an acute illness. Tube feeding does not appear to appreciably prolong survival.

DEPARTMENT ABDOMINAL CT SCANS. WA Messersmith, DFM Brown, MJ Barry, Internal Medicine; Emergency Medicine, Massachusetts General Hospital, Boston, MA PURPOSE: As CT scanning replaces other more focused tests (such as IVP) for common urgent problems such as renal colic, incidental findings are more common. We examined the prevalence and implications of these "incidentalomas," defined as findings unrelated to the original purpose of the scan which may either go unnoticed, or raise concern and generate further diagnostic work-up. METHODS: We reviewed the reports from 321 consecutive emergency department (ED) noncontrast, helical "renal stone" abdominal CT scans obtained between 4 ⁄ 96 -6/97 for the presence of incidental findings. These findings were then subdivided by two independent reviewers into 2 groups based on the level of concern they raised for serious pathology. The hospital records of these patients were then examined to see whether these findings were followed up in any way in the ensuing 2-3 years. RESULTS: The mean age of the patients was 43 (range 16-81); 36% were female, and 67% had evidence of nephrolithiasis confirmed by CT. Of the 321 patients, 14 (4%) were excluded due a clear diagnosis other than nephrolithiasis being made on CT scan. 145 (45%) of the remaining 307 scan reports described incidental findings, of which 61 (45%) were of "moderate" or "serious" concern to both reviewers. Follow-up information was obtained on 80% of the cases with incidental findings, either from PCP notes (29%), specialty clinic notes (47%), or the hospital's computerized laboratory results. Not surprisingly, only 1 of 118 scans where both reviewers had "little or no" concern about the findings had subsequent follow-up at our hospital. However, only 11 (18%) of the 61 cases deemed of "moderate/severe" concern by both reviewers had any evidence of follow-up despite data on 51 (83%) of them. In only 27% of the total cases, and the same 27% of those with "moderate/severe" incidental findings, were they mentioned in the ED record. CONCLUSION: Incidental findings on ED abdominal CT scans are common and sometimes entail further diagnostic testing. Most are never noted or followed up, however, which raises a quality issue, emphasizing the need for effective communication between physicians. effectiveness of an information feedback system on improving health status and satisfaction outcomes. A random subsample of these patients received the Health Beliefs Survey, which contained items adapted from previously published health beliefs scales. The survey included items asking about general faith in doctors and medical science, the harmful effects of medical treatment, and the value of treatments not based on medical science. An additional item asked patients if they felt their beliefs about health were similar to their physicians' beliefs. Response options ranged from 1 (strongly disagree) to 5 (strongly agree). Patient satisfaction was measured with the 12-item humanistic component of the Seattle Outpatient Satisfaction Questionnaire (SOSQ), which is scored from 0 (low) to 100 (high). Additional data gathered included patient age, ethnicity, level of education, and income. We used simple and partial correlation techniques to evaluate the association between patients' reports of their health beliefs and patient satisfaction scores, while adjusting for patient demographic characteristics. RESULTS: We analyzed the survey responses of 127 patients. Patient satisfaction scores had a mean of 75 and were skewed towards high ratings. No significant associations were found between patient satisfaction and items asking about faith and skepticism in doctors and medical science. However, the degree to which patients felt their beliefs about health were similar to their provider's beliefs was positively associated with patient satisfaction (simple correlation coefficient ϭ .25, p ϭ .015), and this association persisted after adjustment for patient demographic characteristics (partial correlation coefficient ϭ .25, p ϭ .02). CONCLUSION: Patient beliefs relating to faith and skepticism in doctors and medical science were not significantly associated with patient satisfaction. However, the degree to which patients felt their beliefs about health were similar to those of their provider was significantly related to patient satisfaction. Further studies are needed to clarify the relationship between health beliefs and patient satisfaction, and to evaluate whether patient-provider discussions of health beliefs could improve patient satisfaction. PURPOSE: The use of oral bisphosphonates for osteoporosis is highly effective but unfortunately limited by gastrointestinal side effects. We examined the effectiveness of pamidronate, an intravenous bisphosphonate, as an alternative therapy. METHODS: We performed a nonconcurrent, prospective study of 78 patients with osteoporosis of varying cause treated in an osteoporosis clinic at an urban tertiary care hospital; 26 were treated with pamidronate, and 52 received other treatments including estrogen, calcitonin, and/or oral bisphosphonates (standard care). Osteoporosis was defined as a T-score of р2.5 at the L-spine, femoral neck, or trochanter on dual energy X-ray absorptiometry (DEXA) scan. Standard care patients were matched 2:1 to the pamidronate group by age, race, and sex. Baseline was defined as the start of pamidronate therapy or the first clinic visit. Pamidronate was administered as a 30 mg infusion over 4 hours every 3 months. Bone mineral density (BMD) was measured at baseline and approximately annually thereafter using DEXA scan. Response to therapy, defined as either stabilization or increase in BMD at one or more years of follow-up, was the primary outcome variable. RESULTS: Mean age was 58 and 54 years and body mass index was 23.5 and 23.8 kg/m 2 for the pamidronate and standard care groups, respectively. Mean follow-up was 1.3 years for pamidronate patients and 1.8 years for standard care patients. There was no statistically significant difference in percent responders between the pamidronate and standard care groups at any anatomical site (see METHODS: Fifty men and thirty-six women, with a mean total cholesterol (TC) level of 246.9 mg/dL (SD ϩ/Ϫ 23.21 mg/dL) and a mean low density lipoprotein cholesterol (LDL-C) level of 165.1 mg/dL (SD ϩ/Ϫ 21.34 mg/dL), were randomly assigned to a treatment or control group. Subjects were administered either a pla-cebo bar or the treatment bar (2.4g/day of Monascus purpureus) and were counseled throughout the study (12 weeks) by a registered dietitian on how to adhere to the American Heart Association (AHA)

Step One Diet. RESULTS: A two-factor repeated measures ANOVA revealed a statistically significant group by time interaction for the treatment group where TC decreased by 14% (34.7 mg/dL p Ͻ 0.0001) and LDL-C decreased by 22% (35.9 mg/dL p Ͻ 0.0001). In contrast, the control group demonstrated a 1.4% increase in TC and no change in LDL-C levels. There were no statistically significant changes in triglyceride levels, high density lipoprotein cholesterol levels, liver function tests, or body weight for the treatment group. Also, there were no statistically significant group by time interactions for caloric intake, dietary cholesterol intake, percent calories from fat, or percent calories from saturated fat. CONCLUSION: This study demonstrates that supplementation with a traditional Chinese food, Monascus purpureus, incorporated into a snack bar, is effective in reducing TC and LDL-C levels beyond the level of reduction obtained through the use of the AHA Step One Diet. In addition to its effectiveness in patients with hypercholesterolemia, Monascus purpureus is well tolerated with relatively few reported side effects. Background: Shared medical decision making (SMDM) is believed to improve adherence in chronic disease management. This has been shown in diabetic populations for Hemoglobin A1C but not for cholesterol levels. Methods: We studied the effect of SMDM on common parameters in diabetes (HbA1c, lipids, visit pattems to Primary Care, Eye and Podiatry clinics and medication usage). Patients in Group A (N ϭ 75) were provided with active SMDM (selfhelp texts and RN initiated telephone counseling) for three months and then followed routinely; Group B patients (N ϭ 73) were matched controls given only routine care. Fourteen months after beginning the intervention we compared the two groups ("post" values) and evaluated the same parameters in the 14 month period prior to the institution of SMDM in both groups ("pre" values).

Results:

There was no significant difference between groups for any of the parameters studied. There was virtually no change over time for the blood levels; the change in clinic visit pattern appeared in both the study and control group. Conclusion: Over the time frame of 14 months (with only three months of active intervention) there was no significant impact of SMDM shown on HbA1c, cholesterol or clinic usage pattern in Primary Care of diabetes. The intervention perhaps occurred over too short a time to produce change. It is possible that factors such as more intense reinforcement, a different type of patient involvement, or a dual intervention directed at both patients and providers may be needed to see an effect. PURPOSE: Whether sex affects the long-term prognosis of acute myocardial infarction (AMI) is controversial. No previous study has formally examined age-related differences in the effect of sex on long-term prognosis. We therefore studied the associations of age and sex with long-term survival in early survivors of AMI enrolled in the Determinants of Myocardial Infarction Onset Study. METHODS: We studied 1935 patients with AMI interviewed between August 1989 and September 1994 a median of four days following AMI. Trained interviewers performed standardized interviews and chart reviews to obtain demographic and clinical information. We searched the National Death Index for patient deaths through December 31, 1995. We used Cox proportional-hazards regression to estimate hazard ratios (HRs) for total mortality, adjusting for previous AMI, diabetes, hypertension, current smoking, former smoking, obesity, usual frequency of exertion, alcohol abstention, use of thrombolytic therapy, use of cardiac medications, and congestive heart failure or ventricular tachycardia during hospitalization. We categorized age into three groups (Ͻ50, 50-69, 70ϩ) and analyzed survival with dummy variables, within individual strata of age, and with an interaction term between age and sex. RESULTS: The table shows adjusted HRs for total mortality with their 95% confidence intervals (CI) according to age and sex. Men under 50 represented the reference group for these comparisons. Women had higher mortality among subjects under 50, but men had higher mortality among subjects aged 70 and older. The interaction between age and sex was signficant (p ϭ 0.04). In adjusted stratified analyses, the HRs and 95% CI for total mortality for women relative to men were 1.4 (0.5-3.9) among patients under 50, 1.1 (0.7-1.7) among patients aged 50-69, and 0.6 (0.4-0.8) among patients aged 70 and older. Exclusion of patients who died within 30 days of AMI (n ϭ 44) did not change our results. CONCLUSION: In this post-AMI population, the effect of sex on long-term survival depended upon patient age. After multivariate adjustment, women seemed to have worse long-term survival than men. At older ages, however, women had better long-term survival than men. Further research is needed to clarify the mechanisms responsible for this interaction.

Center, Boston, MA PURPOSE: Moderate drinkers have a lower risk of acute myocardial infarction (AMI) than abstainers. Whether moderate drinkers have a different long-term prognosis following AMI than abstainers is unknown. We therefore assessed the association of alcohol consumption and long-term survival among early survivors of AMI enrolled in the Determinants of Myocardial Infarction Onset Study. METHODS: We studied 1935 subjects with AMI interviewed between August 1989 and September 1994 a median of four days following AMI. Trained interviewers assessed usual alcohol consumption with a standardized questionnaire. We searched the National Death Index for patient deaths through December 31, 1995. We used Cox proportional-hazards regression to estimate hazard ratios for total mortality, adjusting for age, sex, current smoking, former smoking, educational attainment, usual frequency of exertion, previous AMI, diabetes, use of thrombolytic therapy, and congestive heart failure during admission. We confirmed the proportionality of hazards with time-varying covariates. RESULTS: Of the 1935 patients, 47% reported no alcohol consumption, 36% consumed less than 7 drinks weekly, and 17% consumed 7 or more drinks weekly. The unadjusted risks of death in these groups were 22%, 13%, and 9%, respectively. The table shows adjusted hazard ratios and 95% confidence intervals according to usual alcohol consumption. After adjustment, moderate alcohol consumption was associated with lower mortality following AMI. The effect of alcohol was similar in both men and women. Exclusion of subjects who died within 30 days of AMI (N ϭ 44) did not change our results. CONCLUSION: Moderate drinkers have lower long-term mortality following AMI than abstainers, regardless of sex. Two measures of regular mammography use were created: 1) meeting age-appropriate guidelines, that is, having mammography use since age 40 years equal to or exceed the recommended number of mammograms for one's age; and 2) meeting guidelines for age-appropriate use since the first mammogram (if after age 40 years). Multiple imputation was used to assign values for those women with incomplete data for age at first mammogram or total number of mammograms. Five hundred replicates were simulated for each incomplete observation, using normal random deviates, to estimate compliance distributions. RESULTS: Almost all women had at least one mammogram, however, recent (р2 years) mammography use was lower than ever use. Compliance with both measures of regular use was substantially less than ever or recent use and dropped off with age. Even after women undergo their first mammogram, compliance with regular mammography use is low, particularly in older women.

CONCLUSION: Women are not undergoing mammography according to screening guidelines even after their first mammogram. Efforts to improve regular mammography use are needed. were reimbursed primarily on a fee-for-service basis. We examined whether the for-profit status of HHAs affected length of stay (LOS) in home care in 1996. METHODS: We performed a secondary data analysis of current and discharged home care patients reported in the 1996 National Home and Hospice Care Survey. Data were collected on a sample of current and discharged patients from representative home health and hospice agencies. We analyzed patient demographics, diagnoses, dependency in activities of daily living (ADLs) and instrumental activities of daily living (IADLs), and the for-profit status of agencies. We developed a Cox Proportional Hazards model for LOS (days) in home care, using SUDAAN to adjust for the complex sampling design. RESULTS: In 1996, an estimated 10.2 million patients received home care in the US (64% female, 61% white, 62% Medicare, 68% у65 years). Of these patients, 12% had CHF, 11% COPD, 15% diabetes, and 11% cancer; 40% had one or more ADL dependencies, 34% had one or more IADL dependencies, and 34% received care from for-profit HHAs. The median calculated Charlson-Deyo Comorbidity Index score was 1. The mean LOS for current home care patients was 332 days, while the mean LOS for discharged patients was 98 days. Among those discharged, 68% met their goals and 4% died. The remainder were admitted to a hospital (11%) or nursing home (4%), moved or changed HHAs (6%), or had another reason for discharge (7%). After adjusting for sex, race, living alone, referral source, Charlson-Deyo Comorbidity Index, COPD, insurance type and census region, we found the following factors to be significantly associated with increased LOS in home care: For-profit HHA ( METHODS: Patients were selected from 3 primary care practices in New York City. Patients represented all 5 boroughs of the City and 200 were randomly selected using a random number generator. Sample size was computed for an ␣ ϭ 0.05 and ␤ ϭ 0.20. Chi-square analysis was performed to assess statistical significance. A p Ͻ 0.05 was deemed statistically significant. RESULTS: Of 200 patients screened, 16 (8%; p Ͻ 0.001) were positive for exposure to hepatitis C based on ELISA screening and RIBA confirmatory testing. Only one of the 16 knew they were positive prior to testing. Of those testing positive, 75% were male and the average age was 47yo. All patients testing positive for hepatitis C tested positive for prior exposure to hepatitis A (p Ͻ 0.01) and 13 of 16 had antibody to hepatitis B (p Ͻ 0.01). There were no chronic hepatitis B carriers among those positive for hepatitis C. Two patients were co-infected with HIV. Half of the hepatitis C patients were positive for antiphospholipid antibody. The average hepatitis C viral load was 75,000 IU/cc. Only 1 of the 16 hepatitis C patients (6%; p Ͻ 0.01) had an abnormally elevated ALT. None of the hepatitis C patients ever received a blood transfusion; 38% reported a prior hisotry of illicit drug use; 25% reported a possible exposure to body fluids infected with hepatitis C; 37% reported no known risk factors for acquiring hepatitis C. CONCLUSION: The prevalence of hepatitis C in large urban areas may far exceed that of the national average of 1.8%. Physicians should be aware that many patients may not have biological markers suggesting disease and thus follow-up of historical clues relating to prior drug use, multiple sexual partners, recurrent thrombotic events, or exposure to other viral diseases should prompt testing for hepatitis C exposure. PURPOSE: Past studies have demonstrated a powerful correlation between morbidity, mortality and patient-assessed health-related quality of life in hemodialysis patients as measured by medical outcomes study short form-36, also called SF-36. The aim of this study was to identify which of the clinical parameters routinely measured in ESRD patients on hemodialysis correlate with patient-assessed health-related quality of life. METHODS: A retrospective chart review was done on 74 patients dialyzed at a single dialysis center who had completed a SF-36 between June, 98-March,99. 13 charts were excluded due to incomplete response. Personnel instructed on the use and purpose of the SF-36 (in this study the social worker)administered the survey to new patients and to other patients every 6 months. We scored the surveys by computer and filed the results in the medical records.Only results of the first survey of each patient completed in the study window were used for this analysis. The average hemoglobin, albumin levels,total hours of hemodialysis, urea reduction ratio [URR] of each patient measured during a month prior to the date of SF-36 survey were obtained. We used multivariate regression method to control for all covariates and compare with physical and mental summary scores. Of 61 patients, 24 were women, 37 were men. 39% had diabetes, 72% had hypertension and 21% had coronary diseases. RESULTS: Mean age of the patients was 59 years. Mean hemoglobin was 11.3 g/ dl. Mean albumin level was 4.3.Average time of dialysis in a month was 43.4 hours. Mean URR was 68.3%. Mean physical summary score was 35.4 ϩ/Ϫ 11.5, mean mental summary score was 50.4 ϩ/Ϫ 11.7. In this multivariate regression analysis, there were no statistically significant correlation between mental summary score and albumin, URR, hours of hemodialysis per month. No statistically significant correlation was found between physical summary score and any of the variables. Age and hemoglobin showed positive correlation with mental summary score,with P value of 0.003 and 0.01 respectively. An increase in hemoglobin 1 g/ dl correlated with increase in mental summary score of 3.2 with P value 0.01. Every increase in age by one year correlated with increase in mental summary score of 0.36 with P value of 0.003. CONCLUSION: Association between hemoglobin and mental summary scores suggested that sense of mental wellbeing as measured by SF-36 is higher in patients with higher hemoglobin levels. Whether improving hemoglobin with therapy affects the mental wellbeing needs to be assessed with larger prospective studyneeds to be assessed. A small sample size and effect of other comorbid conditions are our limitations. tree using clinical inputs and cost estimates specific to each strategy was used to determine 1) the likelihood of identifying elevated tHcy levels (if screening occurred); 2) the effectiveness of folic acid at lowering tHcy levels; and 3) initial costs. Next, Markov simulations incorporating age, gender and CHD-specific mortality rates were generated to estimate the effect of tHcy-lowering on survival and ongoing costs. The published literature was reviewed to determine the best estimates for the Base-Case Analysis, while sensitivity analysis was used to evaluate uncertain estimates across a broad range of potential values. Cycles during the simulation were one year in length with subjects followed until age 85 or death. The outcomes we assessed were overall life expectancy in discounted life-years and total costs in discounted 1998 dollars. RESULTS: If lowering tHcy with folic acid and vitamin B 12 reduced excess coronary heart disease risk by 40% (Base-Case Analysis), the two intervention strategies cost between $13,000 and $77,000 per life-year saved when compared to No Intervention. Although the Treat All strategy was slightly more effective overall, the Screen & Treat strategy resulted in a much lower cost per life-year saved ($13,700 in men; $27,500 in women) when compared to No Intervention. Incremental cost-effectiveness ratios for the Treat All strategy when compared to the Screen & Treat strategy were over $500,000 per life-year saved in both cohorts. The Screen & Treat strategy remained the most cost-effective intervention under multiple scenarios evaluated during sensitivity analysis. CONCLUSION: Lowering tHcy with folic acid and vitamin B 12 could result in substantial clinical benefits at reasonable costs. If tHcy lowering is considered, a Screen & Treat strategy is likely to be more cost-effective than universal folic acid supplementation. PURPOSE: Electron-beam computed tomography (EBCT) is a new, non-invasive method of obtaining cross-sectional images of the heart that is increasingly being advocated as a method to diagnose coronary artery disease. Before clinical use of EBCT can be justified, however, its overall diagnostic accuracy must be clearly defined. We thus undertook a formal meta-analysis to estimate the accuracy of EBCT in the diagnosis of significant coronary artery disease. METHODS: English-language studies published from 1979 through 1999 were retrieved using MEDLINE, Current Contents, bibliographies, and consultation with experts. We included a study if it: 1) used EBCT as a diagnostic test for significant coronary artery disease (у50% stenosis); 2) reported results in absolute numbers of true-positive, false-positive, true-negative, and false-negative cases; and 3) used coronary angiography as the reference standard for determining the presence of significant coronary artery disease. Data from the included reports was combined using summary receiver operating characteristic (ROC) analysis. Results from 14 EBCT reports, with a total of 1662 subjects (71% men), were combined. Study groups in all instances consisted of individuals undergoing coronary angiography for evaluation of coronary artery disease. The mean number of subjects per report was 119. Reported sensitivity ranged from 81% to 99% and specificity ranged from 21% to 83%. From its summary ROC curve, the maximum joint sensitivity and specificity rate for EBCT was found to be 75%. Overall discriminatory power for EBCT, summarized as a diagnostic odds ratio, was 9.2 (95% confidence interval, 3.9 to 21.4). The use of different thresholds for a positive test in the included studies had only a small and non-significant association with the test's overall discriminatory power. For a threshold value that results in a sensitivity of 90%, specificity can be expected to be 54%; if sensitivity is 80%, specificity will likely be 71%. We found no evidence that the accuracy of EBCT was different in specific age or gender groups. CONCLUSION: Based on our meta-analysis, EBCT appears to be sufficiently accurate at identifying coronary artery disease in individuals, with sensitivity and specificity rates that are comparable to traditional exercise stress testing. Further studies are needed to determine the precise role of EBCT among the current armamentarium of noninvasive tests for coronary artery disease. Until then, our study provides clinicians with estimates of the overall accuracy of EBCT, allowing them to better interpret results from this rapidly diffusing diagnostic innovation. . We evaluated whether omission of RT adversely affected outcomes among older women who underwent BCS. METHODS: We studied a population-based cohort of 2,781 women aged 65 or older included in the Surveillance, Epidemiology, and End Results (SEER) registry, who underwent mastectomy or BCS treatment in 1986-87 for local or regional breast cancer, and for whom Medicare part A and B claims were available for at least 6 years following diagnosis (or until death). Since SEER does not collect information on disease recurrence, disease recurrence was defined as Medicare claims for mastectomy у6 months after diagnosis, RT у8 months after diagnosis, or chemotherapy у15 months after diagnosis. A proportional hazards model was used, and analyses were adjusted for age, stage, race, and socioeconomic status. RESULTS: Initial treatment was mastectomy in 77% subjects, and BCS in 13% subjects. Of those treated with BCS, 55% underwent RT. Stratifying by stage, women who underwent BCS without RT had a significantly higher risk of recur-rent disease (OR ϭ 1.38, p ϭ 0.037) compared to women who underwent mastectomy. Women who underwent BCS with RT had a recurrence rate that was similar (OR ϭ 0.91, p ϭ 0.45) to women undergoing mastectomy. As expected, there was a higher probability of recurrence among regional stage patients. However, the odds ratios for recurrence for women who underwent BCS without RT vs mastectomy were consistent within each stage (OR ϭ 1.36 for local stage, OR ϭ 1.32 for regional stage.) To help evaluate the face validity of our claims-based methodology for determining disease recurrence, we studied disease recurrence as a predictor of death. In the cohort as a whole, subjects who developed recurrent disease had a significantly greater risk of dying. For all cause mortality, the OR ϭ 2.33 (p ϭ 0.0001) for women with recurrent disease. For breast cancer specific mortality, the OR ϭ 7.84 (p ϭ 0.0001). CONCLUSION: As has been found in the randomized trials of younger women, this study found omission of RT with BCS to be associated with an elevated risk of disease recurrence. This population-based study suggests poorer outcomes among older women who undergo BCS without RT. PURPOSE: Preliminary clinical studies have shown that food insufficiency has adverse health consequences, including hypoglycemic episodes and increased need for health care services, among adult diabetics. The purpose of this study is to determine the prevalence of food insufficiency and describe the association of food insufficiency with health status and health care utilization in a national sample of diabetics. METHODS: We analyzed data from diabetics (n ϭ 1503) interviewed in the Third National Health and Nutrition Examination Survey (NHANES III), a cross-sectional representative sample of the civilian non-institutionalized population living in households. Individuals were classified as food insufficient if a family respondent reported that the family sometimes or often did not get enough to eat or that adults in the family had to cut down on the size of their meals because of financial constraints. Both bivariate and multivariate analysis were used to examine the impact of food insufficiency on self-reported health status, hospitalizations, and physician utilization. RESULTS: Six percent of diabetics reported food insufficiency, representing over 568,600 diabetics nationally (95% CI 368,400; 768,800). Diabetics who were food insufficient were more likely to report fair or poor health status than those who were not food insufficient (63% vs. 43%, OR 2.2, p ϭ .05). In a multivariate analysis, fair or poor health status was independently associated with non-white race, poverty, low educational achievement, and number of chronic diseases, but not with food insufficiency. Diabetics who were food insufficient reported more physician encounters, either in clinic or by phone, than those who were food secure (12 vs. 7, p Ͻ .05). In a multivariate linear regression, food insufficiency remained independently associated with increased physician utilization among diabetics. There was no significant association between food insufficiency and rates of hospitalization. CONCLUSION: Food insufficiency is relatively common in this national sample of diabetics and increased physician utilization and reduced health status. Prospective studies are needed to understand the impact of food insufficiency on the health status of and health care utilization by diabetics. PURPOSE: Many experts define a non-traumatic vertebral compression fracture as an osteoporotic fracture. As treatments are available to reduce the risk of recurrent osteoporotic fractures, we sought to determine whether osteoporosis was diagnosed or treated in primary care patients (pts) with non-traumatic compression fractures. METHODS: To identify pts with vertebral compression fractures cared for in a hospital-based primary care internal medicine practice, we performed a computerized search for the terms vertebral "compression","wedging", or "loss of height" from all chest, thoracic and lumbosacral spine X-rays performed between 1/1/97 and 1/1/99 as part of usual care. We excluded pts with X-rays or medical records indicating trauma (defined as fall from more than standing height) or metastatic cancer. We reviewed pt's outpatient medical records for demographics, risk factors for osteoporotic fractures, comorbid illnesses, and diagnosis or treatment of osteoporosis. To confirm non-prescription therapy, we conducted telephone surveys in 12/99. RESULTS: Of 171 pts (median age 76), 66% were female, 12% had taken corticosteroids, 32% had an X-ray ordered for back pain or a fall, and 56% had multiple compression fractures. Primary care physicians (PCPs) documented fractures for only 32% of pts. As some pts were diagnosed with osteoporosis without documentation of compression fractures, 39% of pts were diagnosed with osteoporosis overall. Bone densitometry was performed for 21% of pts between 1992-99. A minority of pts was treated for osteoporosis: 42% were taking calcium, 35% were taking a multivitamin/vitamin D, and 37% were offered у1 prescription drug (20% alendronate, 12% calcitonin, 23% estrogen). Rates of prescription treatment were higher when osteoporosis was diagnosed (76% when diagnosis made vs 12% when diagnosis not made). In multivariable analysis adjusted for age and using the generalized estimating equation to account for clustering by provider, pt characteristics independently associated with receiving prescription treatment were: female sex (AOR 2.2, 95% CI 1.0, 4.6), steroid use (7.9 [2.9, 21.0]), back pain or fall as indication for X-ray (5.7 [2.4, 13 .6]), у2 X-rays with fracture (2.9 [1.3, 6.9], lower weight (0.9 [0.8, 1.0] per ten pound increase) and fewer comorbidities (0.7 [0.5, 0.9] per unit increase in Charlson comorbidity score). CONCLUSION: Most pts with non-traumatic vertebral compression fractures were not treated for osteoporosis. This suboptimal care is explained partially by PCP's failure to diagnose osteoporosis after a vertebral compression fracture was described on an X-ray report. Education of PCPs and radiologists about the diagnosis and treatment of osteoporosis and better communication between physicians may improve care for pts with osteoporosis. . Inclusion criteria were: 1) English language publications; 2) populationlevel interventions; and 3) outcomes measures for tobacco use. A second search was done using the key terms: Asian, Black, Hispanic, Minority. Only 1 of 62 IUPT papers and 18 of 174 CWI papers focused on minorities. We reviewed the articles using a standardized study abstraction form developed by the Task Force for the Guide to Community Preventive Services. Abstracted data included components of the intervention, characteristics of the study populations, and exposure and outcome measurements. Quality of study design and execution were evaluated by two independent reviewers. Discrepancies were resolved by consensus. RESULTS: IUPT was found to be an effective intervention for decreasing cigarette sales to Hispanic (H) and African American (AA) smokers. After controlling for income and education, price elasticity of demand (% change in demand resulting from 1% change in price) was Ϫ1.89 for H and Ϫ0.32 for AA smokers, compared with Ϫ0.14 for white smokers. This study is consistent with 44 other studies done in non-minority populations. The table below describes CWI in the 6 studies representing the best evidence subset (Good or Fair Quality Rating Score):

CONCLUSION: Increasing the unit price of tobacco is an effective intervention to deter smoking among Hispanic and African American smokers. Community-wide interventions may increase smoking cessation, although only 2 of 6 studies achieved statistical significance. These data suggest the need to increase the unit price of tobacco and to further study community-wide interventions in minority communities. conducted a review of the circumstances surrounding the deaths of 12 homeless men and 1 homeless woman (a second woman remains unidentified in the city morgue) who died on the streets of Boston between July 1998 and January 1999. METHODS: Medical charts of these individuals at BHCHP's two hospital based primary care clinics as well as records from a detoxification program where the primary author serves as medical director were reviewed. Of note this study has not yet been able to review the utilization of the mental health system by this group. RESULTS: The review revealed that not only did these individuals suffer from complex medical conditions but that 100% suffered from chronic substance abuse (EtOH was the primary substance of choice in 12 of the 13) and two-thirds suffered from severe mental illness. In addition the review showed multiple emergency room visits and hospitalizations particularly in the six months prior to death. Strikingly, 9 of the 13 (69%) had been seen in the emergency room or admitted to the hospital within three weeks of the day of death. Many had been seen multiple times, with several seen within 72 hours of death. Half of those who died had also been admitted to the Barbara McInnis House (BHCHP's medical respite facility) with acute medical problems during the six months prior to their deaths. This high frequency of utilization of the health care system was paralleled by a dramatic number of admissions to publicly-funded detoxification programs. Virtually all of the decedents had multiple admissions during the year prior to death. Ten of the 13 (77%) had received detoxification from alcohol and other drugs within 6 weeks of death, with 3 (23%) having died on the streets within one week of discharge from a detoxification unit. CONCLUSION: The homeless persons who died on the streets suffered from an unusually high "tri-morbidity" of acute and chronic medical illnesses, persistent mental illness, and relapsing addictions to alcohol and other substances. Contrary to our original hypothesis that these individuals had fallen through the safety net, this group was not only well-known to homeless providers on the streets, but also had extraordinarily high utilization patterns of the medical care system (especially emergency rooms, acute care hospitals, and respite care) as well as the substance abuse system (primarily the acute detoxificaton units). Additional study is needed to understand the utilization of the mental health system by these individuals. The relationship between depression, anxiety, and hostility has been well studied in clinical coronary artery disease (CAD), but not subclinical CAD. We studied the relationship between multiple psychosocial variables and subclinical CAD, as represented by coronary calcification METHODS: Cross-sectional study of a prospective, consecutive sample of 630 consenting, active duty U.S. Army personnel (39-45 y.o.) without known CAD, who were undergoing a routine physical as required by regulations. Each participant was surveyed with validated instruments to assess for depression, anxiety, somatization, stress, and hostility level. Predicted risk was assessed using the Framingham risk index, based on measured conventional CAD risk predictors. Subclinical CAD was determined using EBCT to quantify coronary artery calcifica-tion (CAC; Agatston method). Assessments of psychosocial factors and CAC were blinded to each other. RESULTS: Of 700 personnel screened, 630 agreed to participate. The mean age was 42; 83% were male, 72% white, 83% college graduates, and 76% officers. The 10 year predicted risk of clinical CAD was 4.5%. The prevalence of CAC was 19% (mean CAC ϭ 10.3). There was a modest relationship between predicted risk and CAC (ROC area 0.61, p ϭ .004). The prevalence of prior or current criteria-based DSM-IV psychiatric disorders was 14.6% (depression: 6.3%, anxiety: 1.4%, somatoform disorder: 9.5%, with some overlap). As a group, 90% of participants had hostility scores lower than the median normative scores. There was no correlation between CAC score and depressive (r ϭ Ϫ0.05, p ϭ 0.25), anxiety (r ϭ 0.01, p ϭ 0.8), somatization (r ϭ Ϫ0.08, p ϭ 0.14), stress (r ϭ Ϫ.04, p ϭ 0.45) or hostility (r ϭ Ϫ0.06, p ϭ 0.24) scores. Similarly, there was no correlation between any of these psychosocial variables and predicted risk. Those without psychiatric disorders had higher CAC than those with prior or current psychiatric disorders (11.6 vs 1.3; p ϭ 0.0001). CONCLUSION: In a relatively young screening population, psychosocial factors were not related to the presence or extent of early, subclinical atherosclerosis. These data suggest that any association between psychosocial factors and coronary events would be through mechanisms other than the development of athersclerosis, such as through interactions with vulnerable plaque. We conducted a survey of all hospital and community based medical attending physicians at a tertiary care hospital to determine whether physicians screen in concordance with current guidelines. The survey included questions about physician gender, specialty, practice setting and patterns regarding colorectal cancer screening. Endpoints of interest were based on current recommendations and consisted of appropriate age (50 years old) and methods of screening of asymptomatic individuals. Appropriate methods of screening were defined as the use of home based fecal occult blood testing, flexible sigmoidoscopy, colonoscopy or barium enema but not digital rectal exam or office based guaiac testing after digital rectal exam. Data was analyzed by logistic regression to determine independent predictors. RESULTS: Surveys were sent to 491physicians. Valid responses were received from 215 physicians, of whom 85% were community based. From these responders, 38.7% were general internists, 14.6% were gastroenterologists and 46.7% were other subspecialists. Fifty-three percent of physicians started screening at the appropriate age and 82% of physicians did not have an upper age limit for screening. Ninety-two percent of physicians utilized inappropriate initial screening tests in addition to or in place of the recommended screening tests. Working in a hospital based practice was the only independent predictor of screening at the appropriate age (OR ϭ 2.47, CI ϭ 1.08 to 5.68) and using appropriate screening methods (OR ϭ 7.86, CI ϭ 2.67 to 23.14). CONCLUSION: Many physicians initiated screening at inappropriate ages or utilized inappropriate screening methods in addition to or in place of recommended colorectal cancer screening methods. Hospital based physicians were more likely to start screening at the appropriate age and screen in accordance with current screening recommendations for colorectal cancer. This may have been due to their positions as role models and educators and their need to remain updated on current issues. Our data suggests that increased efforts are needed to encourage and educate physicians to practice in accordance with these guidelines. Additional studies are necessary to determine whether these early findings are reproducible and whether they apply to other screening recommendations. HIV infected groups, HOSP and ER in the high-era were slightly higher than in the transition-era but were consistently lower than those in the low-era. Furthermore, in the high-era, the HOSP and ER use of the HIVϩ were no different than those in the HIVϪ whose HOSP and ER remained stable over time. CONCLUSION: At the population level, widespread use of HAART has resulted in substantial decrease in the utilization of high-cost health care services. The slowing of progression to AIDS by HAART resulted in health care utilization that was similar by HIVϩ and HIVϪ women.

A Paranjape, LM Sullivan, JM Liebschutz, Section of General Internal Medicine, Boston Medical Center/Bedford VA; Department of Statistics, Boston University, Boston, MA PURPOSE: An important barrier to universal screening for intimate partner violence (IPV) is lack of a well validated, easy to use screening tool. We conducted a study to develop a simple, validated screening tool for the detection of a lifetime history of IPV in women. METHODS: We interviewed 75 women in the emergency room of an urban teaching hospital. Consecutive English-speaking female patients waiting for non-urgent medical problems, between the ages of 18 and 65, were eligible.The initial interview consisted of demographic questions and potential screening items; the items encompassed all aspects of IPV, were drawn from published literature and clinical experience, and had dichotomous response options. A semi-structured clinical interview followed; we obtained details of all previous and current relationships, using probes for specific violent acts. We defined lifetime IPV as any one or more of the following acts committed by an intimate partner: sexual abuse as any sexual act that was coerced or committed under threat of violence, physical abuse as repetitive, intentional violent acts directed at the subject, emotional abuse as threats of serious injury to subject or her children, deliberate denial of access to money, car or house keys. Using reciever-operator curves(ROCs), we compared the estimated sensitivity and specificity for each item with respect to IPV and determined the best 4 items. We compared the estimated sensitivity and specificity for all response options for all combinations of the best 4 items using ROCs and calculated 95% confidence intervals (CIs) of the sensitivity and specificity of the final screen. RESULTS: The prevalence of lifetime IPV was 63%; of those abused 23% reported IPV in the past 1 year. The mean age was 36 years (SD:10 years); 75% of subjects had an annual income Ͻ$20,000. The three items with the best operating characteristics were: "Have you been in a relationship where your partner has pushed or slapped you?", "Have you ever been in a relationship where your partner threatened you with violence?" and "Have you ever been in a relationship where your partner has thrown, broken and punched things?". The sensitivities of these items were 87%, 78% and 83% respectively. For a positive response to any 1, any 2, or all 3 questions, the sensitivity (with 95% CIs) of the three-question screen STaT to detect IPV was 96% (90%, 100%), 89% (81%, 98%) and 64% (50%, 78%) respectively. The corresponding specificity of each response combination was 75%, 100% and 100%. CONCLUSION: A single positive response to a simple three-question screen STaT has a sensitivity of 96% and specificity of 75% to detect lifetime IPV. This screen may improve physician efforts at universal detection. McNemar testing revealed no systematic tendency for surrogates to assess care either more positively or more negatively. CONCLUSION: We conclude that the use of surrogates to assess the symptoms of terminally ill patients is reasonably accurate, except for pain and urinary incontinence. However, with the possible exception of the quality of physician care, surrogates' ratings of satisfaction and the quality of care do not reflect patients' views. While our sample size is small and further study is needed, these results suggest that surrogate ratings of many symptoms can be used as valid outcome measures in interventions to improve end-of-life care. However, for the important symptom of pain and for levels of satisfaction and quality, other instruments or strategies will need to be developed. DD as a first test, followed by a single US only in case of an abnormal DD level, and by phlebography only in patients with a normal US and a high clinical probability of DVT, was the cheapest, and, therefore, the most cost-effective option. This strategy allowed a 15% reduction in incremental costs compared to the most expensive algorithm. CONCLUSION: Combining clinical probability and DD to a single US in the diagnostic workup of DVT is highly cost-effective, allowing to reduce costs without any increase in mortality. Serial US appears less cost-effective. PURPOSE: To identify predictors that determine which women will comply with mammography, using a model that incorporates demographic data as well as attitudinal scales of race, religiosity, and trust. METHODS: We recruited a stratified random sample of 101 women aged 50 to 75 yrs. from church groups, health fairs, and medical clinics in urban areas of North Carolina and Florida. We administered a structured questionnaire to obtain demographic information, to assess the above attitudes, and to ascertain compliance with mammography. RESULTS: Sixty-six percent of the patients interviewed received mammograms within the last two years. 41% of those interviewed were African-American and the remainder were white. In the logistic regression the predictors of whether or not a patient received a mammogram were income, race, and self-reported health status. Religiosity, trust, and perceptions of racism did not predict whether or not a woman received a mammogram. The odds ratio for income less than $20,000 was .199 (Wald confidence limits .082 to .480), for African-American race .237 (Wald confidence limits .08 to .70), and for average or poor self-reported health .287 (Wald confidence limits .097 to .852). CONCLUSION: African-American race remains an independent predictor of decreased breast cancer screening. The specific reason for reluctance to obtain mammography remains unclear. PURPOSE: Epidural scarring resulting from lumbar surgery has been implicated as the cause of post-procedural low back pain and sciatica, though the evidence surrounding this association is unclear. We performed a critical literature review to assess the relationship between scar formation and post-procedural pain. METHODS: A Medline search was performed for all English language articles concerning human subjects published between 1966-1999 first using the key words scar, fibrosis, and cicatrix. In the second stage, identified articles were searched using key words failed back, sciatica, back pain, and adcon (surgical device used in lumbar surgery). Articles were evaluated for data concerning the causal relation between post-procedural scarring and back pain or sciatica. Five criteria were utilized in the evaluation of causation: consistency (a scar, when present, is associated with pain); specificity (a scar is associated with specific type of pain); strength (larger size scars lead to more pain); temporality (scar formation precedes pain); and coherence (a decrease in scar size leads to a decrease in pain). Articles were abstracted by two separate reviewers (RP, SR). RESULTS: Of 262 articles initially identified, only 15 reported data relevant to the causal relation between scarring and post-procedural sciatica and back pain. Articles included 10 case series (8 prospective, 2 retrospective), 3 case-control studies, and 2 randomized clinical trials with a mean of 81 patients per study. Data sufficient to evaluate consistency was reported in 11 studies; scarring did not meet this causal criteria in any of the 11 studies. Of 6 studies providing data to evaluate specificity, only 1 study reported findings sufficient to support causal specificity. Of 6 articles reporting data concerning temporality, none reported that scarring preceded low back pain or sciatica. Only one article reported data concerning coherence and it did not meet the requirement of causality. Of 10 studies reporting data concerning strength, 8 found no association between size of scar and post-procedural pain. Of interest, the two studies reporting a positive strength association between scar size and post-procedural pain involved two randomized trials of a surgical device intended to decrease scar size. CONCLUSION: While there is suggestive evidence of an association between scar size and extent of pain as reported in the surgical device trials, there is little evidence in the literature to support a causal relationship between scar formation and post-procedural low back pain or sciatica after lumbar surgery. The Charlson Comorbidity Index (CCI) has been previously validated in many populations and is used widely to describe and adjust for burden of comorbidity in clinical research. A previous attempt to validate its use in patients with Chronic Obstructive Pulmonary Disease (COPD) failed. The Adult Comorbidity Evaluation (ACE-27), a modification of the Kaplan-Feinstein index, is designed to improve on prior comorbidity indexes by including information on an increased number of comorbid conditions as well as their severity. Using medical record review, the ACE-27 records the severity of 27 conditions. Each is graded absent (0) or mild to severe (1-3) with specific clinical criteria for each grade. The index condition (in this case, COPD) is not counted as a comorbid condition. We evaluated this new instrument and compared its performance to the CCI in hospitalized COPD patients.

We identified a cohort of 74 inpatients with COPD exacerbations in Burlington. Vermont in 1995. Comorbidity scores and covariates were obtained by chart review, and information on mortality was retrieved from charts and the Social Security Death Index. Two methods of scoring the ACE-27 were evaluated: a continuous scale of 0-81 points and a summary "prognostic comorbidity" which is positive if any one grade 3 (severe) or any two grade 2 (moderate) comorbidities were present. Analyses were performed using Cox proportional hazards models and chi-square analysis. The mean age of the 74 inpatients was 69.5 years, 53% were male, 19% were in the intensive care unit and two died in the hospital. Over five years of follow-up, 58% died with average time of observation of 2.5 years. The CCI failed to significantly predict survival in this population. The continuous ACE-27 score performed similarly. However, the presence of prognostic comorbidity was a powerful predictor. The 5-year death rate was 51% in patients without prognostic comorbidity and 76% in patients with prognostic comorbidity (p ϭ 0.047 by chi-square analysis). Cox proportional hazards analysis showed similar results. Correction for age, gender and acute illness in multivariate models had little effect on the hazard ratios (HR). The CCI was not a significant predictor of mortality in patients with COPD. The ACE-27, especially when used to indicate prognostic comorbidity, performed better. Limitations of the ACE-27 include lack of validation in other populations and the necessity for skilled chart review. Future studies should attempt to further validate the ACE-27 in other patient populations. CONCLUSION: Of patients with symptoms suggestive of acute cardiac ischemia, the 10% with ECG LVH or BBB were older and as likely to have acute cardiac ischemia as patients with primary ST-T wave abnormalities. Of note to the clinician, compared to patients without ST-T wave abnormalities, these patients had more false positive admissions, yet more confirmed diagnoses of ACI or CHF, and had higher 30-day mortality rates possibly related to greater age and rates of CHF. OBJECTIVES: First to compare physical activity levels using a validated accelerometer between men and women, with and without PAD. Second, to use a commonly used physical activity questionnaire to measure physical activity. METHODS: 302 participants from a prospective study of lower extremity functioning in PAD completed the Cardiovascular Health Study (CHS) physical activity questionnaire. 75 patients with PAD and 48 patients without PAD wore an accelerometer continuously for seven days to measure physical activity. The accelerometer measures physical activity levels based on vertical movement at the hip joint. RESULTS: PAD participants had lower physical activity levels than non-PAD participants when measured by the accelerometer (820 kcals/week ϩ/Ϫ 397 vs 1045 kcals/week ϩ/Ϫ 483,p ϭ 0.006). Physical activity levels in PAD participants when measured by the CHS questionnaire were 767 kcals/week ϩ/Ϫ 990 compared to 878 kcals/week ϩ/Ϫ 890 in non-PAD participants (p ϭ 0.319). In linear regression analysis, ABI was an independent predictor of accelerometer-measured physical activity (regression coefficient ϭ 325 kcals/one unit ABI, p ϭ 0.0237) after adjustment for age. In multivariable analyses, ABI was independently associated with physical activity levels measured by the accelerometer after controlling for age,sex,race,and comorbid illness (regression coefficient 288.4 kcals/one unit ABI, p ϭ 0.05). CONCLUSION: Men and women with PAD have significantly lower physical activity levels than patients without PAD when measured with an accelerometer. A commonly used questionnaire for measuring physical activity levels was not a sensitive measure of physical activity when compared to the accelerometer. Future studies should be directed towards improving physical activity in patients with PAD in order to prevent cardiovascular mortality and functional decline. that chelation therapy has little or no beneficial effects. Given the fact that patients are nonetheless using the treatment, we conducted a survey to examine sociodemographic and clinical variables related to chelation therapy use in patients undergoing coronary angiography.

Heart Disease (APPROACH) is a cohort study of all patients undergoing coronary angiography in the province of Alberta, Canada. The APPROACH database captures information on sociodemographic, clinical, and treatment variables. Data on quality of life (EuroQoL) and chelation therapy use were ascertained through a survey mailed to patients one year after angiography. For this analysis, we included patients undergoing angiography in 1995-96 and surviving at least one year after angiography. Two logistic models were fit to identify variables associated with chelation therapy use: one for use of chelation therapy at some time (users') and another for use at the time of the survey (current users'). RESULTS: Among 5933 patients responding to the one year survey (69% response rate), 466 (7.8%) reported use of chelation therapy at some time and 212 (3.6%) were current users. The variables significantly associated with being a user were: abnormal coronary anatomy (e.g., odds ratio 3.2 for 3 vessel disease and 2.9 for left main relative to those with normal anatomy), coronary artery bypass surgery (CABG) treatment prior to angiography (1.4), score of 1.0 on EuroQoL scale (0.8), angioplasty (PTCA, 0.7) and CABG (0.6) in the first year after angiography. Current use was associated with abnormal coronary anatomy, diabetes (0.6), PTCA (0.7) and CABG (0.4) in the first year after angiography. CONCLUSION: Chelation therapy users appear to have more extensive coronary artery disease and poorer quality of life than do non-users and are less likely to have been treated with revascularization after coronary angiography compared with nonusers. This suggests that chelation therapy users may be turning to the treatment as a 'last resort' after being judged not suitable for revascularization. Alternatively, some patients may be declining revascularization in favor of this less invasive, but unproven, treatment option.

PURPOSE: Much attention has focused on the high prevalence of distressing symptomatology, especially pain, in hospitalized patients at the end of life. We assessed the prevalence and severity of physical symptoms and the level of psychosocial well-being among seriously ill outpatients with cancer, congestive heart failure (CHF), and chronic obstructive pulmonary disease (COPD) in a general medicine clinic. METHODS: During a 6-month period in 1999, patients in an academic general medicine clinic were identified by their Primary Care Physician (PCP) as having severe COPD, severe CHF, or cancer. Those patients willing to participate in a controlled trial of interdisciplinary care were recruited for participation in an ongoing research project entitled "The Comprehensive Care Team." For administrative purposes, within the clinic, PCPs and their patients are divided into two equivalent practices. Patients in one practice were assigned to the intervention group; patients from the other practice received usual care. To assess patients' baseline level of symptomatology prior to any intervention, all patients completed a self-administered questionnaire regarding their pain, shortness of breath, depressive symptoms, and quality of life. The questionnaire included previously validated instruments, including the Brief Pain Inventory, the UCSD Dyspnea Index, the CES-D, and the Padilla Quality of Life Assessment. RESULTS: 48 patients were enrolled during the 6-month period (31 in the intervention group (64.6%) and 17 in the control group (35.4%)). The average age for all patients was 66 years (standard deviation (SD) ϭ 14.3), with an average education level of 13.6 years (SD ϭ 3.4). 73% of patients were female, 52% white, and 23% African-American. 33.3% of the sample had COPD, 25% CHF, 22.9% COPD and CHF, 14.6% cancer, and 4.2% cancer and CHF. On a 0-10 scale, patients reported an average pain intensity of 4.0 (SD ϭ 2.5) with an average maximum pain during the last 24 hours of 4.7 (SD ϭ 3.2). 88% of patients reported dyspnea. A CES-D score greater than 16 suggests the need for anti-depressant medication. Among patients surveyed here, the mean CES-D score was 18.7 (SD ϭ 13.3). 10% of the sample had a CES-D score of greater than 40. On a 0-10 scale, the average patient quality of life was 7.2 (SD ϭ 2.6), with 23.4% of the sample rating their quality of life as 5 or less. At baseline, there were no statistically significant differences between intervention and control patients. CONCLUSION: As is true among hospitalized patients at the end of life, significant pain, dyspnea, and depression are common among outpatients with serious CHF, COPD, and cancer in a general medicine clinic. Primary care physicians appear to be undertreating physical symptoms and depression, even among patients they identify as having serious illness.

Epidemiology, Kyoto University Hospital, Kyoto, Japan PURPOSE: Bacillus Calmette Guerin (BCG) revaccination has been implemented in Japan among the tuberculin-negative 1st grade primary and junior high school students for decades. Controversies regarding the effectiveness of BCG revaccination and low incidence of tuberculosis (TB) among the children were the impelling factors to conduct this study. The main objective of this study is to compare the costs and benefits of BCG revaccination in Japan.

METHODS: Cost-benefit analysis was conducted for cohort of school children who underwent revaccination during 1996 from the perspective of the payer for health care. The study population comprised of estimated 1.4 million 1st grade primary school and 1.5 million 1st grade junior high school students enrolled in 1996 in all over Japan. Assuming 50% vaccine efficacy for revaccination, 10 years of duration of protection, and 5% annual discount rate, calculation was made on the total hypothetical number of TB averted by revaccination. Both direct and indirect costs were included in the analysis. RESULTS: The revaccination program for 1996 school children cohort would prevent 294 TB cases over a 10-year period at a cost of US $115,147 per case averted. The direct and indirect costs of revaccination were estimated at US $31.4 and US $2.5 million, respectively. The direct benefits, the treatment costs saved for future cases averted by revaccination, were estimated to be US $2.3 million. The major indirect benefit, the averted loss of work output attributable to TB-related morbidity, was estimated at US $1.6 million. The benefit-cost ratio remained at 0.12 with baseline assumptions and ranged from 0.04 to 0.28 for one-way sensitivity analysis. The costs of revaccination were found to exceed benefits even after inputting assumptions which tend to favor revaccination. CONCLUSION: From the economic point of view, BCG revaccination policy is not favorable in the present Japanese situation. , and medicine attending (GM) using predefined standardized criteria in order to classify the PSA test as a screening or diagnostic service. The attending physician (ATTEND) who ordered the PSA test completed a similar questionnaire in which they classified their intent in ordering the PSA test as either screening or diagnostic using identical criteria. Concordance rates were calculated between each rater and ATTEND's classification. Valid responses were received for 95 tests (79%), which constitute the study cohort. RESULTS: Raters demonstrated a high level of agreement with ATTEND's classification as a majority of tests generated complete agreement between all 4 raters and ATTEND (n ϭ 51,53%). Concordance rates by individual rater were 0.75 (GM), 0.77 (ONC), 0.78 (UR), 0.80 (RES) for an average concordance of 0.78. Of note, in 7 tests all 4 raters provided identical classifications, yet disagreed with ATTEND's classification. Follow-up discussions with ATTEND indicated ATTEND misclassified all 7 tests as screening services though the tests clearly met prespecified diagnostic criteria. Correcting the 7 tests inappropriately classified by ATTEND increased concordance rates to 0.84 (ONC), 0.87 (RES), 0.88 (UR), 0.91 (GM), for overall corrected average concordance of 0.88. CONCLUSION: Standardized criteria applied to medical chart data provide a valid method of retrospectively classifying PSA tests as screening or diagnostic. While concordance rates vary based on rater specialty, variances are small and likely not meaningful. Caution, however, should be taken when evaluating results dependent on the retrospective evaluation of PSA ordering intent. Our data provide validation for retrospective evaluation of screening status, as part of case-control studies of PSA screening efficacy currently under way.

MC Reid, ME Tinetti, PG O'Connor, J Concato, VA Connecticut, and Yale University PURPOSE: Although several methods are available to measure alcohol use, the most useful method for older adults remains uncertain. We sought to determine the levels of agreement and concordance between 3 commonly used measures in a study of older (у65 years) community-living adults. We conducted surveys of (1) veterans enrolled in the West Haven, VA primary care clinic (N ϭ 303), and (2) Medicare beneficiaries (N ϭ 511) in the greater New Haven, CT, region. We administered standard quantity-frequency (QF) questions, the CAGE, and the Alcohol Use Disorders Identification Test (AUDIT) to all subjects who reported use of alcohol in the past 12 months: 59% of VA, and 51% of Medicare participants. To assess agreement, we determined the percentage of subjects who screened positive (ϩ) by each measure defined as: QF ϩ ϭ Ͼ14 drinks per week; CAGE ϩ ϭ score у2; and AUDIT ϩ ϭ score у8. We assessed concordance by calculating kappa values for each of the pairwise comparisons. Participation rates were 71% and 81% for the VA and Medicare samples, respectively, and the mean ages were 73 and 76 years. The proportions of current drinkers who screened positive in the VA and Medicare samples, respectively, were: QF ϩ ϭ (14% and 4%); CAGE ϩ ϭ (19% and 4%); and AUDIT ϩ ϭ (16% and 5%). Lev-els of agreement (%) and concordance (kappas) were highest for the QF vs. AUDIT comparisons, but were marginal for the other comparisons as shown below

Our results show that commonly used alcohol measures have limited agreement and concordance in older adults, with the possible exception of the QF vs. AUDIT comparisons. The use of any single measure may fail to detect a substantial number of older individuals with potentially important alcohol exposures. Research is needed to determine the optimal screening strategies for detecting alcohol disorders among older populations.

WHEN SHOULD WE STOP SCREENING? S Rich, H Welch, VA Outcomes Group, White River Junction, VT PURPOSE: Although the age at which to start screening has been the subject of considerable debate, the question of when to stop has received little attention. To help address this question, we calculated the days of life given up by stopping screening at various ages. METHODS: For each of three cancers (breast, cervical and colon), we used life tables to calculate life expectancy at various ages. The days of life given up by stopping screening at a specific age is the difference in life expectancy between two life tables: one for stopping screening at that age and the other for continuing screening. Based on the randomized trial data, we used a 30% reduction in cancer-specific mortality for breast and colon cancer screening. Because there are no comparable data for cervical cancer, we assumed a 30% reduction in the "best guess" analysis and a 70% reduction in the "best case" analysis. We also assumed there was no harm with screening. DATA: All-cause and cancer-specific mortality were obtained from the National Center for Health Statistics and the Surveillance Epidemiology and End Results Survey (SEER). RESULTS: Assuming a starting age of 50, screening throughout life has a maximum benefit in life expectancy of 42.7 days for breast cancer and 28.1 days for colon cancer. Assuming a starting age of 20, screening throughout life for cervical cancer has a maximum benefit of 7.1 days and 46.6 days in the "best guess" and "best case" analyses, respectively. Days of life given up by stopping screening (% of total benefit) CONCLUSION: Even assuming the mortality reduction with screening persists in the elderly, 80% of the benefit is acheived before age 75 in breast cancer, age 80 in colon cancer, and age 65 in cervical cancer. The small benefit of screening in the elderly may be outweighed by harm, e.g., anxiety, additional testing, and unnecessary treatment. PURPOSE: Weight gain, actual or feared, is a major barrier to smoking cessation, especially for women. Effective ways to avoid or treat post-cessation weight gain are lacking. Bupropion sustained-release (SR) is an FDA-approved smoking cessation aid. A 7-week course of the drug temporarily blunts post-cessation weight gain. We hypothesized that a longer course of therapy might have a more permanent effect on post-cessation weight gain. METHODS: We analyzed data from a multicenter relapse prevention trial in which 784 adult smokers took open-label bupropion SR 300 mg/d for 7 weeks. At week 7, the 432 smokers who had quit (no cigarettes for the past 7 days) were randomly assigned to a double-blind trial comparing an additional 45 weeks of bupropion vs placebo. Weight was assessed at end of drug therapy (52 weeks) and at 13, 15, 18, and 24 months. Analysis was limited to smokers who were continuously abstinent for 24 months (n ϭ 97). RESULTS: At the end of drug treatment, patients who were assigned to receive bupropion long-term gained less weight than those assigned placebo (difference, CONCLUSION: Cigar use is common in college students, especially white and black males, cigarette smokers, users of other substances, and students whose priorities are more social (parties, fraternities, attending sporting events) than educational. We find no evidence to support concerns that cigars may serve as a "gateway" to cigarette smoking in college populations. However, cigar use is not harmless and ongoing efforts to reduce tobacco, alcohol, and substance use in college should also address cigars. The study was performed in a 720 bed tertiary teaching hospital. All inpatient-imaging exams are requested on-line using physician order entry with indications chosen from predefined menus. During the 27 month study period (Dec. 1995 -March 1998 , there were 89,450 non-obstetrical admissions. In all, 856 exams were performed to evaluate abnormal (abn) LFTs, including 540 US (63%) and 316 CT (37%) exams. Altogether, 759 image reports were successfully matched to lab results. Reports were coded as: Positive (pos-abn and related to the diagnostic indication), Normal (Nl) and Other. A total of 10,613 LFTs [AST, ALT, total (TBR) and direct(DBR) bilirubin, and alkaline phosphatase (AP)] completed within 3 days of the exam orders were analyzed. LFT results were clasified by severity as follows: nl, mild abn (1-2xnl), moderate (mod) abn (2-10xnl) , and severely abn (у10xnl). The 5 labs were collapsed into 3 categories: AST/ALT, AP, and TBR/DBR. For multiple testing, the most abnormal LFT was selected. Severity of lab category results were defined by the most abn lab test within each category(ie if mild abn AST and mod abn ALTmod abn ranking). To determine the number of "abn lab categories" per patient, lab categories Ͼ2xnl (only mod or severe abn) were considered. Image yield of abn lab categories is the rate of pos exams/all exams. CONCLUSION: While over a quarter of inpatient abd imaging exams can explain abn LFTs, both the severity and frequency of abnormal LFTs correlate poorly with the diagnostic yield. Not uncommonly, the imaging indication of "abnormal LFTs" were associated with normal or near normal labs, and the diagnostic yield was comparably good. To compare how similar patients evaluate hemodialysis (HD) and peritoneal dialysis (PD) a few weeks after beginning dialysis. METHODS: Patients were participants in the CHOICE study, a national study comparing hemodialysis and peritoneal dialysis outcomes. Because patients were sampled within centers, this analysis was limited to centers offering both modalities. 593 of 644 eligible patients completed baseline rating surveys after a few weeks on dialysis which included 3 overall ratings and 20 items rating specific aspects of dialysis care as poor, fair, good, very good, or excellent. Survey items have been published and were developed through exploratory in-depth interviews and importance ratings by 100 dialysis patients of over 100 candidate items. RESULTS: Patients on peritoneal dialysis were younger (22% PD vs. 13% HD Ͻ 40), more educated (19% PD vss. 33% HD Ͻ high school), more likely to be Caucasian (80% PD, 60% HD), working (26% PD vs. 10% HD), and married (66% PD vs. 51% HD). PD patients were more likely than HD patients to live Ͼ30 miles away from the dialysis center (39% PD, 9% HD). PD patients had fewer comorbidities (ICED 0-1 for 45% of PD and 30% of HD patients). There were no differences among PD and HD patients in sex, or SF-36 physical and mental component health status self-ratings (mean PCS 32 and MCS 45 for both groups). Patients on peritoneal dialysis gave higher ratings of dialysis care overall (PD 82% vs. HD 55% excellent) and of all specific aspects of care rated. Items with the greatest differences were in the domain of information provided (average difference of 35 in % respondents rating excellent) and technical quality, such as the amount of fluid removed (difference 36% excellent). Smaller differences were seen in ratings of coordination among nephrologists and other doctors, availability of nephrologists, and center staff's helpfulness and concern. Adjustment for patient age, race, education, health status, marital status, employment status, distance from the center, and time since starting dialysis did not reduce, and for most items increased, the difference between the groups. CONCLUSION: Both with and without adjusting for patient characteristics, after a few weeks on dialysis patients on peritoneal dialysis rate their care higher than those on hemodialysis. Primary care practitioners may wish to recommend this modality preferentially to end-stage renal disease patients who are eligible for both. Future CHOICE analyses will determine if these results are sustained over a longer term of dialysis, and will compare patient ratings of care to clinical outcomes. BACKGROUND: Physical activity is associated with lower total mortality rates among individuals without a history of heart disease. However, it is unclear whether mortality rates continue to decrease commensurately with higher levels of physical activity. METHODS: We examined the relationship between frequency of vigorous physical activity and mortality in a prospective cohort of 86,738 US male physicians from the enrollment cohort of the Physicians' ealth Study. Study participants were 40-84 years old without prior history of myocardial infarction (MI), stroke, cancer, or liver disease at baseline. On the baseline questionnaire, men reported how often they exercised vigorously enough to work up a sweat. Outcome measures were total and disease-specific mortality. RESULTS: There were 3161 deaths during the mean follow-up period of 5.5 years.

We observed an L-shaped relationship between the frequency of vigorous physical activity and total mortality. After adjustment for other predictors of mortality, compared to the men who reported the lowest (rarely/never) frequency of vigorous physical activity, more active individuals had highly significant reductions in the risk of death during follow-up. Those who reported exercising vigorously: one to three times per month, once per week, two to four times per week, five to six times per week, and daily, had relative risks for total mortality of 0.71, 0.62, 0.59, 0.59, and 0.66, respectively (p for linear trend 0.0001). The inverse association between frequency of vigorous physical activity and total mortality was statistically significant among men aged 40-59, 60-69, and 70-84. Lower mortality rates among men who exercised vigorously was primarily due to lower CVD mortality rates.

Men who exercise vigorously also experienced lower mortality rates from total cancer and other (non-CVD, non-cancer) causes. Much of the inverse association between frequency of vigorous physical activity and cancer mortality appeared to be explained by a significant decrease in lung cancer mortality (p for linear trend 0.01). The remaining four most prevalent causes of cancer death in this cohortcolon, prostate, pancreatic and hematologic-were not associated with frequency of vigorous physical activity. CONCLUSION: These data support a L-shaped relationship between frequency of vigorous physical activity and total mortality, even among older men aged 70-84. The shape of the curve appears to be similar for mortality from cardiovascular disease, cancer, and other (non-CVD, non-cancer) causes of mortality. The inverse relationship between frequency of vigorous physical activity and cancer mortality appears to be largely due to lung cancer mortality. No statistically significant associations between frequency of vigorous physical activity and mortality were seen for other cancers. 

for disease on the decision to take HRT for primary prevention. The purpose of this study was to determine the effect of current data on risks and benefits, as well as individuals' preferences for the various outcomes, on the decision. METHODS: We used a Markov decision analysis model for 50-year-old PMW considering HRT for preventive medicine. The outcome measure was the extension of quality-adjusted life months (QALMs) as a result of taking HRT. We used the best available data on risks of breast cancer (BC), coronary heart disease (CHD), and osteoporosis-associated hip fracture (HF) and the effect of HRT on these risks, and the median utility values chosen by 52 women who evaluated the pertinent health states. For the sensitivity analysis, we varied the utilities from the 25th to the 75th percentile of these values. We assumed the effect found in a trial of HRT for secondary prevention of CHD is the same for primary prevention. RESULTS: HRT increases BC incidence and decreases HF. For CHD (from the HERS Study) there is an initial increase in incidence, followed by a decline. PMW at average risk gain 2.2 QALMs by taking HRT. PMW who value CHD and HF as having the worst impact on health and BC as having the mildest impact gain the most from HRT, 7.9 QALMs. PMW who value CHD and HF as having the mildest impact on health and BC as having the worst impact lose 0.2 QALMs from HRT. Depending on the utility values chosen and on individuals' risk factors, HRT results in a gain of up to 22.5 QALMs (for PMW with low risk of BC and high risk for CHD and HF) or a loss of up to 14.1 QALMs (for PMW with high risk of BC and low risk for CHD and HF). HRT results in smaller gains for black PMW than for white PMW, and smaller gains for older PMW than for younger PMW. CONCLUSION: Unless future studies show a larger benefit on CHD mortality or other health states, HRT decisions for PMW should include careful consideration of individual preferences for all of the potential outcomes. The model can readily incorporate data on new treatments and other outcomes as they become available. PURPOSE: A previous decision analysis suggested that, depending on individual risk factors, HRT might increase life expectancy for many postmenopausal women, but the risks and benefits of HRT for BCS have not been studied. METHODS: We used a Markov decision analysis model for 50-year-old BCS considering HRT for preventive medicine. The outcome measure was the extension of quality-adjusted life months (QALMs) as a result of taking HRT. We used the best available information on risks of breast cancer recurrence, coronary heart disease (CHD), and osteoporosis-associated hip fractures, and the effect of HRT on these risks. We assumed the effect of HRT found in a randomized trial of HRT for secondary prevention of CHD (the HERS study) is the same for primary prevention: an initial increase in incidence, followed by a decline. For the base case analysis, we used the median of the utility values chosen by 52 women who evaluated the pertinent health states. For the sensitivity analysis, we varied the utilities from the 25th to the 75th percentile of these values. RESULTS: BCS at average risk for CHD and hip fracture lose 6.6 QALMs by taking HRT. BCS who value the CHD and hip fracture states as having the worst impact on health and the breast cancer state as having the mildest impact on health lose the least from HRT, 3.0 QALMs. BCS who value the CHD and hip fracture states as having the mildest impact on health and the breast cancer state as having the worst impact on health lose the most from HRT, 9. RESULTS: Information from 2678 women, over 50 years old (range 50-90 years), was analyzed. The tables summarize percentage of women that reported testing in the last two years.

CONCLUSION: The proportions of women 50 years and older who reported breast or cervical cancer screening differed on the source of care. The proportions were lowest for women who reported having no usual source of care. Patients with an internist as provider were more likely to be screened than patients with a general practioner/family practice physician as provider. PURPOSE: Because alcohol has been identified as one risk factor for decreased adherence to antiretroviral therapy (ART), we sought to identify all the factors associated with decreased adherence to ART in a cohort of HIV-infected patients with a history of alcohol problems. METHODS: We measured quantity of alcohol consumed in the past month and adherence to ART in 157 HIV-infected patients with a history of alcohol problems (identified by either у2 positive responses to the CAGE questionnaire or by clinical assessment). Adherence was assessed by self-report using a modification of the AIDS Clinical Trials Group (ACTG) adherence instrument. This information was summarized in a continuous variable defined as the percentage of prescribed doses of ART taken over the prior 3 days. We evaluated other variables previously demonstrated to be associated with poor adherence to ART: social support; cocaine use; heroin use; depressive symptoms; and number of ART doses per day. We first examined the crude unadjusted relationship between alcohol and adherence using correlation analysis and then examined the relationship between alcohol and adherence using multiple linear regression adjusting for variables determined to be significant in bivariate analyses. RESULTS: Subjects had the following characteristics: 19% women; 39% black, 21% Latino, 35% white, 5% other; mean age was 41 years; mean CD4ϩ count was 406/L. Sixty-four percent (100/157) of patients reported 100% adherence in the prior 3 days (mean number of drinks 33); 36% (57/157) reported Ͻ100% adherence (mean number of drinks 80). The correlation between the number of drinks in the past month and percentage 3-day adherence was Ϫ0.29 (p ϭ 0.0002). In the multivariable analysis controlling for cocaine and heroin use, social support, ART doses prescribed, and depressive symptoms, we found that number of drinks in the past month was inversely related to and the only variable significantly associated with 3-day adherence to ART (p ϭ 0.02). CONCLUSION: In HIV-infected patients with a history of alcohol problems, the number of drinks consumed was the only significant variable associated with decreased prior 3-day adherence to ART. Based on this, brief interventions for alcohol problems, which have been demonstrated to decrease alcohol consumption, should be part of a comprehensive effort to improve adherence to ART. 1991-1993 vs. 1994-1998 period was conducted. A cost-minimization study of the program was also performed. 811-785 workers per year were studied, on average 51% were immunized, with the recommended vaccine. Influenza like illness (ILI) episodes and days lost to work were measured in immunized (I) and controls (C), all new cases were recorded monthly since 1991. Annual risk difference (RD) for I and C episodes and days lost was computed for period 1994-1998, and for 1991-1993 preimmunization vs. 1994-1998 CONCLUSION: A history of sexual and physical abuse is common among persons admitted for detoxification, and this trauma is associated with greater substance abuse severity in both men and women. These findings support efforts to develop and evaluate substance abuse treatment strategies that simultaneously address SPA and addiction. year and lifetime breast cancer risk perceptions, and 2) to test the effect of presenting tailored risk information on the accuracy of breast cancer risk perceptions among a primary care population. METHODS: A randomized controlled trial of tailored breast cancer risk information versus a control intervention was conducted. Women 40 years and older were randomly recruited from two primary care practices. Women in the intervention group were presented with tailored risk information derived from the computer based National Cancer Institute Breast Cancer Risk Tool. The tailored risk information was presented to the patient in an illustrated pamphlet. Control subjects received an American Cancer Society pamphlet on breast cancer prevention. One month after the intervention follow-up risk perceptions were assessed. Five year and lifetime risk estimate ratios were defined as the subjects' five year and lifetime risk estimates divided by the computer based calculated risk estimates. Risk ratio differences were defined as the five year and lifetime baseline risk ratios minus the five year and lifetime follow-up risk ratios, respectively. A Mann-Whitney test was used to compare the five year and lifetime risk ratio differences between intervention and control groups. RESULTS: Seventy-seven women have been enrolled in the study to date. Fortytwo percent (42%) of subjects had a high school level education or less, 93% were Caucasian, and 16% were less than 50 years of age. At baseline women overestimated their five year risk of breast cancer by a factor of 15 and their lifetime risk by a factor of 4. The table below presents risk estimates for the intervention and control groups. 

As seen in the table above, women in the intervention group significantly improved their 5 year breast cancer risk estimates compared to women in the control group. CONCLUSION: We report for the first time that women overestimate short term breast cancer risk to a greater degree than lifetime breast cancer risk. However, an intervention that presents tailored risk information to patients in a quantitative and graphic format can improve the accuracy of risk estimates. These findings can be applied to patient education efforts in the primary care setting. PURPOSE: Treatment of sepsis syndrome is expensive, often encompassing a number of discretionary modalities such as pulmonary artery catheterization and albumin. The goal of study was to assess intercenter variation in resource use among patients with sepsis syndrome. METHODS: We conducted a prospective cohort study of 1028 adults admitted to 8 academic medical centers between January 1993 and April 1994. This sample represented a weighted random sample of patients who were in an ICU or had a positive blood culture, and all patients who received a novel therapeutic agent or who died in an emergency department or ICU. Information on all other patients admitted to the 8 centers during the study period (n ϭ 248,761) was also obtained. The main outcomes were differences in total hospital charges (CHARGES) and length of stay (LOS) for patients with sepsis syndrome compared to the remainder of patients in each center. Rates of treatment with varying procedures used after sepsis onset were also measured. PURPOSE: Sepsis syndrome is a common and serious condition associated with substantial mortality, and these critically ill patients often receive expensive and intensive therapy. The goal of study was to assess resource utilization associated with sepsis syndrome in 8 academic medical centers. METHODS: We performed a prospective cohort study in which we evaluated 1028 adult admissions between January 1993 and April 1994, representing a weighted random sample of patients who were in an ICU or had a positive blood culture, and all patients who received a novel therapeutic agent or who died in an emergency department or ICU. The main outcomes were length of stay (LOS) and total hospital charges (CHARGES). We compared these outcomes for cases with those of all other patients admitted to the 8 centers during the study period (n ϭ 248,761). In the sepsis group, we also determined correlates of post-onset LOS. All means and their differences were calculated using appropriate case weights. RESULTS: In linear regression models adjusted for age, gender, insurance status, Charlson comorbidity score, ICU stay, discharge DRG weight and study site, the mean LOS for patients with sepsis syndrome was 18.2 days, which was 11.0 days longer than for all other patients (p Ͻ 0.0001; r 2 ϭ 0.19), while the mean difference in CHARGES was $43,942 (p Ͻ 0.0001; r 2 ϭ 0.43). Eight independent correlates of increased post-onset LOS at or prior to onset of sepsis included pre-onset LOS; stay in an ICU; steroid therapy; surgery; burn injury, multiple trauma, bowel perforation or severe pancreatitis; Hickman catheter; transplant of liver, heart or lung; and cardiogenic shock. CONCLUSION: The large total charge and length of stay of patients with sepsis syndrome compared with those without emphasize the need to assess the proportion of this utilization attributable to sepsis syndrome. Correlates of resource use in this population included a number of factors associated with tertiary care. 

all 60% of respondent rated that role extremely important. The second most important domain (rated extremely important by 42% of respondents) CHW attributes, e.g., worker knowledge, communication and interpersonal skills. The next most important domain, rated by an average of 40% of respondents, was benefits from CHWs, e.g., social support and lowered blood pressure. extremely important. Domains rated fourth and fifth in importance were CHW services and visit arrangements respectively. CHW services, e.g., checking blood pressure and providing health information was rated extremely important by 33% of respondents. Visit Arrangement, e.g., call prior to home visit and the location of the visit in the home, was rated extremly important by 27% of respondents. CONCLUSION: Clients note diverse important aspects of CHW care, and most highly value the role that CHWs play in extending medical care in underserved communities. Evaluation instruments obtaining client rating of CHW services should include aspects of care salient to service recipients and should be used to assess care. PURPOSE: Clinical software is having a growing impact on patient care, with an influence ranging from passive information to active monitoring and therapeutic intervention. Since patient outcomes are increasingly affected by clinical software, a process to monitor and improve the quality of clinical software is warranted.

One key source of information about problems and their clinical impact are the end-users. However, user-feedback is typically analyzed in a piecemeal fashion, and is not usually analyzed in an ongoing way. The purpose of this study was to develop a meaningful classification structure of clinical software user-feedback. Further, our goal was to analyze our findings to evaluate potential impact of problems, to evaluate rates of problem reporting, and to extract lessons for developers of new clinical applications. METHODS: We evaluated all user feedback from a newly implemented computerized physician order entry system at Massachusetts General Hospital, a tertiary hospital in Boston, Massachusetts. From the start of the application in October 1998 through December 1999, all of the 888 feedback comments were analyzed.

Comments that related to operation, policy, or were general remarks were removed, leaving 785 user comments. From these, 18 categories were developed to delineate the software function being performed. This scheme appears reasonably complete (all user comments can be readily assigned), and includes functions such as alerts, co-signature, authentication, order activation, screen formatting, medication start or stop, order status, and results viewing. Next, we rated the impact of the problem as either one of cost, safety, or efficiency. In each category, a range of one to four was applied to indicate severity of impact. RESULTS: In all, 634 comments had an impact on efficiency (70%). 233 comments had a potential impact on patient safety (26%), and comments with a cost impact totaled 36 (4%). (The total number is greater than 785 since some comments had impact on more than one category). The rate of feedback declined from a peak of 231 comments per month initially to a steady average rate of 20 per month, suggesting either responsiveness to early suggestions or user fatigue. Several fundamental problems of importance for future developers emerged, including inadvertent order activation, insufficient alerting, potentially confusing templates, and necessity of safety measures for drug ordering boundaries. CONCLUSION: A clear-cut mechanism for assembling and learning from userfeedback is possible, and is a beneficial cornerstone to developing correct and safe clinical software. When user feedback is classified and analyzed useful lessons can be learned and applied to future software. We conducted a cost minimization analysis of telemedicine by building a model to compare telemedicine with face-to-face care. For cost minimization analysis, comparisons of competing options are analyzed and the one that costs least is favored. Analysis was conducted from both societal and health care provider (Veterans Affairs) perspective. We have captured both direct medical costs and lost productivity costs for societal analysis. However, intangible costs were not included. The time frame of the project was set at 5 years. Standard cost analysis procedures were used and annual cost of managing a patient in each program was calculated. All cost data were standardized to 1999 dollars, future costs were discounted at 3% rate. For base-case analysis, personnel and capital cost data were used from the ongoing telemedicine program and face-to-face care. Sensitivity and threshold analyses were conducted for all variables. RESULTS: 74 patients (116 visits) have been seen in the telemedicine clinic. 9 (12%) of these patients subsequently made a face-to-face visit to MVAMC for a procedure. Results of base-case analysis showed an annual societal cost of $441 per patient managed by telemedicine versus $540 per patient managed by face-toface care. From a health care provider's perspective the base-case cost of telemedicine ($253/patient/year) was higher than that of face-to-face care ($216/patient/ year). Sensitivity analysis demonstrated that if the number of patients seen by telemedicine was doubled, or the duration of telemedicine visit was reduced from 45 (base-case) to 30 minutes, or the cost allocated to lodging for face-to-face care was increased from $30/night (base-case) to $58/night (national average), then telemedicine becomes the favored option. CONCLUSION: Pulmonary telemedicine clinic is a cost effective way to deliver outpatient care to patients residing in an area of limited access to pulmonologists. Telemedicine would be even more cost effective if the analysis included the monetary value of intangible benefits of telemedicine, such as improved access and decreased demands related to patient travel. Approximately 70% of current smokers report a desire to stop smoking completely. However, data from the 1991 National Health Interview Survey of the Center for Disease Control (CDC) indicate that only a little more than half of the smokers who had at least one outpatient visit with a physician or other healthcare professional during one year period were advised to quit. In a retrospective chart review of 32 consecutive patients with ICD 9CM code: 410 (Acute Myocardial Infarction) discharged from our hospital's coronary care unit (CCU), documentation of smoking cessation advice was present in 11 patients' charts (34%). This data prompted us to do a prospective analysis of this important quality of care indicator. METHODS: The housestaff on monthly rotation in the the CCU were given a comprehensive overview of risk factors for CAD and the pervasive problem of the lack of documentation of smoking cessation advice was emphasized in a briefing at the beginning of the month for two consecutive months. The charts were reviewed by an independent observer for the documentation of smoking cessation advice prior to the discharge. RESULTS: During this two month study period, 18 patients with the diagnosis of acute myocardial infarction at discharge from the CCU were identified as current smokers at the admission. Thirteen (72%)patients were male and 5 (28%) were female. Smoking cessation advice was documented in 11 patients' charts (61%). When interviewed by the independent observer at the time of discharge, 5 patients (28%) were determined to quit smoking by themselves, 7 patients (39%) expressed to seek help for smoking cessation, and remaining 6 patients (33%) were not sure of quitting. CONCLUSION: Our data show that taking a simple measure like housestaff briefing, the documentation of smoking cessation advice improved from baseline of 34% to 61%. This suggests that housestaff motivation can indeed improve the outcome related with quality of care issues. A significant improvement in many other aspects of patient care may be achieved by conscious efforts in their training.

PRACTICE. G Albertson, J Kutner, L Schilling, CT Lin, E Cyran, L Ware, J Steiner, RJ Anderson, University of Colorado, Denver, CO PURPOSE: To describe a patient population presenting to a primary care clinic with a written list. We sought how often patients bring a pre-written list, the demographics of this population, their health status and the reason for the visit. METHODS: 507 consecutive patients enrolled in an academic internal medicine clinic presenting for primary care were studied. Patients were asked to complete a pre-visit questionnaire asking about demographic data, reason for today's visit, health status, length of relationship with their PCP, whether they usually bring a pre-written list and the reason for the list. Providers were surveyed with a postvisit questionnaire and asked to rate the medical/social complexity of the patient, to list known or suspected mental health diagnoses and to rate their overall visit satisfaction. RESULTS: 111 ⁄ 507 patients (20%) presented during the study period with a pre-written list Patient gender, level of education, annual income and ethnicity did not differ significantly between those with or without lists. Patients with a list self-rated their overall health lower than patients without a list (p ϭ 0.029). Patients with a list tended to be older (age Ͼ45 years) compared to patients without a list (p ϭ 0.08). Duration of the patient-provider relationship did not influence whether patients presented with or without a list (p ϭ 0.56), but patients were more likely to bring a list when their appointment was with a provider other than their PCP.

Lists were more common in encounters scheduled for medication refills (31% vs 15%, p ϭ 0.001), follow up on an existing concern (61% vs 49%, p ϭ 0.028), to discuss health screening (11% vs 5%, p ϭ 0.046) or to request a test (17% vs 7%, p ϭ 0.002). Patients who brought a list to the study encounter were more likely to bring a list when seeing any doctor compared to patients who did not bring a list to the study encounter (69% vs 14%, p ϭ 0.001). Patients cited help with recall (63%), improvement in visit efficiency (46%), and ease of explanation of symp-toms/concerns (26%) as the most common reasons for bringing a list. Providers rated social/medical complexity higher in patients with a list compared to those without (p ϭ 0.029). Known or suspected psychiatric co-morbidity rated by the provider did not differ significantly between the two groups. Providers found lists to be helpful in 75% of patients who brought a list and PCPs felt a list would have been helpful in 10% of patients who presented without a list. PCP rating of overall visit satisfaction was comparable in patients with and without a list. CONCLUSION: Patients commonly bring pre-written lists to encounters with their PCPs. Few demographic data characterize this group of patients. The old adage that once a list writer, always a list writer seems to in this medical setting. Patients use pre-written lists to facilltate communication of symptoms with their doctor and improve visit efficiency. Providers view patient lists as helpful and no association is noted between list-bringing behavior and an underlying mental diagnosis. PURPOSE: The purpose of our study was to evaluate the installation of a webbased electrocardiogram (EKG) archive retrieval system at a major tertiary care academic hospital. Few studies have been conducted to investigate the benefits of these new data technologies and their impact on provider practice. METHODS: In a before and after trial, all consecutive patients seen in the Admissions Evaluation Center at the Hospital of the University of Pennsylvania between February and April 1998 were enrolled as pre-intervention subjects prior to the installation of a web-based EKG archive retrieval system in November 1998. All patients in the AEC between February and April 1999 were enrolled as post-intervention subjects after the installation of the system. We used logistic regression to study EKG orders and adjusted for comorbitities, age, sex, and whether the patient was a walk-in or pre-scheduled visit. RESULTS: A total of 1543 patients were enrolled in the pre-intervention group and 1005 patients in the post-intervention group. Adjusting for age, sex, comorbitities, and scheduling status, 71.0% of patients had an EKG ordered after computer installation compared to 75.3% prior to installation (p Ͻ 0.043). CONCLUSION: Health care institutions continue to install expensive medical information systems. This study showed a statistically significant, although clinically modest reduction in EKGs ordered after installing an expensive new computer technology. Health care institutions must continue to evaluate whether medical information systems are worth the significant expense incurred during their installation. We sought to determine if enhancement of feedback with an Achievable Benchmark of Care (ABC), a data-driven, peer-based, measure of excellence, would increase feedback effectiveness. METHODS: Medical records were reviewed for 1360 diabetic patients from the practices of 70 Alabama primary care physicians. Physician performance on diabetes-related indicators (measurement of HgbA1c, cholesterol and triglycerides, influenza immunization, and foot examination) was calculated. Indicator performance rate was the percentage of patients who received the procedure at least once during the preceeding 18 months. We calculated an ABC for each indicator based on a subset of top-performing physicians. Physicians were divided randomly into a group that received feedback of personal performance and mean group performance and a group that received similar feedback enhanced with an ABC. Baseline measurement was from 1/94-6/95 and follow up from 1/97-6/98. RESULTS: The table shows baseline performance and absolute improvement for physicians receiving ABCs compared to physicians receiving usual feedback for three indicators:

Physicians receiving ABC enhanced feedback improved more than physicians receiving usual feedback (absolute incremental improvement: 5% for HgbA1c, 9% for influenza vaccination, and 2% foot exam, p Ͻ 0.05 for all). Similar results (not shown) were achieved for measurement of cholesterol and triglycerides. Multivariable analyses adjusted for clustering of patients within physicians without loss of significance, and revealed that fewer years in practice correlated positively with performance improvement.

CONCLUSION: The ABC offers a low-cost, effective approach to enhancing provider feedback. With high face validity, the ABC has many advantages over subjectively defined benchmarks.

PURPOSE: Many hospitalist programs have demonstrated significant cost reductions but the reasons for these cost reductions are not clear. The goal of this study is to define specific patient parameters that may contribute to the savings. We hypothesized that hospitalists are more effective in implementing savings in patients with more common diagnoses, those admitted via the ED (as opposed to ICU) and those who are able to get part of their therapy at home. METHODS: During the 98-99 academic year, WVU implemented a hospitalist service (H) which was compared with the two other medicine services, one staffed by a General Internist (GIM) and the other by a subspecialist (SS). A total of 2577 patients were randomly assigned to the teams based on housestaff call schedule. The discharged patients from the three teams were retrospectively analyzed, specifically looking at the following subgroups: 1) Top 10 percentile cost patients 2) Patients admitted via ED vs "other" source (outside referrals, clinics, direct admits) 3) patients who spent any time in the intensive care unit 3) The top 5 DRGs 4) Patients discharged with home health. RESULTS: The CONCLUSION: Hospitalists are more likely to decrease costs in patients who do not spend time in an ICU, patients whose diagnoses are one of those more commonly seen and patients discharged with home health, perhaps reflecting quicker mobilization of resources. Source of admission (ED versus other) does not impact on the savings. They have a lesser impact on the "most-expensive" patients. To determine if variation exists in postoperative pneumonia (POP) and respiratory failure (RF) rates following noncardiac surgery among VA hospitals after adjustment for patient-specific risk factors. METHODS: Cases were selected from prospective observational data collected through the National VA Surgical Quality Improvement Program. POP was defined as positive sputum culture with antibiotic treatment or an infiltrate on chest x-ray diagnosed as pneumonia or pneumonitis following surgery. RF was defined as mechanical ventilation greater than 48 hours postoperatively and/or reintubation subsequent to postoperative extubation. Logistic regression models using patientspecific risk factors to predict POP and RF were developed using Phase I cases (10/91-12/93, 44 VA hospitals) and validated using Phase II cases (1/94-8/95, 123 VA hospitals). The models were used to estimate the probability of POP and RF for each case with the sum within each hospital designated as the expected number of complications. Observed to expected (O:E) ratios and 95% confidence intervals (CI) were computed for each hospital. RESULTS: Phase I POP rates ranged from 0.8% to 8.5%. After risk adjustment, 7 hospitals (16%) were high outliers (CI Ͼ 1), 12 hospitals (27%) were low outliers (CI Ͻ 1), and 9 hospitals (20%) changed rank by more than 5. Phase I RF rates ranged from 1.0% to 6.0%. After risk adjustment, 5 hospitals (11%) were high outliers, 7 hospitals (16%) were low outliers, and 21 hospitals (48%) changed rank by more than 5. No high outlier hospital for RF was a high outlier for POP. Four of the 7 low outlier hospitals for RF were also low outliers for POP. Phase II POP rates ranged from 0% to 10.5%. After risk adjustment, 26 hospitals (21%) were high outliers, 21 hospitals (17%) were low outliers, and 38 hospitals (31%) changed rank by more than 10. Phase II RF rates ranged from 0% to 6.0%. After risk adjustment, 15 hospitals (12%) were high outliers, 23 hospitals (19%) were low outliers, and 44 hospitals (36%) changed rank by more than 10. Widespread collection and reporting of such quality indicators have been advocated as a means to compare health care organizations. We assessed the variation in such processes of care for CAP among 38 U.S. academic hospitals participating in the University HealthSystem Consortium pneumonia benchmarking project. METHODS: Adults with a discharge diagnosis of CAP and no exclusion criteria (e.g., HIV, tuberculosis, cystic fibrosis, immunosuppressed, hospitalized in past 10 days) were eligible. Retrospective hospital records were reviewed for up to 40 consecutive eligible patients discharged between December 1, 1997 and February 28, 1998 . Process measures assessed included: initial administration of antibiotics (ABX) within 8 hours of arrival, collection of blood cultures (BCX) prior to antibiotic administration, and blood and sputum culture (SPX) collection and oxygen saturation (O2sat) measurement within 24 hours. We report the median percentage of patients achieving each process measure by hospital and their range. Differences across hospitals were assessed controlling for the severity of CAP at admission using the Pneumonia Severity Index. RESULTS: The population included 1100 eligible patients admitted to 38 hospitals. The median percentage of patients achieving each process measure by hospital and the range across hospitals is depicted.

After controlling for pneumonia severity on admission, the percentage of patients achieving these each of process measures varied significantly across hospitals (all p Ͻ 0.005). CONCLUSION: In conclusion, more than 2-fold variation exists in achieving markers of quality care among patients admitted with CAP across U.S. academic institutions. METHODS: We performed a telephone survey of American College of Physicians-American Society of Internal Medicine members who identified themselves as generalists. Physician reports of inpatient care arrangements and responses to questions regarding attitudes toward the hospitalist model were primary outcome measures. A summary score of physician attitude was derived from highly correlated survey questions (␣ ϭ 0.84). Multivariable models were employed to determine factors associated with attitude summary score. RESULTS: Four hundred physicians (51%) agreed to participate. The median age of respondents was 45 years; 81% were male, 50% practiced in independent or small group settings, and most admitted patients to private, non-profit teaching hospitals. Most physicians reported they cared for 2 or fewer hospitalized patients each day. Ninety-four percent were familiar with the term hospitalist, "65% reported that the model was available in their community, and 25% used hospitalist services for all or some of their inpatients. The model was more commonly available in Western states (p Ͻ 0.05). Few (2%) physicians reported the presence of the mandatory" hospitalist model in their community. Seventy-three percent felt that the hospitalist model would reduce continuity of care. Only 28% of physicians felt that their patients would prefer care from an inpatient care specialist, but 51% felt that patients might get better care in such a system. In multivariable models, physicians who were in solo practice had a less favorable attitude towards the hospitalist model (p Ͻ 0.01). Physicians who had training in specialties with more inpatient practice had a more negative attitude (p ϭ 0.047), as did physicians who had 15 or more patients hospitalized each month (p ϭ 0.011). However, physicians who had hospitalists in their community had a more positive impression of the hospitalist model (p Ͻ 0.001). Factors such as physician age and gender, or hospital characteristics were not associated with differences in attitude score. CONCLUSION: In this national sample, hospitalist services were widely available; mandatory hospitalist services were rare. Despite feeling that quality of care might be improved, physicians were concerned about the hospitalist model's effects upon patient-doctor relationships and patient satisfaction. Investigations are needed to determine true effects of the hospitalist model upon patient satisfaction and the doctor-patient relationship.

Backus, M Moron, P Bacchetti, AB Bindman, Medicine; Epidemiology & Biostatistics, UCSF, San Francisco, CA PURPOSE: Preventable hospitalization rates are a previously validated measure of primary care effectiveness. To assess managed care's impact on primary care effectiveness, we examined whether an increase in managed care over time was associated with a decrease in preventable hospitalization rates. METHODS: Hospitalization rates and managed care penetration were calculated on an annual basis for adults ages 18-64 years in each of the 394 zip-code defined primary care service areas across California. Annual statewide hospital discharge data provided information on every hospitalization and whether it was for a patient insured by managed care. Preventable hospitalizations were identified by ICD-9 codes as those with a primary diagnosis of asthma, CHF, COPD, diabetes or HTN. Hospitalization rates were calculated using census estimates of the 18-64 year old population for each area. To diminish the impact of inaccurate population estimates, we created "standardized" preventable hospitalization ratios by dividing the combined rates of these 5 conditions by the admission rate for appendicitis and GI obstruction in each area, as these two conditions were assumed to be stable over time. Managed care penetration in each area was determined on an annual basis as the percentage of hospitalizations for that area that were insured by managed care. Change in managed care penetration was calculated as the average of the 1995 and 1997 values minus the average of the 1990 and 1992 values. Changes in standardized preventable hospitalization ratios were log transformed and then calculated similarly. Because these were not normally distributed, rank correlations (r) were calculated, both overall and within each quartile of initial standardized admission ratios. RESULTS: From 1990 to 1997, managed care penetration increased from 24.3% to 42.2% and overall admission rates declined from 956 per 10,000 to 785 per 10,000. The change in managed care was not associated with a change in standardized preventable hospitalizations overall (r ϭ Ϫ0.04; 95% CI Ϫ0.14, 0.06; p ϭ 0.42). However, the negative association between increasing managed care penetration and changes in the standardized preventable hospitalization ratios increased progressively when moving from areas in the lowest quartile of baseline preventable hospitalization ratios towards those in the highest quartile. In the 98 areas with the highest initial ratios, increases in managed care penetration were significantly associated with a decline in the standardized preventable hospitalization ratios (r ϭ Ϫ0.3; 95% CI Ϫ0.47, Ϫ0.11; p ϭ 0.003). CONCLUSION: An increase in managed care, particularly in areas with initially high preventable hospitalization rates, was associated with a decrease in these rates over time suggesting an improvement in primary care effectiveness.

Levy, J Arterburn, Medicine, Univ. of TX Health Science Center at San Antonio and the Audie L. Murphy Memorial Veterans Hospital, San Antonio, TX PURPOSE: SUMSearch (http://SUMSearch.uthscsa.edu) is a new method to find medical evidence that uses meta-searching and contingency searching. As SGIM is considering endorsing this product, we summarize its recent usage. METHODS: We report the results of the last 502 consecutive searches. We optimistically define a successful search as retrieving at least one article from MED-LINE. RESULTS: 421 (84%) of searches were successful. Previously, we reported that 22% of searches used more than two words within any one search term ("long terms") and these searches were significantly more likely to fail. Prior to the current series, SUMSearch was programmed to warn clinicians when they were about to submit long terms. In the current series, 12% of searches contained long terms and the presence of long terms was no longer a predictor of failed searches. SUM-Search now records when the MESH browser is used to select search terms. Searches used the MeSH browser in sixteen percent of searches. These searches were significantly less likely to fail (8% versus 16%). In 50% of the successful searches, clinicians clicked at least one link to a document. The median time was 155 seconds to finish selecting a median of two documents. The shortest time needed was 10 seconds to click a single document. The longest time occurred when a search that was executed, then three hours later 12 documents were clicked over 20 minutes. CONCLUSION: 84% of searches retrieved at least one document. Based on the current and a previous analysis, searches that use the MeSH browser and avoid long search terms are more likely to succeed. Clinicians click documents from only 50% of successful searches. Possible explanations for not selecting documents include: perceived lack or genuine lack of germane citations, overlooking germane articles, and insufficient time to review list of documents. These results for clinicians using SUMSearch are consistent with prior research that showed 50% of literature searches done by medical librarians are able to answer clinical questions. We plan to redesign the list of citations returned by SUMSearch to facilitate review of the many citations. METHODS: A cohort of diabetic enrollees in the state's largest academic Medicaid MCO was assessed longitudinally to compare their exposure to diabetic services before and after TennCare. A total of 88 cohort members met the following inclusion criteria: 1) age 18-64 throughout the study period (92-96), 2) continuous enrollment for at least 320 days in two years of TennCare (95 and 96), and in the Medicaid program in two years before TennCare (92 and 93), 3) at least one diabetes diagnosis or medication prescription in each year of the study, and 4) no Medicare or other third party insurance. Claims data were used to assess baseline characteristics and chart review data were used to assess health services utilization for 82% of cohort members (n ϭ 72) for whom complete medical records were available. Data for 94, the year of transition to the TennCare program, was excluded from the analysis. The paired t-test was used to compare exposures to clinical services and outcomes before (92-93) and after TennCare (95-96). RESULTS: For the cohort participants, average age was 48.5, 76.4% were female, 90.3% were black, 2.8% white and for 6.9% race was unknown, 86.4% were eligible on the basis of disability, 10.2% were eligible on the basis of AFDC, 3.4% were eligible on the basis of blindness. A large percentage were identified to have complications of diabetes during the study period, including nephropathy (14.9%), retinopathy (22.0%), neuropathy (14.9%) and osteomyelitis (2.8%). Subjects had an average of 6.0 vs. 7.9 outpatient clinic visits per year (p ϭ .003), 0.5 vs. 0.7 diabetic eye exams (p ϭ .08), 0.2 vs. 0.5 foot exams (p ϭ .1239), 0.4 vs. 0.6 cholesterol assessments (p ϭ .0042), and 0.4 vs. 0.9 glycosylated hemoglobin assessments annually (p ϭ .0011). No other significant differences were found in rates of cholesterol assessments, urine protein assessments, mammography, average blood pressure, average total and LDL cholesterol. However, average glycosylated hemoglobin decreased from 9.6 to 8.3 (p ϭ .0165). CONCLUSION: Diabetic enrollees experienced increases in utilization of many recommended health services, and despite disease progression, experienced improved glucose control following the initiation of Medicaid managed care. These increases may reflect historical trends or, alternatively, improved chronic disease care in a primary care gatekeeper system. METHODS: This survey-based study of access to care, service utilization and satisfaction with care, was administered at outpatient clinic sites in an academically affiliated group practice consisting of 120 primary care providers. There were no significant differences in terms of age, gender or ethnicity between responders (75%) and non-responders. The current study focuses on the 478 (12% of 3,990) patients who answered "yes" to the question "Do you have asthma?". Categorical variables were analyzed using Fisher's exact test for comparison of proportions. In comparisons of means for continuous variables, 2-sided t-tests were used. Results with a two-tailed p-value less than or equal to 0.05 were considered statistically significant. The independent association of insurance status with perceived access problems and emergency department use was assessed using logistic regression analysis. RESULTS: Compared to patients with other forms of insurance in the same managed care network, patients covered by the state's Medicaid program were more likely to report access problems for asthma-related care, including difficulties reaching a health care provider by telephone and obtaining clinic appointments and asthma medications. Adjusting for relevant clinical and sociodemographic variables, Medicaid patients were significantly more likely to report at least one access problem compared to non-Medicaid patients (AOR ϭ 3.3, p Ͻ .0005). Patients reporting at least one access problem were also more likely to have made at least one asthma-related emergency department visit within the past year (AOR ϭ 2.4, p Ͻ .0001). Reported barriers to care did not translate into reduced patient satisfaction. CONCLUSION: These findings suggest that health insurance alone does not guarantee health care access. Comparison of Medicaid patients to other enrollees within the same managed care organization provides a reliable assessment of perceived barriers to care by controlling for variation in the system of care delivery. Additional study is needed to determine why some Medicaid patients, who have financial access to services and are documented users of primary care services, nonetheless report barriers to obtaining care. The existence of multiple barriers to the use of services in an insured population has important implications for patients, purchasers and community health planners, particularly in light of ongoing efforts to produce the highest quality, most efficient system of care. Method: An assessment was conducted as follows: 1) a 13-page (175 item) self-administered questionnaire was completed, which included questions on demographics, adherence to universal precautions (UP), work stress, and risk factors for tuberculosis (TB) and hepatitis; 2) a 5 unit PPD tuberculin skin test (TST) was placed, and 3) one tube of blood was drawn for hepatitis B and C serologies. RESULTS: Of 102 eligible CHCWs in RI, 81 agreed to participate (79%), and 59 (73%) completed all three parts. Out of 62 TST's, two positive reactions (Ͼ10mm) were found (3.2%), and one borderline conversion at 9mm (1.69%). The two positive TST's included one who had been treated for TB in the past, and one who noted it was his usual BCG reaction. In comparison, Rhode Island hospital tested 3357 employees between January and September of 1999, reporting a conversion rate (Ͼ10mm) of 0.42%. Six CHCWs (7.5%) were positive for hepatitis B, all past infections (hepatitis B surface antigen negative), but two were previously unknown. Of the two unknowns, one had indicated non-occupational risk factors for hepatitis on the questionnaire. The other CHCW had only core antibody (IgG) positive, no surface antibody, suggesting remote infection. CHCWs immunity to hepatitis B due to vaccination was also reported. Five CHCWs (6.8%) other than those positive for hepatitis B, did not receive the entire series of immunizations. Of those who did, 19 (28%) had no hepatitis B surface antibody (no evidence of immunity). Three CHCWs (3.8%) were positive for hepatitis C antibody, two previously unknown infections. One CHCW scored the second lowest on adherence to UP. Although the other CHCW was in the top 25% of UP adherence, previous employment in a high risk job was reported. CONCLUSION: This is the first study to examine the occupational risk of tuberculosis and bloodborne infections in CHCWs. The data suggest that CHCWs may be at increased risk than health care workers (HCWs) for tuberculosis. The prevalence for hepatitis B is lower than that reported for HCWs in the medical literature (10%-20%), which may be due to the observed rate of hepatitis B vaccination. The prevalence for hepatitis C in CHCWs was slightly higher than reported values for HCWs (1%-2%). Ongoing research will determine if these trends are observed in other correctional facilities. . Frequencies of procedure use and crude death rates were calculated (per 100 IHD patients). Race and gender specific adjusted odds ratios and 95% confidence intervals (C.I.) for hospital mortality were calculated, controlling for potential confounders, using multiple logistic regression analyses. Odds ratios (OR) compared death rates among those who had the specified procedure with those who did not. RESULTS: There were 36,683 IHD patients. Their mean age was 65.8 years (S.E. ϭ 12.6). Sixty-one percent of the Whites were men; 55% of the Blacks were women. Frequencies for procedure use, crude hospital death, and the adjusted OR (95% C.I.) for hospital death, respectively are listed below.

Whites underwent angiography (angio), PTCA, and CABG more often than Blacks. However, the crude hospital death rate was highest among White women and lowest for Black men. Using hospital death as the outcome variable and controlling for age, insurance status, and co-morbidities (hypertension, diabetes, heart failure, and hypercholesterolemia), the race and gender specific OR (95% C.I.) revealed that Black women who underwent CABG were more likely to die, compared to Black women who did not have CABG (OR ϭ 2.5; 95% C.I. (1.5-4 .0)). A similar trend was observed among Black men, while CABG was associated with lower death rates among Whites. CONCLUSION: These data suggest that CABG was associated with higher in-hospital mortality among Blacks with IHD. Referral bias or limited access to coronary angiography and PTCA/CABG may preclude Blacks from obtaining the maximum benefit of these procedures. Prospective studies on diverse populations with documented IHD are needed to shed insight into patient outcomes, following PTCA/ CABG. The primary care of patients whose symptoms are medically unexplained (est. 25%-50% of ambulatory visits) often frustrates providers and results in patient dissatisfaction. There are few empirical, quantitative studies that assess the costs of ineffective and maladaptive utilization of medical care by somatizers. METHODS: Consecutive patients presenting for an office visit in a primary care clinic at a large urban hospital were asked to complete a 27-question screening survey. Items adapted from the Somatic Symptom Inventory and the Whiteley Index of hypochondriasis evaluated bodily complaints, health anxiety, and bodily preoccupation. Patients with scores above (N ϭ 212) and below (N ϭ 664) the 86 th percentile were compared. This threshold was selected in order to identify the estimated 14% of primary care patients with a diagnosable somatoform disorder. Demographic characteristics along with morbidity, resource utilization, insurance status and service costs were obtained from hospital records for 12 months preceding the screen. RESULTS: Of the 3344 patients invited to participate, 788 (24%) were ineligible, 720 (28%) declined, 224 (7%) did not complete the survey, and 114 (3%) were excluded for incomplete records. Patients in the below-threshold group were randomly selected from each score tertile. After adjusting for comorbidity (Charleson Index), and sociodemographic characteristics, the top-scoring group had higher total, physician, lab and procedure costs, and more primary care and specialist visits. Emergency ward visits and hospitalization rates were not significantly different. Adjusted differences in selected outcome measures by whether above or below somatization score threshold (N ϭ 876, SE in parentheses)

CONCLUSION: Somatizers are disproportionately high utilizers of medical services. Even after controlling for serious morbidity, they accrued costs 38% higher than other patients, made about 3 more office visits, and saw one additional specialist per year. Early detection and intervention may prevent the development of an enduring pattern of maladaptive and costly utilization. PURPOSE: Physician self-disclosure can be defined as statements made by a physician that describe his/her personal experience in areas that have medical and/ or emotional relevance for a patient. Although the use of physician self-disclosure has been regarded as either a boundary violation or as a means of building trust and rapport with patients, little is known about its prevalence or impact on the physician-patient relationship. METHODS: We utilized an existing database of physician-patient encounters that consists of approximately 10 audiotaped office visits for each of 124 physicians (59 primary care physicians and 65 surgeons). Patients in the study were mostly Caucasian (85%) and female (55%). The mean age was 51 years. A trained coder listened to each audiotape and coded for physician self-disclosure as defined above using the Roter Interaction Analysis System. After the encounter, each patient completed a survey that assessed sociodemographics, health status and various satisfaction measures. The differences between patient response to visits with and without physician self-disclosure were examined using ANOVA and linear regression. RESULTS: Self-disclosure occurred with approximate equal frequency in primary care visits (102/589, 17%) and in surgical visits (93/676, 14%). The relationship between physician self-disclosure and patient-reported outcomes differed by physician specialty. For surgeons, self-disclosure was associated with higher patient ratings in nonverbal communication (p ϭ .000) and in the American Board of Internal Medicine communication behavior scale (p ϭ .003). These patients were significantly more likely to report greater feelings of warmth/ friendliness (p ϭ .007) as well as overall satisfaction (p ϭ .001). For primary-care physicians, selfdisclosure was consistently associated with poorer ratings of these same measures. In fact, following visits with primary care physicians who made statements of self-disclosure, patients were significantly less likely to report that they would recommend their doctor to a friend (p ϭ .05). There was a statistically significant interaction for physician specialty and self-disclosure on all of the above measures. These relationships were not changed when controlling for patient age, gender, race, education, self-rated physical and emotional status or the number of prior visits with that physician. CONCLUSION: Physician self-disclosure appears to be predictive of greater patient satisfaction with surgical but not primary care visits. These surprising findings suggest that the communication between patients and physicians across specialties may carry subtle differences in dynamics and consequences that define the patient experience in ways that merit further investigation. PURPOSE: Although many have investigated the degree to which patients prefer to participate in their own medical decisions, the purpose of this study was to measure surrogates' desire for participation in end-of-life decisions in the intensive care unit (ICU) setting and to compare that desire with their impression of how decisions were actually made. METHODS: All patients who died in any of 3 ICUs (surgical, medical, or cardiac) at a large inner-city academic medical center between April and October 1999 were eligible for inclusion in the study. Trained interviewers conducted telephone or inperson interviews with one closely involved family member or other surrogate within 2 months after each patient's death. Each surrogate was asked about his/ her desire to participate in decisions for the patient and then to describe how decisions were actually made. Responses were categorized into 3 domains: decisions made mainly by surrogate, mainly by physician, or jointly by surrogate and physician. RESULTS: Of 136 deaths over the 6-month period, we identified a surrogate for 104 patients. 49 of 104 surrogates responded to the survey and 35 of 49 respondents indicated that there were decisions to be made for the patient. 43% of respondents were African-American, 52% were Ͼ65 years old, 52% had at least some college and 60% were a spouse, 24% were a child, 10% were a parent, and 6% were a sibling of the patient. 51% of surrogates reported a desire to make joint decisions with the physicians, although 37% reported a desire to make decisions themselves and 12% preferred that physicians make all the decisions. When comparing surrogates' desire for participation with their actual self-reported level of participation, there was only 57% observed agreement with a weighted Kappa statistic of 0.33 (95% confidence interval, 0.06 to 0.59). Of 15 surrogates whose desire did not match their actual involvement in decisions, 12 (80%) were given more responsibility in decision making than they preferred to have. Most of the mismatch occurred when the surrogate desired a joint decision. Of 18 surrogates who reported a desire for joint decisions, only 7 (39%) reported that the decision was made jointly. CONCLUSION: Most families or other surrogates want primary or joint responsibility for decision-making for terminally-ill patients in the ICU setting but a substantial number of surrogates report that they are given more responsibility than PURPOSE: Health care staffing in hospitals tends to be less intense on weekends than on weekdays, despite the relatively consistent day-to-day burden of disease.

We tested whether mortality rates differed for patients with serious emergencies who were admitted to hospital on a weekend rather than on a weekday. METHODS: All acute care hospitalizations resulting from an emergency department visit in Ontario, Canada were analyzed between 1988 and 1997 (n ϭ 3,805,788). Hospital mortality for patients admitted on weekends was compared to hospital mortality for patients admitted on weekdays using a chi-square analysis and logistic regression techniques. RESULTS: We found no difference in aggregate hospital mortality rates for patients admitted on a weekend rather than a weekday (5.9% vs 5.9%, p Ͼ 0.20). Patients with ruptured abdominal aortic aneurysms (n ϭ 6,661), however, were significantly more likely to die if admitted on a weekend rather than a weekday ( To determine if manuscript submissions to U.S. general medical journals have declined relative to those at specialty and international journals since 1994, and to determine trends in representation of U.S. and international contributors. METHODS: We solicited submissions information from 55 medical journals, including 6 general and 25 specialty journals in the U.S., and 5 general and 19 specialty journals internationally. Editors provided data on total manuscript, research, and non-research submissions, and on authors' countries of origin between 1994 and 1998. Trends for individual journals were modeled using linear regression. Trends by journal category and comparisons between general and specialty and U.S. and international journals were analyzed using generalized estimating equation regression for absolute and relative change. RESULTS: All 6 U.S. general medical and 20 other journals responded, for a response rate of 47%. Results are reported as (absolute; percent) change in manuscripts per year. Submissions rose and fell for U.S. general journals individually but no significant change was detected in aggregate for total (20. 3 In a prior study, using evidence-based guidelines created by local breast cancer experts, we found that 1 in 6 (44/275) women did not receive beneficial adjuvant treatment in accordance with the guidelines; 22 had RT and 22 had systemic therapy omitted. To understand the reasons their physicians had for treatment omission, we conducted semi-structured interviews and audio-taped all 13 surgeons at a tertiary academic medical center whose early-stage breast cancer patients did not receive local or systemic treatments in accordance with the guidelines. One investigator reviewed all interview transcripts, a second reviewed 40 transcripts. Based upon content, categories were created and reasons for underuse coded with excellent inter-rater reliability ( ϭ 1.0). Reasons for underuse were classified into categories. The physician thought: 1) the patient should have been treated and could not explain why they weren't (34%); 2) the patient should not have been treated due to an extenuating common clinical circumstance not addressed by the guideline, e.g., age (30%), or an unusual clinical circumstance, e.g., concurrent second cancer (11%); 3) the patient should have been treated but they refused treatment (16%): 4) the guideline did not apply to a patient because they disagreed with the clinical content of the guideline (9%). Categories of root causes of underuse of breast cancer adjuvant treatments were identified. In no instance was the physician unaware of guideline-recommended treatments. Interventions targeted at root causes of underuse such as a tracking system or patient navigator to prompt and follow-up on referrals, and discussion at multidisciplinary rounds of patients with extenuating clinical factors could address the majority of reasons for underuse of beneficial breast cancer treatments. RESULTS: In the UCC, of 17,441 adult patients seen, 5949 (34%) were screened with 263 positive responses (incidence 4.4%). In the ER only 2152 patients were screened (less than 20%) with 263 positive responses (incidence 6.7%). In the UCC, 944 cards (16%) and in the ER 810 cards (38%) were returned marked "unable to screen". CONCLUSION: The prevalence of recent abuse in patients successfully screened at the UCC was 4.4% and in the ER 6.7% (overall 5%). This prevalence is consistent with published data considering our population included both men and women. Our concern is that only 34% patients in the UCC and less than 20% in the ER were successfully screened. Common barriers to screening include 1) patient not alone, 2) mental or physical status of patient, 3) language barrier, and 4) caregiver's lack of comfort in asking the question. Physician support of routine screening is important to compliance with screening. Despite an increased length of the training session and an outside trainer/expert, the ER was less successful at screening. In spite of the barriers, both practice areas are planning to continue universal screening. Physicians were asked to complete a 28 question survey and self-report their own BMI, height and weight. The instrument demonstrated reliability of the knowledge and attitudes subscales of 0.98. The subscales were fitted with probabalistic conjoint models, and linear logit scores were used for summary knowledge and attitudes values. RESULTS: The majority of respondents were male (55%), and most (93%) reported no prior special training in obesity. The average self-reported BMI was 20 (range 1-90) compared with an average actual BMI of 23 (p ϭ .06). Responses did not vary significantly by year of training or physician gender. Despite having reasonable understanding of the medical consequences of obesity, basic obesity knowledge questions were the most difficult for respondents as assessed by probabilistic conjoint models. 66% did not know the threshold BMI for obesity, and 69% did not know that waist circumference is a reasonable measure of obesity. 41% failed to report their own BMI within 10% of its actual value; inability to accurately estimate one's own BMI was correlated with decreased confidence in the ability to treat obesity (p ϭ 0.04). Although 95% of physicians agreed that treating obesity was important, attitudes towards treatment were generally negative. Respondents overwhelmingly reported a lack of treatment success (70%), 59% felt unqualified to treat obese patients, and nearly 30% reported treatment to be futile. 66% indicated confidence in their skills, and confidence was associated with an increased self-perception of qualification (r ϭ .317, p ϭ .004) and success in treating obesity (r ϭ .575, p Ͻ .001). Summary knowledge and attitude scores were not correlated (r ϭ 0.07). CONCLUSION: Despite a good knowledge of the medical outcomes of obesity, the respondents had a poor grasp of the basic tools necessary to identify obesity. Many negative attitudes toward the treatment of obesity were also expressed. If physicians are to be the major change agents for obese patients, better physician educational efforts must be designed to address both knowledge deficits and induce attitude change. METHODS: We performed a cross-sectional study of IDM in private primary care physicians' (n ϭ 59) and surgeons' (n ϭ 65) offices. Completeness of IDM was measured by content analysis of audiotaped physician-patient discussions of clinical decisions, using criteria representing the minimum discussion for basic IDM: discussion of the nature of the decision and either the patient's role in decision making or the patient's preferences. We used the percentage of all decisions for each physician that were complete as an aggregate measure of physician performance of IDM (IDM score). Additional data included physician and patient demographics, number of years in practice and hours per week in practice for physicians, physician perceptions of the how well they knew the patient, attitudes and beliefs towards malpractice risk, and beliefs about the importance of attention to psychosocial concerns. We used multivariate linear regression with generalizing estimating equations to ascertain associations between these characteristics and IDM score. RESULTS: We analyzed 1057 audiotaped office visits, with approximately ten patients for each physician. Physicians averaged 20% complete decisions by minimum criteria, with a range of 0-50%. In multivariate analysis of all physicians, fewer years and fewer hours per week in clinical practice were both independent predictors of higher IDM score (p ϭ 0.005, 0.028 respectively). The physician's perception that they knew the patient "well" had a negative association with IDM score. (p ϭ 0.037). Surgeons had higher IDM scores than primary care physicians, though not statistically significant. (22 vs. 19, p ϭ 0.098) No other physician or patient characteristics, nor visit length, showed significant association with IDM score. CONCLUSION: We found that few physicians meet minimum criteria for IDM in routine office practice. Fewer years in practice and fewer hours per week predicted better IDM, while believing to know the patient well predicted poor performance. Physicians in practice fewer years may have less paternalistic notions of decision making than their senior colleagues. Those working fewer hours per week may percieve less time pressure, despite lack of independent association between visit length and IDM score. Physicians who believe they know the patient well may assume that patients no longer desire involvement, though it is unclear whether their patients would concur with that assumption. Performance of foot examinations was not affected by sociodemographic factors. CONCLUSION: It is reassuring that race and education were not associated with differences in quality of care. However, although all subjects in this study were insured and had identical health benefits, older persons, women, and those with lower income were at greatest risk for not receiving needed care for their diabetes. Despite a higher burden of co-morbidity, patients seen by GIM had significantly lower rates of both mortality and cardiac complications. As well, patients seen by nursing showed statistically significant decreases in LOS and cost. Although the adjusted rates of complications and cardiac complications were higher in patients seen by anesthesia, this may reflect a tendency for anesthesia to see most high risk cases. CONCLUSION: We found that an association exists between pre-operative GIM consultation and reduced cardiac complications and mortality, even though overall complications, LOS, and cost are not improved. Further work is now needed to explore the mechanisms for this encouraging association. The baseline year also demonstrated that only 52% of patients received appropriate antibiotics (range 34% to 72%) based on an expert panel using the 1998 IDSA guidelines for hospitalized CAP. During the implementation phase, improvement efforts focused on two recommendations: initiating antibiotics in the ER and choosing appropriate antibiotics for treatment. Followup data after implemetation showed significant changes from the baseline year. Overall delivery of the first dose of antibiotics in the ER increased from 63% to 89% (OR 4.8, 95% CI 2.3-9.7). Appropriateness of antibiotic choice increased from 52% appropriate to 86% appropriate (OR 5.7, 95% CI 2.9-10.9). All hospitals in the study improved on both measures. CONCLUSION: A program of clinical profiling, academic detailing, and targeted areas for improvement is associated with improvement in the ER care of patients with community acquired pneumonia. Future programs designed to improve care and reduce medical errors may benefit from similar approaches. BACKGROUND: Use of HRT is related to patient demographics, but little is known about physicians' role in women's choices. We investigated the use of HRT in postmenopausal women in Maryland to determine the role of general internists, family practitioners, and gynecologists in prescribing HRT, and the associated reasons for starting HRT. METHODS: In 1999, we mailed questionnaires to a population-based sample of 1000 randomly selected women ages 50 and over living in the 5 largest counties in Maryland. We asked about demographics, source of information about HRT, use of HRT, characteristics of their primary care physician (PCP) and HRT prescriber, and reasons for starting or not starting HRT. RESULTS: 29 surveys were undeliverable; 603 surveys were returned, of which 408 (42% overall) were completed and 195 were blank. The respondents had a mean age of 63.6 years (SD 9.5), 75% were white, 18% African-American, 65% married, and 35% had a household income less than $40,000. 46% were currently taking HRT, 15% took it in the past, and 39% never took it. Overall, 64% of women cited a physician as their primary source of information about HRT, 9% the news media and 7% a woman's magazine; 4% had never heard of it. Those who received the information from a physician were 7.5 times more likely to take HRT than those who didn't, after controlling for age, race, income and education (p Ͻ .001). Although 60% of respondents' PCP's were general internists, 27% family practitioners and 7% gynecologists, 76% of current users were prescribed HRT by gynecologists and only 14% by general internists and 4% by family practitioners. Depending on the specialty of the prescriber, current users had different reasons overall for starting HRT (p ϭ .03); women prescribed HRT by gynecologists as compared to internists were less likely to begin HRT for osteoporosis prevention (18% vs. 36%), or cardiovascular prevention (23% vs. 32%) and more likely for control of menopausal symptoms (22% vs. 18%). Among women who never took HRT, 40% never discussed it with their PCP. Of those who discussed it, 40% chose not to take it because their PCP did not recommend it, and only 8% because of fear of breast cancer. CONCLUSION: In this population-based sample, nearly half the respondents were currently taking HRT. Physicians are the primary source of information about HRT for most women and appear to have an enormous impact on a woman's choice to take HRT. We developed a tool to evaluate four aspects of the quality of these studies: study design and choice of control group, statistical treatment of the results, patient selection and evaluation of the intervention's use. Two trained readers evaluated each article; differences were resolved by consensus. RESULTS: Our search identified 588 articles pertaining to critical pathways. We excluded 493 because they did not have a control group and 19 because they did not meet our definition of a critical pathway, leaving 76 articles for analysis. Most critical pathways involved surgical (67%) or medical (13%) diseases and the vast majority (92%) were in hospital settings. STUDY DESIGN: A randomized controlled trial design was used in one article (1.3% ). An observational design with historical controls was used in 95% of the trials and with concurrent controls in 11% (some trials had more than one control group). TREATMENT OF RESULTS: Costs were reported in 71% of the trials and 91% of these showed a benefit (lower costs). These benefits were supported statistically in only 36%. Hospital length of stay (LOS) was reported in 90% of the trials and 84% of these showed a benefit (shorter LOS). These benefits were supported statistically in only 49%. Accounting for potential differences in case-mix between groups occurred in only 38% of trials. PATIENT SELECTION: Patient selection most commonly used procedure codes (63%) or billing codes (35%). Clinical criteria to define eligibility were reported in 33% of trials. Seventy-five percent had specific enrollment dates for all groups in the study. EVALUATION OF USE: Only 37% of the articles reported some analysis to assure the reader that the intervention (the critical pathway) was used correctly. The most common method reported was training of users (75%) and less often reported was actual frequency of use (39%). CONCLUSION: The preponderance of literature supporting the effectiveness of critical pathways is based upon non-randomized observational studies using historical controls. Most studies do not attempt to assess patient case-mix. Before continuing widespread implementation, we should design trials that address previous methodological pitfalls and validly prove or disprove the effectiveness of critical pathways. METHODS: Cross-sectional post-visit surveys and audiotape analysis of 283 patients seeing 31 physicians in 15 urban primary care practices who identified themselves as being either African American or white. Main outcomes were patient ratings of physicians' participatory decision-making (PDM) style and communication behaviors. Analysis of variance and multiple linear regression were used to measure the associations between race-concordance and the outcomes. RESULTS: Patients, aged 18-88 years, were 58% African American, 68% female, and 17% college graduates. Physicians were 58% African American and 55% female. African American and female physicians had higher PDM scores than their counterparts, p ϭ 0.05 and p ϭ 0.01, respectively. Patients in race-concordant relationships rated their physicians' PDM styles as more participatory (p ϭ 0.07). The length of visits ranged from 13 minutes for African American patients with white doctors to 25 minutes for white patients with African American doctors (p ϭ 0.001). Differences in communication behaviors included: psychosocial, positive, and emotional talk, and patient-centered interviewing by physicians, and psychosocial talk by patients. White physicians were the least patient-centered regardless of patient race, African American physicians with African American patients were intermediate, and African American physicians with white patients were the most patient-centered. There were no differences in physician verbal dominance, physician use of facilitators, or physician and patient biomedical talk by race-concordant status.

CONCLUSION: Patient ratings of partnership with physicians are higher in raceconcordant relationships. Audiotapes confirm differences in several patient-centered communication behaviors by race-concordant status. Future analyses based on underlying theories of partnership between physicians and patients may clarify the nature and context of these differences. PURPOSE: Older patients hospitalized with medical illnesses are frequently discharged with worse ADL function than their pre-illness baseline. Our goal was to describe the course of functional loss in these patients and to examine the impact of increasing age on functional loss. METHODS: On hospital admission (ADM), we asked 2309 patients (mean age 79) or their surrogates about the patient's independence in 5 ADL (bathing, dressing, eating, transferring, toileting) at baseline two weeks before ADM, and at ADM. We interviewed them again at DC. We defined functional decline as the loss of independence in at least one ADL. RESULTS: Overall, 799 (35%) patients declined in function between baseline and DC. Of these, 402 patients declined before ADM and 397 patients declined between ADM and DC. Of the 1510 (65%) patients who did not decline between baseline and DC, 455 patients declined before ADM, but recovered before DC, while 1055 patients did not decline either before or after ADM. As shown in the table, the oldest patients were the most likely to decline between baseline and DC, both in bivariate analyses, and in multivariate analyses controlling for APACHE scores, Charlson comorbidity scores, presence of dementia, and baseline ADL and IADL function.

CONCLUSION: We conclude that functional decline is common in older hospitalized patients, often occurs after hospital ADM, and the risk markedly increases with advancing age. PURPOSE: Over 40 million Americans are uninsured. Among the major obstacles to universal insurance are a lack of consensus about a basic benefit package and concern about cost. To address these barriers to universal coverage a project was designed to provide uninsured patients the opportunity to define a benefit package within cost constraints. METHODS: Groups of uninsured individuals recruited from community settings and ambulatory care sites located in central North Carolina were convened to solicit their insurance choices using a simulation exercise, Choosing Healthplans All Together. In this group exercise, participants design a managed care package by selecting from 15 categories of services, trading between breadth of services and flexibility of restrictions (e.g., formulary, co-payments, utilization review) within the constraints of a fixed monthly premium. The premium is comparable to the typical per member/per month premium paid by U.S. employers for managed care plans and permits coverage of 60% of the services offered in the exercise. RESULTS: Fifty-seven participants in 5 groups, out of a planned total of 240 participants in 20 groups, participated in the exercise. Participants were predominantly male (86%), minority (88%), and socio-economically disadvantaged (90% had an annual income Ͻ$15,000). Participants picked the following services in descending order of frequency: hospitalization (98%), pharmacy (91%), dental (91%), vision care (81%), tests (79%), primary care (74%), long term care (70%), home health (67%), mental health (61%), specialty care (58%), uninsured coverage (56%), last chance (transplantation and clinical trials, 53%), allied health and other services (51%), complementary medicine (49%), and infertility (23%). Those services that were more frequently included in participants' plans were also more likely to have the more flexible option chosen. Participants chose to include an average of 10 services with intermediate flexibility (score of 1.8) on a scale from 1 (tightly managed care) to 3 (more flexible care) during their first selection, and slightly broader coverage (10.25 services) with less flexibility (1.75) during their final selection following role-playing illness episodes. CONCLUSION: These data suggest that socio-economically disadvantaged, uninsured individuals tend on average to pick a package that balances breadth and flexibility of health services when offered the opportunity to design their own health insurance coverage within the constraints currently experienced by many U.S. employees. OBJECTIVE: Many quality improvement programs target specific diseases, but little work has been done to identify structural components of primary care offices related to quality of care such as number and type of staff, or office procedures. In the process of developing a self-assessment tool for primary care practices, we aimed to identify important characteristics of office practice associated with primary care quality from the physician's perspective. METHODS: We recruited internists and family practitioners for two-hour focus groups. The sessions were audiotaped and transcribed, and two physicians (GLD, JAF) reviewed transcripts to identify all office characteristics possibly related to quality of care; a third adjudicated disagreements (DEF). We assembled these 60 characteristics in a questionnaire and distributed them to focus group participants. Participants ranked the characteristics on two scales from 1 to 10 based on the importance to quality and the degree to which the characteristic was under control of the physician. We calculated mean scores for importance to quality and control. RESULTS: Twenty-five clinicians from different practices participated in focus groups and 23 ⁄ 25 returned ranking questionnaires. Sixteen of the 23 were internists, 10 were in solo practice, 11 had more than 1 ⁄ 3 of patients capitated, 5 were women, and 6 were non-white. The table shows the mean (SD) for the five highest and five lowest ranked of the 60 practice characteristics.

We also asked how MDs would spend an extra $50,000 to improve quality. Of the 47 comments, 20 mentioned increasing salary or hiring new staff, 10 mentioned computers, and 7 mentioned new office equipment. CONCLUSION: Physicians rated interpersonal aspects of care and staff at the top of the list for impacting quality. Rankings for specific office procedures were more disparate. These physicians appear not to embrace disease management programs, an evidence-based intervention, for improving quality of office-based primary care. INTRODUCTION: Hypertension is among the most common treatable conditions in primary care and is a risk factor for morbid vascular events. We studied whether a single outpatient blood pressure (BP) reading can predict such events among hypertensive inner-city primary care patients. METHODS: We identified primary care patients with hypertension listed among their outpatient problems during their first scheduled visit in 1993 to the primary care practice affiliated with an urban public hospital. Data were extracted from a 25-year old electronic medical record system and included the systolic BP (SBP), diastolic BP (DBP), and heart rate (HR) routinely recorded during their index visit, prior comorbid illnesses and results of laboratory and imaging tests recorded on or before the index visit in 1993, and hypertensive drugs taken at the index visit. We then searched their next 5 years' records for inpatient and outpatient diagnoses and laboratory tests indicating the first occurrence of myocardial infarction (MI), stroke (CVA), heart failure (HF), ischemic heart disease (IHD), chronic renal insufficiency (CRF), and all-cause death. For each adverse outcome, independent (predictor) variables were assessed via Cox regression. Those variables found to be univariately significant predictors were entered in final multivariable Cox models. RESULTS: We identified 5825 hypertensive primary care patients who kept scheduled visits in 1993; 69% were women and 63% black. Their mean age was 57 years. In the next 5 years, 433 (7%) had an MI, 1017 (17%) a CVA, 1385 (24%) IHD, 1309 (22%) HF, 718 (12%) CRF, and 766 (13%) died. Controlling for clinical factors and prior history, the SBP (but not DBP) recorded at the index visit was found to be a significant multivariable predictor of CVA (p ϭ 0.001), combined CVA/MI (p ϭ 0.002), IHD (p ϭ 0.005), and CRF (p Ͻ 0.001). HR predicted death (p ϭ 0.003), and no vital sign predicted MI alone or HF. A 10 mmHg higher SBP was associated with an increase of 13% (95% CI 6-21%) in the risk of CRF, 9% (3-15%) for CAD, 7% (3-11%) for CVA, and 6% (2-9%) for combined CVA or MI. A 10 beat/min increase in HR was associated with a 16% (5-26%) increase in the risk of death. No significant association with any of the target outcomes was found for mean BP, pulse pressure, the HRBP index, or the ratio of SBP to DBP. CONCLUSION: Vital signs recorded routinely during primary care visits have surprisingly significant prognostic value. Specifically, a single systolic blood pressure reading predicted the subsequence occurrence of CVA, combined CVA/MI, IHD, and CRF while heart rate predicted death. These are thus true vital signs which should be attended to during primary care visits.

Studies, University of Chicago, Chicago, IL PURPOSE: Adherence to medication is a major determinant of therapeutic effectiveness. Recent advances have led to significantly improved methods of monitoring and promoting adherence to therapy, and several statistical approaches have been proposed for analyzing adherence data. Yet, classic intent-to-treat (ITT) analyses of randomized controlled trial (RCT) data intentionally ignore subjects' adherence; this approach may leave clinical decision-makers with more limited information than they would desire. The purpose of this study was to examine recent, clinically influential RCTs, to assess how the latest studies measure, report, and interpret data regarding adherence. METHODS: High-impact RCTs within the fields of general internal medicine and medical subspecialties were identified through a review of all articles presented in the 1997-98 and 1998-99 "Update" series in the Annals of Internal Medicine. Cited articles were excluded if they described: non-drug treatment, non-self-administered study drug(s); therapy related to perioperative care or outpatient procedures; or a comparison of inpatient vs. outpatient therapy. 65 articles cited in the "Update" series and published in 1997 or 1998 qualified for analysis. Assessment included: size and site of study; duration of active therapy; measures of daily adherence to therapy; and report and analysis of adherence data. RESULTS: Among recent influential RCTs, there was a broad range of study size (range ϭ 20-18,790 subjects; median ϭ 450) and duration of active therapies (range ϭ Ͻ1 week-10 years; median ϭ 48 weeks). Only 35 articles (54%) included at least one measure of daily adherence to study therapy; 20 of these utilized frequently biased measures such as pill counts or patients' self-report. Measures of adherence data were more commonly employed in smaller (Ͻ450 subjects Little is known about the frequency of illness in persons with addictions who lack primary medical care. Our objective was to assess the burden of medical illness in this population and to identify substances of abuse and other patient characteristics associated with worse health. METHODS: Patients without primary medical care who were admitted for alcohol, heroin or cocaine detoxification were enrolled in the Health Evaluation and Linkage to Primary Care (HELP) study. Subjects completed standardized interviews assessing demographics, social support, substance use and medical problems. Burden of illness was assessed by self-report of physicians' diagnoses, and by the Short Form Health Survey (SF-36). To characterize subjects with the most illness, we assessed the relationship between patient characteristics and the SF-36 Physical Component Summary (PCS) score in bivariate and multivariable analyses. RESULTS: A total of 470 persons were enrolled. The mean age was 35.7 (SD 7.8) years, 76% were male and 46% were African-American; 57% had an annual income of less than $20,000 and 38% were unemployed. Almost half of all patients (45%) reported being diagnosed with a chronic illness, 80% had prior medical hospitalizations and 21% had been prescribed medications for a chronic physical illness. The most common diagnosed chronic conditions were asthma/chronic obstructive lung disease (20%), hypertension (16%) and chronic liver disease (12% The objective of this study was to determine whether early achievement of quality process markers in CAP is associated with i) higher proportions reachng clinical stability within 48 hrs of admission by the criteria of Halm et. al., ii) decreased hospital length of stay, or iii) fewer inpatient deaths. METHODS: Retrospective chart review identified 1100 eligible patients discharged with a primary or secondary diagnosis of CAP between 12/1/97 and 2/28/98 from among the 38 U.S. hospitals in the University Health System Consortium pneumonia benchmarking network. Logistic regression was used to assess the univariable and severity-adjusted relationship between each process measure and the 3 clinical outcomes listed above. Test of trend assessed the relationship between process markers and admission pneumonia severity (PSI). RESULTS: All 3 process markers were achieved more frequently the higher the PSI class (p Ͻ 0.05).In univariable and PSI-adjusted analyses, process measures were not significantly associated with better outcomes (Table) . However, the PSI-adjusted ORs for early ABX and BCX were compatible with shorter lengths of stay and fewer inpatient deaths, respectively.

CONCLUSION: In our analysis, process markers were not significantly associated with improved outcomes. Our results may be subject to residual confounding by pneumonia severity, and their interpretation may be limited by insufficient sample size. However, they suggest these process markers are not strong or uniform determinants of outcome in the context of care academic hospitals provide. The American College of Physicians guidelines for perioperative care recommend the use of beta-blockers for certain patients to improve outcomes after surgery. Studies of physician behavior, with respect to guidelines and recommended practices, have shown that beta-blockers have been underutilized after myocardial infarction. We evaluated physicians' concordance with the perioperative use of beta-blockers. METHODS: We reviewed the medical charts of adult patients having a cholecystectomy at a tertiary care medical center from 1997 through 1998. Two trained chart abstractors reviewed the medical records for demographic and clinical information and determined whether a cardiologist or non-cardiology medical specialist performed a preoperative evaluation. Patients met criteria for perioperative beta-blocker use if they had a history of coronary artery disease or if they had the presence of two or more risk factors: age greater than 65 years, history of hypertension, diabetes mellitus, hypercholesterolemia, or current smoking. We also noted emergency operations and the presence of contraindications to beta-blocker use. RESULTS: Among the first 87 cases of cholecystectomy we reviewed, criteria for beta-blocker use were met in 42 (48%) of patients. Among these, 4 (9%) had emergency operations and 2 (5%) had contraindications to beta-blocker use. For the remaining 36 patients, the average age was 68 years, 55% were female, 26% had coronary artery disease, 50% had a documented preoperative medical evaluation in the chart by a non-cardiologist, 22% had an evaluation by a cardiologist, and 28% had no documented evaluation. Of the 9 patients with documented coronary artery disease, 6 were evaluated by non-cardiologists and 3 by cardiologists. There were 5 (14%) patients on beta-blockers prior to admission and 31 patients not on beta-blockers. Patients evaluated by cardiologists were more likely to be on betablockers (38%) than those evaluated by non-cardiologists (7%, p ϭ 0.06). Of the 31 patients not on beta-blockers at admission but meeting criteria for their use, none were started on beta-blockers preoperatively. The only postoperative cardiovascular complication was one case of congestive heart failure. CONCLUSION: We found that perioperative beta-blockers are underutilized in patients with risk factors for coronary artery disease, despite evidence that their use in appropriate individuals may be life-saving. Our findings are consistent with previous surveys of physician practice on beta-blocker use. If larger studies confirm these findings, quality improvement initiatives are indicated to reduce perioperative and postoperative complications of coronary artery disease. naire and a health habits survey. The intervention comprised 1) mailed dietary recommendations focusing on fruits and vegetables, meats, whole fat dairy, and low fat dairy, tailored to patients' baseline intake and stage of change; 2) verbal endorsement by the PCP of the benefits of these recommendations; 3) two telephone motivational interviews with study staff to set dietary goals; and if needed, 4) a telephone consultation with the study nutritionist. To account for cluster randomization, we used a mixed model analysis of variance. RESULTS: We recruited 230 subjects in the intervention and 274 in the control group. Demographics for the intervention and the control groups were, respectively: mean age, 50 and 57 years (p Ͻ 0.01); proportion female, 77% and 63% (p Ͻ 0.01); race/ethnicity, 78% white, 13% African American and 95% white, 1% African American (p Ͻ 0.01). 89% of the participants completed the final survey. The baseline mean (SD) intake of fruits and vegetables was 3.0 (1.7) servings/day in the intervention group and 3.3 (1.6) servings/day in the control group. At 3 months, intake of fruits and vegetables had increased by 1.1 servings/day in the intervention and by 0.3 servings/day in the control group. Using an intent-totreat analysis and adjusting for age, sex, race, and baseline intake of fruits and vegetables, we estimated that the change in fruit and vegetable intake in the intervention group was 0.56 (95% CI 0.31, 0.81) servings/day higher than in the control group. The change in consumption of red and processed meats, whole dairy, and low-fat dairy, did not differ between the control and intervention groups. Process data suggested that most subjects focused on their intake of fruit and vegetables rather than the other 3 food groups. The intervention was well received by both participants and PCPs. CONCLUSION: This low intensity nutrition intervention was effective in helping patients increase their fruit and vegetable intake. Clinical interventions that combine practice supports and brief physician endorsement may be effective and efficient in producing behavior change among patients. METHODS: Medical residents completed a standardized patient data collection form for the "first" and the "most interesting" patients in a given three hour clinic session. For each case, the residents designated the follow-up interval and identified new diagnostic questions. The next day, the group of residents (average 4-5) met with an attending for "follow-up interval" rounds at which randomly selected cases were presented. For each case presented, each resident in the group independently chose a follow-up interval and identified new diagnostic questions.

To quantitate the residents' responses regarding the assignment of follow-up interval and the identification of new diagnostic questions, we defined the variable, "agreement." For each case, the residents were considered to have "agreement" in follow-up interval if: (1) the same follow-up interval was chosen by у75% of the respondents, or (2) all respondents chose a follow-up interval that was within Ϯ50% of the mean. "Agreement" in identification of new diagnostic questions was defined as at least two-thirds of the respondents identifying the same diagnostic question in a given case. RESULTS: A total of 82 cases were presented. Agreement in the assignment of follow-up interval was seen in 46% (38/82) of the cases, with the average length of follow-up interval of 4.2 weeks. Agreement in the identification of new diagnostic questions was seen in 82% (67/82) of the cases. In cases where residents agreed on the identification of new diagnostic questions, the agreement on follow-up interval was 52%; however, in those cases where residents did not agree on the identification of new diagnostic questions, there was only a 20% agreement on follow-up interval. CONCLUSION: When presented with the same clinical data there is poor agreement (46%) between resident physicians in the assignment of follow-up interval.

One factor which appears to be associated with the level of agreement in follow-up interval is the level of agreement in identifying new diagnostic questions. However, even at high levels of agreement (82%) in the identification of new diagnostic questions, there is poor agreement in the assignment of follow-up. This suggests that there are other factors which influence residents' decision making regarding follow-up interval. In Press). Physicians were ranked according to the mean RVI from the clinical vignette study and this ranking was used as a predictor for actual practice patterns. Provider specific predictors were sequentially entered into a mathematical model to determine their role in explaining the variance in revisit intervals. RESULTS: The characteristics that contributed to the shortening of the revisit interval included physician perceptions of patient stability (p Ͻ .001) and patient compliance (p ϭ .012). Physician decision to order diagnostic tests (p ϭ .15) and to change the treatment of the main diagnosis (p Ͻ .001) also resulted in shorter RVI. When studied in actual practice, these physician characteristics accounted for 29% of the variance in the assignment of the revisit interval (r ϭ .53). When the provider's rank from the clinical vignette data was added into the model, these five variables accounted for 88% of the variance in assignment of revisit intervals (r ϭ .93). CONCLUSION: These data suggest that clinical vignettes assessing physician behavior can strongly predict physician practice patterns. Combined with other physician level data, information from properly constructed vignettes gives excellent correlation with observed physician practice variation. This model needs to be validated in other physician populations and expanded to other practice parameters. Level of training had no effect on RVI. Physiologic markers of disease severity contributed to physician perceptions of the patient's stability but did not significantly contribute to the assignment of the RVI. Several physician specific characteristics were associated with shortening of the RVI. Patients perceived to be non-compliant were assigned a shorter revisit interval (10 v. 13 weeks, p Ͻ 0.012) as were those considered unstable (8 weeks v. 15 weeks (p Ͻ .001). Ordering diagnostic tests shortened the RVI from 14 to 12 weeks. When therapy for the primary diagnosis was changed, the RVI was significantly shortened (15 v. 10 weeks, p Ͻ .001). 29% of the variance in assignment of RVI was accounted for by physician perceptions about patient stability and compliance and practice parameters (r ϭ .53). 71% of the encounters studied were with male physicians. Disease severity for patients of male and female residents did not differ. Female physicians were more likely to overbook themselves to see their patients (p ϭ .016). The sex of the physician was a significant predictor of RVI. Females assigned a shorter RVI (10 v. 13 weeks, p ϭ .013) irrespective of level of training, perceptions of the visit, or other practice parameters monitored. CONCLUSION: Among physicians trained at the same institution, variation exists in the assignment of the RVI that is not explained by patient characteristics. The timing of the RVI was also predicted by physician gender. This may reflect the female physicians inclination to provide more health education and preventive care. A total of 32% reported having postponed needed medical care during the twelve months preceding the interview. Overall 12% had postponed needed care because they could not take time off work, 12% did not have any way to get to a medical appointment, 11% had responsibilities for taking care of others, and 10% were too sick. Moreover, 21% of the sample had gone without needed health care during the past 12 months because they had to spend the money on food, clothing, or rent. Additionally, 45% had gone without needed medical care at a County facility during the past 12 months because they thought it would have been too expensive. Women, adults, Latino/Hispanics, patients without health insurance, and patients at contracted community clinics were the most likely to report having gone without care because of competing needs. Barriers to needed health care continue to exist, even for patients who rely on Safety Net Providers to fulfill their medical needs. Competing needs for other basic necessities are an important factor contributing to delays in care. Identification of these barriers is important for eradicating disparities in access to and receipt of medical care. PURPOSE: To identify patients' reasons for seeking health information on the Internet, type of information sought and perceptions of the quality of information available; and to establish whether medical information obtained is shared with medical providers. METHODS: We mailed a confidential survey to 1000 patients, age 18-65, randomly selected from billing files of a general internal medicine practice. The survey consisted of demographics, questions about patient use of the Internet for medical information, frequency of use, reasons for use, and respondents' perception of quality of information. The survey also assessed whether medical information acquired on-line is shared with medical providers. RESULTS: The response rate after the first mailing was 27%. Respondents' mean age was 47.6 (SD 12.1) and 62% were female; 74% indicated that they had used the Internet for medical information. Common reasons for using the Internet were to obtain information on: nutrition/diet (64%), side effects and/or complications of medical therapy (61%), alternative/complementary medicine (44%), second opinions about medical conditions (40%), cancer (17%), diabetes (9%) and women's health (9%). Of those using the Internet, 59% felt that the quality of information was "excellent" or "very good", and 41% assessed the quality of information as "good" or "fair". Of Internet users, 65% rated information on the Internet as the same or better quality than that from their doctor. Most (60%) indicated that they do not discuss this information with their physicians, and 12% indicated that they went "on-line" for health information instead of seeing or speaking with their physician. CONCLUSION: Patients use the Internet to seek information on a broad range of medical questions, and a majority are confident that the information is as good or better than the information that they receive from their physician. Physicians need to understand the benefits and limitations of this extensive medical resource available to patients and to recognize the possibility that patients may use the Internet as an alternative to consulting their physician. PURPOSE: How disease-state management programs affect patient and physician satisfaction is largely unknown. We compared the satisfaction and knowledge of 2 cohorts of patients: (1) patients randomized to have access to a multi-disciplinary telephone-based anticoagulation service (AS), and (2) patients receiving their anticoagulation care from their primary-care physician. We also assessed physician satisfaction with the AS. METHODS: We surveyed 300 patients (mean age ϭ 73) about their satisfaction with their anticoagulation care and their knowledge about warfarin therapy. We also surveyed the 17 referring physicians about their experience with the AS. RESULTS: In an intent-to-treat analysis, patients randomized to the AS sites were more satisfied with the timeliness of getting their blood test results (P ϭ .02), were more likely to know what a safe international normalized ratio (INR) value was (P ϭ .001), and felt safer taking warfarin (P ϭ .04) as compared with patients at the control sites. In an on-treatment analysis of patients enrolled in the AS versus control site patients, AS enrollees were more satisfied with the timeliness of getting their blood test results (P ϭ .001), with the availability of the staff (P ϭ .03), and with the overall quality of their anticoagulation care (P ϭ .02). AS enrollees also were more likely to know what a safe INR value was (P ϭ .001), what a low blood test or INR meant (P ϭ .01), and felt safer taking warfarin (P ϭ .02). Physicians reported that using the AS saved, on average, 4 minutes of their time and 13 minutes of their staff's time, per INR. All physicians were willing to recommend the use of the AS to a colleague. CONCLUSION: Patients and physicians are satisfied with anticoagulation care provided by a multi-disciplinary, telephone-based AS. These findings suggest that a disease-state management program can be endorsed by primary-care physicians and improve patients' satisfaction with and knowledge about their antithrombotic therapy.

Services Research, Durham VA Medical Center, Durham, NC PURPOSE: To compare patient outcomes with rhinosinusitis managed by primary care providers (PC) versus otolaryngologists (ENT) with respect to disease severity, comorbidities, general and disease-specific quality of life (QOL), patient satisfaction, and utilization of VA services. METHODS: On 165 consecutive patients treated for rhinosinusitis from the PC and ENT clinics we collected demographic and comorbidity information. At enrollment, 6 weeks, and 12 weeks later we recorded responses to surveys on general QOL (Short Form-12, SF-12), rhinosinusitis QOL (Sinonasal Outcome Test-20, SNOT-20), disease severity and duration (Chronic Sinusitis Survey-Duration-Based, CSS-D), and patient satisfaction (Visit Specific Questionnaire, VSQ). Prescriptions, radiology, and laboratory data, and number of visits for rhinosinusitis were also collected at the same intervals. Group differences were compared using Wilcoxon analysis and t-tests as appropriate. RESULTS: Baseline demographic variables between groups (86 PC and 79 ENT) were equal except for a higher proportion of patients with a history of nasal polyposis in the ENT group (41% ENT vs. 20% PC, p ϭ 0.004). At baseline, ENT patients had lower physical health scores (PCS-12, p ϭ 0.02), but PC patients had lower mental health scores (MCS-12, p ϭ 0.03). There was no difference in SNOT-20 scores, or symptom severity scores (CSS) between the groups. However ENT patients had a longer duration of nasal congestion (p ϭ 0.006), previous antibiotic use (p ϭ 0.006) and nasal spray use (p Ͻ 0.001). Overall patient satisfaction was higher in patients treated by ENT (p ϭ 0.01) due to differences in time spent with the provider (p ϭ 0.02) and waiting time (p Ͻ 0.001). At 6-and 12-week follow-up, there were no differences in the proportions of patients who were cured, improved, the same, or worse. At 6-and 12-weeks, SF-12 scores were unchanged. In both groups, SNOT-20 scores improved from baseline to 6 weeks (p ϭ 0.007 PC, p ϭ 0.03 ENT), but did not change between 6 to 12 weeks. Symptom severity scores decreased in both groups at 6 weeks, however ENT patients had continued decreases in severity scores at 12 weeks. ENT patients had higher resource utilization (more visits, labs, radiological tests, and surgical procedures) during the 12week follow-up period (p ϭ 0.001). CONCLUSION: We conclude that outcomes of patients treated by PC or ENT improve by 6 weeks and remain stable at 12 weeks as measured by disease-specific QOL and symptom severity surveys. Variations in antibiotics prescribed, followup visits, labs, radiological tests, and surgical procedures between PC and ENT are due to differences in type of sinusitis. PC subjects are more likely to have acute or recurrent sinusitis that responds to initial treatment. ENT subjects are more likely to have chronic symptoms or are treatment failures. In the data is the patient's residental zip code at initial clinic presentation. For the period 1989 to 1999 yearly counts and percentges of new patients for each zip code were assembled using the SAS statistical package release 6.12. The study has approval from the Mercer University IRB and the Bibb County Health Department. RESULTS: The table shows the results for the three most frequently encountered zip codes (31201, 31204, 31206) during the study period. The only other zip code with notable percentages of new cases was 31030 which had 7% of the new patients in 1996 and 11% in 1998. Zip code 31030 is located thirty miles from the core zip codes. The median number of zip codes in a year was 11 with a minimium of 3 in 1989 and a maximum of 28 in 1997. Generally over the study period the number of new patients increased yearly and the number of zip codes also increased; but the three core zip codes always contributed at least 50% of the new cases. CONCLUSION: The three core zip codes, located in the largest population concentration of the region (Macon, Georgia) and representing 45% of the thirteen county population of 250,500, contributed at least 50% of the new cases in any year. One possible confounding factor explaining this could be the location of the clinc in one of the core zip codes. Because of transportation or other potential issues, HIV patients in outlying zip codes may be presenting to clinics outside of the region. Alternatively before diagnosis patients may be relocating to the core zip codes to access improved health and social services.

HYPERTENSION. SM Dy, JH Young, K Roberts, M Lawrence, A Jacobsen, D Brotman, S Putman, MJ Klag, Division of General Internal Medicine, Johns Hopkins, Baltimore, MD PURPOSE: Admissions for severe, uncontrolled hypertension (HTN) are common at our urban teaching hospital, often resulting in unnecessary morbidity. To evaluate how best to prevent these admissions, we assessed correctable risk factors for poor control in these patients, including lack of knowledge, access to care, and medication-related issues. METHODS: We prospectively screened all African-American urban residents admitted to our internal medicine services between 8/99 and 1/00 for the presence of severe HTN. Patients were eligible if the average of 2 blood pressures in the emergency room was at least 180 ⁄ 110 . Patients were excluded if a secondary cause (e.g., pregnancy) or other medical condition (e.g., pulmonary edema) explained the blood pressure elevation. Physicians abstracted medical information from charts and patients completed a detailed interview. RESULTS: Of 62 eligible patients, 57(92%) agreed to participate. The mean age was 51 years (range, 28 to 81 years), 45% were female, 44% had finished high school, and 52% were employed. Half were smokers, and 21% reported active cocaine use. The mean blood pressure was 209 ⁄ 130 , and the mean duration of HTN was 10 years. Half had been previously admitted for HTN within the last five years, and 80% rated their health as fair or poor. Only 13% had not been on antihypertensives and only 10% had not seen a doctor within the past year. Two-thirds reported having a primary care physician; of these, 80% were able to give his or her name. All but one patient was able to correctly name at least one adverse consequence of uncontrolled HTN, and 80% could name two or three. Two-thirds of patients had medical insurance, but only half of these had full pharmacy coverage. Forty-six percent had difficulty paying for medications, and 79% of these said that they sometimes missed medications because of financial problems. However, 65% of patients said that they took their medication every day when they had it. Seventy percent of patients were not actively taking blood pressure medication at the time of admission, and half of the rest were poorly controlled at baseline. Difficulty with cost was the most common reason why patients were not taking their medication. Eighty-six percent of patients with cost problems were not taking their medications on admission, as compared to 47% of those without cost problems. CONCLUSION: Most of these severely affected hypertensive patients had a regular physician and knew about the consequences of HTN. Most were trying to take medications, but many had difficulty because of the expense. Interventions to reduce hospitalizations for severe HTN should include improving access to antihypertensive medications.

1996. LE Egede, X Ye, MD Silverstein, General Internal Medicine; Center for Health Care Research, MUSC, Charleston, SC PURPOSE: An estimated 16 million people in the United States (US) have diabetes mellitus. Diabetes cost the US $98 billion in 1997. 42% of the US population was estimated to use alternative care in 1997 at a cost of $21 billion, and alternative care was most commonly used for chronic medical conditions. We evaluated the prevalence and pattern of alternative care use in persons with diabetes in 1996. METHODS: The Medical Expenditure Panel Survey (MEPS) is a survey of the United States civilian and non-institutionalized population, drawn from the National Health Interview Survey database, and it provides a nationally representative estimate of health care utilization, expenditure, and insurance coverage. We did a cross-sectional study using 883 MEPS diabetic respondents that reported using any form of alternative care treatment in 1996 to determine patient characteristics of alternative care users compared to non-users and patterns of use. RESULTS: The prevalence of alternative care use in diabetics was 7.8% compared to a prevalence of 4.3% in the general MEPS sample in 1996. Among all diabetics, alternative care users differed from non-users. Users were more likely to be between the ages of 45-64 years (57% Vs 38%), have Ͼ12th grade education (54% Vs 24%), and have employment other than self-employment (52% Vs 32%). Users of alternative care were more likely to receive nutritional advice (41%), to receive alternative care from a nurse provider (14.5%), and were more likely to have discussed it with their primary care provider (55%). CONCLUSION: In this sample of diabetic persons, the prevalence of alternative care use is lower than estimated in the literature (7.8% vs. 42%). Diabetic alternative care users were more likely to be older, have Ͼ12th grade education, and be employed when compared to diabetic non-alternative care users, and pattern of use involved nutritional therapy which is consistent with the disease manifestations. Zip 1989 Zip 1990 Zip 1991 Zip 1992 Zip 1993 The purpose of this study was to determine the prevalence of common psychiatric disorders among patients during their first visit to a neurology clinic, and the likelihood of establishing a neurologic diagnosis if a psychiatric disorder is present. METHODS: We administered a prospective diagnostic survey to a consecutive sample (n ϭ 234) of new referrals to a hospital-based neurology clinic. Before their visit, all patients completed the PRIME-MD. After the visit, the study neurologist, blinded to the PRIME-MD results, completed a questionnaire regarding their diagnostic assessment, which included the presence or absence of psychiatric disorders. RESULTS: Psychiatric disorders were detected in 39% of patients. The most common diagnoses were somatoform disorder (24%) and depression (23%) (some patients had multiple disorders). A neurologic diagnosis was made in 85% of cases, 79% of which were felt to be primarily organic, 5% psychiatric and 16% both organic and psychiatric. The presence of any psychiatric disorder was associated with the lack of a neurologic diagnosis (OR 0.5, 95% CI 0.3-0.9). Subgroup analysis found specifically the diagnosis of somatoform disorder (OR 0.5, 95% CI 0.3-0.98) or anxiety disorder (OR 0.3, 95% CI 0.1-0.8) associated with the lack of a neurologic diagnosis. CONCLUSION: Thirty nine percent of new patients seen in a general neurology clinic had a psychiatric disorder, higher than the rate found in primary care (20-29%). The presence of any psychiatric disorder, somatoform disorder, or anxiety disorder reduced the likelihood that an organic cause would be found for the patients' symptoms. Prospective studies are needed to determine if screening for psychiatric disease prior to referring patients with unexplained neurologic complaints would reduce costs or improve recognition of potentially treatable psychiatric disorders. The purpose of the present study, the Primary Care Anxiety Project is to document the psychopathology, clinical course, psychosocial functioning, treatment received and economic burden of anxiety disorders in subjects presenting to a general medical setting for treatment of a non-psychiatric medical illness. This is a pro-spective, naturalistic (i.e., observational), longitudinal follow-up study and subjects will be followed for a minimum of two to three years. METHODS: We report on the intake data from 282 subjects who screened positive for anxiety and who met DSM-IV criteria for Panic Disorder, Generalized Anxiety Disorder, PTSD, Social Phobia, or Mixed Anxiety Depression Disorder on a structured clinical interview. The screening measure was an adaptation of DSM-IV criteria into self-report questions. This is the first prospective follow-up study of anxiety disorders in a general medical setting. This paper reports findings on the mental health and medical treatment utilization of PCAP subjects. RESULTS: Fourty percent of all subjects with anxiety disorders also had a comorbid major depressive episode (MDE). While the presence of MDE did not increase the likelihood of psychosocial treatment (e.g., psychotherapy), the presence of depression clearly increased the likelihood of receiving any psychotropic medication (Chisq ϭ 24.6, p Ͻ .001). For example, 43% of anxiety-disordered subjects with MDE versus 14% of anxiety-disordered subjects without MDE received an SSRI (Chisq ϭ 29.6, p Ͻ .001) and 23% of subjects with comorbid MDE received an anxiolytic compared to 14% without an MDE (Chisq ϭ 3.21, p Ͻ .07). Multiple regressions indicated that, even while controlling for the number of medical illnesses and other predictors, the presence of Post-Traumatic Stress Disorder was strongly related to high numbers of general medical visits (OR:1.93, p ϭ .04) and hospitalizations (OR:2.12, p ϭ .03). CONCLUSION: These data indicate that anxiety disorders present a major "cost" to the individual and society in the sphere of treatment utilization. Patients with anxiety disorders in primary care settings, particularly those without a comorbid MDE, receive psychotropic medication at relatively low rates. Primary care physicians may benefit from further education regarding the assessment and treatment of anxiety disorders. METHODS: A survey instrument was developed and pretested which asked how willing physicians would be to perform 8 American Medical Association (AMA) proscribed actions and 2 allowed actions involving lethal injection, based on a 4 point Likert scale. Questions assessing reasons for participation in capital punishment, and attitudes toward the death penalty and assisted suicide were included. The impact of attitudinal and demographic variables on the number of proscribed actions respondents would be willing to perform were analyzed via Analysis of Variance (ANOVA) and multiple logistic regression analyses. RESULTS: Of the 962 surveys which were received by subjects, 413 (43%) were completed and returned. Forty-one percent of respondents indicated they would be willing to perform at least one proscribed action, while 25% were willing to perform 5 or more actions, and 14% were willing to perform all 8 proscribed actions. Perceived duty to society (p Ͻ 0.001), and approval of the death penalty (p Ͻ 0.001) and of assisted suicide (p ϭ 0.015) all correlated with an increased number of actions willing to be performed by respondents. Previous membership in the AMA also correlated with an increased willingness to be involved in lethal injection actions (p ϭ 0.031). Only 3% of respondents knew of any guidelines on this issue. CONCLUSION: It is disturbing that many physicians would be willing to be involved in the execution of adults against their will, despite clear opinions against such involvement in cases of capital punishment by the medical profession. However, these opinions have not been well disseminated among physicians; medical societies and ethicists must therefore better inform the profession about these issues. The issue of physician involvement in the process of capital punishment should also be addressed at a societal level, given the impact that societal norms have on physicians' decisions in these cases. No differences between the two groups were seen in the presence of leukocytosis, the frequency of comorbidities (diabetes, peripheral vascular disease, immunosuppression), prior antibiotic use and consultation requests, the duration of antibiotic therapy, or readmission rate. In multivariate analysis, patient age and the presence of fever retained a highly statistically significant association with physician base specialty. Linear regression showed that duration of fever, but not physician base specialty, was independently associated with length of stay. CONCLUSION: At our institution patients admitted to medical services tend to be older and are more likely to be febrile than patients admitted to surgical services. Surgeons tend to use cloxacillin more commonly as first-line therapy, whereas internists prefer cephalosporins. Length of stay is comparable on surgical and medical services, and is significantly associated with duration of fever. Surgeons are far more likely to see patients in follow-up than internists. Our data show considerable practice variation, and suggest that the development of more standardized approaches to management would be desirable. In an era of managed care, patients are increasingly concerned about the physician-patient relationship and access to needed services. In this study, we assessed the impact of physician (MD) behavior and process of care on patients' intention to leave their usual site of care and unwillingness to recommend that site to family or friends. METHODS: All patients aged 20-75 with at least one visit in the prior year to a staff physician at 11 ambulatory care sites were randomly selected for telephone survey and chart review. Chi-square and t-tests were used to examine associations between the outcomes and demographics, health status, reports and ratings of care, duration and site of care, number of visits and type of provider. Potentially modifiable aspects of MD behavior and process of health care delivery were analyzed with logistic regression, adjusting for other univariate predictors of unwillingness to return or recommend. RESULTS: Of 2669 respondents (response rate 55%), 5.8% were not willing to return and 6.9% would not recommend their site of care. In multivariate analysis, patients unwilling to return were more likely to be dissatisfied with the amount of time spent with the MD (OR 2.0, 1.3-3.3), report that the MD did not listen to what they had to say (OR 12.5, 5.0-31.2), and to report not receiving all needed services in the past year (OR 3.1, 1.7-5.4). Similarly, patients unwilling to recom-mend were more likely to be dissatisfied with the amount of time spent with the MD (OR 1.7, 1.0-2.8), to report that the MD did not listen (OR 6.4, , and that they did not receive needed services (OR 7.6, 4.5-12.7). Furthermore, these patients were more likely to be dissatisfied with the in-office waiting time (OR 4.3, , and to report that the MD did not address the reason for their visit (OR 4.0, 2.1-7.7). In a subgroup analysis of patients who had clinical tests, failure to explain their purpose also predicted unwillingness to recommend (OR 9.9, 1.9-52.4). CONCLUSION: Our findings suggest that the extent to which MDs inform and involve patients may influence their intention to return or refer others to their usual source of care. Perception of the amount of time spent with the MD was predictive of both outcomes. As rationing of time and services under managed care increases, inattention to these potentially modifiable aspects of patient care could result in worse doctor-patient communication and increased attrition in primary care practices. PURPOSE: An intensive nurse practitioner-community health worker(NP-CHW)intervention improved blood pressure modestly among young urban black hypertensive men. We investigated the impact of the intervention on patient satisfaction with medical care for high blood pressure (HBP). METHODS: 309 men were randomly assigned to receive a less intensive (LI) intervention (one-time HBP education and referral) or a more intensive(MI)intervention, including HBP education and HBP management by a NP-CHW team. At 24 month follow-up, participants received a satisfaction questionnaire containing rating of overall satisfaction (1,not at all, to 5,extremely satisfied) and 37 ratings of qualitity of 10 aspects of care (1,poor, to 5,excellent (E)). RESULTS: 82% (N ϭ 255) of the men completed the questionnaire. MI and LI groups were similiar at baseline in age (mean MI 42, LI 41 years), education (mean MI 11, LI 12 years), and the proportions using illicit drugs (51% MI, 48% LI),reporting excellent health status (8% MI, 4% LI), and without health insurance (58% MI, 48% LI). Both groups rated items about provider technical quality (35-37% E), information (33-37% E) and interpersonal treatment (35-37% E) better than access issues (20-26% E). More than twice as many MI patients were extremely satisfied with the way their MD or NP took care of them (55% vs. 27%), and with "medical care for HBP" (55% vs. 26%). Domains with the largest differences between groups in E ratings were information (24-28% difference, e.g., 49% vs. 21% E rating of helpfulness of advice on HBP and how to keep healthy, difference 28%, 95% CI (17%, 39%)); technical quality (27-28% difference, e.g., 46% vs. 19% E rating of the MD or NP's ability to figure out what's wrong with you, difference 27%, 95% CI (17, 37); and interpersonal treatment (25-27% difference, e.g., helpfulness, friendliness, and caring of the MD or NP, 46% vs. 20% E rating, difference 26%, 95% CI (16, 36)). There was less marked but still substantial differences between the MI and LI patients in ratings of access issues (e.g., ease of getting medications, 30% vs. 14% E rating, difference 16%, 95% CI (6, 26). Groups were similar in a few items. CONCLUSION: An intensive NP-CHW intervention that modestly lowered blood pressure also dramatically improved several dimensions of patient satisfaction with medical care for HBP in young black urban men, a group with low ratings of care at baseline and at high risk for treatment dropout. Future research should explore whether these better ratings predict adherence to medical treatment and lower risk of preventable cardiovascular disease. . On univariate analysis there was no significant difference between physician specialty, level of training and practice size or setting, AND sub-scale scores or willingness to accede to patient request for referral, except that attending generalists working in HMO settings had lower career satisfaction scores than those working in non-HMO settings (p ϭ 0.04). Generalists with greater levels of cost-consciousness were more willing to accede to patient request for referral (p ϭ 0.01), In response to a request for written comments, 62 of 269 respondents [23%] complained of malaise and burnout, excessive patient loads, inadequate visit length, poor compensation from insurance companies, and loss of autonomy. CONCLUSION: Other than the lower career satisfaction of non-resident generalists in HMO settings and the increased willingness of cost-conscious generalists to accede to patient request for referral, there were no significant differences between generalist's field and practice demographics AND cost-consciousness, comfort with uncertainty, career satisfaction, and willingness to accede to patients' request for referrals. Even so, written comments by almost 1 ⁄ 4 of respondents revealed an undercurrent of dissatisfaction with the practice of primary care today. Further study of generalist physicians' demographics and attitudes, complemented by analysis of practice patterns, is warranted to determine the effects of the contemporary health care system and of medical training on the provision of primary care in the U.S. PURPOSE: To assess the effectiveness and safety of a practice guideline (PG) in reducing the duration of intravenous (IV) antibiotic therapy and LOS for patients (pts) hospitalized with community-acquired pneumonia (CAP). METHODS: A randomized trial of PG implementation was conducted at 1 university, 3 community teaching, and 3 community non-teaching hospitals in western PA. The PG recommended conversion from IV to oral antibiotics and/or hospital discharge (DC) when pts met specified clinical criteria. Randomization to the intervention arm (I-A) or control/usual care arm (C-A) was at the physician (MD) group level for internists, family MDs, and pulmonary or infectious disease MDs in each hospital (709 MDs in 173 groups). Pts admitted by these MDs were identified prospectively using standard eligibility criteria for CAP. When pts met all criteria for IV antibiotic conversion and/or DC, only physicians in the I-A received the PG intervention, including (1) placing a detail sheet with PG recommendations in the medical record, and (2) nurse-mediated physician reminders and concurrent feedback. Primary outcomes were LOS and duration of IV antibiotic therapy. Secondary outcomes were 30-day mortality, readmission, and return to usual activities. Multivariate analyses of outcomes were performed using a clustered discrete proportional hazards model, adjusting for hospital site and CAP severity risk classes I-V at baseline. RESULTS: Overall, 608 eligible pts were enrolled (median age 71, 46% male, 83% white) from 116 physician groups. No baseline differences existed for I-A and C-A pts. Hazard ratios (HR) and 95% confidence intervals (CI) for the primary outcomes in I-A and C-A are tabled below (larger HR ϭ shorter LOS or duration IV).

There was a significant interaction (p ϭ 0.01) between hospital site and intervention effectiveness for LOS, with HR by site ranging from 0. PURPOSE: To assess the effectiveness and safety of a practice guideline (PG) in reducing the duration of intravenous (IV) antibiotic therapy and LOS for patients (pts) hospitalized with community-acquired pneumonia (CAP). METHODS: A randomized trial of PG implementation was conducted at 1 university, 3 community teaching, and 3 community non-teaching hospitals in western PA. The PG recommended conversion from IV to oral antibiotics and/or hospital discharge (DC) when pts met specified clinical criteria. Randomization to the intervention arm (I-A) or control/usual care arm (C-A) was at the physician (MD) group level for internists, family MDs, and pulmonary or infectious disease MDs in each hospital (709 MDs in 173 groups). Pts admitted by these MDs were identified prospectively using standard eligibility criteria for CAP. When pts met all criteria for IV antibiotic conversion and/or DC, only physicians in the I-A received the PG intervention, including (1) placing a detail sheet with PG recommendations in the medical record, and (2) nurse-mediated physician reminders and concurrent feedback. Primary outcomes were LOS and duration of IV antibiotic therapy. Secondary outcomes were 30-day mortality, readmission, and return to usual activities. Multivariate analyses of outcomes were performed using a clustered discrete proportional hazards model, adjusting for hospital site and CAP severity risk classes I-V at baseline. RESULTS: Overall, 608 eligible pts were enrolled (median age 71, 46% male, 83% white) from 116 physician groups. No baseline differences existed for I-A and C-A pts. Hazard ratios (HR) and 95% confidence intervals (CI) for the primary outcomes in I-A and C-A are tabled below (larger HR ϭ shorter LOS or duration IV in I-A).

There was a significant interaction (p ϭ 0.01) between hospital site and intervention effectiveness for LOS, with HR by site ranging from 0.69 (CI ϭ 0.40, 1.17) to 2.24 (CI ϭ 1.54, 3.25). HRs were larger in subgroup analyses that excluded 61 (10%) pts admitted from nursing homes and 96 (16%) pts in risk class V (HR for LOS ϭ 1.18, CI ϭ 0.97, 1.42; HR for IV antibiotic duration ϭ 1.22, CI ϭ 0.97, 1.52). No significant differences in any secondary outcomes existed among I-A and C-A pts. CONCLUSION: Dissemination of a PG did not compromise pt outcomes and resulted in small yet statistically significant site-specific reductions in LOS. The larger reductions in duration of IV antibiotic therapy and LOS in less severely ill non-nursing home pts suggest that future PG implementation should focus on this more homogeneous low-risk subgroup of pts with CAP. OBJECTIVE: Literature shows that many patients taking long term medication regimens have difficulty with compliance. One of the important predictors for medication adherence is past adherence behavior. DESIGN: As part of a larger study to increase medication adherence in Latino adolescents with latent TB (tuberculosis), 130 PPDϩ teens were surveyed concerning past medication-taking behavior. RESULTS: Baseline data indicate that 54% reported that they did not finish taking some type of medicine in the past. Ninety-three percent used compliance aids to help them take their medicine. Popular aids included: calendars (n ϭ 34); taking medication at a particular time (n ϭ 71); having someone remind them (n ϭ 71); putting their medication where they can see it (n ϭ 84); and putting their medication near something they use daily (n ϭ 71). Barriers to missing doses of medication were also identified by 75%. Barriers related to time issues, such as oversleeping and coming home late as well as running out of medication. Preliminary findings from monthly self-reports on INH adherence indicate that 75% of subjects are not adherent with medication-taking. CONCLUSION: These findings underscore the need to develop adherence behaviors and to decrease barriers which affect INH therapy.

PURPOSE: To examine the relationship between outpatient drug abuse treatment (ODAT) program's managed care (MC) involvement and potential patients' access to 'treatment on demand' (TOD). METHODS: As part of a panel study, we surveyed a nationally representative sample of 618 ODAT programs in 1995. The dependent variable, TOD, was defined as a director's report of an average wait of zero days for treatment. They reported MC involvement as the percent of HMO-/PPO-insured patients, of patients requiring prior authorization, and of patients whose ongoing treatment is subject to concurrent review; these MC variables were categorized a priori into none (referent), 1 to 25%, and Ͼ25%. Logistic regression models evaluated the association between MC involvement and TOD. RESULTS: Of 589 programs with wait time data, 66% reported TOD. Controlling for profit status, concurrent review for Ͼ25% of patients (OR, 0.43; 95% CI, 0.20-0.92), and prior authorization for 1 to 25% (OR, 2.0; 95% CI, 1.0-4.0) or Ͼ25% (OR, 3.2; 95% CI, 1.3-7.7) of patients were related to TOD. HMO-/PPO-insurance of 1-25% of patients may also influence the availability of TOD (OR, 0.59; 95% CI, 0.34-1.0). Private for-profit programs (87%) were also more likely to report no wait than not-forprofit (63%) or public (60%) programs (P ϭ 0.001). Ongoing analyses will further evaluate the effect of managed care involvement on access from 1988 to 1995. CONCLUSION: Preliminary results suggest a stepwise relationship between prior authorization and greater access to treatment on demand. Conversely, concurrent review and patients' HMO/PPO involvement were associated with decreased access to immediate treatment. Private, for-profit programs also appear to have more readily available treatment capacity than public or not-for-profit programs.

PD Friedmann, MD Stein, S Lemon, TA D'Aunno, Brown University and University of Chicago PURPOSE: To examine whether on-site service delivery, referral mechanisms, case management and transportation promote patients' medical service utilization in the context of drug abuse treatment. METHODS: We analyzed the Drug Abuse Treatment Outcomes Study (DATOS), a 12-month longitudinal study of 6148 patients enrolled in a purposive national sample of 71 drug abuse treatment programs from 1991-1993. Director's surveys provided information about medical services delivery. Analyses evaluated the effect of on-site location of services within the drug treatment program, external referral (formal and informal), the program's emphasis on case management, and the availability of transportation on the log-transformed number of medical visits reported by patients at a three-month in-treatment personal interview. Multivariate, three-level hierarchical linear models addressed patient clustering within programs, and programs within the four treatment modalities, methadone maintenance (MM), long-term residential (LTR), short-term inpatient (STI), and outpatient drug-free (ODF), as well as patient-level case-mix variables from the intake interview. Other variables that increased medical utilization included programlevel factors (treatment modality and the extent to which referrals are monitored), and baseline patient-level factors (worse self-reported health, more medical comorbidity, greater treatment motivation, greater medical utilization in the 12 months prior to addiction treatment, and black race). CONCLUSION: Patients in addiction treatment programs that focus on centralized delivery mechanisms (on-site delivery only) most reliably utilize medical services, while those in programs with a mix of on-site and referral receive medical services least reliably. In the context of addiction treatment, transportation services are a more important facilitator of medical service delivery than is case management. These findings suggest that initiatives, grants and organizational policies to promote linkage of addiction treatment clients to medical services should emphasize focused, integrated delivery mechanisms, and transportation. We conducted a survey to study the attitudes of primary care physicians towards the use of LMWH for outpatient DVT management and to identify common and potentially correctable barriers to the implementation of this form of treatment. METHODS: Cleveland Clinic Health System primary care physicians (n ϭ 344), including general internists (n ϭ 229) and family physicians (n ϭ 115), were sent confidential surveys between September and November 1999. The survey consisted of 18 questions. Physician training background, practice type, attitudes towards LMWH and barriers to the outpatient treatment of DVT with LMWH were assessed. Barriers were assessed using a 5-point Likert scale (1 ϭ strongly disagree; 5 ϭ strongly agree). Results were tabulated and intergroup comparisons performed using appropriate statistical tests. RESULTS: A total of 151 (44%) physicians responded to the survey. Thirty-eight physicians (26%) who responded had used LMWH for the outpatient treatment of acute DVT. Physicians who had used LMWH in this setting felt significantly more knowledgeable about and comfortable with LMWH (p Ͻ 0.001) than physicians who had not used LMWH in the treatment of DVT. There was a trend towards more LMWH use by general internists (p ϭ 0.06) than family physicians. General internists also felt more knowledgeable about LMWH and its use compared to family physicians (p ϭ 0.014). Each proposed barrier to outpatient LMWH use for DVT treatment received a mean score greater than 4 (4 ϭ agree). Barriers to treatment included an ability to get insurance approval to cover the cost of LMWH, a lack of ready access to duplex ultrasound, a lack of a nurse for patient education, a perception of patient non-compliance and availability of LMWH at a local pharmacy. CONCLUSION: According to our study, LMWH has not become widely used by pri-mary care physicians for the outpatient treatment of DVT despite its potential for cost savings and reduced hospital length of stay. Further educational efforts need to be undertaken to expand the use of LMWH in the outpatient setting because of its many benefits over unfractionated heparin. and corresponding BCBSM claims data from members continuously enrolled in the responding practices from 1/1-12/31/95 (n ϭ 76,389). Bivariate analysis revealed that larger group practices were significantly more likely to have primary care gatekeepers (p Ͻ .001) and outpatient case managers (p Ͻ .001). Multivariate linear regression analysis was performed to evaluate the independent effects of these group practice strategies on the natural log of patient pharmacy expenditures PMPY. This analysis controlled for patient and group practice characteristics. RESULTS: The following strategies were independently associated with lower pharmaceutical expenditures PMPY: employment of outpatient nurse case managers (p Ͻ 0.008), primary care physician gatekeeping (p Ͻ 0.001), policies to control pharmaceutical detailing (p Ͻ 0.000), and medication lists and outpatient charts (p Ͻ 0.000). Log linear conversion of drug charges indicates 8.3% lower costs related to office and medication records, 5.6% lower for pharmaceutical detailing, 5.2% lower for pharmaceutical detailing policies, 5.2% lower for outpatient nurse case managers, and 3.8% lower for primary care physician gatekeeping. These data reveal substantial variations in prevalence of administrative techniques to managed pharmaceutical costs within primary care group practices in an HMO network. CONCLUSION: Groups which use these strategies have significantly lower pharmaceutical costs PMPY. Over half of the group practices surveyed had some policy in their clinic regarding physician interaction with pharmaceutical sales representatives (PSRs). Most policies related to managing the PSR visit times or limiting visits to groups rather than to individual physicians. Very few clinics reported more stringent policies, such as, not accepting sample medications or prohibition of gifts. Nonetheless, the presence of these modest policies was associated with lower prescription drug costs. PURPOSE: Medication errors are common and preventable in the hospital setting and can lead to adverse drug events (ADEs). However, comparatively little is known about these events in the outpatient setting. We sought to examine the frequency and type of medication errors and potential ADEs (medication errors with potential for patient harm) that occur among outpatients. We prospectively collected copies of prescriptions given to patients by 13 primary care providers at two general medicine practice sites in Boston, each over 4 weeks. Site 1 was an academic hospital practice with electronic prescription writing (but no error checking). Site 2 was a community-based practice with handwritten prescriptions. Prescription copies were screened by a pharmacist for errors. Among 384 prescriptions, there were 68 medication errors (rate, 18%), and of these, 14 were potential ADEs. The most common errors were dose (60%) and frequency (16%) errors. The rate of medication errors at Site 1 (electronic) was significantly lower than at Site 2 (handwritten) (6% vs 34%, p Ͻ .0001). At Site 2, 51 ⁄ 54 (94%) of medication errors could have been prevented by electronic prescribing linked to the medical record. However, the frequency of potential ADEs at the 2 sites was similar. In addition, there were significantly more medication errors and potential ADEs related to new prescriptions than to refills. We conclude that medication errors are common in the outpatient setting and that new prescriptions more often contain medication errors and potential ADEs. Outpatient computer systems that include electronic prescribing may be an effective way to reduce the number of medication errors, although error checking may be necessary to reduce serious error rates. (1996) (1997) (1998) (1999) . Study participants completed a self-administered questionnaire at baseline (68.5% response rate) and follow-up (69.4% response rate). The baseline questionnaire measured 4 structural features of care (access, continuity, integration, duration of the primary care relationship) and 4 interpersonal features of care (communication quality, interpersonal treatment, doctor's knowledge of the patient, and patient trust). Multivariable logistic regression was used to examine each scale's role in predicting adherence to the primary physician's advice, and voluntary disenrollment from the physician's practice. There were 2 indicators of adherence-one measuring attempts to change any of 5 behavioral risk factors (smoking, alcohol use, obesity, sedentary life-style, nonuse of seatbelts) based on the physician's advice ("process" adherence), and the other measuring actual change in the risk factor ("outcome" adherence). Voluntary disenrollment was determined using patient-provided information linked to a statewide physician registry. Regressions controlled for patients' sociodemographic profile, health status, and ambulatory utilization. RESULTS: Trust was the leading predictor of voluntary disenrollment from a primary care physician's practice (OR ϭ 1.58, p Ͻ .001), but was followed closely by the other 3 indicators of relationship quality (OR ϭ 1.49 to 1.53, p Ͻ .001). Patients with baseline trust scores in the 5th percentile were 3.3-times more like to disenroll than those with 95th percentile trust scores. "Process" adherence (attempted behavior change) was best predicted by 2 measures of relationship quality: doctor's knowledge of patient, and patient trust (p Ͻ .001). "Outcome" adherence (successful behavior modification) was predicted only by patient trust (p ϭ .05).

CONCLUSION: The quality of doctor-patient relationships, particularly patients' trust in their primary physician, is strongly predictive of adherence to medical advice, and of loyalty to a physician's practice. In an era marked by increasing pressure on physicians to attend to such factors as market share, productivity, and efficiency, these findings remind us that the very essence of medical care delivery remains the interaction of one human being with another, and suggest that medical practices cannot afford to ignore this in the race to the bottom line. CONCLUSION: In spite of regulatory hurdles to writing and dispensing, and concerns about the appropriateness of opiates, use of these drugs is common. Managing patients on opiates may be a substantial part of outpatient care. If these results generalize to other care settings, they suggest the need to recognize the importance of chronic pain and opiate management, and develop programs accordingly to optimize and rationalize pain care. . We evaluated whether differences in use by these factors have diminished with increased adoption of BCS use over time. METHODS: The Surveillance, Epidemiology, and End Results (SEER) database was used to select a cohort of 158,496 women у30 years diagnosed with local or regional stage breast cancer from 1983 to 1996 who received surgical treatment. Income and education were estimated using census data. Logistic regression models were developed to predict BCS use by age, education, income and population density, controlling for other factors. Models for each of the factors were tested for changes over time in ordering of groups, spread between highest and lowest use groups, and changes in slope. Differences in ordering were determined by groupwise comparisons of the average standard error for each group at four points in time. Change over time in spread was assessed by year-wise comparisons of the difference of the adjusted means of the highest and lowest use groups RESULTS: The use of BCS increased overall by 344% from 1983 to 1996, and increased among all groups examined. Women у80 yrs. had highest BCS use compared to other age groups in 1983 (p Ͻ 0.001), yet had the lowest use by 1996 (p Ͻ 0.001). Although BCS use increased in all age groups, the increase in use over time was significantly lower for the oldest age group. For local stage disease, the spread between the highest and lowest use age groups decreased over time (p ϭ 0.01). For regional stage disease, the spread increased (p ϭ 0.01). During the period of study, there was a persistently higher use of BCS among higher income groups. The increase in use of BCS was slowest among women in the poorest quartile leading to a widening of the spread over time between the highest and lowest income groups (p Ͻ 0.00001). Results by educational status were similar to the results by income status. Women residing in more urban areas had persistently higher BCS use over time. There was a significant increase in the spread between the most urban and least urban groups from 1983 to 1996 (p ϭ 0.00007). CONCLUSION: Despite an overall increase in the use of BCS over time, there remain significant differences in the use of BCS among the groups examined. Factors associated with lower use of BCS continue to include lower income, lower education and more rural residence. Older women, who previously had the greatest use, now have the lowest use. This analysis suggests that differences in BCS use by age, SES, and population density will perisist, or even increase, for the immediate future. The risk of cardiovascular disease in patients with hypertension is determined by the level of blood pressure and also by the presence or absence of target organ damage (TOD) or other risk factors (RF). The JNC VI provides a practical classification by level of blood pressure (stage 1, 2, 3) and presence of TOD or RF (risk group: A, B,C). This report makes emphasis on absolute risk and benefit and uses risk stratification as a guidance to the treatment strategy. We aimed to evaluate whether the processes and outcomes of care were associated to the baseline risk group in addition to the initial level of blood pressure. METHODS: We examined the medical records of 922 randomly selected patients with hypertension who received their regular care in our primary care clinic with at least one year of follow up. Based on the JNC VI we obtained the information to assign each patient to his/her corresponding stage and risk group. We recorded information on demographics, treatment, and level of control (Ͻ140-90 mmHg).

We also defined intensity of therapy (increasing the dosage or adding a new drug) as a marker of the process and the mean decrease of systolic blood pressure (SBP) and diastolic (DBP) as a marker of outcome of care. RESULTS: The mean (SD) age was 59. The mean SBP at the initial visit was 155 Ϯ 20.0 mmHg and the mean DBP was 98.9 Ϯ 40.6 mmHg. The mean decrease in SBP and DBP between initial and last visit was 5.6, 15.8 and 30.9 (p ϭ 0.0001) and 5.2, 11.3 and 20.2 (p ϭ 0.0001) for stage 1, 2 and 3 respectively. Adjusting for stage, the mean decrease in SBP and DBP was 15.6, 14.1 and 16.7 (p ϭ NS); and 10.5, 9.9 and 11.4 (p ϭ NS) in group A, B and C respectively. The intensity of therapy by stage was 53%, 66% and 81% for stage 1, 2 and 3 respectively (p ϭ 0.09).

Adjusting for stage, we did not observe differences across groups (p ϭ NS). CONCLUSION: Although BP was reduced as a consequence of more intensive treatment according to initial stage, it appears that physicians are not considering the patient's global risk profile when making decisions about the management of hypertension. Doctor's need not only to incorporate level of BP but also overall risk to improve the effectiveness of their interventions.

Goldberg, EZ Oddone, DL Simel, Ambulatory Care, Durham VA Medical Center, Durham, NC PURPOSE: National guidelines for the management of chronic pain with narcotics recommend establishing a written pain contract, limiting the number of providers, and providing consistent, scheduled supplies of narcotics. It is not known how adoption of these interventions affects provider or patient behavior. This study measured the effect of an explicit pain management program on unscheduled patient visits, narcotics prescribing behavior, and oxycodone use in a large ambulatory care clinic at a university-affiliated VA hospital. METHODS: In 1997, the Durham VA Ambulatory Care Clinic initiated a comprehensive pain management program to help primary care providers manage patients whose treatment involved chronic narcotics. This program involved a specific pain contract signed by the provider and the patient and a clerical mechanism that assisted with scheduled prescribing of controlled substances. All patients with evidence of such a pain contract were identified retrospectively. A matched control group of patients followed in the ambulatory care clinics during the same period that received comparable amounts of narcotics, but did not have a pain management plan on their chart, was also identified. Demographic information, all outpatient visits, and narcotics prescriptions were abstracted and summarized for both groups. Visits and narcotics prescriptions from other nearby VA hospitals were collected for intervention patients. Emergency department visits, the number of separate providers issuing narcotics prescriptions, the number of unique narcotics prescriptions, and the amount of oxycodone consumed were summarized by calendar quarter for intervention patients before and after the execution of their pain management contract, and for control patients throughout the study period. RESULTS: 91 patients with a pain management contract were identified. Compared to the time period prior to the contract, visits to the emergency department, the number of providers issuing prescriptions, the number of separate prescriptions for narcotics, and the number of dispensed oxycodone tablets all decreased significantly (p Ͻ 0.001 for each measure) after implementation of the contract. Visits to other nearby VA hospitals did not increase. In a matched group of 224 patients receiving narcotics, emergency department visits decreased during the observation period, but to a lesser degree. The number of separate providers issuing narcotics and the number of unique prescriptions did not change over time, although the number of oxycodone tablets consumed increased steadily for this control group. CONCLUSION: A pain management program made explicit by a provider-patient contract can result in fewer unscheduled visits, a consolidation of narcotics prescribing, and lower overall use of narcotics. CONCLUSIONS: Adherence interventions which decrease failure rates by only 10 to 25% will be cost-effective, even if expensive. Studies evaluating such interventions should report results in terms of both change in failure rate and cost, in order to allow rapid assessment of cost-effectiveness for development of HIV health policy. In the first year of a patient and clinician educational intervention, we decreased antibiotic treatment of adults with uncomplicated acute bronchitis from 80% to 45%. The purpose of this study was to evaluate the impact of reduced antibiotic use for acute bronchitis on duration of illness and patient satisfaction. METHODS: We conducted a telephone survey of adults diagnosed with acute bronchitis at a clinic where antibiotic treatment of acute bronchitis had been decreased, and a usual care clinic of similar size and demography, belonging to a group-model HMO in the Denver metropolitan area. Patients were contacted 2-3 weeks following their office visit. Major outcomes were time to illness resolution, and satisfaction with the episode of care. A proportional-hazards model was used to evaluate duration of illness (relative risk (RR) and 95% confidence intervals (CI) reflect probability of symptom resolution). Satisfaction between groups was compared using the Wilcoxon rank sum test. RESULTS: Antibiotics were prescribed to 63% and 84% of respondents from the intervention and usual care clinics, respectively (p Ͻ 0.001). Using multivariate analysis, increased age was associated with longer duration of illness (age Ͼ65: RR ϭ 0.57 95% CI ϭ 0.34-0.95), but not gender, comorbidity, duration of illness prior to the visit, or antibiotic treatment. After adjusting for these factors, there was no difference in duration of illness between intervention and usual care clinics. Satisfaction with the visit did not differ between patients from intervention and control clinics, nor between patients receiving and not receiving antibiotics. CONCLUSION: Reducing the rate of antibiotic treatment of adults with uncomplicated acute bronchitis does not prolong illness duration or produce dissatisfaction with care. METHODS: We calculated visit and antibiotic prescription rates for incident, uncomplicated acute bronchitis from administrative claims data from a group-model HMO. We compared utilization between sites from the initial post-intervention study period (11/97-11/98) to the follow-up period (11/98-2/99). The positive results of the initial intervention were publicized system-wide in the fall of 1998, and during a single staff meeting at the full intervention site. No other patient or staff education took place in the follow-up period. Bronchitis visit rates and antibiotic prescription rates were compared between sites using a mixed-effects logistic regression model, adjusted for patient and provider characteristics. RESULTS: Antibiotic prescription rates remained high at the control sites, and remained low at the full intervention site during the follow-up period (78%, 80% and 53%, respectively; p ϭ 0.001). 30-day return visits for bronchitis or pneumonia also remained constant during the follow-up period. The likelihood that a patient with an office visit for acute bronchitis during the first intervention period sought care again during the follow-up period was lower at the full intervention site compared to control sites (OR ϭ 0.4, 95% CI ϭ 0.2,0.8). CONCLUSION: During the subsequent winter, a successful intervention to reduce antibiotic treatment of uncomplicated acute bronchitis was sustained with minimal reinforcement. This intervention also reduced subsequent visits for uncomplicated acute bronchitis among patients who previously sought care for this illness. CONCLUSION: Use of herb/nutr supps is very common in general medicine patients who take prescription medications. Potential for interaction between herb/ nutr supps is common, although most potential interactions are minor. PCPs need to inquire about use of herb/nutr supps in their patients, and consider potential adverse interactions with prescription medications. N ϭ 1292 ). Quality of care was assessed with disease-specific explicit criteria for the process of hospital care. Criteria were developed for the admission work-up, evaluation and treatment, and readiness for discharge for each disease by panels of expert physicians. Adherence scores (percent of applicable criteria met) were produced for admission (ADM), evaluation and treatment (TRT), and discharge (DC) for each patient. Bivariate associations of race with quality (adherence score) were examined with the t-test. After controlling for severity of illness (APACHE II) and other covariates, the independent association of race with quality was determined for ADM and TRT with linear regression and for DC with logistic regression, because DC scores were not normally distributed. RESULTS: Mean age and mean APACHE score were similar (P Ͼ .05) in Black and White patients for DM, CHF, and COPD. Mean ADM scores were higher (82.0 vs. 79.4%; P Ͻ .001; CHF and 71.7 vs. 68.7%; P Ͻ .001; COPD) in Blacks and were also higher (P Ͻ .001; CHF and COPD) in teaching hospitals. Mean TRT scores were higher (P ϭ .05; COPD) in Blacks and were higher (P Ͻ .001; CHF and COPD) in teaching hospitals. Mean DC scores did not vary by race or teaching hospital. Blacks were more likely (P Ͻ .001) to be admitted to teaching hospitals. After controlling for age, APACHE score, comorbidities, and other covariates Blacks had higher ADM scores than Whites with CHF (␤ ϭ 2.15; P ϭ .006), COPD (␤ ϭ 2.89; P Ͻ .001), but not with DM (␤ ϭ 1.25; P ϭ .20). Adjusted TRT score was higher (␤ ϭ 3.64; P ϭ .03) in Blacks with COPD. Blacks had lower (OR 0.67, 95% CI 0.47-0.95) adjusted DC scores for CHF. However, when we controlled for teaching hospitals ADM and TRT scores did not differ by race; only the DC score for CHF was significant (0.67, 95% CI 0.47-0.95) in Blacks. CONCLUSION: We found that adjusted quality was significantly higher in Blacks at ADM (CHF, COPD) and at TRT (COPD) and was significantly lower in Blacks at DC (CHF). After adjusting for teaching hospitals, quality was similar by race at ADM and TRT for each diagnosis, but remained lower in Blacks for one of three diagnoses (CHF). Differences in quality by race may be influenced by higher quality at teaching hospitals and a higher likelihood that Blacks were treated in teaching hospitals.

In order to assess the effect of various cost control and managed care tools on Workers' Compensation costs at a self insured University Hospital, a study was conducted on six cohorts of injured workers spanning calendar years 1993-1998. Cost control techniques were started in 1993 and included institution of an early return to work program, internal administration of legal cases, institution of an injury prevention program, and utilization of modified duty assignments in the workplace. Managed care techniques were instituted in 1997. Such strategies included aggressive case management and tailoring of the provider panel.

Each successive intervention resulted in a reduction in total expenditures, both medical and indemnity. In the early years (1993-96) a reduction in expenditure of more than 50% was realized for both medical and indemnity payments. This trend continued when managed care techniques were later introduced (1997) as indemnity payments were reduced almost 40% and lost time cases were reduced over 50% from calendar year 1996 to 1997. In addition, the percentage of patients who chose to remain in the system increased from 50% to over 90% from 1993 to 1998. The high quality of the provider panel, measured by provider reputation, training and board certification status, was maintained. Sequential application of cost control and managed care techniques can result in significant reductions in Worker's Compensation costs, both medical and indemnity, in this setting without a decrease in patient satisfaction or a compromise in the quality of care. Our results imply that time to surgery is a proxy for the burden of medical illness and that the delay in surgery in prior studies was a marker for active medical problems. These results suggest that it may not be as important to rush patients to surgery as previously recommended. Recent studies have verified the clinical benefit of COX-1-sparing agents compared with non-selective NSAIDs in reducing peptic ulcer incidence, yet the higher relative cost of the newer agents has raised questions of whether their use is cost-effective. METHODS: In this study, a Markov decision model was designed to compare the projected costs and effectiveness of four different therapeutic strategies for highdose NSAID use. These were: 1) ibuprofen alone; 2) ibuprofen plus misoprostol; 3) ibuprofen plus omeprazole; 4) celecoxib or rofecoxib. Similar clinical efficacy of each regimen at the respective maximum dose for the treatment of inflammation and pain was assumed. The analysis was conducted from a payer perspective and the time horizon was ten years. Each strategy's ability to prevent NSAID-induced ulcers was modeled from published evidence in the medical literature. Utility values for the quality of life effects of peptic ulcer were estimated from previous costutility studies in ulcer disease. Cost data on ulcers were obtained from a recently published study of the economics of ulcer disease in a managed care setting. Drug costs were obtained from the 1999 Drug Topics Red Book. All cost and utility estimates were maximally "biased" to be economically favorable to the COX-1-sparing agents. RESULTS: For the base-case analysis (assuming an ulcer incidence of 0.5% per year), the marginal cost-effectiveness ratio (CER) for COX-1-sparing agents was $1,100,000 per quality-adjusted life year (QALY) gained compared to the use of ibuprofen alone. Ibuprofen combined with omeprazole was similar in cost-effectiveness to celecoxib or rofecoxib. The misoprostol/ibuprofen strategy was dominated unless the annual ulcer rate exceeded 3%. Sensitivity analyses indicated that the marginal CER for COX-1-sparing agents fell below $100,000 per QALY when the annual risk of peptic ulcer was greater than 5% per year. The marginal CER was less than $50,000 per QALY only when the annual ulcer risk was 8% or greater. CONCLUSION: This study suggests that even in a model that maximally favored the clinical and economic performance of COX-1-sparing NSAIDs, such drugs are only cost-effective in patients who are at substantially increased risk of developing NSAID-induced peptic ulcers and their complications. PURPOSE: While advance directives (AD) have been widely promoted as a means to ensure that patients' treatment preferences are followed, there is extensive evidence for their apparent failure. We conducted a survey of physicians to examine how they would respond to difficult clinical situations with an explicit AD and the reasons for their treatment decisions. METHODS: We mailed surveys to all internal medicine faculty and resident physicians at an academic medical center. The survey consisted of 6 hypothetical clinical scenarios. Each described a seriously ill patient with an explicit AD. The scenarios contained a potential conflict between the AD treatment preference, clinical prognosis, and/or family/friends. Three scenarios described a fair prognosis with an AD declining treatment; the remaining 3 described a poor prognosis with AD preferences for aggressive treatment. Respondants were asked to make a treatment decision and then indicate the rationale for their decision. RESULTS: A total of 117 ⁄ 250 (47%) analyzable surveys were returned. Overall, decisions by faculty and residents were NOT consistent with the AD in 65% of cases (see Table. ) Physicians cited a variety of reasons for their treatment decisions including the AD (37%), family/friends (when present in scenario, 15%), likelihood of survival (18%), and quality of life (34%). When physicians made decisions inconsistent with the AD, they were more likely to list other reasons for their decisions (89%, p Ͻ 0.001.) Similarly, of those making decisions consistent with the AD, a majority of faculty and residents cited the AD as their rationale in those cases (86%, p Ͻ 0.001.)

CONCLUSION: Despite the presence of an explicit AD, internists frequently make treatment decisions that are not consistent with the AD. In difficult clinical situations, internists appear to consider other factors such as prognosis, perceived quality of life, and family/friends wishes as more determinative than the AD. Given the potential for conflict between ADs and clinical/social factors, future work needs to elucidate how strictly patients might want their AD preferences followed. PURPOSE: Clinical decision making for patients with influenza-like illness has become more complex with the recent introduction of rapid office tests for the identification of influenza infection and the recent approval of a new class of antiviral drugs, the neuraminidase inhibitors. These drugs are active against both influenza type A and type B viruses, in contrast to the older antiviral drugs which are active only against influenza type A. The purpose of this study was to determine the utility of rapid testing for influenza followed by antiviral therapy and the utility of empirical antiviral therapy. METHODS: We performed a decision analysis and cost-effectiveness analysis, using results of clinical trials of rimantidine, an older antiviral drug, and zanamivir, a neuraminidase inhibitor, for the treatment of average-risk (for complications of influenza) and high-risk patients with influenza-like illness. Four strategies were compared to symptomatic treatment alone: empirical rimantidine treatment, empirical zanamivir treatment, test for influenza A and treat positives with rimantidine, and test for influenza A and B and treat positives with zanamivir. Outcome measures were reductions in the number of days of fever and influenza symptoms and the cost per quality-adjusted life year (QALY) saved. RESULTS: Compared to symptomatic treatment, the empirical rimantidine treatment strategy results in the greatest reduction in the number of days of fever, 1.4 for average-risk persons and 2.0 for high-risk persons, relative to symptomatic treatment alone. The strategy of testing for influenza A and B and treating positives with zanamivir results in the smallest reductions, 0.7 and 0.9 days, relative to symptomatic treatment alone. Testing for influenza adds cost but does not improve outcome. Empirical zanamivir treatment is the optimal strategy if influenza type B causes more than half the documented influenza cases in the general population, an uncommon occurrence. The cost of the four strategies ranges from $37 to $63 per person, including the cost of treating symptoms and bacterial compli-cations if they occur. The cost per QALY saved ranges from $11,800 and $8,600 for average-risk and high-risk patients with the empirical rimantidine treatment strategy to $42,100 and $32,900 for average-risk and high-risk patients with the strategy of testing for influenza A and B and treating positives with zanamivir. CONCLUSION: Antiviral therapy results in modest reductions in symptoms of patients with influenza-like illness. The cost for these benefits is also modest. Testing for influenza adds cost without additional benefit. The most effective and costeffective strategy is empirical treatment with rimantidine, an older antiviral drug. These patients may not necessarily perceive that they need a primary care provider. This study's objective was to determine whether young adults who recieve care within a primary care system have different perceptions of the adequacy of their health care than those who do not recieve primary care services. METHODS: The study was conducted among patients who receive care at an urban VA medical center with an integrated primary care system. Subjects included all individuals who were discharged from military service after August, 1990 and had received medical care between October 1997 and September 1998. In January 1999, a questionnaire was mailed to 1113 patients who met these criteria. The questionnaire included demographic items, the SF-36V health status instrument and the VA Customer Satisfaction Survey (developed in collaboration with the Picker Commonwealth Foundation). Respondents were classified by whether they had received care from an assigned primary care provider. Comparisons of satisfaction items between patient groups were adjusted for age, gender, and health status by using analysis of covariance and logistic regression. RESULTS: Of 467 respondents, 15 were excluded because their age was greater than 60 years. Median age of the remaining subjects was 38 years, 86% were male, 84% were Caucasian, and 69% had an assigned primary care provider (PCP). Patients having a PCP had higher satisfaction with their most recent clinic visit (p ϭ .05), but they were not more satisfied with overall care. Patients having a PCP did not consistently have more favorable perceptions of specific provider behaviors. Among patients with a PCP, 18% reported that they were not involved in decisions about their care as much as they desired, 23% reported not having complete confidence in their provider, and 23% reported not receiving as much information about their condition as they desired. These proportions did not differ significantly from those reported by patients without a PCP. However, more of the patients with a PCP (75%) than of those without a PCP (55%) reported that their providers were always familiar with their recent medical history (p Ͻ .001). CONCLUSION: Among this young adult population, a primary care system appeared to affect some perceptions of medical care but not overall satisfaction. For some physician communication behaviors, primary care providers do not perform better than other providers. Primary care providers may benefit from initiatives to improve communication behaviors. tion at the provider level was found for obtaining a HgbA1c or LDL measurement in the past year and varied substantially across facilities (percent of variation due to provider ranging from 2-20%). Eye exam rates varied most at the clinic site level (9% of the variance in rates) and little or none at the physician level. CONCLUSION: Clinic site level profiles for important processes of care related to diabetes can be feasibly developed and will be reasonably reliable given the number of patients and the magnitude of facility differences. While physician level profiles will usually have too low reliability to be useful, for some indicators at some sites investigating provider level differences in performance may be possible.

TP Hofer, R Hayward, HSR&D, VA; Internal Medicine, UM, Ann Arbor, MI PURPOSE: Physician implicit review has been used to assess whether hospital deaths are "preventable." Studies using such reviews have suggested that the number of deaths in U.S. hospitals that are due to medical errors is shockingly high. We sought to uncover what physicians mean when they describe a death as "preventable" in terms of the probability of a patient surviving had care been optimal. METHODS: We conducted 344 reviews of 102 different eligible hospital deaths, at seven Department of Veterans Affairs Medical Centers. Cases with a terminal diagnosis who were admitted to the hospital predominantly for comfort care were excluded. A total of 15 physician reviewers were trained in the use of the implicit review instrument for the study. Each was a board certified internist with extensive experience in inpatient medicine. A previously reported instrument was used. The question on preventable death was rated on a 5 point scale (definitely yes, probably yes, uncertain, probably not, definitely not). In this study we added to the review instrument several questions asking the probability of certain events, such as the likelihood that the death would have been prevented if care had been optimal, rated from 0%-100%. The unit of analysis was the review although we accounted for the clustering of reviews by patient and reviewer using multilevel statistical models. RESULTS: 7% of the reviews rated overall care as being substandard and 14% reported care as of borderline quality or worse. Similar to previous studies, in 8% of the reviews death was rated as probably or definitely preventable by optimal care, and in 17% of the reviews the preventability of death was rated as uncertain. In addition, the inter-rater reliability of ratings of the preventability of death was similar to previous studies (ICC ϭ 0.22 for two reviewers).

If a death was rated as probably or definitely preventable, the probability of preventing death with optimal care was estimated to be 57% (95% CI 47%, 66%), however, the likelihood of surviving three months with good cognitive and physical function was estimated at only 11% [4%, 19%] . Amongst all deaths reviewed, the physicians estimated that, had care been optimal, the average probability of living 1 year or more with good congitive and physical function was 0.9% [0.08%, 0.17%]. CONCLUSION: The alarming statistics about preventable deaths in US hospitals should be tempered by an understanding that the deaths are identified, by physician review, in very sick patients whose chances of surviving to leave the hospital would be uncertain, even with optimal care. A further implication relevant to this finding is that hospital mortality rates can not capture much of the difference in preventable death rates since many such patients will die before discharge even with optimal care. We describe trends in antibiotic use and resistance among GNR isolates. METHODS: We collected data on inpatient antibiotic (fluoroquinolones, third-generation cephalosporins, and aminoglycosides) prescribing from our hospital pharmacy records. Volume of prescribing was expressed as defined daily doses (DDD) per year. Data on the most commonly prescribed outpatient antibiotics from 1997-1999 were obtained for patients belonging to a large local insurance plan. The annual proportions of resistance among major GNRs were extracted from microbiology reports. We used simple linear regression(SLR) and chi-square for trend to detect significant changes over time. RESULTS: Inpatient fluoroquinolone prescribing rose from 11,140 to 43,384 DDD per year during 1993-1999. The increase in use was greatest from 1998-1999 (25,871 to 43,384) . Cephalosporin use increased from 15,845 to 26,258, and aminoglycoside use dropped from 18,249 to 11,319 (all p Ͻ .01 by SLR). Among outpatient antibiotic prescriptions, ciprofloxacin was the 8th and 10th most prescribed drug for 1998 and 1999; levofloxacin emerged as the 6th leading expense in 1999. GNRs had significant increases for inpatient quinolone resistance:

CONCLUSION: Inpatient fluoroquinolone use at our institution has risen markedly since the introduction of levofloxacin. There was a concomitant increase in inpatient GNR resistance to fluoroquinolones, notably for Pseudomonas aeruginosa, with little change in resistance for the other 2 antibiotic classes. As fluoroquinolone use and resistance are increasing, strategies are needed to define and promote the optimal use of these new broad spectrum antibiotics. PURPOSE: Influenza vaccination is recommended for elderly people and chronic patients of any age but is also available for healthy adults aged under 65. During a community campaign, we implemented a multifaceted intervention in a university-based primary care centre to promote influenza vaccination of high-risk people. Study aims were : (1) to measure influenza immunisation rates among ambulatory patients younger than 65 with or without a chronic disease; (2) to assess the impact of this intervention on vaccination coverage. METHODS: In this pre-/post-intervention study, we analysed data from a random sample of patients aged below 65 who attended our academic primary care clinic during autumn trimesters of 1995 and 1996. Respectively 405 and 445 patients were included before and during intervention trimesters. The intervention comprised patient information by leaflets and posters, a walk-in vaccination clinic, a physician training workshop, reminders on medical records and periodic feedback to physicians on vaccination performance. Data were collected from the computerised billing file and the review of medical records. RESULTS: At baseline, immunisation coverage of outpatients younger than 65 was 5.2% with higher rates among those aged 50-64 (11%) or with a chronic illness (19.2%). Vaccine uptake was lower among diabetics (11.5%) than in patients with a pulmonary (30%), cardiac (20%) or another chronic disease (e.g., renal failure, immunosuppression) (28.6% approached statistical significance to be independent predictors of influenza immunisation. CONCLUSION: At baseline, few ambulatory patients younger than 65 received the influenza vaccine, even those at high risk. Vaccine uptake doubled after an intervention targeting high-risk people and combining strategies directed at patients, physicians, clinic organisation and the community. The largest effect was observed among chronic patients for whom influenza immunisation is recommended. However, as their immunisation coverage is not optimal, we must develop new strategies and intensify our efforts to promote influenza vaccination of high-risk patients aged below 65. 1.7) . A desire to gain expertise in a particular CAM therapy was associated with younger age (OR ϭ 1.07), percent time in clinical practice (OR ϭ 2.9) and family member use during childhood (OR ϭ 2.0). Gen-eralists were no more likely than subspecialists to desire to gain expertise in a particular CAM therapy. CONCLUSION: Personal use and recommendation of CAM therapies to patients is common among university physicians at these teaching hospitals. Similar to general population surveys, physicians do not inform their primary care providers that they use such therapies. This survey identifies characteristics of physicians likely to use CAM in treating their own health problems or prescribe CAM in clinical practice. 1.7) . A desire to gain expertise in a CAM therapy was associated with younger age (OR ϭ 1.07), percent time in clinical practice (OR ϭ 2.9), and family member use during childhood (OR ϭ 2.0). Generalists were no more likely than specialists to desire to gain expertise in a particular CAM therapy. CONCLUSION: Personal use and recommendation of CAM therapies to patients is common among university physicians at these teaching hospitals. Similar to general population surveys, physicians do not inform their primary care providers that they use these therapies. This survey identifies characteristics of physicians who use CAM to treat their own medical conditions and who prescribe CAM in their clinical practice. of these (18%) desired to "do everything" for their family member and 9 (56%) desired to "go a little longer". In contrast, MDs wanted to "do everything" for 1 patient (5%) and to "go a little longer" for 6 (33%). Agreement between the family member's wishes and the MD's perceptions of those wishes was 50% (8:16, kappa ϭ .22). Of the 3 patients whose family members desired to "do everything", 1 received "everything", 1 had mid- PURPOSE: Race affects the delivery of healthcare but the effect of race on health outcomes has been inconsistent. Racial disparities are at least partly due to differences in access to healthcare. The VA operates a nation-wide, government-funded healthcare system with few financial barriers and may allow for improved access to care for black patients. Therefore, we studied the effect of race on mortality for patients admitted to Veterans Affairs (VA) hospitals, a healthcare system that offers potentially equal access to care. METHODS: We studied 39,190 patients admitted to one of 147 VA hospitals during the fiscal year 1996 for one of six common medical conditions: heart failure, angina, diabetes mellitus, pneumonia, chronic obstructive pulmonary disease, or chronic renal failure. Due to relatively small numbers, we excluded women and other racial groups. For patients with more than one hospitalization, a single admission was randomly selected. We used logistic regression and proportional hazard models to adjust for differences in patient demographic characteristics, comorbidity score (based upon number of concurrent illnesses), length of hospital stay, and hospital characteristics. The primary outcomes were unadjusted and adjusted 30-day mortality rates for blacks and whites; we also compared unadjusted and adjusted inpatient and 6-month mortality rates. Finally, we compared adjusted mortality rates after stratification by variables that we hypothesized might interact with race (age, financial status, comorbidity score, length of stay, coronary bypass surgery availability, urban location, region, and disease. (7%), gastroenterology (7%), pulmonary (6%) and hematology/oncology (6%). Almost three-fourths indicated board certification in their subspecialty with onethird certified for both internal medicine and pediatrics in their subspecialty. Twothirds are women, and most practice in cities greater than 50,000 people (85%) in a group setting (94%). Almost half of SS graduates spend greater than 80% of their time in direct patient care and one-third indicated a university based practice. Of the 58 procedures, only 13 are performed by more than half of SS graduates, vs 31 for PC graduates. Five of the top 10 procedures performed by the largest proportion of Med-Peds graduates were similar for SS and PC: chest xray interpretation, EKG interpretation, venipuncture, cerumen removal, and peripheral intravenous catheter placement. For no procedure did more than half of SS graduates desire additional training, in contrast to 16 procedures identified by more than half of PC graduates. Although SS graduates perform fewer procedures, they desired more training during residency in procedures that PC graduates also indicated were important. Eight of the top 10 procedures which both SS and PC graduates desired more training are the same and are shown.

CONCLUSION: Med-Peds subspecialists appear to perform procedures less frequently than primary care Med-Peds physicians unless the procedure is in their area of expertise. However,physicians who pursue subspecialty careers still indicate a desire to learn primary care procedural skills in residency training. PURPOSE: In the last 20 years concerns with cost-effectiveness and the rise in the penetration of managed care have produced changes in the type of diagnosis that are treated as outpatients. However, these changes may have been disproportional between metropolitan and non-metropolitan areas. The purpose of this study is to determine the regional differences in the most common diagnosis seen by internists. METHODS: The National Ambulatory Medical Care Survey (NAMCS) is an annual probability sample survey conducted by Centers for Disease Control on patients' office visits. The 1996 NAMCS dataset was used to generate lists of the most common diagnoses seen by internists in outpatient setting in both metropolitan and non-metropolitan areas. Metropolitan and non-metropolitan areas were defined per NAMCS protocol. Sampling weights were applied to generate national estimates of total number of visits. Diagnostic clusters (version 4.1 1998) developed by Schneeweiss et al were used to group diagnostic codes representing similar conditions. Clinically significant difference was defined as a relative difference of at least 50% between metropolitan and non-metropolitan areas. RESULTS: The estimated number of patient visits to internists was 100,977,712 in 1996. We evaluated the top 12 most common diagnostic clusters, which constituted more than 50% of the reason for the outpatient visits in both metropolitan and rural areas. Six diagnostic clusters were common to both rural and metropolitan areas: hypertension (13.7%), diabetes mellitus (5.1%), acute upper respiratory infection (4.6%), acute lower respiratory infection (4.4%), ischemic heart disease (3.2%), and asthma (2.4%). In non-metropolitan areas chronic diseases were also common: COPD (6.0%), chronic rhinitis ( Our results show a difference between metropolitan and rural areas in terms of the most common diagnosis seen. General medical exams and minor injuries were more commonly seen in the metropolitan areas and could be related to an increased penetration of managed care in these areas. The impact of these differences on health care delivery is unknown and needs to be further studied. PURPOSE: Opportunities for appropriate screening, early and accurate detection and management of coronary heart disease (CHD) in women are frequently missed because risk factor evaluation and modification, diagnosis, and treatment protocols have been based on the traditional non-gender specific model of disease. This is of particular concern to physicians who routinely care for populations of women at risk for cardiovascular disease.

Participants will receive vital information on identification and treatment of risk factors, the role of diet therapy for both prevention and treatment, symptoms and outpatient diagnostic testing. This seminar will enable physicians to educate their patients about their personal risk factors, understand how to intensify their prevention methods for their patients, and to improve the critical evaluation of potential cardiac symptoms of CHD in women. Participants will receive a comprehensive packet that includes information on key points discussed for providing complete patient education: presentation summary; references; copies of the slides; case studies; and a patient education brochure, "Guide to Heart Healthy Eating." LEARNING OBJECTIVES: This program is flexible as it is divided into four segments-risk factors, diet, symptom issues, and diagnostic testing-each with their own learning objectives. Using different modules, the presentation can touch on one or two aspects or provide a comprehensive understanding of CHD in women. Outline of program content: I. Welcome, introduction of presenter, workshop rationale, review of learning objectives, description of comprehensive participant handout, and appropriate disclosures. II. Slide lecture: A. Risk factors and their modification B. Diet therapy as prevention and treatment C. Symptoms and treatment issues D. Diagnostic testing III. Summary of physician, testing, and patient-related issues IV. Question and answer session RESULTS: Our objective is to provide current information on the prevention, detection, and management of CHD, focusing on gender specific guidelines for health care professionals. Participants are expected to share this information with their colleagues and female patients, thereby decreasing the incidence of and mortality from CHD among a large population of women. CONCLUSION: Participants will be provided with a complete guide they may use in presenting this information to students and physicians. The material included in the participant handout will be expanded upon by the presenter. Equipment needs: Preferred: LCD projector, PC compatible computer (if available), slide projector, 3 carousels, screen, podium with speaker light, lavaliere microphone, and laser pointer. PURPOSE: Guidelines have been developed to assist physician decision making, however abundant data demonstrate that even generally accepted guidelines are often not followed. To improve physician adherence to guidelines we developed a patient-specific computerized outpatient reminder system. Although many reminders were followed, many were not. Structured review of the medical record was used to classify potential reasons why reminders were not followed. METHODS: Reminders were in 4 categories: 1. diabetes care (eye exams, HbA1C and cholesterol studies, nursing education visits); 2. therapeutic recommendations (ACE-inhibitor use in diabetic hypertensives, and ASA, beta-blocker, and statin use in coronary artery disease(CAD) patients); 3. expensive medication use (NSAIDS, H2-blockers, statins, ACE-inhibitors), and 4. health maintenance (Pap, mammogram, cholesterol, Pneumovax). During a 1 year period reminders were delivered to 197 PCPs affiliated with a tertiary care institution. All reminders were printed on the ambulatory face-sheets that PCPs receive at the time of patient office visits. We assessed how often patients received the recommended care items after reminders were delivered. When recommended care items did not occur, we performed chart reviews on a random sample of records to identify potential reasons why reminders were not followed. We evaluated charts spanning 5 reminder types: HbA1C, diabetic eye exams, CAD and no ASA, CAD and no Beta-Blocker, and Expensive ACE-Inhibitor. classes. Expensive NSAIDs were not switched to inexpensive NSAIDS at a significantly increased rate however they were discontinued more often in the intervention group (10.9 vs 6.0%, p Ͻ 0.02). Health maintenance reminders led to increased compliance rates in all areas studied (all p Ͻ 0.001).

CONCLUSION: Computerized patient-specific outpatient reminder systems significantly improve physician compliance with many recommended guidelines. Patterns of adherence varied by guideline and reminders for lab tests and ophthalmology referrals were followed more often than those related to medication prescribing. shown to be effective in randomized controlled trials. One of the key findings of the guideline is that counseling by nurses (and the involvement of multiple types of providers) is associated with significantly higher quit rates. This study describes 1) the participation of nurses and medical assistants (MAs) in the routine identification and counseling of smokers and spit tobacco users at each clinic visit, and 2) overall concordance with AHCPR guideline recommendations approximately 3 years following their dissemination. METHODS: We conducted exit interviews of 4071 consecutive adult patients who presented for routine, non-emergency care at 8 community-based primary care practices (7FP, 1IM) during the pre-intervention phase of a controlled guideline implementation trial. RESULTS: Overall concordance with guideline recommendations at the time of the clinic visit is shown in the table (% concordance by nurses/MAs in parentheses):

CONCLUSION: The majority of smokers who present for routine primary care in our region are asked about their smoking status; nurses (and MAs) are integrally involved in this assessment. Of those patients who receive any counseling, the primary care clinician is typically the only team provider to offer advice to quit. Spit tobacco use is generally not assessed routinely and few users receive any counseling. Increased involvement of nurses in the implementation of smoking (and spit tobacco) cessation guidelines requires clinic leadership and support, identification of cessation as a high priority preventive care objective, and training of nurses in effective brief intervention techniques. (16) Ask about willingness to quit* 23 (5) 8 (0) Advise pt to quit* 28 (5) 16 (3) Help pt to set quit date 2 (0) 2 (1) Discuss pharmacotherapy* 11 (Ͻ1) 3 (0) Arrange follow-up 3 (0) 2 (0) *p р 0.05 for comparison between smokers and spit tobacco users whereas other factors without prognostic importance (gender, race, insurance) were incorporated in these decisions. Cognitive feedback to contrast how ED physicians use certain cues with how well these cues actually predict ACI-related short term outcome may improve ED decision making. Previous studies suggest that medical oncologists are more likely than surgeons to favor breast conserving surgery (BCS) over mastectomy for treatment of early stage breast cancer. We examined whether discussing surgical options with a medical oncologist was related to type of surgery and satisfaction with treatment decision-making in two states where we have demonstrated differences in rates of BCS. We identified women with stage 1 or 2 invasive breast cancer at 17 Massachusetts hospitals and 30 Minnesota hospitals. Rates of BCS among women eligible for BCS were 74% in MA and 48% in MN. We abstracted type of surgery and other clinical characteristics from medical records. We surveyed 792 women in MA (61%) and 1633 women in MN (83%) two months after surgery and asked about the number of physicians with whom they discussed treatments, the specialty of the physicians, and, for the 3 physicians they considered most important, if surgical treatments were discussed prior to final decisions about surgery. We also asked about satisfaction with treatment decision-making. We assessed the effect of discussing surgical options with a medical oncologist on both type of surgery and satisfaction using propensity score analysis to adjust for patient, hospital, and surgeon characteristics. Women in MA saw a mean of 3.5 physicians compared to 2.8 in MN (P Ͻ 0.001). 58% of women in MA and 28% in MN discussed surgical options with a medical oncologist (P Ͻ 0.001). In MA, discussing surgical options with a medical oncologist was not related to type of surgery but was associated with greater satisfaction with treatment decision-making (Table) . In MN, where the rate of BCS was much lower, discussing surgical options with a medical oncologist was associated with BCS but was not associated with greater satisfaction. The effect of discussing surgical options with a medical oncologist on treatment and satisfaction is complex. When BCS is standard therapy in a community (MA), seeing a medical oncologist does not influence treatment but is associated with greater satisfaction, possibly due to providing women with more information. When BCS is not standard care, seeing a medical oncologist is associated with greater use of BCS but does not affect satisfaction, possibly because surgeons' and medical oncologists' recommendations conflict. To better understand the effect of specific experiences during patient-physician interactions, we asked patients about ambulatory care experiences and assessed the relationships between problem experiences and both patients' trust in their physicians and whether they had considered changing physicians. We surveyed 2086 patients (response rate 61%) insured by a large national health insurer who had at least one visit with a primary care provider in 1995. To assess problem experiences, we asked if their physician always (1) gives them enough time to explain the reason for the visit (2) gives answers to questions that are understandable (3) takes enough time to answer questions (4) asks about how their family or living situation affects their health (5) gives as much medical information as they want, and (6) includes them in decisions as much as they want; and we identified a problem experience if their physician did not. We also asked about trust in their physicians (using a 10-item scale) and if they had considered changing their doctor due to concerns about their care. We used linear regression to assess the effect of problem experiences on trust and logistic regression to assess the effect of problem experiences on considering changing physicians. In both models, the independent variables of interest were the 6 problem experiences, and control variables included patient, physician and plan characteristics. Most (78%) patients reported at least one of the six problem experiences noted above and 49% reported at least two. Patients with more problems had lower trust scores (P Ͻ 0.001) and more often reported having considered changing physicians (P Ͻ 0.001). In the linear regression model, each of the six problems remained independently associated with lower levels of trust (all P Ͻ 0.005). In the logistic regression model, three of the problem experiences were independently related to considering changing physicians. Patients whose physicians did not always give enough time to explain the reason for the visit (OR 1.9; 95% CI 1.3 to 2.9), did not always take enough time to answer questions (OR 3.2; 95% CI 2.1 to 5.0), and did not always give enough medical information (OR 4.0; 95% CI 2.4 to 6.9) more often considered changing physicians. A majority of patients report problem experiences in the ambulatory setting. These experiences are strongly related to lower levels of trust in their physicians and several are also strongly associated with considering changing doctors. Efforts to improve patients' experiences may promote more trusting relationships and greater continuity and therefore should be a priority for physicians and health care organizations. We selected all adult members of two United Healthcare plans who had at least one outpatient or inpatient claim with a diagnosis of depression during the years 1993-1995 and were continuously enrolled for 12 months. Pharmacy claims data were used to construct measures of duration of treatment, dose, and type of antidepressant, We examined the effects of two different definitions of a new episode of depression (using a 4-month versus 9-month clean period prior to the index depression visit) on the likelihood of being treated with antidepressants at therapeutic doses. Further, we compared two different ways of identifying an episode of depression (requiring 1-visit versus 2-visits with a code for depression) with respect to their influence on antidepressant prescribing practices (dose and duration). We also examined whether antidepressant type was related to the likelihood that antidepressants were prescribed at therapeutic doses. RESULTS: The duration of the clean period was not related to conclusions about antidepressant prescribing practices. Patients with two or more visits with depression diagnosis codes were significantly more likely to receive antidepressants than those with only one visit and were more likely to receive therapeutic doses at each time period (86% vs 72% at 1 month of treatment and 23 vs 15% at 5 months, p р 0.0001). Among persons receiving antidepressants, those receiving selective serotonin reuptake inhibitors (SSRIs) were more likely to receive therapeutic doses and to continue treatment for at least five months than those prescribed other classes of antidepressants. In multivariate analysis, being prescribed an SSRI versus another class of antidepressants was significantly associated with receiving both one month (OR ϭ 7. 3 [5.7-9.3] ) and five months (OR ϭ 2.0 [1.6-2.5]) of therapeutic treatment. CONCLUSION: Our results indicate that CONCLUSION regarding appropriateness of antidepressant prescribing can vary markedly depending on how the quality measure is specified. Given that administrative data are and will continue to be used for both monitoring and quality improvement purposes in the short run, it is critical that we understand how variations in measurement specifications influence the CONCLUSION that are drawn about treatment of depression in health plans. CONCLUSION: These data demonstrates that an elevated serum Po4 at the time of base line echo is a strong independent predictor of progression of heart valve calcification in chronic HD patients. Ca-Po4 product derives its influence from serum Po4. Age is an independent predictor of progressive heart valve calcification but serum Ca, iPTH and duration on HD are not. Performance for Medicaid patients tended to start out lower and remain lower. Similar patterns were noted for AAs versus whites and, to a lesser extent, for women versus men. For example, for AA women, adherence to beta blockers went from 32% to 38% to 46%. Multivariable adjustment for hospital characteristics, and for patient clustering within hospitals, did not change these contrasts. CONCLUSION: Care improved from the year before to the year after release of guidelines, and was even more improved two years later. This was also true for low-income, minority, and female patients, although differences for disadvantaged populations persisted. The extent to which the improvements are attributable to guideline release is the object of further study. Nevertheless, care is evolving towards guideline adherence. PURPOSE: Because of previous reports in the literature of poor physician compliance with National Cholesterol Education Program (NCEP) guidelines, we designed a retrospective cohort study to assess patients' achievement of their target low density lipoprotein cholesterol (LDL-C) and physician compliance with NCEP guidelines in our practice. METHODS: Investigators randomly selected 70 patients with a diagnosis of hyperlipidemia seen in our academic Internal Medicine practice in 1996 and followed them for 2 years. We collected data including demographics, past medical history, medication use, and lipid profiles by chart review. We assessed compliance with NCEP guidelines by 1) identifying those patients that reached their target LDL-C based on risk profile and 2) determining the number of cholesterol management changes made during the study period. RESULTS: Of 70 patients reviewed thus far, 57 met eligibility criteria for the study. At baseline, the cohort had the following characteristics: 94% white; 47% male; average body mass index 31.7 kg/m 2 (sd ϭ 5.7 kg/m 2 ); average cholesterol 219 mg/dl (sd ϭ 27 mg/dl); 95% had health insurance; 26% had coronary heart disease; and 61% had two or more cardiac risk factors. Lipid analysis was performed at least once during the study on 68% of patients. Of these, 54% attained their target LDL-C. Patients without coronary heart disease were more likely to be at their target LDL-C compared to those with coronary heart disease, but the difference was not significant (56 vs. 33%; p ϭ 0.17). Physicians had 35 opportunities during the study period to make adjustments in dietary or medication management for those patients not at their target LDL-C. Eighteen management changes (51%) were made, with 60% and 48% taking place in those with and without coronary heart disease, respectively. CONCLUSION: We found that only half of our patients with a diagnosis of hyperlipidemia were at their target LDL-C. Additionally, physicians missed half of the opportunities to adjust cholesterol management. Further study of an intervention to improve cholesterol management in our practice, such as feedback of the results of this study to our physicians, is needed. Health care workers who are poorer and have less access to care have even lower rates. This is not a site-specific problem but a national one. Exhortation has not caused a majority of health care workers to comply with recommendations. The focus must change to education programs, stringent accreditation and regulatory standards, and improving low-income health workers' personal access to care. In the multivariate model, admission to an intensive care unit and length of hospital stay were positively related to the proportion of criteria for evaluation and treatment during the stay that were met; however, this did not significantly modify the negative relationship with age. CONCLUSION: Greater age was significantly associated with less adequate in-hospital management of patients with CHF. This result is disturbing because the incidence of CHF increases with age and CHF is among the leading causes of hospitalization among the elderly. In the multivariate model, living alone at home was positively related to the proportion of quality criteria met assessing readiness for discharge, while non sinus cardiac rhythm on admission affected this proportion negatively; however, both these relationships did not significantly modify the association with LOS. CONCLUSION: CHF patients with LOS Ͻ10 days are less likely to meet explicit quality criteria of readiness for discharge than patients staying longer. Therefore, because insufficient readiness for discharge may have a negative impact on patient outcomes, reductions in LOS in these patients should be accompanied by measures assuring that readiness for discharge is reached faster. total of 1296 patient months were reviewed. The mean age of patients in both groups was 70. There was no statistical difference in mean prothrombin time (PT) or international normalized ratio (INR) between groups. In addition, there were no differences in indications for anticoagulation between groups. Although there was no obvious indication of complications associated with under-or excessive-coagulation in the charts, data related to INR monitoring was either missing or uninterpretable in the UC patient charts twice as likely as in the AC charts. One of the differences in the charts was the utilization of an anticoagulation flowsheet developed by the medical group. It was present in all of the AC charts but either not present or not utilized by physicians in the UC group. CONCLUSION: Nurse or pharmacist-directed anticoagulation clinics have been shown to faciiltate close monitoring of patients on oral anticoagulation therapy in various clinical settings. While there was a trend toward more consistent monitoring in the AC group, it was not statiscally significant. Despite the development of an anticoagulation flowsheet patient information related to therapy was often missing. Other health-care providers would have difficulty managing anticoagulation for a patient if that physican were unavailable, increasing the likelihood of complicaitons. The presence of an nurse-directed anticoagulation clinic is one strategy to address this problem. We previously compared the appropriateness of coronary angiography as judged by expert panel ratings vs. decision-analytic models for the post-discharge period for patients after acute myocardial infarction. The purpose of this study was to compare expert panel ratings vs. decision-analytic models for angiography prior to hospital discharge. METHODS: We constructed a decision-analytic model based on the clinical variables utilized by an expert panel to develop appropriateness ratings for the use of pre-discharge coronary angiography after myocardial infarction. Variables considered were age, use of thrombolytic therapy, and whether or not the infarction was complicated by persistent chest pain, mitral regurgitation or ventricular septal defect, cardiogenic shock, persistent pulmonary edema, severe left ventricular dysfunction, stress-induced myocardial ischemia, or recurrent ventricular tachycardia or fibrillation at least 24 hours after myocardial infarction. We used the decision model to simulate the 32 clinical scenarios evaluated by the panel to calculate the gain in quality-adjusted life years (QALYs) and incremental cost-effectiveness ratio of pre-discharge coronary angiography compared with no angiography, and compared these outcomes with the appropriateness score and appropriateness classification estimated from the expert panel ratings. RESULTS: The Spearman correlations were 0.29 and Ϫ0.62 for appropriateness score vs. QALY gain and incremental cost-effectiveness ratio, respectively. Of the clinical scenarios that were categorized as "appropriate," the average gain in QALYs was 0.62 years and the average incremental cost-effectiveness ratio was $27,000 per QALY gained, where averaging was based on the likelihood of a patient being in each clinical scenario. Of the clinical scenarios that were categorized as "equivocal," the average gain in QALYs was 0.28 years and the average incremental cost-effectiveness ratio was $42,000 per QALY gained. None of the clinical scenarios was categorized as "inappropriate" by the expert panel. Cardiogenic shock was the most influential variable for both the decision-analysis outcomes; age was the most influential variable in determining the appropriateness score. CONCLUSION: There was good agreement between the appropriateness score and the incremental cost-effectiveness ratio of pre-discharge coronary angiography compared with no angiography after acute myocardial infarction. The agreement between the appropriateness score and the gain in QALYs was much weaker. PURPOSE: Clinical practice guidelines are tools designed to improve the test-ordering system. However, not all guidelines are effective in meeting this objective because they commonly address systems rather than the behaviors underlying these systems. The Theory of Planned Behavior (TPB) has been used to identify factors motivating people to perform certain behaviors and may be helpful in better understanding factors that influence test-ordering behavior. The primary behavioral concepts of this theory include behavioral beliefs (an action will have a positive or negative outcome), normative beliefs (identifying people who influence an action), and control beliefs (factors that promote or deter an action). The objective of this study was to determine the applicability of the TPB to physician decision making with regard to test ordering. METHODS: Echocardiograms (ECHOs) are a commonly ordered and perhaps over utilized test; therefore, we applied the TPB to identify behaviors influencing physicians to order them. Using a modified focus group approach, eight major behavioral beliefs, five normative beliefs, and six control beliefs that influenced ECHO ordering among physicians were identified. A survey based on these belief domains was developed and published on the World Wide Web. 1327 members of the SGIM were invited to participate through an electronic mail message. The survey required each respondent to rate responses on each belief using a seven point Likert Scale (1 ϭ very strongly disagree, 7 ϭ very strongly agree). Descriptive statistics without subgroup analysis were reported as part of this exploratory study. RESULTS: 320 Physicians completed the survey of which 282 (88%) were staff physicians, 24 (8%) were Fellows, and 14 (4%) were residents. Behavioral Beliefs most influential in the process of ordering an ECHO included clarification of equivocal physical examination findings and affirmation of diagnostic decisions.

Cardiologists were identified as the group most likely to influence the ordering of an ECHO reflecting a Normative Belief. Reminders of recently done ECHOs were most likely to prevent a respondent from ordering another reflecting a Control Belief.

CONCLUSION: This study showed that the TPB might be relevant in understanding physician decisions with regard to test ordering. An approach addressing behavioral beliefs bolstered by recommendations of specialists (those influencing normative beliefs), and incorporating reminders of duplicate testing may be more effective in modifying test ordering behavior than guidelines alone. Incorporation of the TPB may be helpful in designing future strategies to modify physician behavior. (1) the frequency of meetings between legislative assistants and physicians (2) the issues discussed, and (3) the perceived effectiveness of lobbying. RESULTS: Senate and House legislative assistants met with an average of 10.0 and 4.0 physicians per month, respectively. This suggests that approximately 29,000 such meetings occur annually. The most common issues discussed were Medicare reimbursement (mentioned by 80% of subjects), managed care reform (75%), and funding for medical research (25%). Other issues such as access to care for the uninsured, tobacco control, abortion rights, and violence prevention were rarely discussed. Most legislative assistants rated physicians as very effective (44%) or somewhat effective (46%) as lobbyists. The most common suggestion for improving physician lobbying was to focus less on reimbursement and to address a broader range of health care issues. CONCLUSION: Physicians are frequent and effective lobbyists on reimbursement, managed care, and research issues. Policy makers appear receptive to increased physician input on a broader range of health care issues. Including these issues in lobbying efforts has the potential to shape health policy in a way that improves both patient care and public health. RESULTS: In multivariate analyses, regional effects were strong predictors of performance on the CAHPS measures. Health plan type, size, and age of plan were related to health plan performance. For-profit (FP) health plans had lower adjusted performance on the customer service and access composites and the plan ratings while health plans with national affiliations had lower adjusted scores in both the customer service and delivery composites as well as in ratings of care, specialists, and the plan. Controlling for other variables, the plan ratings and customer service composite scores of national plans, regardless of tax status, were more than one standard deviation lower than those of local not-for-profit (NFP) plans and the ratings of local FP plans were about one-half standard deviation lower. In addition, the scores on the customer service composite were over one standard deviation worse for both national FPs and national NFPs and the scores on the access The CONSORT statement was designed to assist "peer reviewers in evaluating reports of trials submitted for publication". We conducted a pilot study to evaluate CONSORT's utility for improving the methodological aspects of reviews. METHODS: A single-blind randomized control trial (RCT) was conducted by enrolling consecutive RCT manuscripts submitted to JAMA during a six-week period. Up to ten content reviewers per manuscript were recruited and randomized to receive either a "standard" or "CONSORT" packet. A study instrument measured main outcomes. Section one contained the Feurer format and section two used the review quality instrument (RQI). Section three adapted the Chalmers and Jadad RCT manuscripts scales to grade reviewers' reviews for methodological completeness (number of 22 methodological features mentioned), and accuracy (percent of manuscripts' methodologic flaws identified). RESULTS: Five of six possible manuscripts were enrolled. Forty of 46 randomized reviewers returned reviews. CONSORT reviewers did not differ from standard reviewers on the Feurer scale or RQI (Table) . Non-significant trends suggested CONSORT reviewers had improved methodological completeness and accuracy. Four reviewers reporting epidemiology or statistics backgrounds were more methodologically complete (background 7.26, no background 3.06; p-value ϭ 0.008) and accurate (background 25.0%, no background 10.5%; p-value-0.08). Editors' rated more favorable reviews that were more methodologically complete and accurate (p-value Ͻ 0.001 for both trends). Post-hoc calculations found that our sample size had the power to detect a difference in means for methodological accuracy of 15 percent, and would have required a sample size of 330 to detect a difference of ten percent. The study instrument had an inter-rater correlation coefficient of 80 percent. CONCLUSION: We created and evaluated a unique study instrument which may have utility in measuring the frequency of which methodological concerns are addressed by peer reviewers. Our study instrument or sample size may be unable to measure existing differences. Reviewers may have not used the CONSORT statement because its inclusion was under emphasized to avoid the Hawthorne effect. Low methodologic accuracy and completeness scores suggest reviewers do not focus on methodology or lack sufficient background to comment on it. Editors may favor more comprehensive methodologic commentary. PURPOSE: Placement of gastric feeding tubes has increased despite 50% mortality at six months in those who receive them. This suggests that physicians are unaware of the poor prognosis of patients with feeding tubes, perceive other benefits besides survival, or place them for other reasons. METHODS: We interviewed the attending physicians caring for patients who had recently had a feeding tube placed to determine their perceptions of the benefits of gastrostomy placement. Patients older than 21 who had a new gastric feeding tube placed were identified by daily procedural logs of interventional radiology and the gastroenterology suites at a major teaching hospital and a community hospital. Physicians responsible for ordering the placement of these gastric feeding tubes were interviewed by telephone or were sent a written interview. Patients with diagnoses of trauma or non-head and neck malignancy were excluded. Physicians were interviewed independent of patient participation in the study. RESULTS: We contacted 157 physicians and interviewed 140 (89%). Patients' major diagnoses were 28% stroke, 19% neurodegenerative disorders, 16% head and neck cancer, and 37% other serious medical illness. The mean age of the patients was 66 years, 65% were white, and 49% were women. Most (54%) of the physicians had cared for the patient for two weeks or less. Physicians' expected benefits of gastrostomy placement included: improvement in nutritional status (94%), improvement in hydration (67%), prolongation of life (66%), administration of medications (61%), and prevention of aspiration (55%). A minority reported (25%) that it would facilitate placement of the patient, and 15% reported conflict with family members over the decision. Physicians perceived the mean quality of life to be 33 before the gastrostomy was placed and 45 afterward (p Ͻ .05, 100 ϭ perfect health; 0 ϭ death). They estimated that 11% of the patients would be alive in one year without the feeding tube, but 72% would survive more than one year with the feeding tube and 20% would survive 10 years or more. CONCLUSION: Physicians expected improvement in nutrition and hydration in patients who received gastric feeding tubes. A minority reported conflict over the decision with family members or pressure from long-term care facilities. Physicians perceived that placing the gastrostomy would substantially prolong their patients' lives, which is not supported by the current empirical literature. This misperception by physicians may contribute to the increase in utilization of this procedure. RESULTS: In all diseases, the model and the physicians' estimates exhibited good calibration and discrimination. However, patients' estimates exhibited poor calibration and discrimination, the latter index reflecting few differences between the most severely ill and those with more reserve.

CONCLUSION: The SUPPORT prognostic model and physicians calibrate and discriminate well in each of these diseases. However, patients do not recognize the difference between those with very short expected lifespans and those who are expected to live somewhat longer. This is especially true of lung cancer, CHF, and COPD patients. Perhaps patients with these conditions so often survive multiple life-threatening exacerbations, and thus these patients have few cues to inform them when the risk of death is increasing. Perhaps also, the time course of colon cancer and cirrhosis give more reliable cues, such as known limited aggregate survival, weight loss,and discovery of metastases. The poor calibration and discrimination for patients/surrogates might direct improvement efforts at sharing prognosis better between patients and physicians or at ensuring that appropriate advance care plans are made for all with serious disease.

LG Smith, Mount Sinai School of Medicine, New York, NY OBJECTIVE: To implement an innovative educational experience in managed care for senior internal medicine residents. The specific focus at this institution was to initiate a pilot program that tried to impart competency in managed care for trainee residents. The larger objective was fostering of mutual understanding and working relationships between academic health centers (AHC) and managed care (MC). DESIGN: An administrative elective was created from one of two yearly outpatient blocks of PGY3 residents. It comprised: a) Rotation at the corporate headquarters of a for-profit-MC organization, b) experience at a full risk office practice, c) a practical introduction on setting up and managing a medical office, and d) membership in a committee at the AHC corresponding to that of the MCO, designed to help understand health-care delivery models from different perspectives. Results of pre and post-training evaluation among PQE residents and comparison group (CG) were collated from a total of 56 grantee programs. RESULTS: (A). When perceived mastery of managed care competency in 18 topics was compared PQE residents had better 'well prepared to perform' scores in following 13 asterisked (*) topic areas versus CG group (p Ͻ 0. METHODS: Developed and administered a 45 minute telephone survey to measure structure in the following areas: 1) immutable characteristics of provider groups such as size and profit status; 2) compensation arrangements for physicians; and 3) the process of monitoring referral care and resource use. RESULTS: Medical Directors from 54 of 57 physician groups who are participants in the 1998 Physician Value Check Survey project completed the FOS (96% response rate). These physician groups which provide the majority of managed care in these markets are geographically dispersed with 48% in Southern California, 39% in Northern California, and 13% in the Pacific Northwest. The percent of revenue derived from capitated contracts, where groups are compensated at a fixed amount per member per month, was highest in Northern California at a mean ϩ/ Ϫ SD of 86 ϩ/Ϫ 24% and lowest in the Pacific Northwest at 33 ϩ/Ϫ 14% (p ϭ .0001). The dominant form of primary care physician (PCP) compensation also varied by region:

With full risk contracting, physician groups are motivated to monitor the appropriateness and volume of referral care. The most frequent management strategies used in all regions at similar proportions include: mandatory pre-authorization from the group for specialist visits (43%) or procedures (76%); pre-authorization by the PCP for specialist visits (72%); PCP-level profiling with feedback (74%); and implementation of formal guidelines that provide criteria with regard to appropriate use of referral care (74%). CONCLUSION: Financial and organizational structure varies widely for the groups which provide the majority of care to the managed care populations in these regions. Linkage of these structural characteristics to patient level data will be critical to determine whether physician group characteristics are associated with better outcomes of care. CONCLUSION: Patients who achieved abstinence without treatment (more than 10% of all those with AD) have health and health status comparable to those without alcohol disorders. Our results suggest that patients with the greatest burden of illness are those who receive treatment and that abstinence improves mental health status.

PURPOSE: Antibiotic resistance is a growing health problem. Little is known about the attitudes and perceptions of physicians regarding its causes, and no study has ever been published comparing internal medicine residents to attendings. It is perceived that residents don't have an adequate appreciation of antibiotic usage and its ramifications. METHODS: A survey containing a 19 item self-administered questionnaire was sent to 488 internal medicine physicians in 4 Chicago hospitals, of which 424 (87%) were returned completed. The survey asked physicians to rank potential causes of antibiotic resistance, ranging from "unimportant" to "extremely important." The sample included 243 internal medicine residents and 181 internal medicine attendings. The residents included 143 from a public hospital and 100 from a university hospital, and the attendings included 114 general internists, 21 ID specialists, and 46 other specialists. RESULTS: Resident physicians consistently ranked all the potential causes of antibiotic resistance as high or higher (in terms of potential importance in causing the problem) than attending physicians. Specifically, residents ranked 6 potential causes at a statistically significant higher level compared to attendings. These included: 1. Prescribing antibiotics for self-limited bacterial infections. 2. Prescribing antibiotics for self-limited non-bacterial infections. 3. Prescribing antibiotics for shorter than recommended duration. 4. Prescribing antibiotics empirically without a definite diagnosis of bacterial infection. 5. Prescribing broad spectrum antibiotics when equally effective narrower spectrum antibiotics are available. 6.

Having poor access to good information on local antibiotic resistance patterns. There were 12 potential causes ranked at similar levels by residents and attendings. Only one cause was ranked at a statistically significant lower level by resi- .8 b a differs from NOAD, b differs from CADT dents compared to attendings (on the influence of pharmaceutical company advertising and drug promotion). There were no significant differences between residents training in a public hospital (predominantly from international medical schools) and those training in a private university hospital (predominantly from US medical schools). CONCLUSION: Residents in internal medicine programs surveyed in this study attribute as much or more importance to the various potential causes of antibiotic resistance as attending physicians. Whether this implies that current residents in training are more aware of this problem than attendings, or that perceptions may change with time, remains a question for further study.

KW Mayer, JS Kutner, GIM, UCHSC, Denver, CO PURPOSE: The objective was to describe physician behavior changes in response to physician profiling and to explore possible reasons why physicians change their behavior. METHODS: Primary care physicians (internal medicine and family medicine) who are members of the Colorado Medical Society (n ϭ 1740). A survey regarding physician profiling was mailed to identified physicians. RESULTS: 337 of 1651 eligible physicians completed the survey (34% internal medicine, 62% family medicine, 4% other), for a 20.5% response rate. 99 retired physicians were excluded. 93% of respondents work in an outpatient setting. Practice location and types varied (31% urban, 46% suburban, 22% rural, 60% single-specialty, 17% multi-specialty, 18% solo practice, and 4% academic/administrative). The most common type of patient insurance was HMO, followed by PPO, medicare and medicaid. 81% of respondents currently receive profiling data. The figure describes how often physicians reported changing their behavior based on the type of profile received. Seventy-nine percent reported that data from different sources was contradictory, 71% found the data confusing and difficult to interpret, 68% did not believe the data were statistically valid,and 50% did not believe the data were collected correctly. Fifty-three percent agreed that they received too much data to process meaningfully, and 62% did not believe that data helped them practice medicine better. Only 38% of respondents found the data useful in improving patient care, 40% found the data useful in controlling costs at their practice, and 42% used the data to identify areas to expand their knowledge base. Forty-nine percent reported that they wanted their profiles to resemble those of their colleagues, but 65% reported changing their behavior to keep their managed care contracts.

CONCLUSION: Despite the proliferation of profiling data, physicians report infrequent behavior change based on profiles received. Physicians appear to be skeptical of the data quality and application to clinical practice. Pressures related to managed care contracts seem to be the primary motivator for behavior change based on profiling data. Current physician profiling practices do not appear to change physician behavior.

Kutner, GIM, UCHSC, Denver, CO PURPOSE: To describe physician profiling patterns and the perceived usefulness of profiling among primary care physicians. METHODS: Primary care physicians (internal medicine and family medicine) who are members of the Colorado Medical Society (n ϭ 1740). A survey regarding physician profiling was mailed to identified physicians. RESULTS: 337 of 1651 eligible physicians completed the survey (34% internal medicine, 62% family medicine, 4% other), for a 20.5% response rate. 99 retired physicians were excluded. 93% of respondents work in an outpatient setting. Practice location and types varied (31% urban, 46% suburban, 22% rural, 60% single-specialty, 17% multi-specialty, 18% solo practice, and 4% academic/administrative). 81% of respondents currently receive profiling data. Types of profiling information received included: pharmacy/cost per member (66%), pharmacy versus formulary (62%), preventive care (57%), number of referrals (44%), types of referrals (42%), disease managment (42%), patient-specific data (42%), number of tests ordered (27%), cost per episode (27%), length of stay (24%), and clinical episode by diagnostic category (15%). Format of profiling data received included: physicians compared with self (51%), physicians compared with group and individual names given (31%), physician groups compared with each other (25%), physicians compared with group and no individual names given (13%), and hospitals compared with each other (10%). The figure describes how useful physicians found each type of profile. Physicians who received profiles were more likely than those who did not receive them to find data on numbers of referrals (59% vs. 45%, p ϭ 0.022), tests ordered (68% vs. 46%, p ϭ 0.001), and disease management (52% vs. 32%, p ϭ 0.001) NOT useful. Those receiving profiles, compared with those who did not, found data comparing the physician to self (55% vs. 39%, p ϭ 0.008) NOT useful, while physicians compared with group-names given-WAS useful (51% vs. 36%, p ϭ 0.018).

CONCLUSION: Physician profiling is common, but not necessarily perceived as useful by those who are being profiled. Overall, preventive care profiles received the most favorable response from physicians. Profiles were more useful in theory than in practice; for certain types of profiles, those who received them were less likely than those who did not to find them useful. BACKGROUND/AIMS: Missed appointments result in disruption of care and decreased office productivity. Published average "show rates" (SR) are 58% and are lower in younger and lower income patients. "I forgot" is the most common reason for a no show. Computerized telephone reminder systems have not been studied in primary care settings with high percentages of capitated Medicaid patients (CMA). We studied the effectiveness of such a system, Smart Talka (ST), in 2 inner city practices to determine: 1) what percentage of our patients could be reached to confirm or reschedule and 2) the effect of ST on show rates. METHODS: We implemented ST at 2 inner city, general internal medicine practices (GIM) serving 14,000 capitated patients with 81% African Americans and 70% CMA insurance. We programmed ST to call each patient up to 6 times between 10 AM and 8:30 PM on Mondays, Wednesdays, and/or Fridays. ST recorded the number of homes reached with or without confirmation (confirming or rescheduling via touch-tone) and those homes not reached. Reasons for not reaching homes include those with no phone, wrong number, disconnected (no number), and called but could not complete the number of programmed tries before the appointment, the latter usually because the phone was busy or not answered (past date). We collected SR results via our patient scheduling system. Pamphlets on ST were available to patients at the practices. RESULTS: In a 1-month period, ST attempted calls to 2,728 GIM patient numbers and reached 1,762 (64%). The average show rate for all patients was 54%. Sixty nine per cent of CMA patients v 72% of capitated commercial patients confirmed and showed (ns).

CONCLUSION: Computerized telephone reminders can be effective in contacting inner city, primary care patients and improving show rates. This results in improved continuity of care and office productivity. Reports can help practices improve by identifying problem areas, e.g., no confirmation, wrong phone number. Based on initial program success, we now call patients five days a week and plan more intensive education to improve our confirmation rate. We also feed back wrong numbers to the practices, asking them to find alternative contact numbers. PURPOSE: Clinicians in busy medical practices often do not have access to updates in treatment of abused patients; lack of time and access to educational materials are major barriers. We developed a video for clinicians on screening and treatment of patients for child, domestic or sexual, or elder abuse. We used preand post-viewing tests to evaluate the effectiveness on improving knowledge and changing attitudes. METHODS: Physicians, social workers, researchers, and legal staff worked with a professional filmmaker to develop a 35 minute movie which included epidemiology, typical patient presentations, and recommended treatments (including legal reporting requirements) for child, domestic and sexual, and elder abuse. Experts associated with the host institution (Johns Hopkins Medical Institutions) gave interviews and role-played typical patient-physician visits for each abuse type. The scripts included information required by JCAHO. Our pre-and post-test included 24 knowledge (11 circle correct answer and 14 T/F) and 24 attitude (5-point scale, strongly agree to strongly disagree) items. Knowledge items were balanced between the abuse areas. Attitude items tested clinician comfort with, and understanding of, abuse issues and clinician reaction to common myths about abuse. We included a 7 item, 5-point Likert scale evaluation of the video on the post-test. RESULTS: In our first month we received 102 surveys from eight sites in two academic medical institutions: ages 20-65ϩ; 40% graduated since 1990; 57% MD, 28% RNs, 10% other. The MDs were 60% GIM or FP, and 15% PEDS or OB (25% other). 64% had received some training in the area of abuse, 40% 1-2 lectures and 11% a 2-3 day short course. The number giving the correct answer increased for 23 of 24 knowledge items, 12 (50%) of them significantly. The proportion of "agree" or "strongly agree" responses increased on all but two items, 19 of 25 (76%) significantly. We found a impressive 41% positive attitude shift in both "likelihood to consider abuse when faced with mental decline in an elder" (p Ͻ .0001) and in "avoiding discussing abuse with patients so as not to offend" (p Ͻ .0001). On the evaluation section, 32% rated the video "excellent", and 55% indicated they were very likely to recommend the video to a colleague. CONCLUSION: This videotape was an effective means of standardizing education and rapidly disseminating it within and across medical institutions. We found clinicians had adopted a number of myths and other incorrect information. Clinicians showed significant improvement in knowledge and positive attitude shifts after viewing clinical role plays and reliable informant presentations. They also rated the video highly. We recommend videos for clinical updates for busy medical practices. PURPOSE: Although 50% of Argentine women are estimated to experience domestic violence (DV) at some point in their lives, clinicians have historically not been formally trained in detecting and evaluating women at risk for DV. We surveyed practicing generalists to assess their level of knowledge and practice styles regarding DV and their perception of need for training in this area. METHODS: Surveys were distributed to 291 generalists from 11 medical centers in 4 regions of the country. We used a 10-point scale to assess their knowledge of DV. We asked them about the frequency with which they discussed DV with their patients, and had them rate need for training. RESULTS: 175 surveys were completed (60% response). 66% were men, mean age was 46 y, the average number of women patients seen per month was 143. Mean knowledge score was 4.6 (Ϯ1.8). On average, the respondents had discussed DV during the last month in 1.2 (Ϯ0.22) opportunities and DV was diagnosed 0.8 (Ϯ0.18) times during the previous month or in 0.6% of women seen. 78% of physicians reported no previous training about DV; 65% considered themselves "not properly equipped to diagnose or treat" DV victims. 47% showed an interest in undergoing some type of formal training. 70% wanted written materials. CONCLUSION: Practicing generalist clinicians in Argentina score poorly on an assessment of knowledge of domestic violence and perceive a need for additional training in this area.

Medicine, University of Chicago, Chicago, IL PURPOSE: Hospitalists are increasingly being used in place of primary care physicians to care for hospitalized patients. However, patients may prefer to be cared for by their primary care physician, with whom they may already be familiar. The objective of this study is to determine patients' preferences and willingness to pay to be cared for by their primary care physician versus a hospitalist during a hospitalization for a general medicine condition. METHODS: A willingness to pay survey was administered by telephone to a nationally representative sample of 308 persons in July and August 1999. The main outcome measures were preferences and willingness to pay for care by the patients' primary care physicians versus a hospitalist over the length of a hospitalization for a general medical condition. RESULTS: 87% of patients currently have a primary care physican. Among patients with a current primary care physician, 70% prefer care by the primary care physician, 23% state no preference, and 7% prefer care by a hospitalist. Preference for and willingness to pay for care by the primary care physician increased with the duration of and patients' satisfaction with the relationship with the primary care physician. Across all patients, mean willingness to pay for care by a primary care physician versus a hospitalist was $200, but 8% of patients reported willingness to pay of $1000 or more. Eliminating those patients for whom willingness to pay was $1000 or more reduced the mean willingness to pay to $95. Eliminating the 16% of patients for whom willingness to pay was $500 or more reduced the mean willingness to pay to $51. Eliminating the 20% of patients for whom willingness to pay was $250 or more reduced the mean willingness to pay to $34. CONCLUSION: Most patients prefer to be cared for during a hospitalization by their primary care physician, but willingness to pay is modest compared to the $300-$700 per admission cost savings found in randomized trials of hospitalists. Some patients, especially with strong relationships with their primary care physician, report much larger willingness to pay. The design of hospitalist programs should be sensitive to these concerns. The option to have care by one's primary care physician at a modest copayment might substantially mitigate the adverse effects of hospitalists for the fraction of patients who highly value care by their primary care physician while decreasing costs for the majority of patients. Analysis of audiotaped patient-provider verbal interaction using standard talk categories from the Roter Interaction Analysis System and specific measures of depression counseling coded for sequences of depression talk. Analysis: Analysis of variance and covariance to evaluate differences in depression type adjusted for encounter length and patient characteristics (demographics and casemix). RESULTS: PCP encounters with depressed patients who had a structured diagnosis of major depression were shorter in overall duration but duration of talk focused on depression was longer by 1-2 minutes, and included significantly more of this depression-focused talk (76.9 utterances, p ϭ .0219) and less biomedical talk (p ϭ .0028) compared with patients who had only subthreshold depression (53.8 utterances) or no depression (27.2 utterances). CONCLUSION: PCPs delivered more psychotherapeutic depression counseling to their patients with major depressive disorder compared with patients who had no depression or symptoms but no disorder. Findings suggest that PCPs are addressing the needs of their depressed patients and practice structures that support more counseling time may increase the likelihood that patients recieve approapriate care for depression.

PURPOSE: Patient satisfaction has much to do with areas not formally addressed during house staff rounding and in the traditional daily "SOAP" progress note. A pilot survey done at our hospital found that patient concerns were addressed infrequently. House staff have the opportunity to engage patients regularly around their fears, worries or uncertainties. We set out to examine the hypothesis that a simple intervention of the house staff asking for and documenting patients daily concerns would improve patient satisfaction. METHODS: Half of the interns on ward service (one team of 3 interns) were assigned to amend the traditional dialy SOAP note to a daily SCOAP note: Subjective, Concerns (open-ended inquiry into major concerns for the day), Objective, Assessment, Plan. The other team of 3 interns continued with the traditional daily SOAP note. The setting was an adult inpatient ward in a a community hospital and the trial was conducted over two months (two consecutive pairs of intern teams). Daily concerns of the intervention arm were tabulated and all patients were asked to complete a patient satisfaction survey at the time of discharge. RESULTS: Data were collected from 41 patients (24 from the intervention arm and 17 from the control arm). Patients whose concerns and worries were elicited by house officers during rounds and documented in the chart (the SCOAP note arm) were more likely to be satisfied with their care (92% vs. 71%) and were more likely to feel their day-to-day concerns were met (96% vs. 59%) than the control patients. Overall, 48% of the patients had concerns about discharge planning and 48% had concerns of pain and discomfort, followed by food and diagnosis (each 32%), sleep (24%), and concerns of upcoming tests/procedures/surgeries (20%). CONCLUSION: The traditional SOAP progress note should be changed to a SCOAP note (where daily concerns are inquired about and documented) to improve patient satisfaction. In addition, every effort should be made to keep patients abreast of any and all planning around their discharge.

PURPOSE: Primary care physicians and health insurers have been the focus of efforts to improve the quality and delivery of preventive health care such as smoking cessation counseling. We describe a collaborative effort between a primary care practice and an insurer to address the mutual goal of helping patients stop smoking. METHODS: The goal of the project was to increase patient smoking cessation rates and documentation of smoking cessation status and counseling by instituting systematic methods to identify smokers. We designed and implemented a three-armed trial that compared two smoking cessation interventions and usual care. Both interventions met the Agency for Healthcare Research and Quality's Clinical Practice Guidelines for instituting a systematic method to identify all patients who smoke. The interventions differed in their primary purposes: one intervention was aimed at only identifying and documenting the smoking status of patients; the other intervention also reminded clinicians to provide smoking cessation counseling and counseling documentation. RESULTS: The design phase of the project involved: 1) obtaining agreement by the health insurer and clinical practice leadership on clinical objectives, standards for documentation, and payment for smoking cessation counseling; 2) jointly developing a research protocol and implementation plan; 3) obtaining the endorsement of clinicians on the intervention teams; and 4) training clinicians and their staff to use the interventions. During the implementation phase, the insurer provided: clinician education seminars on effective and efficient smoking cessation counseling, patient information materials, and all intervention materials. CONCLUSION: This project illustrates a collaborative approach between a primary care practice and a health insurer that leads practices to institute clinical practice guidelines. It also demonstrates how a practice can implement improvements in patient care that meet mutually shared goals between a health insurer and a primary care practice (improving identification of patients who smoke and reducing patient smoking rates), as well as goals that are more important to one party, such as improved counseling documentation rates (insurer), and time-efficient smoking cessation counseling techniques (clinicians).

PURPOSE: Axillary lymph node status remains the best prognostic indicator for patients with invasive breast cancer. Recently, the role of axillary lymph node dissection in patients with T1a or T1b tumors has been debated. Some authors have suggested that the incidence of node positivity is so low in these patients that routine axillary dissection should be eliminated; others maintain that it should continue to be the standard of care because of its role in the determination of prognosis and systemic therapy decisions. The purpose of this study was to establish an institutional database of axillary lymph node status in patients with early breast cancer, to determine whether incidence of nodal positivity justifies axillary dissection in all patients regardless of tumor size. METHODS: All patients with T1a (less than or equal to 5 mm) and T1b (6-10mm) breast cancer who underwent both primary breast surgery and axillary node dissection at Mayo Clinic Jacksonville from 1992 through 1998 were identified through the tumor registry. The following data were collected for each patient: tumor size, biological grade, estrogen/progesterone receptor status, number of nodes harvested, and number of positive nodes. RESULTS: Of 163 patients evaluated, 39 had T1a and 124 had T1b tumors. Node positivity was 0% for T1a and 11.3% for T1b tumors (p ϭ 0.028). There was no relationship between lymph node involvement and estrogen receptor status (p ϭ 0.29). However, there was a significantly higher risk of lymph node positivity for progesterone receptor negative (p ϭ 0.012) and for estrogen receptor negative/ progesterone receptor negative lesions when compared to those with hormonally receptor positive tumors (p ϭ 0.040). Further, we found a significantly higher risk for lymph node positivity as tumor size increased (p ϭ 0.002). Finally, higher tumor grade conferred a higher risk for lymph node involvement. CONCLUSION: Our data demonstrate that, while T1a lesions have a minimal risk of nodal positivity and may not necessitate routine axillary dissection in the future, T1b lesions should be managed with routine analysis of axillary lymph node status. Whether sentinel node mapping can change this standard awaits further study. METHODS: We utilized 1998 medical and prescription claims data from an employer-based capitated insurance plan for an intergrated health delivery system covering 126,000 members in Eastern Massachusetts. We evaluated potential predictors of the concurrent use of multiple antihypertensives. We identified and analyzed 80,692 antihypertensive claims made by 11,383 members. Using pre-scription date and pill quantity combined with an assumed defined daily dose for each antihypertensive, we calculated the start and end dates for each prescription. We then ascertained whether 3 or more (3ϩ) antihypertensives were used concurrently. This information was merged with medical claims to determine if selected co-morbidities were present during the year. We also assessed other factors, including age, gender and whether the patient had been to three or more providers. Logistic regression was used to evaluate the independent impact of these predictors on the likelihood of receiving multiple antihypertensives. RESULTS: Of all members receiving antihypertensive medications, 35% concurrently received 3ϩ antihypertensives at some point during the year. Cardiac comorbidities significantly increased the likelihood of antihypertensive polypharmacy: heart failure (OR:3.1, 95% CI:2.4-4.1), coronary artery disease (1.9, 1.7-2.1), atrial fibrillation (1.3, 1.0-1.8), and hypertension (2.1, 1.9-2.3). Diabetes also predicted the use of multiple antihypertensives (1.7, 1.5-1.9). Increasing patient age made polypharmacy more likely (OR 2.2, 1.9-2.7 for Age у65 vs. 18-44). Female patients also were more likely to be on multiple antihypertensives (1.2, 1.1-1.3). The strongest independent predictor of antihypertensive polypharmacy was visits to more than three providers (3.9, 2.8-5.5). CONCLUSION: Antihypertensive polypharmacy may be legitimate, even necessary in patients with complicated caridac diagnoses. In addition to clinical predictors, nonclinical factors such as increasing age and female gender independently increased the likelihood of mutiple antihypertensives. Our study may suggest that better coordination is needed between multiple physicians caring for the same patient. In addition, consolidation of therapeutic regimines may be possible. While administrative claims data have several weaknesses, these data can identify predictors of polypharmacy. In addition, such information can provide physicians in capitated plans with useful feedback on polypharmacy. While such feedback may identify legitimate polypharmacy, it also may identify potentially dangerous, but unnecessary drug combinations.

General Hospital, Boston, MA; Howard University College of Medicine, Washington, DC PURPOSE: To examine patterns and quality of hypertension treatment from administrative claims data from a health care delivery system in comparison with national prescribing practices. METHODS: We analyzed 1998 prescription claims data from a commerical plan of an integrated delivery system that covers 126,000 members in 11 geographic regions in Eastern Massachusetts. From over 600,000 prescription claims, we identified 80,692 for antihypertensive drugs. We categorized these prescriptions by drug class: Angiotensin Converting Enzyme Inhibitors (ACEI, including A2 Receptor blockers), Calcium Channel Blockers (CCB), Beta Blockers (BB), Diuretics (DIUR), and others. Percent totals of each class were calculated for individual regions and in total. Costs per antihypertensive prescription also were computed and then compared to both national treatment guidelines and national data on patterns of hypertension treatment. RESULTS: Among all regions, 29% of all antihypertensive prescriptions were ACEI, 31% for BB, 19% for DIUR, 16% for CCBs, and 6% for others. This pattern was substantially different from published national patterns: 33% for ACEI, 11% for BBs, 8% for DIUR, 38% for CCB, and 9% for others. Within the delivery system there were wide variations in precsribing by region. CCBs ranged from 9.5% of prescriptions in one region to 22% in another. ACEIs ranged from 24% to 34% between regions, while the combination of BBs and DIURs ranged from 39% to 57%. The average cost of antihypertensives per prescription was $32.90, but varied between $24.60 to $33.80 by region. Difference in costs between drug classes, ranging from $58.35 for CCBs to $6.63 for diuretics, were a major contributor to these regional differences. CONCLUSION: Overall antihypertensive prescription patterns from this capitated plan appear are more consistent with JNC VI national guidelines than published national statistics. Consistent with these guidelines, the use of BBs and DIURs predominantes and results in lower treatment costs in comparison to national patterns. The wide regional variations between antihypertensive class and average prescription cost indicate room for further change. ACEIs and CCBs remain a large share of prescriptions and significantly increase the cost of therapy. For example, an estimated $300,000 in pharmacy costs could be averted within this delivery system by reducing CCB use to the low level (10%) found in one region. The magnitude of health care costs involved suggests the value of interventions that result in prescribing that is more consistent with national guidelines.

PURPOSE: In the managed care era, insurance companies often dictate primary care practice characteristics and allowable practice divisions between primary care and subspecialty providers. The purpose of this study is to seek patient preferences in this area. METHODS: A total of 700 patients of a multispecialty (primary and subspecialty care) academic internal medicine practice were randomly surveyed. The survey was divided into two sections. The first asked for patient preferences in terms of primary care versus subspecialty management for 26 different conditions and procedures. The second section asked for patient preferences concerning individual primary care physician characteristics and primary care practice characteris-tics. Usable surveys were returned from 652 patients, including 443 females and 209 males ranging in age between 18 and 84. A total of 597 patients reported having a regular primary care physician with 555 of these patients visiting a primary care office more than once in the last year. RESULTS: Of the conditions/procedures that were listed, primary care management was strongly favored in 12 (pneumonia, arthritis, asthma, back pain, stomach ulcer, fatigue, hyperlipidemia, fracture, suturing, joint injection, lancing a boil, rectal exam). Subspecialty management was strongly favored in 10 (Alzheimer's disease, angina, depression, personal problems, mole removal, sigmoidoscopy, cardiac testing, diaphragm fitting, pelvic exam, prostate biopsy). Four items (impotence, vaginal infections, diabetes, and chest pain) received equal preferences for primary and subspecialty care. The primary care physician characteristics that were identified as most important were compassion, accessibility, and spending time during the visit. The least important were liking the patient personally and being the same sex as the patient. The most prefered practice characteristics were answering all questions, spending time during the visit, sending letters about results, same day accessibility for acute problems, and always seeing the same physician. Answering phone calls within an hour and high med school class rank were viewed as least important. In a final dichotomous question, 66% of respondants prefered a primary care physician who ranked in the bottom 50% of his/her med school class but was compassionate, whereas 64% prefered a subspecialist who lacked compassion and a good bedside manner but who finished in the top 50% of his/her med school class. CONCLUSION: Patients have surprising and specific preferences for their primary care providers and their practices. These preferences should be routinely assessed and be considered, along with physician preferences, by insurers when forming policies. PURPOSE: Despite the growing numbers of refugees and asylum seekers and their complex medical, mental health, social, and legal needs, little descriptive data exists about this patient population. We describe the patients enrolled through primary care in a refugee/survivor of torture program. METHODS: We reviewed the charts of all patients seen between 12/98 and 12/99. RESULTS: 50 patients were enrolled of which 33 (66%) were males. The mean age was 41 years for women and 40 years for men; age ranges were 30-70 years and 18-65 years, respectively. 47% of the patients were from Central America and the Caribbean, 34% from Africa and the Middle East, 10% from South America, and 8% from Europe. The main reasons for uprooting and torture were civil war (34%), political activities (14%), ethnic persecution (14%), relative of a victim (10%), humanitarian activities (6%), refusal to cooperate (6%), and religious activities (6%). Patients were subjected to threats (86%), harassment (40%), beatings (38%), battlefield conditions (22%), destruction of property (22%), sexual torture (16%), and electric shocks (12%). After uprooting and torture, 32% of the patients continued to experience significant psychological trauma. Common psychiatric conditions were post-traumatic stress disorder (54% met all Diagnostic and Statistical Manual (DSM) IV criteria and 10% had some DSM IV criteria), depression (38% met all DSM IV criteria and 4% had some DSM IV criteria) and anxiety (40%). 28% percent of the patients had scars and another 8% showed evidence of organ dysfunction or limb deformity. 30% of the patients were active or former smokers and 22% had a CAGE score equal or greater than 3. No patient had recreational drug abuse. 36% of the patients were enrolled in mental health care and 16% declined. 6% of the patients declined to continue their primary care with the program. 15 affidavits were written and 3 court testimonies were made on behalf of these patients. 21 patients had active legal needs; attorneys working with the Center took 38% of these cases. 20% of the patients were referred for the first time to a social worker. 33 patients had previously seen a healthcare provider of which 82% were never asked if they had suffered uprooting or torture. The mean time between arrival to the country and first contact with the program was 4 years, a range of 3 months-15 years. CONCLUSION: Chronic physical and mental sequelae are common among this patient population, as the continuation of psychological trauma after uprooting and torture. Another important problem is the non-recognition by healthcare providers of this vulnerable population, which in part explains the long delay between arrival to the country and access to specialized care. PURPOSE: Resident physicians are often required to perform invasive procedures at the bedside on patients with disordered hemostasis. We wished to determine the threshold levels of International Normalized Ratio (INR) and platelet count at which resident physicians felt comfortable proceeding with three common procedures without prior correction of the hemostatic derangement. METHODS: As part of a seminar on quality assurance, 19 residents in internal medicine (PGY1 ϭ 7, PGY2 ϭ 7, PGY3 ϭ 5) were presented with three clinical scenarios, each of which portrayed a patient in urgent need of one of the following procedures: paracentesis(PC), thoracentesis (TC), or lumbar puncture (LP). In the first iteration of each scenario, respondents were asked to indicate the threshold level of INR above which they would not perform the procedure without first infusing fresh frozen plasma (FFP). In the second iteration, they were asked to record the threshold platelet count below which they would not proceed without platelet transfusion. RESULTS: Threshold values (see table for median values and ranges) showed wide variation in each iteration of all three scenarios. For example, while four respondents stated they would perform a paracentesis even if the INR were above 3.5, almost a third would give fresh frozen plasma prior to the procedure if the INR exceeded 1.5. Although the data show that residents tend to exercise greater caution with LP than PC or TC, four residents indicated that they would perform LP without platelet transfusion even in the face of severe thrombocytopenia (Ͻ20). With increasing resident seniority (PGY1 and PGY2 vs PGY3), a trend toward a more cautious approach to LP and a more liberal approach to TC, was also evident (data not shown).

CONCLUSION: There is wide variation in resident physicians' perceptions of the indications for correction of hemostatic derangements prior to performing invasive procedures. Potential adverse consequences of inadequate knowledge in this area include the unnecessary delay of procedures, or inappropriate utilization of blood products. Although this pilot-study is small, it strongly suggests that there is a need for the dissemination (and development, where necessary) of guidelines for the transfusion of blood products prior to performing these common procedures. PURPOSE: Breast self examination (BSE) is a patient-centered, convenient, noninvasive and inexpensive screening modality for breast cancer. Having been taught BSE by a medical provider predicts whether it is performed. The purpose of this study was to determine which patient and provider characteristics are associated with the teaching of BSE. METHODS: Medical records were reviewed for documentation of whether BSE was taught to 2245 randomly selected women, ages 40 to 69, who were enrolled continuously for 10 years in an HMO, had at least one screening clinical breast exam and had no history of breast cancer. Using univariate, logistic and generalized estimating equations multiple regression analyses, patient and provider variables were tested for their association with BSE teaching. RESULTS: Over 10 years, 68% (n ϭ 1534) of the women had BSE teaching documented at least once (median 1.6, range 0-13). Documentation of BSE teaching was more likely among women under age 65 than those over (70% vs. 64%, p ϭ 0.005), women of income greater than $35,000 than those under (69% vs. 63%, p ϭ 0.025) and women who had been prescribed estrogen replacement therapy than those not (72% vs. 66%, p ϭ 0.004). White women were taught more commonly than black women (70% vs. 62%, p ϭ 0.012) and were also more likely to be taught at least twice (45% vs. 29%, p ϭ 0.001). Body mass index, marital status and having a relative with breast cancer were not associated with documentation of BSE teaching. BSE teaching was directly related to the number of clinical breast exams (p ϭ 0.001) and screening mammograms performed (p ϭ 0.001). Multivariable GEE logistic regression analysis found patient age, frequency of clinical breast exams and frequency of screening mammograms to be independent correlates of BSE teaching. Of the primary care providers who saw at least 3 patients in the study sample (n ϭ 248), 88% of nurses, nurse practitioners and physician assistants, 75% of internists and only 17% of the obstetrician-gynecologists documented BSE teaching at least once (p ϭ 0.001). Female providers taught more than did male providers (79% vs. 64%, p ϭ 0.01). These associations remained significant in multivariable logistic regression. CONCLUSION: Two-thirds of female patients in this HMO setting were taught BSE at least once during the 10-year period. Patient age, general breast cancer screening frequency as well as provider type and gender were associated with the documentation of BSE teaching. CONCLUSION: While adherence to NCEP guidelines for cholesterol screening and treatment was higher for patients with CAD, overall adherence was low. Several factors, including frequent contact with PCP and CAD risk, were significantly associated with adherence. These data suggest the need for targeted efforts to increase adherence with NCEP guidelines, particularly for at-risk patients. PURPOSE: Disease management (DM) utilizes coordination of care across the disease spectrum and the settings where care takes place, generally is based on a capitated payment and risk-taking, and has been successful in the management of patients with a variety of chronic diseases. Since DM is a form of proactive patient care, it is essential to be able to predict which patients are likely to have adverse outcomes and to target care to prevent complications where possible. Unfortunately, however, there are no generally accepted, validated risk assessment tools for patients undergoing dialysis. METHODS: Renal Management Strategies (RMS), an affiliate of Baxter Healthcare, has been providing DM for over 1200 patients enrolled in a large national managed care organization in 16 cities. We developed a unique risk assessment tool with 10 items based on a variety of patient characteristics and previous history of resource utilization, and also used the ICED index to evaluate patient risk. RESULTS: The following correlations have been found between elements of these two instruments and patient outcomes:

CONCLUSION: Although additional refinement of the risk assessment tools is needed, particularly further simplification, such instruments are of considerable value as more dialysis patients are cared for using a DM approach. The multi-faceted risk tool which includes utilization data is more predictive of risk than the ICED, which includes only co-morbid and physical conditions.

Rabow, S Dibble, Internal Medicine; Institute of Health and Aging, University of California, San Francisco, San Francisco, CA PURPOSE: Traditionally, primary care clinicians provide care to patients near the end of life. The quality of patient care often depends upon the ease of access and the willingness of primary clinicians to refer to ancillary services such as pharmacists, home care, hospice, psychologists, spiritual leaders and social workers. The purpose of this study was to investigate referral patterns and challenges of primary care clinicians for patients near the end of life. METHODS: Using a written questionnaire with 29 items, we surveyed primary care clinicians in a university-based general internal medicine practice regarding their referral patterns for seriously ill outpatients with chronic obstructive pulmonary disease, congestive heart failure or cancer. Five-item Likert scales were used to quantify the ease of providing care (1 ϭ difficult, 5 ϭ easy), number of referrals made (1 ϭ none, 5 ϭ Ͼ3 referrals) and areas in which the primary clinicians needed help (1 ϭ no help, 5 ϭ a lot of help). We calculated mean scores and compared them between residents and faculty using the t-test statistic. RESULTS: Of the 81 clinicians surveyed, 64 responded (79%). The clinicians were predominantly female (56%), Caucasian (59%) residents (63%) and had a mean age of 33.3 (S.D. 7). Overall, primary care clinicians reported difficulty in arranging homecare (mean score ϭ 2.2), providing psychosocial care to patients (mean score ϭ 2.4 ) and families (mean score ϭ 2.2), and providing adequate access by phone for patients (mean score ϭ 2.2 ). Residents indicated greater difficulty than faculty with conducting advanced directive discussions (mean score 2.1 vs. 3.1, p Ͻ 0.01) and with providing psychosocial care to patients (mean score 2.0 vs. 3.0, p Ͻ 0.01). Compared with faculty, residents indicated more help was needed with hospice referrals (4.0 vs. 3.0, p Ͻ 0.02). CONCLUSION: To maximize the quality of care for patients near the end of life, faculty and residents in the university setting need increased access to home care services and psychosocial supports. This assistance may most easily be provided through an integrated multidisciplinary team designed to provide end of life care. There was also a significant decrease in the frequency of the following common diagnostic clusters: Ischemic Heart Disease (9.1% to 2.8%), Degenerative Joint Disease (3.7% to 1.8%), Depression/Anxiety/Neurosis (3.2% to 2.1%), Medical Surgical Aftercare (1.5% to 0.5%), and Obesity (1.5% to 0.7%). CONCLUSION: Our analysis shows an increase in acute illnesses seen in the outpatient setting. This result could be attributed to the recent shift of acute care from inpatient and emergency departments to the outpatient setting due to the increased penetration of managed care. The increase in general medical exams may be due to the need to establish a primary care physician in many health plans or the increased emphasis on preventive medicine. The reasons for the three-fold decrease in the frequency of ischemic heart disease needs to be investigated further. In view of these results it may be beneficial for internal medicine residency and CME programs to ensure that proper attention is being given to the diagnosis and management of acute illnesses. risk classes I-II (low risk) and should have been managed as outpatients. These patients had more comorbidities (1.6 vs 0.7, p Ͻ 0.001) and signs/symptoms (3.6 vs 2.5, p Ͻ 0.001) than the 62 class I-II patients who were managed as outpatients. No deaths or rehospitalizations occurred in these patients compared with no deaths and a 3.2% rehospitalization rate for the class I-II patients who were managed as outpatients. No deaths occurred in any patients in risk classes I-III, regardless of treatment setting. Mortality rate was 14% in risk class IV and 33% in risk class V. CONCLUSION: Agreement between observed clinical practice and the prediction rule was 86%. Deviations from the prediction rule were skewed toward over-hospitalization, as 26% of our low-risk patients were managed as inpatients, while 4% of our high-risk patients were managed as outpatients. Some of this deviation may be attributed to factors such as compliance or social issues, patient preferences, etc. that are not addressed by the prediction rule and may not be captured in the medical record. We believe most of it is attributable to clinicians overestimating the severity of illness in low-risk patients. Prospective use of the prediction rule can help identify these low-risk patients and triage them to outpatient management, which should result in cost savings without adversely affecting outcomes. PURPOSE: Internet surveys are an increasingly common way to access large populations at minimal costs in obtaining research data. Although many factors make this approach attractive, the reliability of responses obtained is unclear, even if a paper-based questionnaire is placed on the Internet unaltered. The purpose of this study was to establish reliability in administering a questionnaire by both traditional and computerized interfaces. METHODS: A questionnaire consisting of the Rhinoconjunctivitis Quality of Life Questionnaire (RQLQ) and the allergy-specific Work Productivity and Activity Impairment index was developed for three different interfaces: paper form (P), touch screen (TS), and web-based (WB) form. Volunteers answered this questionnaire in two of the three formats, assigned randomly and in random order. No supplemental instructions were provided by the research assistant, who administered all questionnaires sequentially in an office setting. Responses were transformed to a summary score for the RQLQ, compared for each individual, and differences by survey format were explored using one way ANOVA. Test-retest responses for each individual were compared on an item-by-item basis using Pearson's correlation and kappa coefficients. Completion time for each test was monitored and compared for each format pair using Student's t test. Respondent acceptance was ascertained using an open-ended question upon completion of the second questionnaire. RESULTS: 75 volunteers completed 150 questionnaires in one of three possible format pairs (P-TS, P-WB, and TS-WB). 54 (72%) of respondents were female and the median age group for respondents was 40-49 years. Responses for most items were highly correlated for all formats (p Ͻ 0.001) although correlation coefficients ranged from .575 to .960. Kappa coefficients suggested excellent concordance in responses obtained by each format pair. Interestingly, the RQLQ score was modestly, although significantly higher for the paper-based form, while similar for both TS and WB formats. Completion time did not differ by questionnaire format and acceptance of each testing mode by respondents was similar. CONCLUSION: Although concordance and correlation of responses to a questionnaire administered using different interfaces is similar, a systematic bias may have resulted in higher scores for respondents using the paper-based form. This study suggests alternate methods of data acquisition may have significant value in obtaining research data without affecting the reliability of results. Nonetheless, careful testing prior to implementation is important to identify other factors that may lead to small, but significant systematic differences. METHODS: Patients were recruited from a university hospital urology clinic, a university hospital internal medicine clinic, and from university-affiliated community primary care clinics. Letters were sent to 236 eligible patients by the 25 primary care providers and 2 urologists in the study asking their patients if they would participate in the study. Respondents (n ϭ 140) who returned a postcard were sent the survey materials. 124 (or 52.5% of the eligible sample) completed and returned a survey; the 85 patients who reported current use of sildenafil were included in this analysis. Sexual and erectile function were assessed using the International Index of Erectile Function (IIEF). Items from the marital interaction scale from the Cancer Rehabilitation Evaluation System Short Form (Cares-SF) were used to assess relationship with sexual partner. The survey also included the SF-12 and the emotional well-being scale from the SF-36. Respondents were asked to report about their status for both before and after treatment with sildenafil.

RESULTS: The mean age of the respondents was 60.5 years (range 36-80). 53% were married; 95% were white; 62% had a 4-year college degree; mean household income was $75,000. Physical and mental health for the sample at baseline was similar to the U.S. population: the average SF-12 physical health summary score was 52.9; the average mental health summary score was 51.4. Users of sildenafil reported a 54% increase in overall IIEF scores, 88% increase in erectile function scores, 60% increase in overall sexual satisfaction, and 36% increase in intercourse satisfaction (all p Ͻ 0.01). 38% of respondents indicated that using sildenafil had definitely improved the quality of their life. Likewise, 29% of respondents indicated that using sildenafil had definitely improved their relationship with their partner. With the use of sildenafil, there was a statistically significant improvement in scores for erectile function (p Ͻ 0.01), relationship with sexual partner (p ϭ 0.01), and emotional well-being (p Ͻ 0.01). In a multivariate model, improvements in erectile function and relationship with sexual partner were each significantly associated with improvement in emotional well-being (R 2 ϭ 0.19, p Ͻ 0.01). CONCLUSION: Users of sildenafil reported dramatic improvements in erectile and sexual functioning. These improvements were associated with improvements in their scores on emotional well-being and relationship interaction measures. Prospective studies are needed to further investigate these potentially important associations. , a specific inhibitor of Cox-2, showed osteoarthritis (OA) patients treated with VIOXX had significantly fewer clinically significant gastrointestinal (GI) adverse events than those who received non-selective NSAIDs. We estimated the economic implications of these results. METHODS: A model-based decision analytic approach was used, focusing on events related to GI problems that imply healthcare resource use (gastroduodenal perforations, symptomatic gastroduodenal ulcers or upper GI bleeding (PUBs) and minor GI problems, e.g., dyspepsia). Event probabilities came from clinical trials of VIOXX. To extrapolate trial results to practice, resource utilization profiles for GI events were developed using the literature, updated by chart review studies and costed using current published sources. Average wholesale prices were used for base case drug costs, weighted by 1999 market share for generic and branded products. Efficacy of GI co-medications (co-meds) and mortality risk associated with major GI problems were inferred from the literature. A 75% reduction in future prophylactic GI co-med prescription rates with VIOXX was assumed for the base case. Base case 1 year analyses were done with the PUB data obtained from a prespecified pooled analysis of the VIOXX clinical trials. Analyses were also performed using pooled results of two 12 week endoscopic surveillance trials, with adjustments for silent ulcers of 40% and 85%. Sensitivity analyses explored the effects of varying costs, hospitalization and surgery rates for PUBs, efficacy of GI co-meds and rates of GI co-med use. We calculated iatrogenic cost factors, which express a drug's economic burden as the ratio of the total expected cost of drug treatment (drug cost plus costs of secondary drug effects) to drug cost alone. RESULTS: The iatrogenic cost factors implied by the base case were 1.86 for NSAIDs and 1.18 for VIOXX. The NSAID iatrogenic results are consistent with previously published results. Under base case conditions, the expected cost savings in GI problems and co-meds averted with VIOXX versus NSAIDs was $0.81 per day, representing a 95% offset of the difference in drug price. In analyses based on endoscopic data, therapy with VIOXX was less expensive than therapy with NSAIDs regardless of silent ulcer adjustment. Results were most sensitive to prophylactic GI co-med rates, and robust over a range of model assumptions and costs. CONCLUSION: In this analysis based on differences in clinically significant GI events and dyspepsia for OA patients, VIOXX had a markedly lower iatrogenic cost compared to NSAIDs due to cost savings in GI problems and co-meds averted. When endoscopic data alone were considered, VIOXX was cost-saving across all assumptions about silent ulcer rates. With the exception of race, discomfort with FFS was also significantly related to all of these characteristics. As shown in the table, multiple logistic regression models found education, insurance type, and income to be independent correlates of discomfort with both FFS and CAPITATION:

CONCLUSION: A majority of patients wanted more information about how their PCP was paid. Discomfort with fee-for-service and capitation were evident compared to salaried compensation. Patient concerns about financial incentives are important to consider in the design and disclosure of physician compensation. PURPOSE: Despite convincing evidence that breast-conserving surgery (BCS) followed by local irradiation produces equivalent cancer cure rates compared to mastectomy in early stage breast cancer, the utilization of BCS for patients who meet clinical criteria remains low. METHODS: Using a structured questionnaire, we interviewed 101 women aged 50 to 75 yrs. to assess their strength of preference for BCS and mastectomy when presented with scenarios offering these therapies for their newly diagnosed breast cancer. The scenarios were presented in a standard gamble format and preferences are reported as utility scores (1 ϭ perfect health, 0 ϭ death). Bivariate analyses to identify predictors of utility scores and to compare same subject utility scores for BCS vs mastectomy were performed using non-parametric techniques. RESULTS: For the new diagnosis of early stage breast cancer followed by BCS then irradiation the mean utility score was .83 (95% CI. 77, .89) compared to .82 (95% CI .76, .88) for this diagnosis followed by mastectomy. The .01 mean utility score difference was not statistically significant. In bivariate analyses, "good" selfreported health predicted higher utility scores for both procedures while income greater than $50,000 per year predicted a higher score for mastectomy. Age, race, education, health insurance status, and marital status did not predict any significant differences in utility scores for either BCS or mastectomy. CONCLUSION: We conclude that the very limited use of BCS is driven by factors other than patients' preferences. (63). Analysis of consultant recommendations identified trends among specialities, with internists (especially GIM) serving more of a management and evaluative role, sub-specialists more of a diagnostic role and surgeons and gastroenterologists (GI) more of a procedural role. Chi-square for a gradient in the proportions tests were used to test the significance of the associations between ordered physician groups and their recommendations.

Consults and the Recommendations Recorded by Specialty Groups CONCLUSION: Chart review found that hospitalized patients receiving consults have a longer than average LOS despite rapid response by the consultants. The type of recommendations differed by speciality. Internists (especially GIM) provide more management and evaluative guidance, sub-specialists offer more diagnostic suggestions and surgeons and gastroenterologists propose more procedures. Determination of the impact of these differences on patient outcome and the reasons for choosing a particular consultant requires further study.

K Ramasubbu, HS Gurm, DG Litaker, Internal Medicine Residency Program, The Cleveland Clinic Foundation, Cleveland, OH PURPOSE: Differential enrollment into clinical trials by gender has been previously described. Because study conclusions arising from a predominantly male population must be extrapolated to all patients, a concern for generalizability has been raised. In 1993, recommendations were made by the Food and Drug Administration specifically encouraging the inclusion of women to address this potential issue. The purpose of this study was to review clinical trial (CT) enrollment among studies published in a major medical journal to reassess changes in this practice. METHODS: We conducted a systematic search of all articles published in the Original Articles section of the New England Journal of Medicine from 1994-1999. All randomized CTs in which the primary end point was either total mortality or included mortality in a composite endpoint were considered for abstraction. We excluded all studies concerning gender-specific diseases, subset analyses of other trials, and metaanalyses. Trials were characterized by primary specialty and the primary factor of interest was percentage of women enrolled. Data were analyzed using Student's t test and ANOVA. RESULTS: 1322 original articles were published in the period 1994-1999 including 423 CTs. Although 121 met initial inclusion criteria, we focused attention on the four specialities responsible for publishing the majority of these reports. 94 CTs were identified for abstraction and represented publications related to cardiovascular disease in 50 (54.3%) articles, treatment of cancer in 21 (22.3%) articles, infectious diseases in 12 (12.8%) articles, and the gastrointestinal system in 10 (10.6%) articles. Considering all study results together, only 11 (11.7%) studies reported the primary outcome mortality by gender. groups. Interestingly, of 13 studies funded by the NIH, the percentage of women enrolled was lowest (20.7%). CONCLUSION: At a time when evidence-based medicine is emphasized in residency training and medical practice, there are surprisingly few gender-specific data available to guide the treatment of a substantial segment of the population. Despite the recommendations of a federal agency, enrollment of women into CTs published in a major medical journal is disproportionately low. These results suggest an ongoing need for a conscious effort at enrollment of women in future research. PURPOSE: The benefit of multiple psychotropic medication use (polypharmacy) among patients with schizophrenia is uncertain. However, clinicians may attempt to manage psychotic patients' residual symptoms with additional psychotropic agents. Despite this practice, the prevalence of polypharmacy, and characteristics of patients receiving multiple psychotropic agents in clinical practice, is unknown. METHODS: We evaluated patients treated at a hospital-based ambulatory psychotic disorder clinic (Schizophrenia and Treatment Evaluation Program) at the University of North Carolina Hospital, Chapel Hill, NC to determine the incidence and characteristics of patients receiving psychotropic polypharmacy. Clinic patients' physicians were contacted and asked to provide data including patients' demographic characteristics, baseline psychiatric diagnoses, aspects of medical history, and current psychotropic medication use. Patients receiving polypharmacy were identified and polypharmacy was evaluated for associations with patient characteristics using t-test and chi-square analysis. Of the 130 patients enrolled in the clinic, we obtained data on 106 patients (82%) whose results are reported here. RESULTS: Patients were 34 Ϯ 12 years of age, predominantly male (63%) and Caucasian (79%), though the clinic contained a substantial population of African-American patients (19%). Primary patient DSM IV axis I diagnoses were schizophrenia (56%), schizoaffective disorder (35%), non-organic source psychosis (7%) and bipolar disorder (6%); few patients (4%) presented with axis II disorders. Patients had received psychiatric treatment for an average of 10.5 Ϯ 9.1 years and most (80%) were treated with an atypical antipsychotic agent. In addition to antipsychotic agents, patients were treated with antidepressants (12%), mood stabilizers (5%), antianxiety agents (4%). More than a third (34%) were treated with an antipsychotic agent and two or more psychotropic agents, while 8% were treated with two antipsychotic agents. Psychotropic polypharmacy was noted in 62% of patients, with an average number of 2.3 medications. Patients prescribed multiple medications were older (mean: 33.6 vs 26.7, p Ͻ 0.01), had been in psychiatric treatment longer (mean: 10.0 vs 4.5, p Ͻ 0.02), and were more likely to be white (86.2% vs 70.6%, p Ͻ 0.05); gender and a usual source of medical care were not associated with polypharmacy. CONCLUSION: A significant portion of patients undergoing treatment at a hospital based ambulatory psychotic disorder clinic receive psychotropic polypharmacy. Further research is needed to elucidate the appropriateness of this polypharmacy. To examine the relationship of experience (e.g., intern, resident, attending) with physicians' emotional reactions to dying patients this study assessed grief reactions, burnout, and coping behavior in a sample of physicians who recently experienced a patient death on a general medicine unit in an academic hospital. METHODS: Of the 61 physicians eligible for the study, 89% (n ϭ 11) agreed to participate. Participants included 18 interns, 18 residents, and 19 attendings the majority of who were male (64%), married (63%), and caucasian (76%). Each week two patients who died on the medical service of two large teaching hospitals were randomly chosen, and the attending, resident, and intern who cared for the patient at death were selected for study inclusion. The identified physicians were contacted by phone and asked to participate in the study which involved completing self-administered questionnaires as well as participating in an interview about thier experience in caring for the patient who died. Interviews were transcribed and read for common themes. The Maslach Burnout Scale, the Grief Reaction Inventory, and 6 items from the Brief Ways of Coping Scale assessed burnout, grief, and coping, respectively. Questionnaire data were analyzed with SPSS-9.0 software. One-way ANOVAs tested the relationship between level of training and reactions to patient deaths, and hierarchical multiple regression tested the effect of experience (in years) on the relationship between grief reactions and burnout. RESULTS: Attending physicians reported significantly fewer grief reactions (F(2,52) ϭ 4.65, p Ͻ .02) and burnout symptoms (F(2,52) ϭ 3.40, p Ͻ .05) than interns and residents; however, groups did not differ in the number of coping behaviors used to manage grief reactions (F(2,52) ϭ 1.9, p ϭ .16). Regression analyses indicated that number of years experience mediated the relationship between grief reactions and burnout (␤ ϭ Ϫ.36, p Ͻ .01); as physicians gained experience, grief was less likely to be associated with burnout symptoms. Interview data suggested that interns were more affected by patient deaths because they were more intimately involved with patients and their families. Furthermore, interns indicated that they were less likely to receive the support they needed when compared to the reports of residents and attendings. CONCLUSION: A greater sense of emotional closeness to patients and families as well as a lack of support from colleagues appeared to make interns more vulnerable to the effects of patient deaths when compared to more experienced physicians. PURPOSE: Prior to implementing a standardized decision aid for chest pain patients in our Emergency Department (ED), we sought to understand how our physicians diagnosed and triaged such patients without it. METHODS: We presented 20 written cases to 147 physicians specializing in internal medicine, emergency medicine and cardiology at one institution. Each case contained clinical data needed to estimate the patient's probability of coronary artery disease (pCAD), acute myocardial infarction (pMI), and major complications (pComp) as reported in large published cohort studies. Respondents were asked to estimate these probabilities, make a provisional diagnosis (MI, unstable angina, other) and a triage decision (CCU, non-CCU telemetry, other). Accuracy of probability estimates was measured in two ways: 1) as proportions within 5 percentage points of the best point estimates reported in the literature, and 2) as proportions within clinically relevant threshold ranges (e.g., pMI Ͻ5%, 5-25%, Ͼ25%). Agreement was measured as interquartile ranges (IQR) and reliability coefficients (R) for continuous variables, and as kappa statistics (k) for categorical variables. RESULTS: Overall, agreement among physicians was moderate for pCAD (R ϭ 0.49) and pMI (R ϭ 0.45) but only fair for pComp (R ϭ 0.30), diagnosis (k ϭ 0.31), and triage decisions (k ϭ 0.36). Wide variation persisted after controlling for physicians' specialty and experience. Accuracy of estimates of pCAD, pMI, and pComp compared with the literature (ϩ/Ϫ5%) were only 18%, 26%, and 97%, respectively. Median probability estimates exceeded clinically relevant threshold ranges in all cases except pCAD when pCAD was high. Among the lowest risk cases, physicians overestimated pCAD, pMI, and pComp in 81%, 62%, and 67%, respectively. CONCLUSION: Our physicians varied widely in their diagnoses and triage decisions in simulated cases of ED patients with chest pain. They consistently overestimated pCAD, PMI, and PComp. Achieving consensus in decision-making will require physician education and, perhaps, standardized decision aids.

Hospital; Rush Medical College, Chicago, IL PURPOSE: Before implementing an outcomes-based triage aid for chest pain patients in our Emegency Department (ED) (derived from Goldman et al's prediction rule; NEJM 1996), we compared the triage aid with our physicians' decision-making in simulated cases. METHODS: In 20 simulated cases, we compared trichotomous triage decisions (CCU, non-CCU telemetry, other) by 147 physicians at one institution with decisions recommended by the decision aid. Using recursive partitioning, we modeled the multivariate predictors of physicians' triage decisions: clinical variables of the cases (eg, electrocardiogram (ECG) changes, congestive heart failure (CHF), hypotension, prior myocardial infarction (MI) or revascularization); physicians' probability estimates (for coronary artery disease (CAD), MI, and cardiac complications); physicians' diagnoses (MI, unstable angina), specialty (Medicine, Emergency Medicine, Cardiology), level of training (resident, attending), and experience. Using the decision aid as the gold standard model, we measured the accuracy of physicians' decisions. We also examined the impact of physicians' diagnoses and probability estimation on the accuracy of both the gold standard model and the recursive partitioning model. RESULTS: In contrast to the decision aid (which is based on probability of complications), the physicians' decision model involved 3 sequential branch points: 1) diagnosis of MI or unstable angina; 2) presence of CHF; and, 3) abnormality of ECG. Compared with the decision aid, physician's triage decisions were inaccurate in 42% of cases. Perfecting physician's diagnoses (in their model) reduced inaccuracy slightly, to 40%. Conversely, using physicians' probability estimates in the gold standard model actually increased inaccuracy, to 56%. CONCLUSION: Our physicians use a diagnosis-based triage strategy for ED patients with chest pain. Adoption of an outcomes-based strategy will require physician's acceptance of both its logic and its empirically derived probability estimates.

PURPOSE: Managed care plans have a major role in selecting health care providers, but there is little evidence that quality plays a role in this process. We surveyed health plan executives to assess whether and how they use publicly available or internally generated hospital outcomes data. METHODS: The CA Department of Corporations provided a list of managed care plans with active Knox Keene licenses as of June 1998. We attempted to contact all 47 HMOs that contracted with hospitals to pay for acute inpatient care. We mailed a questionnaire to a key respondent at each HMO, and followed up with telephone calls. Thirty HMOs provided a usable questionnaire or interview (63.8%). Several HMO characteristics were examined, including size, profit status, model type, and accreditation. RESULTS: Health plan executives identified the following factors as most important in contracting with hospitals (in descending order): JCAHO accreditation, geographic location, and negotiated price. Respondents from for-profit HMOs assigned greater importance to price than respondents from nonprofit HMOs (p Ͻ .02). Other factors deemed very or extremely important by at least 80% of respondents were disciplinary actions against the hospital, the hospital's reputation, and its commitment to quality improvement. By contrast, specific outcome measures (e.g., mortality, complication, or readmission rates; incidence of potentially overused procedures; transplant success rates) were deemed less important. Respondents from staff model HMOs viewed these outcome measures as more important than those from other HMOs (p Ͻ .03). Although most respondents (70%) had reviewed at least one source of publicly available information on hospital outcomes, the usefulness of each source (except www.healthscope.org) was rated "poor" or "fair" by at least 40% of those who had used it. Ten respondents, including 7 of the 9 accredited HMOs, reported conducting internal studies of hospital performance. Health plan executives saw collecting and disseminating hospital quality information as primarily the responsibility of government agencies and accrediting organizations (e.g., JCAHO). About 73% of respondents said that HMO's should collect and analyze their own information on hospital quality. The majority of our respondents agreed that hospital outcome studies would ultimately lead to improved quality of care (87%) and less unnecessary and inappropriate care (70%). CONCLUSION: Managed care leaders in CA base hospital contracting decisions primarily on accreditation, geographic location, and price. Objective data on hospital quality were not viewed as extremely important in this process. However, respondents expressed optimism that such information may become more useful and important in the future. PURPOSE: Preliminary data from animal and human studies suggest that hyperglycemia during acute ischemic strokes is associated with higher morbidity and worse short-term sequellae. If true, this might identify a potential intervention to improve stroke outcomes. Using data from a comprehensive electronic medical record in an inner-city public hospital, we sought to assess (1) the proportion of patients with acute stroke who are hyperglycemic when admitted (2) the severity and treatment of the hyperglycemia, and (3) its association with stroke outcomes. METHODS: Each patient's first hospitalization for ischemic stroke (i.e., primary discharge ICD9 code 434 or 436) during the 5-year period from 7/93 through 6/98 was identified. From patients' electronic records we extracted data on demographics, prior diagnoses, all blood glucose (BG) results performed by the inpatient lab (finger-stick BG results were not available), hypoglycemic drug treatment, inpatient duration and charges, and mortality within 30 days of discharge. Stroke patients with admission hyperglycemia (first BG on admission у130 mg/dL) were compared with euglycemic stroke patients. RESULTS: We identified 671 patients with acute stroke. Their mean age was 62 years, 56% were women, 55% were African-American, and 51% had prior diabetes. Of the 656 (98%) stroke patients with inpatient BG results, 269 (41%) were hyperglycemic on admission, and 153 (23%) had an admission BG у180. Patients who were hyperglycemic on admission were more often women (64% vs 50%, p Ͻ .001) and diabetic (80% vs 30%, p Ͻ .001). An average of 8 additional BG tests were obtained for 183 (68%) of the patients with admission hyperglycemia, with a mean value of 202. Inpatient drug treatment of patients with admission hyperglycemia was suboptimal: 44% received no diabetes drugs, 10% received only oral hypoglycemic drugs, and 13% received only sliding scale insulin. Patients with admission hyperglycemia had 20% longer inpatient stays (7.1 vs. 6.0 days, p ϭ .02), 24% higher total inpatient charges ($10,661 vs $8,5812, p ϭ .03), and twice the risk of dying within 30 days of discharge (11% vs 5.7%, p ϭ .02). Controlling for age, sex, race, and prior diabetes, the odds ratio for 30-day mortality among patients with admission hyperglycemia was 1.9 (95% CI 1.0-3.7, p ϭ .049). The highest BG obtained during hospitalization also predicted 30-day mortality: a 50 mg/dl increase in the highest BG increased the odds of death by 20% (95% CI 10-25%, p Ͻ .001). CONCLUSION: Admission hyperglycemia is common, costly, and morbid among inner-city patients with acute ischemic strokes. It is also inadequately treated. These data support the implementation of clinical trials of intensive management of hyperglycemia in patients with acute ischemic strokes. PURPOSE: Most US hospitals use critical pathways for at least some of their patients. A goal of pathways is to maximize efficiency of care by reducing patient length-of-stay and resource utilization; however, pathway effectiveness in improving efficiency has been inadequately evaluated. We therefore assessed whether or not critical pathways as a clinical management technique have been successful in reducing patient length-of-stay and resource utilization at our academic medical center. METHODS: We evaluated all critical pathways begun in our medical center between 1993 and 1996 in which at least 50 adult patients could be evaluated in the year preceding and in the year succeeding pathway implementation. Each pathway was developed separately, based on the recommendations of a multidisciplinary team. Using a before-and-after design, we evaluated each pathway's effect on patient length-of-stay and resource utilization (as determined by relative value units). We constructed separate models for each pathway using each outcome and included patient data one year before and one year after pathway implementation. We accounted for both secular trends and DRG case-mix changes within pathway groups in our analyses. Statistical significance was based on a two-sided P value of less than 0.05. RESULTS: 13 pathways satisfied inclusion criteria. The diagnoses (or procedures) covered by these pathways were: acute myocardial infarction (AMI), acute pancreatitis, asthma, breast surgery, caesarian section, community-acquired pneumonia (CAP), hip arthroplasty, kidney transplantation, knee arthroplasty, liver transplantation, lung lobectomy, percutaneous transluminal coronary angioplasty (PTCA), and radical nephrectomy. 4 of the 13 pathways significantly decreased patient length-of-stay more than what would be expected based on the secular trend: the pathways for AMI (by 10%), caesarian section (by 7%), kidney transplant (by 12%), and PTCA (by 11%). 4 pathways also significantly reduced ancillary resource utilization more than expected: the pathways for caesarian section (by 26%), CAP (by 10%), kidney transplant (by 14%), and PTCA (by 12%). 3 pathways-caesarian section, kidney transplant, and PTCA-were successful in significantly reducing both patient length-of-stay and ancillary resource utilization. The majority of pathways did not significantly decrease either patient length-of-stay or resource utilization. CONCLUSION: Vast amounts of resources are currently expended on the development and implementation of pathways; to justify this expenditure, pathway efficacy should be demonstrated. Our evaluation reveals that though some pathways significantly reduced length-of-stay and resource utilization, the majority did not. Thus, efforts should be placed on critically assessing the effectiveness of pathways and understanding the reasons behind successful and unsuccessful ones. Catheter-related UTI leads to increased morbidity and healthcare costs. In addition, the majority of patients given indwelling catheters find the catheters uncomfortable and restrictive of their daily activities. Indwelling catheterization is inappropriate in about one-third of catheterized patients; however, the reasons behind inappropriate catheter use are unclear. We hypothesized that catheters are commonly used inappropriately because physicians are often unaware of their presence. We sought to assess: 1) how often physicians and medical students know if their own hospitalized patients have a catheter; and 2) whether catheter awareness depends on catheter appropriateness. METHODS: The physicians and students responsible for patients admitted to the medical wards at four U.S. hospitals were the study subjects. Each provider was given a list with the names of the patients on their service and for each patient was asked: "As of yesterday afternoon, did this patient have an indwelling urethral catheter?" Respondent's answers were compared to the actual patient situation. Chart review assessed catheter appropriateness. RESULTS: Of 469 patients, 117 (25%) had indwelling catheters. Catheterization was considered appropriate in 70% of these patients. Among all catheterized patients, 18% of medical students, 22% of interns, 28% of residents, and 35% of attending physicians were unaware that their patients were catheterized. Among inappropriately catheterized patients, 24% of medical students, 31% of interns, 41% of residents, and 51% of attending physicians were unaware that their patients were catheterized. After adjusting for patient age and gender, hospital, and level of respondent training, respondents were significantly more likely to be aware of catheters that were appropriate (OR ϭ 3.6; 95% CI, 1.6 to 7.8). CONCLUSION: Physicians are commonly unaware that their own patients are catheterized. Inappropriate catheters are more often "forgotten" than appropriate ones. Since the best method of limiting catheter-related UTI is to limit inappropriate urethral catheterization, system-wide interventions aimed at altering the process of catheterization are needed. We plan to initiate innovations limiting urethral catheterization, thereby improving the safety of hospitalized patients. METHODS: PCPs in an urban academic practice were randomized to an intervention or control group. During office visits, intervention PCPs received patient's alcohol screening results (CAGE and consumption), a readiness to change measure, and action recommendations based on problem severity; control physicians received no such information. Eligible patients visiting these PCPs were explicitly defined current hazardous or harmful drinkers. Patient demographics, comorbidity, alcohol consumption, and problems were assessed in standardized interviews using validated instruments. PCP counseling was assessed by patient report immediately after the visit, and was defined as specific advice (regarding safe drinking, to quit, or cut down) or referral (to alcohol treatment or a mutual help group). Power was 80% to detect a 50% increase in counseling rates between PCP groups. RESULTS: Of 41 PCPs, 20 were in the intervention group, 48% were residents, 55% male, and 65% white. PCPs were similar by randomized group, including demographics, training, and attitudes towards patients with alcohol problems. The 312 patients were 64% male, 56% black, 19% white, 16% Latino, 63% high school graduates, 40% unemployed. Mean age was 44. Median annual income was $7500. For most (66%), routine care was the reason for the visit, though 67% had 2 or more concerns to discuss. Patients drank a mean of 6 drinks per drinking day, and 36% were in Precontemplation, 32% Contemplation, and 32% the Action stage of readiness to change. A consistent trend favored intervention group PCPs: they were more likely to provide alcohol counseling (mean physician counseling rate 49% vs. 40%, P ϭ 0.32); to give patients any advice about drinking (56% vs. 44%, P ϭ 0.19); to have a discussion with the patient about drinking (68% vs. 57%, P ϭ 0.23); and to be the initiators of such a discussion when one occurred (64% vs. 52%, P ϭ 0.25). Intervention group PCPs were significantly more likely to give safe drinking limit advice (19% vs. 6%, P ϭ 0.04). Adjusted analyses considering the hierarchical nature of the data, and physician, patient, and visit characteristics, yielded similar results. CONCLUSION: A simple intervention, providing screening results and recommendations for action, was associated with modest effects: a trend towards a small increase in physician counseling for alcohol problems and a greater likelihood of safe drinking advice. To have greater impact on alcohol counseling by PCPs, more substantial interventions are needed. We studied a prospective cohort of subjects in a randomized trial of an intervention to engage adults with addiction in primary care, the Health Evaluation and Linkage to Primary care (HELP) study. Eligibility criteria were: admission to a detoxification unit; alcohol, heroin, or cocaine as drug(s) of choice; and no primary medical care. Exclusions were: inability to provide 3 contacts to facilitate follow-up; pregnancy; plans to leave the area; dementia; and an inability to speak English or Spanish. Subjects were evaluated by interview using standardized measures. Linkage, defined as seeing a primary care clinician at least once, was assessed at follow-up by interview. We included clinically and statistically important predisposing, enabling and illness variables in multivariable logistic regression models. RESULTS: Most subjects were male (76%); the mean age was 36; 46% were black, 37% white, 11% Hispanic; 38% were unemployed; 60% uninsured; 54% recently incarcerated; 47% homeless; and 47% had a chronic medical illness. Most (81%) reported health care utilization in the past 6 months, including addictions, mental health and episodic medical care. Most (87%) had polysubstance abuse problems; 86% had alcohol problems, 75% cocaine, 69% marijuana, and 38% heroin problems. Of 470 subjects, 348 (74%) completed at least 1 follow-up interview over a 2-year period; 144 ⁄ 348 (41%) did not link with primary medical care. In a multivariable model adjusting for age and randomization assignment, the following factors identified at enrollment were independently associated with a failure to link with primary care: male (Odds Ratio 2.6, 95% Confidence Interval 1.5-4.5), white (OR 1.7, CI 1.1-2.8), friends and family that do not support abstinence (OR 2.1, CI 1.3-3.5), no recent medical visits (OR 1.7, CI 1.0-2.9), and absence of chronic illness (OR 1.8, CI 1.1-2.8). Health insurance, employment, homelessness, past incarceration, addiction severity, psychiatric illness, and recent addictions and mental health utilization were not significant predictors. CONCLUSION: This cohort of young addicted patients without primary care had very high utilization of health care services but many failed to link with primary medical care. Characteristics and experiences identified in this study suggest that linkage with primary medical care will require aggressive strategies that make use of episodic health care encounters, and address patients' perceptions of need for primary care. Medicine has attempted to raise public awareness of the capabilities of internists through their "Doctors for Adults" campaign. We wished to determine whether patients presenting to general internal medicine clinics understand the capabilities of internists and whether they could discern important differences between internists and other primary care providers. We also wanted to see if patients felt most confident in internist treatment of specific diseases, symptoms, or preventive medicine skills. METHODS: Adults over age 17 enrolled in internal medicine continuity clinics in Georgia, Iowa, and Vermont were given a written survey to complete before seeing their physician. Patients were asked eleven questions using a 5 point Likert scale on perceived internist characterisitcs. The subjects also selected which of 24 diseases, symptoms, or preventive medicine skills they felt were within an internist's scope of care. Patient age, gender, and education level were also collected. RESULTS: Of 911 surveys distributed, 601 (66%) were completed. Very few (19%) respondents confused an internist with an intern. Nearly half of patients (45%) confused internists with a general or family practitioner and 39% thought internists could treat children. Only 50% of respondents thought internists were trained in women's health and only 38% felt they could perform a gynecologic exam. Patients had more confidence (p Ͻ 0.05) in an internist's ability to treat symptoms than their ability to treat specific chronic diseases or perform specific preventive medicine skills. College educated patients displayed significantly fewer (p Ͻ 0.05) misconceptions about internists in most categories than those with less education. CONCLUSION: Patients lack understanding of the capabilities of internal medicine physicians, especially on how they differ from other primary care specialties. Although patient confidence in internists' ability to treat symptoms is high, confidence in treatment of chronic diseases and preventive medicine skills could be improved. Continued effort is required in public education designed to promote better understanding of the role of the internist as a specialist in adult medicine. PURPOSE: There is increasing use of anticoagulation clinics (ACC) to care for patients on warfarin. This study assessed the manner of health care delivery in ACCs and the correlation of these characteristics with patient anticoagulation control. METHODS: A survey was mailed in 1999 to 581 anticoagulation clinics that were members of a professional ACC association. The survey asked about clinic characteristics including age and size, staffing and sources of referrals; clinical practices including initiation of therapy, frequency of INR testing, use of Point-of-Care monitoring and home monitoring; and patient outcomes. Two measurements of control were the proportion of INRs in range of those performed on a single designated day (cross sectional), as well as over a 6 month period. RESULTS: 233 centers (40%) responded, the majority (66%) had been in operation for less than 5 years and 34% monitored more than 400 patients. While 64% had a physician director, over half reported being staffed by nurses (63% full-time), and two-thirds by pharmacists (48% full-time). Half were affiliated with either teaching or community hospitals, and 28% were affiliated with group practices. General internists and cardiologists accounted for a two-thirds of the referrals, 44% of which were for atrial fibrillation, 15% for heart valve replacements and 11% for deep vein thrombosis. Among patients under 60 years old, 61% of ACCs initiate warfarin therapy with a 5mg dose, 34% with larger doses, and 6% with smaller. Among patients over 60 years old, 40% of clinics start with 5mg, 20% with larger doses and 40% with smaller. Half of the clinics were utilizing Point-of-Care testing in the clinic, and 18% had patients using instruments at home, but only 3% of patients were managing their own therapy. The 2 measures of control were highly correlated; 62% in range over a 6-month period and 68% in range on a single day (p Ͻ 0.001). There was no relationship in the percent of INRs in range and the frequency of ordering INRs or the size of the ACC. The major bleeding complications in a 1-year period were gastrointestinal bleeds (0.6%), genitourinary bleeds (0.3%), soft tissue bleeds (0.2%) and cerebral bleeds (0.1%). There were 1.5% thrombotic events in the same period. Clinics affiliated with the Veteran's Administration (VA) had more INRs in range, over a 6-month period, compared to non-VA affiliated centers (74% vs. 62%, p Ͻ 0.02). CONCLUSION: ACCs monitor warfarin therapy for patients often cared for by primary care providers. These clinics are large and staffed predominantly by nurses and pharmacists. Control rates are high and complication rates are low. The use of these clinics needs to be considered by primary care providers. PURPOSE: To develop and assess the effectiveness of a novel multidisciplinary medical clinic for linking patients from a residential detoxification program to primary care. METHODS: We enrolled 470 subjects undergoing inpatient detoxification from alcohol, heroin or cocaine in a randomized controlled trial, the Health Evaluation and Linkage to Primary Care (HELP) Project. The intervention consisted of a medical evaluation at the detoxification unit by a nurse, social worker and physician trained in motivational interviewing and a specific referral to primary care. Prior to randomization all subjects were interviewed by a research associate concerning demographics, substance abuse, medical problems, social support, utilization, and HIV risk behaviors. The primary outcome of interest was a visit with a primary care clinician after discharge from the detoxification unit. RESULTS: Of the 470 subjects enrolled, 235 were randomized to the intervention. Baseline characteristics included the following: 76% male; 46% black, 37% white, 15% Hispanic; mean age 36 years. Alcohol was identified as the primary or secondary drug of choice by 298 ⁄ 470 (63%). Heroin or cocaine were identified as primary or secondary drug of choice by 354 ⁄ 470 (75%). Comparison of control and intervention subjects revealed no differences in terms of demographic, substance use and medical severity. Among all subjects with follow-up within two years of enrollment (348/470, 74%), linkage was achieved in 65% of the intervention group and 52% of the control group (p ϭ 0.01). Results among those with follow-up at 6 months, found 61% and 36% linkage in the intervention and control groups respectively (p Ͻ 0.001). The findings of increased linkage were consistent in subgroups of subjects who identified alcohol, cocaine or heroin as their drug of choice. CONCLUSION: Linkage of addicted patients to primary medical care occurs among patients in a detoxification unit and can be enhanced by a multidisciplinary medical clinic. Effective linkage of substance-dependent patients to primary care is the first step in engaging medical providers to address these individuals' health and addiction issues. respectively stated medication lapse was a major factor in their decompensation. 24.% of patients also identified physical stresses and 49.% identified emotional/ social stresses as factors. Twice as many pts increased water intake in response to worsening as self-increased diuretic (22 vs. 12)! CONCLUSION: Significant opportunities exist to prevent admission-requiring exacerbation of CHF. Most patients demonstrated a pattern of worsening over several weeks with a paucity of self directed effective measures taken to reverse the worsening. Medication lapses, physical and social stresses, care access barriers, and missed appointment were identified at major potentially ameliorable factors. Distinct patterns of worsening were identified that we are now applying to a preventtion/early treatment strategy, as part of National IHI Collaborative.

General Hospital, Boston, MA PURPOSE: To determine the extent to which patients with coronary artery disease (CAD) receive adequate secondary prevention at outpatient practices associated with a large teaching hospital. METHODS: As part of a Primary Care Operations Improvement (PCOI) project at Massachusetts General Hospital (MGH), we identified 3,933 patients with a diagnosis of CAD (ICD-9 codes 410.0-414.99) who had been seen in one of ten MGH outpatient practices between 10/1/96 to 9/30/97, were younger than 85, living at home, and were free of metastatic cancer. Outcomes of interest for this cohort included aspirin use, beta-blocker use, blood pressure control, LDL testing, and LDL control. Information on lipid testing and results were collected using electronic laboratory data. For the 1,799 patients in Internal Medicine Associates (IMA), we abstracted information from the CoSTAR electronic database on aspirin and beta-blocker use, related medications, medical problems and allergies. On a randomly selected subset of 163 patients, we performed a manual chart review to determine average blood pressures during the two-year study period. RESULTS: Of the 3,933 patients included in the study, 808 (21%) had not received an LDL test during the two-year study period. Among the patients who had been tested, 2016 (65%) had a last LDL level above 100 mg/dL, and 854 (27%) had an LDL above 130 mg/dL. Among a sample of patients without an LDL test or an LDL Ͼ130 mg/dL, 52% were not on a HMG-CoA reductase inhibitor ("statin"), and only 2% were on a maximum dose of a statin. The vast majority of patients not on a statin were nevertheless on aspirin or a beta-blocker. Among 1630 IMA patients with at least 3 recorded medications, 444 (27%) were not taking aspirin and 479 (29%) were not on a beta-blocker. Among the 129 patients with CAD from the manual chart review, 40 (31%) had a recorded systolic blood pressure of 140 mm Hg or greater. Age over 75 had no effect on LDL testing, but did predict decreased statin use. Beta-blocker use was also less common in the elderly, but was not affected by the presence of relative contra-indications to their use. CONCLUSION: Our findings suggest that the biggest deficiency in the secondary prevention of CAD among outpatients at MGH is in the management of hyperlipidemia. Both the lack of lipid testing and the limited aggressiveness of controlling LDL cholesterol to nationally recommended levels may contribute to sub-optimal patient outcomes. On the other hand, aspirin use, beta-blocker use and blood pressure control, while not optimal, seem to be significantly better than that shown from previous national and international studies. Efforts aimed at improving the quality of care in patients with CAD, such as the PCOI project at MGH, should consider lipid management a vitally important area for potential improvement. PURPOSE: Increased copayments for prescription medication with higher prices are frequently required by health plans without conclusive data about unintended effects. We analyzed the impact of differential cost-sharing for ACE inhibitor on drug utilization, savings and potential substitutions by other medications. METHODS: We analyzed 36 months of claims data in British Columbia, which implemented reference pricing in January, 1997. The study patients (119,074) were all non-institutionalized Pharmacare beneficiaries 65 years of age or older who used ACE inhibitors. Time series analyses were conducted to determine the effect of the policy on drug utilization and expenditures. Logistic regression was used to identify predictors for switching or stopping medications. RESULTS: We observed a steep 29% decline in use of cost-sharing ACE inhibitors (ACEI) immedately following initiation of the policy (p Ͻ 0.0001). A longitudinal analysis of medication switching dynamics showed that 80% of all policy related switching to no-cost medications took place within the first three months. During a transition period of three months following the policy implementation overall utilization of ACEI was reduced by 18% compared to the predicted pre-policy trend. Afterwards the overall ACEI utilization rate was still 11% lower than the extrapolated pre-policy trend but increasing at the same rate (p ϭ 0.14). Utilization rates of other anti-hypertensives were unchanged, suggesting no net substitution by alternative medications. The policy saved Can $7.2 million in pharmaceutical expenditures during 12 months. Patients with low income status, heart failure or diabetes were among the most likely to stop all anti-hypertensive therapy (n ϭ 1463) compared to those switching to the reference drugs (OR ϭ 1.6, 1.3, 1.4, respectively). CONCLUSION: The implementation of reference pricing in British Columbia is effective in terms of a sustained reduction in drug expenditures. However, differential cost-sharing affects anti-hypertensive drug therapy disproportionately, particularly in those with low income status, who appear to stop therapy in some cases. However, these impacts raise questions about ways of implementing such policies that avoid inequities in access to essential medications. Further research is needed on the effects of such policies on health outcomes and total expenditures. PURPOSE: The use of autologous blood transfusion, which lessens the risk of blood-borne infectious diseases and immunological reactions, has increased over the last decade. Little is known about the patients who receive autologous blood compared to those who receive packed red blood cells from anonymous donors. METHODS: We conducted a cross-sectional study of patients with hospital admissions in 1996. We used the Nationwide Inpatient Sample, which is a stratified probability sample of hospitals from 19 states that includes information on 6.5 million admissions. From among the patients who received autologous blood transfusion (ICD-9 procedure code 9902), we identified the five most frequent diagnoses. Within each diagnostic category, we compared patients who received autologous transfusion to patients who received packed red cells but not autologous transfusion. Chi-square tests and multivariate logistic regression were used for analysis. RESULTS: The five most frequent diagnoses for which patients received autologous blood were osteoarthritis, prostate cancer, complications of a device or implant, spondylosis, and bone disease. For each diagnosis except device complications, age was inversely associated with the use of autologous blood (p Ͻ .0001).

Only for osteoarthritis were women less likely to use autologous blood than were men, when controlled for age and race, with an OR of 0. PURPOSE: Immunization trials in reasonably healthy institutionalized elderly (Ͼ50 years) have documented reductions in morbidity and mortality rates due to pneumonia. There is no study to document the current profile of standard policy and procedures in place or to evaluate the impact of educational intervention on the immunization rate in long term care facilities (LTCFs). This study was designed to evaluate (i) the current pattern and variables of immunization rates; (ii) the effect of educational intervention on immunization rates in LTCFs. METHODS: Surveys were designed with the focus on variables including standard policy and procedure, standing order for immunization in place, standard policy for healthcare workers and the medical records of the annual assessment of all the residents. The survey was sent to 415 LTCFs in the State of Kansas. Written material informing about immunization was then sent to the directors of the nursing (DON) of all 415 LTCFs and post intervention survey was requested from all LTCFs. We also randomly selected 55 of the LTCFs for review of their medical records to assess flu and pneumonia IR. The study was conducted over a period of 14 months. RESULTS: Only 105 out of 415 LTCF responded to the survey. For influenza, 87% and for pneumonia, 34% of the responding LTCFs had a Policy/Procedure in place. Of these LTCFs, standing orders for immunization for influenza and pneumonia were present in 79% and 74% respectively. Policies for influenza immunization of healthcare workers was present in 78.1% of LTCFs. We received matching immunization rates and DON survey data from 45 LTCFs. Post-intervention immunization rates were available from only 26 LTCF. Pre and post intervention immunization rates for influenza were 74% and 79% respectively. However pre and post intervention immunization rates for pneumococcal vaccination were 10% and 75% respectively. Educational intervention had minimal effect on influenza immunization rates but a significant increase in pneumococcal immunization rates (p Ͻ 0.001). CONCLUSION: We conclude that a tremendous variability in the policy and procedure profile for influenza and pneumococcal vaccination exists and educational intervention can improve the rate of immunization for pneumonia in the LTC facilities in the State of Kansas. Limitations: There was poor response rate as only 25.9% DON filled out the survey. Only 47.27% of randomly selected LTCFs responded to post-intervention survey. We could send video tapes to 75 of the 405 LTCFs due to limited resources.

To keep the uniformity of intervention (written material only), we used data from only 45 of the 55 randomly selected LTCFs. CONCLUSION: In our study, no statistical significance was seen in avoidable admission rates by insurance status. These so-called "soft admissions" for ambulatory sensitive conditions may be more thoroughly scrutinized, especially for the uninsured patients due to their potential non-payor status. We found no evidence to support this increased scrutiny. Of note, a higher avoidable admission rate was evident in the uninsured indigent group, proving that payment mechanism was not an important factor in the decision making process of a potentially "soft" admission. To determine how health plan practices contribute to physicians' overall ratings of health plan quality. METHODS: The Physician's Evaluation of Health Plans (PEHP) Project surveyed by telephone a probability sample of 1757 generalist physicians in 16 health plans in 5 metropolitan areas nationwide. Physicians reported only on the plan from which they were sampled. 70% responded. Three multi-item scales (internal reliabilities: 0.75 to 0.88) assessed physicians' perceptions of facilitators and barriers to high quality care in the plans and the clinical capabilities of plan physicians. Facilitator items included the presence of guidelines, patient education and flagging patients for preventive care. Barrier items included authorization procedures, cost and time constraints, and access to specialists. A single item ("I would recommend this plan to a friend or family member") rated overall health plan quality. All items used fivepoint scales ("excellent" to "poor," or "strongly agree" to "strongly disagree"). Our analytic goal was to explain generalist physicians' overall rating of health plan quality using the three multi-item scales and their demographics (age and gender). We compared the explanatory power of each variable using R-squared. RESULTS: Among univariate models, the facilitators scale explained the most variation in physicians' overall ratings of health plan quality (R-squared ϭ 0.44). In the multivariate model, the listed variables explained 55% of the variation in generalist physicians' ratings of overall health plan quality. The explanatory power of the multivariate model decreased the most when the facilitators scale was removed (R-squared loss ϭ 0.09), followed by the barriers and clinical capabilities scales.

CONCLUSION: Generalist physician ratings of health plan quality are more highly related to what health plans do to facilitate care than the barriers created by plans in managing care. PURPOSE: Excessive reliance on broad-spectrum antibiotics is costly and can exacerbate anti-microbial resistance, yet over-use of such antibiotics is common. The goal of this research was to test a targeted one-on-one educational program ("academic detailing") designed to improve the appropriateness of broad-spectrum antibiotic use. METHODS: Using block randomization, we assigned 17 general medical, oncology, and cardiology teams in a large urban teaching hospital to intervention or control status. During an 18-week study period from January 1999 through May 1999, an infectious diseases consultant or clinical pharmacist discussed rational antibiotic use with house officers who prescribed either levofloxacin or ceftazidime to patients on intervention teams for possibly inappropriate indications. Alternative antibiotic choices were explained. The outcome of interest was the number of days of levofloxacin or ceftazidime administered on each team during the study period, adjusted for baseline prescribing. RESULTS: Prior to the trial, intervention and control teams had similar prescribing patterns for the target antibiotics, and used the drugs for similar indications during the study period (all p-values Ͼ 0.1) During the intervention, days of levofloxacin and ceftazidime prescribed for patients on the intervention teams were 58.8% lower than for patients on the control teams (adjusted p-value ϭ 0.0003). Length of stay, intensive care unit transfers, re-admission rates, and in-hospital deaths were similar in both groups (all p-values Ͼ 0.1). Hospital-wide use of these agents was stable during the study period. CONCLUSION: A targeted one-on-one educational intervention can be a practical, effective, and safe method for reducing excessive broad-spectrum antibiotic use.

Wider use of such programs may help slow the growth of nosocomial anti-microbial resistance. Wisconsin residents access to all FDA approved anti-HIV medication. The WADRP does not impose expenditure caps or have waiting lists, all physicians can participate and patients can receive total reimbursement for medications. The purpose of this study was to identify patient and physician characteristics associated with substandard antiretroviral prescriptions provided to a state-wide cohort of HIV-infected patients. METHODS: WADRP patient data (birth date, gender, risk factor for HIV, ethnicity, income, antiretroviral medications) were linked to physician prescriber data obtained from the AMA Master File. Antiretroviral regimens were classified according to the DHHS guidelines as "optimal" (HAART) or "suboptimal". Chi-square analyses were conducted on the classification of prescribed anti-HIV medication for each patient variable and the likelihood of prescribing "optimal" (HAART) regimens by physician characteristics. RESULTS: Of 369 new WADRP enrollees from 1997 to mid-1999, 116 were prescribed "suboptimal" and 200 were prescribed "optimal" therapy. Patients who were not White were less likely to receive any antiretroviral treatment and less likely to receive "optimal" regimens, 2 (2) ϭ 8.40, p Ͻ .05. Patients with incomes below the federal poverty level were more likely to receive "suboptimal" regimens, 2 ϭ 6.20, p ϭ .05. Ninety six physicians wrote antiretroviral prescriptions for the WADRP patient cohort. Chi-square analyses indicated that physicians who had more experience (6 or more WADRP patients) were more likely to prescribe "optimal" regimens, 2 (3) ϭ 16.77, p Ͻ .01. Physicians who practiced in sparsely populated areas (Ͻ250,000) were less likely to prescribe "suboptimal" regimens ( 2 (1) ϭ 5.73, p Ͻ .05), and physicians whose practices were office-based were less likely to prescribe "suboptimal" regimens, 2 (1) ϭ 4.55, p Ͻ .01. Physician age, ethnicity and specialty were not significant predictors. CONCLUSION: Disparities exist for HIV-infected patients enrolled in a state-wide program which provides antiretroviral medications to all enrollees. Individuals who were poor and not White were more likely to receive either "no" or "suboptimal" antiretroviral therapy. Physicians with less prescribing experience, practicing in more urban and non-office-based professional settings were more likely to prescribe "suboptimal" regimens. Additional methods to limit these disparities are needed. PURPOSE: National guideline implementation programs have substantially increased use of thrombolytic therapy for elderly patients with acute myocardial infarction (AMI). However, evidence of benefit to support use of thrombolytic therapy in the old-old (age 75 and older) is lacking. We studied a large community population of elderly AMI patients in order to answer the following questions: (1) What proportion of eligible and ineligible elderly AMI patients receive thrombolytic therapy? (2) Controlling for baseline risk status and demographics, what is the association between thrombolytic use and in-hospital mortality by age, and among patients with or without medical contraindications to thrombolytic treatment? METHODS: We abstracted the medical records of 2,659 elderly patients admitted with AMI at 37 Minnesota community hospitals for the periods 1992-93 and 1995-96. We measured in-hospital mortality, thrombolytic use among eligible and ineligible patients, and demographic, comorbidity and severity-of-illness variables at presentation. We conducted multiple logistic regression analyses to determine whether age and presence of medical contraindications modified the effect of thrombolytics on mortality, controlling for and stratifying by propensity scores (based on all demographic, comorbidity and severity-of-illness variables predicting thrombolytic use). RESULTS: Sixty-two percent of eligible patients (n ϭ 684) and 16% of ineligible patients (n ϭ 1972) received thrombolytic therapy. Thus, 42% of thrombolytic recipients were ineligible for treatment, mostly due to medical contraindications. Among the old-old with absolute medical contraindications (n ϭ 371), 25% of thrombolytic recipients died as compared with 12% of non-recipients (p Ͻ 0.001).

In the total sample, there was no significant association between thrombolytic use and in-hospital mortality (O.R. ϭ 0.90, p ϭ 0.70), controlling for propensity to receive thrombolytics. However, for every one year increase in age, the adjusted odds of dying for thrombolytic recipients vs. non-recipients increased significantly by 1.04 (p ϭ 0.03). Among patients with absolute medical contraindications (n ϭ 457), the adjusted odds of death among thrombolytic recipients was 2.5 as compared with non-thrombolyzed patients (p ϭ 0.008) CONCLUSION: Overall, there was a significant association between the age of elderly AMI patients and in-hospital mortality following thrombolytic therapy. Most of this association was explained by use of thrombolytic therapy in older patients with absolute medical contraindications. Our findings suggest the need for more care in the selection of elderly patients for thrombolytic therapy.

ORGANIZATION OF CLINICAL PRACTICE. M Spar, D Carlisle, C Damberg, C Mangione, A Brown, KL Kahn, University of California, Los Angeles; RAND, Santa Monica, CA PURPOSE: We used systematic literature review, clinical judgment, and administrative experience in the clinical and management arena to develop a conceptual framework to understand how the clinical structure of organizations influence patient care and outcomes. Donabedian's traditional model of health care quality describes structure, process, and outcomes as essential elements. The last decade has resulted in the centralization of many aspects of clinical care so that medical organizations are implementing plans to use resources to systematically effect the care of cohorts of patients. Whereas the process of medical care pertains to a oneon-one interaction between patients and providers, the structure of clinical care pertains to ways in which resources are devoted to the care of cohorts of patients. We propose that a conceptual model that articulates ways in which medical organizations can influence the care of patients will allow us to better predict patient care and outcomes. METHODS: In the development of this conceptual frame-work, five domains emerged as unique types of structure that might influence patient care and outcomes: patient support, provider support, continuity of care, locus of control, and provider-patient match. These five domains are noted to effect specific components of the clinical structure of care including structure aimed at optimizing: a) patient care (including patient education, use of physician extenders, training of clinical staff); b) the use of evidence based medicine (including use of patient and provider reminders, use of clinical practice guidelines); c) the patient-provider encounter (including use of multidisciplinary teams, medical record availability); d) data exchange (including access to lab data and measures of communication); and e) involvement of the central medical organization (including the extent of their support for assessing and giving provider feedback). We used the conceptual model to generate a focused survey, the "Clinical Structural Survey" or CSS of the corporate clinical leader of medical organizations. Challenges in the devel-opment of the survey included: a) type of organization to survey regarding structure; b) type of respondent to interview c) definition of "policy" pertinent to structure, d) dis-crepancies between reported "policies" and actual practice and e) changes in corporate policy with time. RESULTS: Our survey has been conducted with the leaders of 56 medical organizations. They were able to complete the survey (average time 45 minutes), and repeatedly noted that the survey captured important aspects of the structure of care that their organization considered or implemented. The survey worked equally well in the interview of directors from medical groups and Independent Practice Associations (IPAs), but IPAs had more "not applicable" responses consistent with our conceptual model that medical groups had more centralized policy pertinent to structure than did IPAs, on average. CONCLUSION: Our clinical structure model helps to frame how medical organizations allocate resources to optimize the care of cohorts of patients. We have articulated five key domains that can be analyzed to understand which domains of structure actually do impact patient care and outcomes in the real world of managed care. regularly adjusting insulin, 44% following a meal plan qd, 5% checking blood sugar qd, 62% checking their feet. 47% were non smokers. 26 patients (28%) reported better SC. Using logistic regression in SUDAN 7.5.2, patients were more likely to report better SC, at p Ͻ .05, if they had Ͼ7 visits in the past 3 months (OR 92.6) or were older (OR 1.2). Patients were more likely to report worse SC if they had a DMQOL of 5-13 or 14-17 (OR .13, .15) or were male (OR .2). CONCLUSION: Diabetes is inherently a disease of self management, in which complications can be prevented by judicious practice of self-care behaviors. However, these behaviors involve lifestyle changes that are difficult to practice consistently, even in the most motivated patient group. Our data suggest that DMI patients in low income populations who report a lower quality of life, who are male, or who are younger may need special attention to enhance self care behaviors. Patients with more encounters with their provider report markedly better SC. These data provide some evidence to suggest areas in which patient education and reinforcement of self-care behaviors can have significant impact on future morbidity from DMI. These pts are less likely to have referrals/imaging studies, and more likely to have medications ordered. Therefore, our SC index provides an interesting look into the provision of OHP non-reimbursed services, with potential over-utilization of resources in select pt groups with LBP. This data suggests that paradigms for SC should include tracers for both potential over-and under-utilization of medical services. The outcome measures were compared with ANOVA and paired t-tests. RESULTS: Wait time for a preceptor ranged from less than 1 minute for Preceptor:Resident (P:R) ratios of 1:1 to 43.5 (ϩ/Ϫ19.1) min for a ratio of 1:8. Time spent precepting ranged from 10.9 (ϩ/Ϫ6.2) min across all permutations once precepting was initiated. The total number of patients processed during regular working hours increases on average at ‫1.3ف‬ patients/additional resident with one precep-tor, at 5.5 with 2 preceptors, at 5.5 with 3 preceptors and at 4.5 with 4 preceptors. The baseline number of patients processed rises with the number of preceptors.

CONCLUSIONS: As expected in a model that allows queuing for resources, waiting time and delays rise rapidly when processing time becomes greater than the arrival rate. When the P:R ratio falls below 1:4 there is a resulting rapid rise in time spent waiting for precepting and a decreased proportion of total registered patients seen within regular working hours. Scenarios with 2 or 3 preceptors benefit most from additional residents. Therefore, the optimal Preceptor:Resident ratio should not fall below 1 to 4. CONCLUSION: Preliminary analyses indicated that brief MET lowers needle sharing, but not significantly more than an assessment only. Subgroup analysis on the full sample may reveal IDUs who benefit most from MET. , an affiliate of Baxter Healthcare, has implemented a nationwide renal DM program and has accumulated over 15,000 pt months experience over the past 2 years. RESULTS: Complications of vascular access (e.g., thrombosis, poor blood flow, infection) are the most common causes for hospital admissions, followed in order by cardiovascular (CV) diseases, congestive heart failure (CHF), sepsis (usually related to the vascular access), and pneumonia. Introduction of targeted RMS DM programs utilizing the expertise of Health Service Coordinators (HSCs) to collect data and implement the CQI process, has resulted in a progressive decrease in hospitalization rates:

The RMS vascular access initiative used an educational program for local nephrologists and dialysis units based on NKF-DOQI guidelines; CV admissions decreased in conjunction with a comprehensive anemia management program; CHF admissions fell coincident with a comprehensive educational program focused on achieving true dry weight for dialysis pts. CONCLUSION: DM for ESRD improves the quality of care by decreasing hospital admissions for the 5 most common causes for hospitalization using CQI processes. PURPOSE: Academic detailing is an effective method to change physicians' prescribing practices in the ambulatory setting. However, few controlled trials have studied methods to change prescribing in the hospital setting. The objective of this three armed study was to compare the effects of academic detailing versus a mandatory drug order form versus control on prescribing of histamine 2 receptor antagonists (H2) in the hospital setting. METHODS: A RCT was implemented at an academic medical center between September 1993 and April 1994. Subjects were resident physicians in all departments who wrote prescriptions. Clinical services were randomly assigned to one of three arms: 1) face to face education by pharmacist detailers, 2) a required drug order form or 3) a control intervention of printed material about H2. An expert panel devised appropriateness criteria for the inpatient use of H2. The primary outcome measure was the mean proportion of appropriate prescriptions per physician before and after the one month intervention.A secondary outcome measure was the change in the total number of prescriptions. RESULTS: Of 62 physicians who completed the trial, 15 did not prescribe H2 in either the pre or post intervention period leaving 47 physicians with complete data. Six of 17 physicians in the academic detailing group did not complete the intervention due to time constraints, despite an average of 26 minutes spent by the detailers trying to contact and accomodate the physicians. The mean proportion of appropriate prescriptions per physician did not change in the academic detailing (p ϭ 0.06) or the printed material control group (p ϭ 0.39) and decreased after the intervention in the order form group (p ϭ 0.003). There were no differences between groups in either the mean proportion of appropriate prescriptions or total number of prescriptions per physician after adjusting for baseline differences. CONCLUSION: This RCT comparing academic detailing, a required drug order form, and a control intervention of printed material was unable to demonstrate an improvement in prescribing of H2 by resident physicians in the hospital setting.

The unique difficulties of performing this kind of study in the hospital rather than ambulatory setting are highlighted. These difficulties include the challenge of identifying heavy prescribers, the limited time for follow-up due to one month clinical rotations, and the lack of receptivity of resident physicians to participate. Well designed trials are needed to determine the most effective means of improving physicians' prescribing in the hospital setting. The Institute of Medicine recently concluded that preventable errors were the 8th most common cause of death in the United States and that promoting effective team functioning, possibly using an aviation model, is one way to prevent errors. We surveyed operating room and intensive care unit (ICU) personnel about attitudes concerning error, stress, and teamwork, and compared these attitudes to those of airline cockpit crew members. METHODS: Cross-sectional surveys were administered to medical personnel (physicians, fellows, residents) from urban teaching and non-teaching hospitals in the United States, Israel, Germany, Switzerland, and Italy; and cockpit crew members (captains, first officers, second officers) from several major United States airlines. Common items from four different surveys administered between 1992 and 1999 are reported here. RESULTS: Surveys from 1,033 medical personnel and over 30,000 cockpit crew members were analysed. Typical response rates from cockpit crews were 45%; for operating room personnel (n ϭ 851) 40% to 100%; and for intensive care unit (ICU)personnel (n ϭ 182) the response rate was 59%. 70% of surgical attendings agreed with the statement "even when fatigued I perform effectively during critical times" compared to 46.6% of anesthesia attendings and 22% of airline captains, p Ͻ .001. 55% of surgical attendings disagreed with the statement "Junior team members should not question decisions by ssenior team members" compared to 92% of ICU attendings and 92% of captains p Ͻ .001). In addition, surgeons perceive teamwork and communication in the OR team to be of a higher quality than the rest of the team. Finally, although 94% of ICU personnel believed all errors were important, only one-third reported that medical errors were handled appropriately in their ICU and over 60% agreed that errors were difficult to discuss because of each of the following factors: threat of malpractice suits, threat to job security,personal reputation, expectations of the patients' family/society, and possible disciplinary actions by licensing boards. CONCLUSION: Medical personnel reported that error is important, but difficult to discuss and not handled well in their hospital. These barriers are even more significant given that there appears to be a culture in medicine that denies the impact of stress on performance. Further barriers include differing perceptions of teamwork among various team members, and for operating room teams, support of hierarchies that do not accept input by junior team members. Current asthma guidelines recommend regular assessment of peak expiratory flow rates, symptoms, and quality of life (QoL). We assessed the clinical utility of these recommendations among patients who filled prescriptions for inhaled or oral breathing medications at 36 CVS pharmacies. Patients were eligible if they admitted having "a breathing problem," were taking at least one inhaled or oral asthma drug, and were receiving у70% of their asthma drugs at a single study pharmacy. Patients with asthma (rather than COPD as defined by American Thoracic Society criteria that use age and smoking history) completed the McMaster Asthma Quality of Life (AQoL) Questionnaire. Peak flow was measured at enrollment and during monthly follow-up calls; patients were also asked about recent exacerbations (ER or hospital visits for asthma or lower respiratory infections). We then assessed (1) the correlations between baseline AQoL scores and peak flows and (2) the ability of these baseline measures and monthly follow-up peak flows to predict exacerbations. We enrolled 954 (71%) of 1352 eligible patients. This report is limited to the 562 patients (59%) with asthma. Their mean age was 53 years, and 75% were women. Their mean baseline peak flow rate was 351 L/min (71% of predicted maximum). Mean AQoL scores on a scale from 1 (worst) to 7 (best) were: 4.5 for activity, 4.4 for symptoms, 4.3 for emotion, 4.3 for environment, and 4.4 for overall QoL. Baseline AQoL scores were correlated with baseline peak flows: highest for the scales assessing symptoms (r ϭ 0.24, p Ͻ 0.0001), activity (r ϭ 0.19, p Ͻ 0.0001), and overall QoL (r ϭ 0.21, p Ͻ 0.001), but lower for emotion (r ϭ 0.15, p ϭ 0.0004), and environment (r ϭ 0.11, p ϭ 0.009). Over a mean of 6 months of follow-up, there were 52 asthma exacerbations. Univariate Cox regression showed that having an exacerbation was significantly predicted by each baseline AQoL scale (hazzard ratios [HR] ϭ 0.62 to 0.76, p ϭ 0.0002 to 0.005) but marginally by baseline peak flow (p ϭ 0.05). By multivariable Cox regression, exacerbations were predicted by black race (HR ϭ 0.38, p ϭ 0.001) and overall AQoL (HR ϭ 0.69, p ϭ 0.004), but not baseline peak flow (p ϭ 0.36). Patients had a mean of 5 monthly follow-up peak flows (range 1-13). Controlling for baseline peak flow and AQoL scores, follow-up peak flows did not predict asthma exacerbations within the following 30 days (p ϭ 0.12). Baseline peak flow was correlated with baseline asthma symptoms and QoL measures, yet only overall QoL predicted subsequent asthma exacerbations. Baseline peak flow did not, nor did monthly assessments of peak flow. These results do support the NIH guideline's recommendations for the routine assessment of QoL in patients with asthma, but not the routine assessment of peak flow. Eligibility criteria included randomized controlled studies that looked at populations requiring pap smears for screening purposes. The intervention studied was in the form of a re-minder letter only and there were no restrictions as to country of origin. The studies retained were scored for quality by an adaptation of the method reported by Chalmers. The data were extracted from eligible articles in order to compare the proportion of women screened between the intervention and the control groups during the study period. Two authors independently extracted the data. The Mantel-Haenszel method was used to measure the summary effect size of the intervention. A test for homogeneity using the Mantel-Haenszel method was also performed to test the hypothesis that all of the studies were measuring the same effect (null hypothesis). RESULTS: Of the initial 421 articles found on literature search, 23 articles that dealt with the primary question were retained. Four additional articles were obtained by bibliography review. Fourteen of the studies were randomized controlled trials. The inquiry for unpublished studies yielded two studies. Nine articles were left after excluding those articles that did not fit all of the inclusion criteria, including one unpublished study. Most of the exclusions were due to using combined interventions such as letter and physician reminders. Individual odds ratios and confidence intervals were calculated. The test for homogeneity showed a nonsignificant chi-square value of 5.78 with 8 degrees of freedom. The hypothesis of homogeneity was accepted. The pooled odds ratio showed that patients who received the intervention were significantly more likely to have had cervical cancer screening than those who did not (OR 1.86, ). CONCLUSION: The use of patient reminders in the form of mailed letters is an effective method of increasing the rate of cervical cancer screening. with these preselected DRGs to urban, nonfederal, nonpediatric acute care hospitals were examined. This resulted in 169,168 patients in 422 hospitals in 13 states. The level of housestaff experience was measured as the number of months from July in which the patient was admitted. Multiple regression analyses were used to evaluate the different effects of experience on patient care in teaching hospitals compared with nonteaching hospitals, controlling for patient, hospital, and market characteristics. Our variable of interest was the interaction term, "Teaching/Nonteaching" ϫ "Experience", and when significant, indicates that the passage of one month since July is associated with a different change in outcome (i.e., total adjusted charges or LOS) in teaching hospitals than nonteaching hospitals. RESULTS: The "July Phenomenon" was observed for total charges with the subset of patients with medical diagnoses-i.e., over the course of the academic year, these patients at teaching hospitals experienced a significant decrease in total costs relative to similar patients at nonteaching hospitals (b ϭ Ϫ2.20E-3, p ϭ .003). There were no meaningful affects of housestaff experience on total charges for the surgical and Ob/Gyn diagnoses, and length of stay for all three groups of diagnoses. CONCLUSION: Housestaff training is significantly related to resource utilization with patients admitted with medical diagnoses. The training of inexperienced physicians may represent an important source of fiscal inefficiency for teaching hospitals in a competitive market.

PA Ubel, C Jepson, DA Asch.

PURPOSE: Physicians are increasingly being faced with situations in which one screening option is both more effective and more expensive than another. Economists say that the proper way to make such a choice is to look at the incremental cost (IC) of the more costly option over the less costly one. IC is essentially the extra cost of the given option divided by the extra return-that is, the cost of getting each unit of return that would not be achieved by the less costly option. IC is often counterintuitively high. For example, the IC of annual Pap smears, compared to Pap smears every three years, is nearly $1 million per year of life saved. The counterintuitive expense of some screening tests raises questions about whether physicians' screening recommendations would be influenced by the type and amount of cost information they receive. METHODS: We created a set of three scenarios, each asking the respondent to make a screening recommendation for a hypothetical patient; the scenarios involved decisions about breast, cervical, and colorectal cancer screening respectively. Three versions of each scenario were created, containing either no cost-effectiveness information, average cost-effectiveness information (e.g., the average cost-effectiveness of yearly Pap smears is $39,000 per life year), or incremental cost-effectiveness information (incremental cost-effectiveness of yearly Pap smears is $1 million). In addition, for each scenario, a parallel scenario was created in which the cost information was the same, but the nature of the screening was unfamiliar (gastric cancer, Pap tests using a new procedure to increase yield, and an unlabeled cancer and tests). Average and incremental cost versions of the unfamiliar scenarios were created. We randomized 1,500 U.S. primary care physicians to receive one of the resulting 15 scenarios. RESULTS: A total of 950 physicians responded (62%). Overall, subjects were less likely to recommend the most expensive screening test in unfamiliar scenarios than in familiar ones (p Ͻ 0.001). Their willingness to recommend expensive tests in unfamiliar scenarios was significantly reduced by receiving incremental cost information (p Ͻ 0.001). However, the type of cost information provided had no effect on recommendations in familiar scenarios (p Ͼ 0.2). CONCLUSION: Subjects' screening recommendations are influenced by cost information in unfamiliar screening scenarios but not in familiar ones. .001 ). Using the HARS as criterion, this makes claims data sensitivity ϭ 68% and specificity ϭ 99% for HIV surveillance. The ability to identify HIV cases was limited in regards to time, however, with 5.4% of cases being identified within 30 days before or after the HARS date, 16.8% within 6 months in either direction (one year interval), and 25% within 1 year before or after HARS identification (2 year interval). Of the 788 cases identified by claims and not by HARS, a limited chart review confirmed that 187 of these cases were true positives. CONCLUSION: Claims casefinding methods appear to be useful for HIV surveillance during the period 1992-7 in the Tennessee Medicaid population. However, the ability to assign time of disease onset is less accurate, making this method more appropriate for period prevalence estimates over longer time spans than for incidence or period prevalences over short time spans. For this study, it appears that longer exposure to the claims collection system increases the likelihood of identification. Research that depends on assessment of time at-risk should not rely solely on administrative casefinding algorithms, but should attempt to validate data from other outside sources such as chart or laboratory data. Claims data may be a useful adjunct to other surveillance systems. Much is known about low B12 levels, but elevated levels has not been systematically investigated even though they are common. So we prospectively compared the clinical associations of elevated B12 levels with those of normal B12 levels. We also explored the mechanisms by studying the distribution of B12 in the blood among its transport proteins or transcobalamins (TC). METHODS: The clinical laboratory provided all 94 sera that had B12 level Ͼ900ng/l, out of 670 consecutive assays (14% incidence) over 75 days. Results were confirmed in follow-up blood samples, after excluding duplicate submissions and patients without confirmatory specimens or with normal repeat B12 levels. Clinical and laboratory data from 60 patients (group I) were compared with those of 75 cosecutive controls that had normal B12 levels (group II).The distribution of B12 between TCI and TCII was determined by the QUSO method; normally Ͼ75% of serum B12 is carried by TCI (holoTCI), which has only a storage function, and Ͻ25% by TCII (holoTCII), which rapidly delivers B12 to cells. RESULTS: (1) Among all the clinical conditions, only renal failure (p ϭ .01) and diabetes mellitus (p ϭ .04) were significantly more frequent in group I than II. Multivariate linear regression (MLR) analysis showed only kidney disease to be independently associated with a high B12 levels. None of the patients with high B12 levels, had myeloproliferative disease or leukemoid reaction, the best known causes of high B12 levels, and only 3 had recent B12 injections as possible explanations.

(2) Among all laboratory data, only high creatinine (p ϭ .0001), hypoalbuminemia (p ϭ .0001), anemia (p ϭ .001) and high alkaline phosphatase (p ϭ .01) were significantly were common in group I than group II. MLR showed only high creatinine level and hypoalbuminemia to be independently associated with a high B12 level. Serum B12 levels didn't correlate with WBC counts. (3) Holo TCI was inversely and holoTCII directly correlated with serum B12 levels (p ϭ .0001) and mean % holo TCI was significantly lower in groupI than II (69.6% Ϯ 19.5 vs 87.5% Ϯ 10.0, p ϭ .001). CONCLUSION: Elevated B12 levels are common in clinical practice (14%prevalence). Our data shows renal failure to be the only disorder independently associated with high B12 levels, while better known causes are actually rare. Ours is the first demonstration that B12 seems to be abnormally distributed in the patients with high B12 levels, with less of it being carried by TCI and more by TCII. This suggests that the probable mechanism involves diminished delivery of B12 to cells by TCII. For example, many high B12 levels may be due in part to the failure of normal uptake of TCII-bound B12 by TCII receptors in the kidney in renal failure. were tabluated with and without screening for depression. One, two, and threeway sensitivity analyses were conducted in order to determine the robustness of the model to variations in the estimates and to determine the variables that were most influential in defining the cost-utility of screening for depression. RESULTS: In the base case, annual depression screening had an incremental cost-utility (C/U) ratio of about $225,800 per QALY gained when compared to usual care. Opportunistic screening had an even higher C/U ratio. However, onetime screening (at entry to the health care system) had a C/U ratio of $48,900 per QALY gained. Estimated C/U ratios were most sensitive to the utilities assigned to major depression, prevalence of depression, costs of screening, treatment initiation once depression is diagnosed, remission rates with treatment, and rates of diagnosis in usual care. Sensitivity analyses indicated that annual screening would achieve a C/U ratio below $50,000 per QALY gained only under very specific multiple conditions. A population with a high prevalence of depression (Ͼ13%) and high treatment success rates (Ͼ70%) were both required, along with any one of the following: high treatment rates for those diagnosed with depression (Ͼ70%); very low utility of depression (Ͻ0.54); or low cost of screening (Ͻ$3.00). In contrast, a change in nearly any single parameter increased the C/U ratio of one-time screening above $50,000 per QALY gained. CONCLUSION: Annual or periodic screening for depression is unlikely to be a cost-effective intervention in primary care unless substantial improvements are also made in the quality and effectiveness of depression treatment. However, onetime screening for depression may be reasonably cost-effective even with current practice patterns. The database consists of modules on common topics of importance for primary care physicians (e.g., diabetes or asthma), of which seven have been developed so far. For each topic, information from one hundred or more original journal articles was entered in the database. For each article, the database includes the abstract of the article, study type, clinical topic, medical subject heading (MeSH) terms assigned to the article, and a clinical "pearl" developed by a generalist physician domain expert. The "pearl" summarizes the main points of interest for primary care physicians needing to apply the information to patient care. Physicians access the database through the search interface, allowing retrieval of articles meeting criteria requested by the physician, including study type, evidence quality, and subtopic (e.g., epidemiology or diagnosis). The database is dynamic, as new high-quality journal articles can be added to the database as they become available. The search interface also accesses information on the specified topic from a complementary database, which contains audio-visual segments from recorded lectures given by experts on the topic of interest. We designed a multi-site randomized trial to examine the impact of brief advice for risky drinkers on behavioral, health, economic and satisfaction outcomes in managed care settings. METHODS: Three primary care clinics in each of three HMOs (9 sites) were randomized to one of three conditions: brief advice by their provider (P), brief advice by another trained staff member (S), or usual care (C). Patients in each clinic completed a self-report questionnaire. Those screening positive as risky drinkers in the P and S sites received brief advice. A two-month follow-up telephone interview included the HEDIS 3.0 Member Satisfaction Survey, the PCAS Interpersonal Treatment Scale, and global satisfaction items. Satisfaction scores were compared using Wilcoxon Rank Sum tests and linear regression analysis. RESULTS: Over 900 patients (544 Males, 375 Females; P 248, S 242, C 429) were enrolled from the nine sites, beginning in March 1998. In univariate comparisons, satisfaction scores were slightly lower for participants in intervention (P, S) vs. control (C) sites. In linear regression analysis, adjusting for age, gender and overall health status, participation in intervention sites (P, S) was associated with slightly lower satisfaction. However, when we adjusted for enrollment from one of the P sites where specific local management problems had occurred, there were no significant satisfaction differences between intervention and control sites. There were no significant differences in satisfaction scores of participants in S vs. P sites, in univariate or multivariate comparisons. CONCLUSION: Although intervention (P, S) participants had slightly lower satisfaction scores than controls, our preliminary data suggest that these differences are due primarily to specific local management problems at one of the P sites. Controlling for this one site, it appears that screening and brief advice does not negatively affect patient satisfaction. Further data collection and analyses will provide more information on patient satisfaction with these interventions, and on how satisfaction measures relate to health and behavioral outcomes. The study population was comprised of general internists and family physicians participating in a 90,000 member mixed HMO affiliated with a major academic medical center serving metropolitan Washington, D.C. One hundred seventy six physicians completed the 1993 survey (response rate of 76%). After adjustment for 35 physicians who had retired or could not be located, 108 1998 responses were received (response rate of 76%). Physicians were asked whether they agreed, disagreed or were uncertain regarding statements about PSA test use in PC screening. Query statements included acceptance of PSA tests as a standard, malpractice liability for not ordering, PSA screening test characteristic acceptability and statements regarding the impact of screening and treatment on outcomes. Responses in 1998 were compared to 1993 results with a two-tailed p Ͻ 0.05 as criteria for statistical significance. RESULTS: Significant increases in agreement with attitudes favoring PSA testing were observed from 1993 to 1998. The percentage of physicians expressing favorable attitudes toward PSA testing showed an absolute increase of 11% to 17% (relative increase of 21% to 52%) across the questions repeated over time. In both periods, fewer physicians indicated agreement with statements related to improved clinical patient outcome from early intervention (efficacy of radical prostatectomy, benefits of treatment outweigh risks) than with non-patient related statements. Community-based physicians reported higher levels of agreement with attitudes favoring PSA testing than academic-based physicians in 1993 and 1998. To determine whether a site-defined practice improvement program would increase the proportion of patients achieving AHA/ACC guideline-defined goals for coronary heart disease. METHODS: Nine site-defined quality improvement programs used a variety of provider and patient-based disease management tools to improve care. Cardiovascular risk factors, laboratory values, and medication use were collected prospectively (August 1997-June 1998) from these sites (2 internal medicine, 2 academic, 2 managed care, 3 cardiology; N ϭ 688 patients). RESULTS: Compared to baseline, at 6 months total cholesterol Ͻ200 mg/dL increased from 57% to 74% (P Ͻ 0.001), LDL-C Ͻ100 mg/dL increased from 33% to 74% (P Ͻ 0.001), HDL-C Ͻ35 mg/dL decreased from 27% to 19% (P Ͻ 0.001), rigorous exercise of 30 minutes, 3 times a week increased from 37% to 51% (P Ͻ 0.001), lipid medication use increased from 72% to 79% (P Ͻ 0.01), and aspirin use increased from 84% to 90% (P Ͻ 0.001). No differences in LDL-C goal achievement were seen in patients treated with the most commonly used statins (i.e., 82% ϭ simvastatin; 77% ϭ atorvastatin; P ϭ NS) at 6 months. Despite some improvement in TG levels and blood pressure control, differences at 6 months compared to baseline were not statistically significant. There were no significant changes in diabetes management, smoking cessation, body mass index, and betablocker use at 6 months. Successful implementation of this program was related to 1) the total number of patients enrolled and 2) the use of a nurse study coordinator. No differences in LDL-C goal attainment were found between patient and provider interventions. CONCLUSION: These data appear to demonstrate an opportunity to improve patient care by the use of disease management principles using a variety of interventions and measures to promote best practices in the treatment of CHD patients. The amount of improvement demonstrated correlated more with the dedication of each site to quality improvement (evidenced by the use of a nurse study coordinator and total enrollment achieved) than any particular patient or provider centered intervention. There is a need for additional studies to further evaluate optimal interventions to improve care of patients with CHD. PURPOSE: The quality and intensity of inpatient hospital care should be the same regardless of the day of the week the care is being provided. However, weekend days, particularly Sundays, seem less active in terms of clinical interventions and discharges. This study examines impact of this "Lazy Sunday" effect on length of stay (LOS) and hospital charges and shows how the phenomenon has changed between 1988 and 1995 in the face of increasing cost-containment pressure. We hypothesized that admissions on Monday would be more likely to have a LOS Ͻ5, enabling a pre-weekend discharge. Patients admitted on Thursday would be less likely to be discharged in Ͻ5 days since the typical peak discharge day would fall on the "Lazy Sunday." METHODS: We utilized data from the Nationwide Inpatient Sample (NIS) collected as part of the Healthcare Cost and Utilization Project. To ensure comparisons of patients with similar characteristics, we only selected admissions among patients over 65 from Subsample 1 of the NIS from 1988 and 1995 having the 4 most common DRG's. The NIS also names the weekday of each admission and the corresponding LOS and charges. Since the "Lazy Sunday" effect is hypothesized to impact typical admissions, outlier admissions (LOS Ͼ25 days, charge Ͼ$25,000) were excluded. The significance of differences in average LOS and average charges was assessed with the Student t-test. The Chi square test was used to assess variation in the cumulative percent discharge by day 5. RESULTS: In 1988 and 1995 CONCLUSION: Although the "Lazy Sunday" effect occurs whenever an admission spans a Sunday, the analysis shows that, consistent with our hypothesis, the effect is greater when the admission occurs later in the week, resulting in a 0.5 day increase LOS and increased charges of about $400. While the average LOS has decreased and charges increased over the period between 1988 and 1995, the impact of the "Lazy Sunday" as measured by the absolute difference in LOS on Monday and Thursday has not changed. Given the Monday-Thursday charge difference in 1995, this study suggests that more aggressive discharge planning for the nationwide set of patients over 65 with common DRG's could save at least $90 million in charges annually. PURPOSE: Risk adjustment is beginning to include factors based on specific clinical conditions. Although Alzheimer's disease (AD) is a major cause of morbidity and death among older adults, we know little about how expenditures and utilization for persons with AD change over time. We sought to compare 1994-1996 patterns of resource utilization for Medicare beneficiaries with AD to those for beneficiaries without AD and to test the hypothesis that AD is associated with increased expenditures independent of demographic factors. METHODS: We examined a 5% nationally random sample of paid fee-for-service Medicare claims for services rendered to beneficiaries between 1994 and 1996, using diagnostic coding for AD to define cases. We identified a 1994 cohort and then assembled data reflecting the cohort 's 1994, 1995, and 1996 claims. We assessed usual sources of care, comorbidity, and Medicare expenditures by service and physician specialty. RESULTS: Of 1,138,445 beneficiaries in 1994, 1% had a 1994 diagnostic code for AD. In 1994, per capita expenditures for beneficiaries with AD were $8,676, or 2.1 times those for beneficiaries without AD. Expenditures for those with AD declined after 1994 ($6,744 in 1995 and $7,163 in 1996) , whereas expenditures rose slightly and steadily for the non-AD population ($4,617 in 1995 and $5,250 in 1996) . In 1994, 48% of beneficiaries with AD were hospitalized, compared to 23% of beneficiaries without AD; hospitalization was 1.5 times more common with AD in 1995, and it was 1.3 times more common with AD in 1996. Ratios of per capita expenditures for Medicare beneficiaries with AD to per capita expenditures for Medicare beneficiaries without AD were high for services provided by psychiatrists, geriatricians, or neurologists, relative to those provided by internists. Absence of comorbidity is much less common among beneficiaries with AD (9%), compared to those without AD (27%). Compared to beneficiaries without AD, a smaller proportion of beneficiaries with AD had usual sources of care, and this difference increased over the 3-year period. In multivariate regression analysis, age, race, gender, time enrolled in Medicare, comorbidity, and AD diagnosis were independently associated with total 1996 expenditures. CONCLUSION: AD was independently associated with greater Medicare expenditures from 1994 to 1996, and for AD, expenditures in psychiatry, geriatrics, and neurology are relatively higher than in internal medicine. Medicare beneficiaries with AD may also lack continuity of care. These findings suggest that the role of cognitive function in risk adjustment requires further investigation.

Year (15), bleeding risk or history (15), alcohol abuse (10), or compliance (3). PCP surveys provided reasons for 11 more persons, leaving 21 with no or unusual reasons (e.g., asymptomatic status, history of ulcers without bleeding). Thus, 21 ⁄ 445 (4.7%) of apparent non-use may be inappropriate. CONCLUSION: In contrast to studies suggesting that warfarin is underused in AF, we found high rates of use. Moreover, persons not on warfarin had credible reasons for non-use.In the majority of cases either chart review or PCPs justified nonuse. These findings should be considered in light of the setting-a university affiliated tertiary referral center. Nonetheless, we believe that the primary reason for our more favorable findings is our use of physician review of the entire chart, and direct communication with PCPs, which allowed for judgement regarding such cases as multiple very brief episodes of AF, lack of reliable follow up or documented patient refusal. In the first year of a patient and clinician educational intervention to decrease antibiotic use for adults with uncomplicated acute bronchitis, antibiotic prescribing decreased from 80% to 45% of incident visits at the full intervention clinic, and this effect was sustained during the subsequent year with minimal reinforcement. The purpose of this study was to evaluate clinician characteristics associated with decreased antibiotic prescribing behavior. METHODS: We calculated antibiotic prescription rates for uncomplicated acute bronchitis, based on administrative claims data, for individual clinicians at the control, limited and full intervention sites belonging to a group-model HMO. The primary outcome measure was percent change in antibiotic prescription rate, from baseline winter to year 1 and year 2 winters, for each clinician. Only clinicians with 10 or more visits represented in consecutive years were included in the analysis. We compared antibiotic prescription rates between different clinician types-physicians (MD), nurse practitioners (NP); and different specialties (Family Medicine (FM), Internal Medicine (IM)) during each period. Statistical significance of differences between clinician type and specialties were assessed using Student's t-test. RESULTS: Sixteen, 14, and 11 clinicians from the control, limited and full intervention sites, respectively, were eligible for analysis. The proportion of clinicians with Ͼ10% decrease in antibiotic prescription rates in year 1 were 25%, 42%, and 73%. At the full intervention site in year 1, antibiotic prescription rates decreased a greater amount among NPs (n ϭ 6) than MDs (n ϭ 5) (Ϫ51% vs. Ϫ5%, p ϭ 0.02). However, by year 2, the percent change in antibiotic prescription rates were similar between NPs (Ϫ48%) and MDs (Ϫ36%) (p ϭ 0.46). During either period, there was no difference between clinician specialty in percent change in antibiotic prescription rates. CONCLUSION: Physicians may be slower to adopt changes in prescribing practices in response to quality improvement efforts than mid-level providers. Evaluating the impact of such efforts should include a sufficient time interval to detect changes by all providers. conditions. A total of 2346 patients aged 65 or older who were attending appointments at these clinics completed the 15-item Geriatric Depression Scale (GDS). Primary care physicians in the intervention clinics were notified of their patients' GDS scores. We suggested that participants with severe depressive symptoms (GDS score Ͼ ϭ 11) be referred to the Psychiatry Department and participants with mild to moderate depressive symptoms (GDS score 6-10) be evaluated and treated by the primary care physician. Depressed (GDSϾ ϭ 6) intervention group participants were also offered a series of organized educational group sessions on coping with depression led by a psychiatric nurse. Primary care physicians in the control clinics were not notified of their patients' GDS scores (usual care CONCLUSION: Case-finding for depression did not affect patient outcomes at two years. Perhaps case-finding followed by more intensive interventions than those applied in our study will improve the outcomes of patients with depression. METHODS: To meet an expected 10 million dollar budget deficit, the members of each clinical department of a large tertiary VA medical center were presented with detailed data about their costs; they prioritized 25% of these as least essential to good patient care. Two hundred and ninety eight items for potential cost reductions were proposed. All 2000 employees of the medical center were invited to provide feedback on the items. RESULTS: Twenty eight percent (n ϭ 565 ⁄ 2000 )of the employees participated in the poster feedback session and indicated support or opposition for items. Nineteen percent of responders were "practitioners" (MDs, PhDs, NPs and PAs), 55% were employees who provide clinical support (RN, secretaries, therapists and other staff in direct contact with patients) and 26% were non-clinical support staff. Employees supported, defined as more than 50% endorsement by all groups, 80% of the proposed cost reduction measures. There was universal support for more than 87% of items within the categories of drug restrictions, improved efficiency, revenue generation, test restriction or new services proposed to save money. One quarter of items of service restriction (items proposing restriction of a disease specific therapy) or drug substitutions were opposed. Opposed items included elimination of palliative services for terminal or chronic conditions, treatment or diagnosis of politically popular diseases, items lacking rigorous evidence based standards, and limited access to users outside our immediate geographic area.

The clinical support staff tended to endorse cuts less often (p Ͻ .05) than practioners in the categories listed above, as well as for items shifting workload, those implying medicolegal risks, those proposing decreased length of stay and items proposed solely to satisfy the quota of 25% of departmental costs. Service restriction items were opposed more often than all other types of items (28% vs. 15%, p Ͻ .01). CONCLUSION: Using financial information, clinicians can identify inefficient practices and non-obligatory services they will sacrifice to maintain adequate funding for essential care, and there is employee consensus for most proposed reductions. Employee consensus might be improved through increased participation in item identification and education on evidence based and community standards. Respondents were asked about a list of common medical conditions, their use of conventional medical services, and their use and insurance coverage of 20 AM therapies. We defined high frequency AM use as 5 or more visits annually to a specific type of AM provider and examined use in those who gave information on visits to AM providers in the last year. We used logistic regression to identify factors independently associated with high frequency AM provider use.

RESULTS: Of those eligible, sixty percent completed the interview. Of 2055 respondents, 19.1% gave information regarding visits to AM providers. The median and mean number of visits to a given type of AM provider in the last year were 4 and 9.9, respectively. Mean number of visits to an AM provider varied by the degree of insurance coverage as follows: 12.6 for full, 9.9 for partial, and 8.3 for no coverage. Adjusted odds ratios (AOR) for the likelihood of having made five or more visits to a specific type of AM provider are shown below.

The effect of full insurance coverage was not significantly different from partial coverage when compared in the model. Having visited a conventional medical doctor 5 or more times in the last 12 months had borderline significance as a correlate of high frequency AM provider use, OR (1.61 (0.99, 2.64) ). CONCLUSION: Lack of insurance coverage appears to be a limiting factor in determining the number of visits made to an AM provider. As insurers extend coverage of AM services, the number of visits made to AM providers might be expected to increase. Respondents were asked about a list of common medical conditions, use of conventional medical services, and use of 20 AM therapies. A random subset of respondents who had used at least one AM modality or self-prayer in the last year were also asked whether, if presented with a choice of two insurance plans that were otherwise equivalent, they would choose the insurance plan offering AM benefits. We used logistic regression to determine factors associated with being more likely to choose the insurance plan offering AM benefits. RESULTS: Of those eligible, sixty percent completed the interview. Of 2055 respondents, 54.6% had used at least one AM modality or self-prayer in the last year. Of those, half were surveyed about their likelihood of choosing an otherwise equivalent insurance plan offering AM benefits: 68.9% reported they would be more likely to sign up with that plan, 24.7% were indifferent, and 6.5% would be less likely to sign up with that plan. Adjusted odds ratios (AOR) for being more likely to choose a plan with AM benefits are shown below.

Having a greater number of medical conditions or a greater number of visits to conventional medical doctors in the last year were not associated with choosing an insurance plan with AM benefits. CONCLUSION: Insurance benefits for AM were desired by the majority of respondents who had used any type of AM in the past year. Extrapolation of these results to the US population conservatively estimates that there are at least 75 million Americans for whom the presence or absence of AM insurance benefits may impact the selection of their insurer. RESULTS: The respondents were predominately African American (90%), females (75%), non-smokers (86%), with BMIs of 33.5 ϩ/Ϫ 9.4, 55.7 ϩ/Ϫ 8.9 years of age, completed 8.9 ϩ/Ϫ 3.0 years of education, permanently disabled (63%), with annual incomes of less than 10k dollars (58%). Their mean age when diagnosed with diabetes was 39.8 ϩ/Ϫ 11.3 years, and most (65%) were currently prescribed insulin. During the implementation of TennCare most patients choose their MCO, and kept the same physician (63%). Twenty-eight percent rated their current healthcare under TennCare as excellent, 10% very good, 40% good, 14% fair, and 5% rated their current healthcare as poor. Thirty-eight percent felt the quality of their healthcare before Tenncare was the same; 22% felt it was better or much better; and 40% felt it was worse or much worse. Respondents reported that their blood sugars were controlled most of the time now compared to before TennCare (45% vs. 29%, p ϭ .05), and that they were more likely to check finger stick blood sugars now than before Tenncare (74% vs. 43%, p ϭ .05). They were less likely to check urine dipsticks (10% vs. 19%, p ϭ .05). Patients also reported out of pocket expenses for transportation and medications prior to TennCare to be the same (35%); better or much better (10%); or worse or much worse (55%). More reported that they did not have a test performed because of lack of approval under TennCare (63% vs. 23%, p ϭ .06), and seeing a diabetes specialist at least once a year (45% vs. 33%, p ϭ .05), but there were no significant differences in the percentage of patients who reported receiving detailed information about diabetes (92% vs. 75%), or receiving dietary information(93% vs. 78%). CONCLUSION: After the initiation of Medicaid managed care patients were more likely to report home glucose monitoring, and control of their blood sugars most of the time. The majority was satisfied with their MCO, and reported fewer out of pocket expenses. From the patient perspective enrollment in Medicaid managed care had several positive outcomes, and appeared to result in improved access to appropriate diabetic care. There is concern that physicians may withhold therapy based on their (not necessarily accurate) assessment of patients' likely adherence. We surveyed physicians to assess whether prescribing protease inhibitors to patients is affected by perceived patient ability to adhere to treatment regimens. METHODS: Health care providers of a nationally representative sample of 2864 HIV-infected individuals in the U.S. were surveyed in 1998. Using a 5-point Likert scale, they were asked whether they agreed that a patient's likelihood of adherence is a very important factor in their decision to prescribe protease inhibitors and whether they prescribe protease inhibitors only if a patient is likely to be adherent with the regimen. We examined the association between providers' attitude toward prescribing protease inhibitors and their knowledge of HIV treatment and personal characteristics. The response rate from the provider survey was 72%. RESULTS: Among 411 providers, 89% agreed that adherence is an important factor when deciding to prescribe protease inhibitors and 69% percent agreed that they prescribe protease inhibitors only if a patient is likely to be adherent. In a stepwise logistic regression analysis, providers with better knowledge of HIV treatment and general internists (as compared with infectious disease specialists) were more likely to consider their prosepective assessments of patient adherence in their decision to prescribe protease inhibitors (OR ϭ 4.18 p ϭ 0.04 and OR ϭ 1.56 p ϭ 0.06, respectively). Factors not associated with consideration of patient adherence were physician's gender, ethnicity, years in practice, sexual preference, preference not to treat intravenous drug users, and the belief that many of their patients cannot adhere. CONCLUSION: Among physicians caring for a nationally representative sample of HIV-infected individuals in the U.S., most said they would prescribe protease inhibitors only if the patient is likely to be adherent. This finding raises concern because existing evidence suggests that physicians are poor judges of patients' adherence. Efforts to increase delivery of protease inhibitors to undertreated groups need to address physician perceptions of patients' adherence. reports of how often patients asked that they mislead payors to get coverage for uncovered services. RESULTS: Respondents were 36% primary care physicians, 13% internal medicine subspecialists and 23% surgeons. They had been in practice for a mean of 17 years, 20% were women, and 40% were foreign born. Overall, 31% of physicians reported not offering useful but uncovered services to patients at least sometimes over the last year. Those who reported this were less likely to be satisfied with medical practice (OR 0.58, CI 0.38-0.86) and more likely to report recent patient requests to mislead their insurers to obtain coverage for uncovered services (OR 2.26, ). Other weaker, but statistically significant, correlates were having more than 25% of one's patients under Medicaid (OR 1.7, CI 1.07-2.89) and being American born (OR 1.5, CI 1.02-2.34). Non-significant variables included age, gender, specialty, and practice characteristics such as practice setting, primary method of payment, number of managed care contracts, and proportion of income from managed care. Of those reporting not offering useful services, 76% reported doing so more often in 1998 than 5 years previously. CONCLUSION: While not offering useful services to patients is an undesirable response to the dilemma of such services not being covered, it is an increasingly common one. Offering such services may prompt patients to request that MDs manipulate reimbursement rules, which in turn may motivate MDs not to discuss useful but uncovered services in the first place. In a cross-sectional analysis the direction of causation cannot be known, but such "no-win" scenarios for MDs may contribute to reduced MD satisfaction. PURPOSE: To investigate the correlation between patient satisfaction with care and patient trust of physicians at two sites: A university hospital outpatient (general internal medicine) clinic and a private primary-care clinic. METHODS: A pilot study of patient trust was conducted in November and December of 1999. Adult patients who were cared for by the same physicians for more than 3 months and who had at least 4 visits during this time, were selected for this study. The patient trust of physician ratings were obtained by self-administered questionnaires, the Trust in Physician Scale ( modified by the Stanford Trust Study Group). In addition to patient trust, we obtained patients' characteristics, the lengths of the patient-physician relationships, the reasons for the patients' choice of physicians, the satisfaction with care, and a general trust scale (Yamagishi, Japan). 275 patients used the questionnaire to rate 11 different physicians during their visits. RESULTS: The Trust in Physician scale (5 Likert scales: 1 ϭ totally disagree to 5 ϭ totally agree) were transformed to a 0 to 100 scale, and showed high internal consistency (Cronbach's ƒ¿ ϭ .82) for Japanese patients. Mean scores of male and female were 86.3 (SD: 10.3) and 85.9 (SD: 10.5) respectively. The trust scores were not statistically associated with gender, age, education levels and the lengths of the relationships. However, the trust scores were strongly associated with the reasons for the patients' choices, the general trust scores, and the satisfaction with care after adjusted by patient characteristics. CONCLUSION: The patient trust was a useful concept to assess the relationship between patient and physician, especially the satisfaction with care. The patient trust, one of the health care seeking behaviors, was also associated with the societal trust in general in our study. However, we need further investigations such as the relationship between patient trust and compliance and consequence of medical care. We also surveyed physician groups (response rate 96%) to adjust for the independent effect of characteristics of MC physician groups that may influence selected CAHPS scales. Independent variables from the physician groups adjusted for in multivariate models included: linkage of the PCP's compensation to volume of specialty care; use of a bonus system in the PCP's compensation; experience with capitation; amount of specialty care available within the group; and whether the dominant form of PCP compensation was capitation. All regression analyses were also adjusted for the independent effects of gender, race, and Medicaid status. RESULTS: In the older/sicker subset (N ϭ 261), the mean age was 80 ϩ/Ϫ 4 years and the mean number of comorbidities was 2.9 ϩ/Ϫ 1.2, compared to 73 ϩ/Ϫ 5 years and 1.4 ϩ/Ϫ 1.2 comorbidities for the remainder of the sample. After adjustment for demographic and provider group characteristics, we found that older patients with more chronic medical conditions did not have significantly different CAHPS sub-scales scores from a younger and healthier subset with one exception: older, sicker patients reported more difficulty finding a personal doctor (OR ϭ 0.6, p ϭ 0.03). CONCLUSION: Despite the current economic and administrative constraints of Medicare managed care plans, older patients with multiple chronic illnesses appear to have similar levels of satisfaction as younger, healthier Medicare managed care beneficiaries. PURPOSE: It is generally accepted that an understanding of bio-ethics is crucial to delivering good patient care. Despite its importance in medical education, bioethics is either not taught at all or not taught well in many residency training programs. Reasons include the lack of clinical applicability of some teaching methodologies and the fact that ethical concerns or disputes are often viewed as having no discrete answer or solution. Literature, film and theater create an atmosphere that provides an intimate, fascinating and safe glimpse of situations that would otherwise be impossible to replicate. They allow residents to become deeply invested in the characters and outcomes without confusing the issues with medical jargon and the defensive posturing of case presentations. The goal of our project is to create a curriculum in bio-ethics and professionalism by using the arts as a means for small group discussions of complex and controversial topics. METHODS: The framework consists of didactic lectures, small group discussions and theater/film events that highlight various ethical issues encountered in the practice of medicine. Each discussion or media event is designed to enhance the educational development of physicians in training by augmenting the individual's knowledge, skills and attitudes in medical ethics and professionalism. For example, various short stories, poems or plays are chosen to evoke reflection on topics such as death/dying, patient suffering, patient-doctor relationships and altruism. The unique feature of the program consists of monthly events for residents to experience plays, films or public lectures that also raise awareness and knowledge that is applied to ethics and humanism. Each event is followed by a discussion with faculty and a Likert-type scale survey to assess the educational value of the experience. The housestaff were asked four questions about the event to evaluate the specific learning objectives, the influence on one's attitude toward doctoring, the sense of humanism and the relevance to clinical training. RESULTS: Surveys from three events were obtained for a total of 25 responses. Responders assigned a weight of one point to strongly disagree up to a weight of five points to stongly agree to each question. The average response for each question was as follows: learning potential 4.32, career reflection 3.77, humanism 4.0 and relevance to training 4.75. CONCLUSION: Our survey results support that residents view a curriculum in bio-ethics and professionalism as a highly relevent component of their training. With the help of the arts, educators can further augment post graduate medical education by enabling residents to attain skills and attitudes in bio-ethics and professionalism.

"TURFING" PROFESSIONALISM: THE PATIENT'S PERSPECTIVE. CV Caldicott, RM Frankel, Internal Medicine, SUNY Upstate Medical University, Syracuse, NY; Primary Care, Strong Health/Highland Hospital, Rochester, NY PURPOSE: Medicine residents who receive "turfs" are known to feel abused and overpowered by other doctors who elect not to care for these patients. This pilot study examines whether the patients caught in such transactions of social triage and rejection can discern that they are perceived as "turfs." We further wondered if the unprofessional conduct of their physicians adversely affects their medical care. METHODS: Study subjects included 1) transfers to medicine teams in University Hospital (UH) either from an outside institution or from another specialty service within UH, and 2) patients evaluated in the Emergency Department by a specialty service but ultimately admitted to medicine. Audiotaped interviews were conducted to elicit patient experiences and perspectives on care. Major themes were identified by reviewing the tapes. In brief audiotaped interviews, medicine housestaff caring for these patients explored feelings and experiences regarding appropriateness of transfer or triage. Based on their comments, and applying previously described criteria, the patient interviews were assigned either "appropriate" or "inappropriate" (i.e., "turf") status. Where the housestaff were ambivalent in their feelings about the patient, an "equivocal" status was assigned. RESULTS: Ten of 18 patients were considered appropriate by their housestaff, five inappropriate, and three equivocal. The most common themes among the appropriate patients were a general desire for more communication between doctors and patients, satisfaction with their medical care and hospital experience at UH, and complaints about their care at the outside facility. Among the "turfed" patients, communication was also a theme, but it was expressed as a complaint about communication both between doctors and patients and among doctors themselves: "I got a different story from every doctor I talked to. There was no MEDICAL HUMANITIES/CONCEPTUAL ETHICS communication between doctors." In another theme, "turfs" detected "turf battles" between doctors of different specialties involved in their care even though the patients were not openly informed of such: "There was an obvious dispute between two groups of doctors, a turf battle. [It was] very negative." In the "turfed" patients' last theme, they questioned doctors' motivations, specifically financial remuneration and insurance status: "[not having insurance] was probably half the reason she didn't get taken care of like she should." CONCLUSIONS: "Turfed" patients understood their unfortunate status and perceived unprofessional motivations and conflicts among their doctors. Although less satisfied with their care than appropriately transferred controls, actual consequences of "turfing" on outcomes of care should be investigated in future studies. PURPOSE: Diagnostic procedures have several potential secondary uses in medical practice beyond the identification of disease. They may, for example, provide reassurance to a patient or physician. An extreme form of such secondary use is placebo diagnosis for reassurance. An influential physician, Dr. Richard C. Cabot, in the early twentieth century explored the use of placebo diagnostic procedures. METHODS: Random number sampling of patient records from the private practice of Richard C. Cabot, 1900 to 1920 demonstrated instances of placebo diagnosis and treatment. A follow-up examination of individual records, including letters from Cabot's patients, revealed detailed information about these practices. Historical context was provided by review of early twentieth-century medical literature on the use of diagnosis and placebo, and by review of Cabot's own published writings on the subjects of diagnosis, placebo, and relationships with patients. RESULTS: Cabot experimented with the use of a diagnostic procedure as a placebo. For example, he collected sputum for staining for acid-fast bacilli, and then reassured a patient about the absence of tuberculosis without examining the sputum. Cabot's use of placebo diagnosis coincided with his early experimentation with the self-conscious use of placebo treatments for his patients. These practices were consistent with Cabot's stated concern about the relative over-valuation of laboratory diagnosis and specific therapeutics, and about the influence of these biases on professional identity and autonomy. Cabot's published views on placebos, however, soon hardened in response to a different set of concerns about professional obligations and image. If the public recognized the wide use of placebos, Cabot soon warned, they would denounce the practice. If they did not recognize it, they would increasingly come to believe inappropriately that "every symptom needs a drug." Cabot subsequently abandoned his own use of placebo treatments and diagnostic procedures, consistent with his changing views. CONCLUSION: For a thoughtful and innovative practitioner like Richard Cabot there was flexibility in the secondary use of novel diagnostic procedures flowing from the laboratory in the early twentieth century. Cabot's exploration of placebo diagnosis, for example, waned seemingly in response to changing concerns about appropriate ideals for professional behavior. Concepts of ideal professional obliga-tions, identity and autonomy seem to have deeply influenced his use of new diagnostic procedures.

PURPOSE: To trace the evolution of EBM in the history of medical education. METHODS: Historical analysis of primary and secondary documents. RESULTS: EBM educators can trace their ancestry to a reform of the American system in the late 19th century, which saw the rise of university-based medical schools, raising of admissions standards, expansion of curricula, and the embrace of "progressive education." The latter recognized the importance of experiential and life-long learning and the stifling effects of deference to authority. These notions found expression in the use of laboratories and clerkships to supplement lectures. One of the many forces driving this reform was the influence of Europe. German scientists pioneered experimental laboratory science, while the French systematically documented clinical observations. In 1836, Dr. Pierre Louis's observational study challenged the accepted efficacy of blood-letting. His empirical approach established clinical research as a pathway to the understanding of disease. From this point on, medical knowledge would evolve and expand. American educators recognized physicians' need to continually update their knowledge. Medical students must graduate with the thirst, critical skills, and "informatics" for lifelong learning. Introduced by Osler in 1875, the journal club became an important venue for information exchange. Its primary purpose was to help physicians keep abreast of the emerging world's literature. In the early 20th century, therapies for dramatic acute infections were developed in the laboratory without the need for the persuasion of a p-value. Later, extending Dr. Ernest Codman's "end result" idea, investigators turned to patient-based research to study the subtler statistical outcomes of chronic diseases. Practitioners needed a framework to appraise and apply this new literature. Critical appraisal eventually found its way into medical education curricula, often in journal clubs. Now students in this forum not only "kept up" with articles, but also critically scrutinized them. Although educators also intended journal clubs to impact on clinical practice, patients were rarely discussed. In the last 10 years, educators folded critical appraisal into the larger process of evidence-based medicine, which begins and ends with an individual patient. This required new technologies to serve practitioners on the front lines of patient care. In addition to informing clinical decisions, EBM represented an alternative approach to life-long learning in which the patient encounter cues the acquisition of new knowledge. CONCLUSION: EBM's experiential, self-directed educational approach extends the century-old "progressive education" mandate to include recent advances in clinical epidemiology, medical informatics, and decision science. Historical analyses may inform the current EBM debate. That same year Lincoln Hospital was accredited as one of 14 black hospitals in the country approved for internship training. Close examination of racial policies are instructive. The medical staff was composed of African-American and white physicians from Durham with both groups operating on and caring for patients, and both involved in teaching interns and residents, nurses and allied health professionals. The chiefs of service for nearly every area were white because of the requirement of board certification for accreditation, and thus, they participated in monthly staff meetings and continuing education programs held at Lincoln Hospital. In 1949 the By-Laws of Lincoln Hospital were changed to reflect a policy of open integration for patients and the medical staff. In addition, involvement of Duke Hospital surgery residents and faculty in the teaching program at Lincoln Hospital beginning in the 1930s, and the active engagement of Duke faculty and family members as trustees of the hosptial helped strengthen the facility and its teaching programs for the next four decades. The policy and practice of racial integration of the Lincoln Hospital medical staff continued until the hospital closed in 1976. CONCLUSION: Cobb's assessment of Lincoln Hospital as an oasis resits on its modern facilities when compared to other predominantly all-black hospitals, its residency, nursing and allied health training programs, the high quality of African-American physicians and surgeons at Lincoln Hospital, and its teaching fac-ulty and clinicians who included many of the preeminent white physicians and academic scholars in Durham and at Duke Hospital. PURPOSE: We reviewed narratives written over three years of residency training by our initial cohort of primary care residents to assess them for common themes and evidence of self-reflection. METHODS: Four times, twice as interns and then once each of the next two years, the residents met with 2 faculty members to write narratives and share their stories, voluntarily, with the group. The residents were instructed at the beginning of each session to spend thirty minutes writing about any event in their lives that they viewed as important and that had influenced who and where they are today. Using the constant comparative method, all three authors reviewed the narratives-first as a whole, then each one was reviewed and compared to prior and subsequent narratives-to identify emerging themes and concepts. A series of team meetings were used to refine the meaning of each content area (early PGY1, late PGY1, middle PGY2, late PGY3) discuss alternative interpretations, and reach consensus on common themes. RESULTS: Early PGY1 narratives focused on the interns' experiences with patients and/or their illnesses and shared a common theme-a search for professional identity and values. By the end of the PGY1 year, the stories took a more varied path. Some continued clarifying their professional identities and retained a connection to the idealism with which they began their training. Others, however, began to describe a darker side of residency training, full of fear and hopelessness. PGY2 trainees described a sense of despair and detachment and seemed lost in the depths of very agonizing experiences; the idealism of the internship was gone. The PGY3 narratives reflected on the experience of residency and how it influenced the trainees' perspectives, with half of the papers centered around patient encounters and the other half reflecting on the overall training experience; from the PGY3 stories emerged common themes of hope and reconciliation.

CONCLUSION: The residents were able to come to know each other at a level not often achieved by casual discourse. Residency was a journey that began with an idealistic search for professional identity and values, often fell into a darkness where these came into question, and then, for most, returned "into the light" where their values and identities were affirmed by their experiences. Residency is a difficult time given its temporal, mental, and emotional demands. Written narratives may be one method to encourage residents to process their experiences in order to make them better physicians, more compassionate caregivers, and more fully developed human beings.

LEARNING OBJECTIVES: 1) Highlight the medical literature on elderly drivers and motor vehicle accidents. 2) Review the medical evaluation of elderly drivers involved in motor vehicle accidents. CASE PRESENTATION: A sixty six year old white male is brought to the office by his wife who is concerned about his poor concentration. The patient has also noted difficulty with concentration during the last two weeks. Twenty days ago he was found by the local police to have left the scene of an accident he caused and was ordered to appear before the municipal judge. The judge ordered a complete medical evaluation prior to consideration of reinstatement of driving privileges. Paperwork for a physician statement from the board of motor vehicles has been presented by the couple at this visit. The patient is unable to recall the date of the accident. He recounts that he was having difficulty concentrating at the wheel of his vehicle and collided with another vehicle on a straight stretch of road. He recognizes that he did not stop at the scene but does not know why he did not stop. He sought no medical attention following the accident. His wife, with him in the office for the first time, reports a change in personality over the last five months. She attributes this change to his difficulty concentrating which has been progressive during this time. She is most concerned about what she identifies as imbalance which she has noted over the last three weeks. She has prevented him from falling several times in the last week when his gait was unsteady. His past medical history is remarkable for resection of Dukes A colon cancer two years ago. He had coronary revascularization ten years ago and makes regular office visits for blood pressure control. He continues to smoke cigarettes. Medications include: atorvostatin 10 mg po qd, aspirin 325 mg po qd, HCTZ 25 mg po qd, atenolol 50 mg po bid. Review of systems is remarkable for a productive cough for two weeks. He denies dyspnea, hemoptysis, orthopnea, and chest pain. He reports seeing television in shades of gray, despite watching a color monitor. He has had anorexia and nausea for two weeks, but no vomiting. Physical examination reveals an elderly man with difficulty registering questions from the examiner. (4) patients adopt a "day-at-a-time", "what is to be will be" approach to life. However, seemingly paradoxically (5) most informants have made "final arrangements" for their death (e.g., estate wills, funeral plans, cemetery plots). Similarly, significant percentages of patients have completed some form of advance directive (health care agent, 35%; living will, 55%), but these directives are poorly understood, and intended for use only in limited situations: "dire" states when the patient is pre-terminal, the problem irreversible, and the condition intolerable. CONCLUSION: These patients do plan for certain aspects of the future: they consistently make final arrangements for their death. To a lesser extent they are also willing to plan, through the completion of advance directives, for dire states. But they are reluctant, or frankly opposed, to planning for the end-of-life if they find it upsetting, if they consider it to be God's domain, or if it involves discussing scenarios of serious future illness associated with substantial uncertainty. These findings raise questions about the usefulness of the advance care planning model for addressing end-of-life issues with patients like our informants. (2), nurses-paramedical (5), administrative-logistic staff (3), board of directors (2) . RESULTS: Several meanings of quality cohabit in an organization and are controlled in different ways. For the professionals, the "historical" meaning is that of excellence, but a new meaning has appeared, that of values to be negotiated. For the administrative staff or that of logistic functions the main meaning of quality is that of standards to be respected. Methods of regulation differ according to the meaning given to quality. For people in the medical or paramedical professions, they are passing from implicit regulation based on the excellence of training to scientific regulation, which is based on guidelines or on evidence based medicine. For the non-medical managers, regulation is based on the same principles as those used to regulate quality in the industrial world. CONCLUSION: Quality has different meanings and regulations, which are not univocal. In addition, there exists a true dynamic of quality with the emergence of new meanings and new regulations. Our study suggests that the different regulations of quality extend its "determinants" largely beyond the care itself. Thus, managers but also other health professionals should pay more attention to the evaluation of the relations which link quality of care to management modes and organizational structures.

PURPOSE: To investigate how perception of drug errors influences management of errors. METHODS: Qualitative analysis of semi-structured individual interviews. Audio recording. Coding and analyzes were made separately by two analysts. Any discrepancies were resolved by discussion. Software Atlas-ti was used. Participants: 15 professionals selected according to their hierarchical position and to their function (Nurses: 8, Nursing Management staff: 5, Doctors: 2). RESULTS: Frequency of errors is underestimated because only errors, which present a serious risk to the patient, are taken into account. Drug errors are perceived as being consecutive to individual failure (lack of information, inattention). Means to prevent errors are formulated in the same terms (better training, better information, more vigilance, sanctions). Thus, discussions about errors are rare, informal, limited to some individuals and aim at pointing out the lack of attention. CONCLUSION: Perception of drug errors concerns pre-industrial logic, which is associated with an individual learning pattern based on three principles: training, vigilance and sanctions. This learning pattern runs counter to the logic of the drug dispensing process, which is done according to industrial logic. Therefore, the mechanisms of control devised for individual learning cannot apply with success to the industrial method of dispensing drugs used in most modern hospitals.

Threats.

LEARNING OBJECTIVES: (1) Diagnose Domestic Abuse when it presents as somatic complaints. (2) Recognize that mood disorders may be complicated by domestic abuse.] (3) Recognize culturally biased wording on medical questionnaires. CASE PRESENTATION: A previously healthy 31 year old African-American woman,Gravida 3, Para 3, presented for an initial office visit with a chief complaint of irregular menses. The patient had regular monthly menses every 28 days, with an average duration of 7 days, until two months ago when her menses lasted for 16 days. The following month her menses lasted ten days. The genitourinary review of systems was unremarkable except for pelvic discomfort with menses, chronic dyspaurenia and chronic sexual disinterest. The patient had a normal pelvic exam four months ago. Her history was otherwise unremarkable. Social history was notable for the following: She was a Nigerian born US citizen. She had been married for nine years and her spouse, an Air Force captain, was also a native Nigerian. For the past five years, the patient has been a labor and delivery nurse. Six weeks, prior to this visit the patient had suddenly quit her job. She denied smoking, alcohol use or illicit drug use. A general health questionnaire revealed mutiple complaints in every organ system, with positive answers to every cardiac, abdominal and genitourinary question. Psychiatric review of system was notable for anxiety and depression. She wrote: "I'm stressed out, typical working mother things, maybe I need some Prozac". On the question regarding domestic abuse: Are you now or have you ever been a victim of violence? The patient checked no. Her physical exam, including a pelvic, was unremarkable. When told that she had a normal exam, the patient became tearful and said, "Then I must be just crazy, I'm falling apart." The patient admitted that for six months she has had a disrupted sleep pattern, poor appetite,a ten pound weight loss, poor concentration, anhedonia, feelings of shame, guilt, irritablity, poor self-worth, hopelessness, and despair. The patient denied suicidal or homicidal ideation. The examination then progessed to the following: PHYSICIAN: "And your husband, how's that relationship?" PATIENT: "I should kill him. Just joking. We've been fighting alot. He spends his check on the boat(gambling). We have a lot of bills and I pay them all. I tried to talk to him, but he just calls me stupid, dumb, unsupportive." The patient also admitted to years of forced sexual contact and hitting which she was taught was normal. Recently, the hitting had become more violent. The patient admitted to quiting her job because a coworker had seen her bruises. DISCUSSION: The above case illustrates that underlying domestic abuse needs to be considered in the differential diagnosis of mutiple somatic and mood complaints.Brief questionnaires may be inadequate to distinguish the presence of domestic abuse, especially in certain cultures. whether or not they had undergone coronary angioplasty or bypass. Verbatim transcripts were analyzed qualitatively by a team of behavioral scientists and general internists in order to identify significant dimensions of communication and patient-provider relationships. RESULTS: We identified three principal domains of communication regarding cardiac testing: the substance of the information communicated, the meaning of that information for the patient, and the quality of the patient's relationship with the physician. In reviewing the substance of the communication, patients revealed ambiguity in the information received. One patient stated "They said the heart was fine but there was some blockage in the artery." One patient who was advised to have a heart catheterization described the communication as "They just said that they felt like that it needed to be done. I asked questions that morning when I was there. I wasn't satisfied, but I had to get this done. It was real vague." Focusing on the meaning of the information, patients felt that the terms used to describe their heart disease lacked clarity. A patient discussed being told that he should have a catheterization "because there was something going on, he described it then as a leaky heart. I'm not aware just what a leaky heart is or why its called a leaky." While there was no apparent variation by race in patients' perception of the information received, black patients repeatedly expressed a preference for building a relationship with physicians before agreeing to an invasive cardiac procedure. Statements made by black patients were "to make a decision on a major operation, even to angioplasty for me, I maybe would have agreed to that if I had known the physician doing the treatment," "They didn't know too much about me and my health status." CONCLUSION: We found problematic aspects of the patients' experiences regarding communication of cardiac testing. Our findings suggest that although patients desire clarity from physicians, they are often confused regarding the information received. Both a lack of substance and vagueness of the information received may be linked to feelings of mistrust toward physicians when considering further diagnostic testing. Black patients appeared more concerned with building a relationship with physicians before agreeing to invasive testing. The data for this study comprise 16 videotaped consultations from an Internal Medicine clinic and 21 audiotaped consultations from an Oncology clinic, both based at large university medical centers. The visits ranged in length from 3 to 58 minutes. Combining the data, fourteen physicians and thirty-seven patients were recorded. Each clinic visit was transcribed in detail. Analysis was undertaken using the qualitative methodology of conversation analysis which allows understanding of the sequential and substantive organization of talk and interaction. RESULTS: We find that patients verbally respond to physician recommendations in three different ways. First, patients may respond to clinical recommendations regarding their care with unconditional agreement such as "Okay. Fine". Patients thereby ratify the physician's suggestion and let it stand as the plan to be followed. Second, patients may use weak displays of resistance, such as silence, after a recommendation to signal that something is problematic about the doctor's suggestion. Physicians then commonly reformulate the recommendation to make it more acceptable to the patient. Third, patients may offer counterproposals which show disagreement with the physician's recommendation. These counterproposals may be accepted by the physician or may prompt him/her to justify the original recommendation. CONCLUSION: We demonstrate that on the level of individual medical encounters, patient agency is both apparent and operative, and that physicians are not unilaterally determining treatment plans. Physicians may be contributing to differential treatment approaches by recommendations which are based on their knowledge of standard therapies combined with their perceptions of patients, but patient preferences also affect selection of disease management. What emerges from our data is a demonstration of how treatment plans are negotiated during the medical encounter, how physicians are sensative to subtle interactional cues, and thus how patients contribute to the final formulation of recommendations. Hence, the outcome or product of the medical encounter (the treatment plan) can be associated with patient as well as physician actions. In clinical practice, physicians attuned to these different patient responses will be able to benefit by understanding how to more actively engage patients in negotiating final treatment plans. Open-ended individual 60-90 minute interviews were audio-taped, transcribed, then analyzed and coded through repeated close readings. RESULTS: The 17 participants (9 men, 8 women), had a mean of 3 years of postgraduate training. Eleven participants were primary care residents and 6 were surgical or subspecialty residents. Five themes emerged from the analyses. 1) All participants were continually aware of the low numbers of Black physicians. As a result, participants felt highly visible and closely watched, which led some to feel "uptight" and "self conscious." In addition, 13 ⁄ 17 participants reported feeling socially isolated in their training programs, which led to depression for some and for others, a heightened focus on family and community for support, rather than their professional colleagues. Also, participants reported incidents of mistaken identity (e.g., nurse or orderly). Some reacted by ensuring they could easily be identified as physicians. 2) Fourteen participants believed that if they performed poorly, they would be assessed unfairly or treated more harshly than their majority colleagues. Fearing harsh treatment, participants worked harder than expected and carefully avoided making mistakes. 3) Fourteen participants perceived themselves as representing all Blacks to non-Blacks, which motivated them to set higher standards for themselves to pave the way for future Black trainees. Our participants were viewed with pride by Black patients and received special treatment by the Black support staff, which reinforced their belief that they represent their race. 4) The 15 participants who had mentors found them to positively influence their self-esteem and career choices. Black mentors' familiarity with Black experiences were particularly helpful. 5) Lastly, 13 participants described specific unpleasant interactions with their majority colleagues which they perceived as discriminatory yet they doubted their perceptions. CONCLUSION: Black trainees' unique experiences draw from their small numbers, their visibility, a sense of expectation from majority physicians as well as a drive to exemplify their race. These parallel experiences and perceptions of Blacks in American society. Residency program directors need to be aware of these unique pressures in order to improve medical training for Black physicians. Suggestions to improve the reliability and validity of chart review include training and sharing multiple viewpoints to increase mutual understanding of a case. "Structured implicit" methods direct an experienced reviewer's attention and judgement to specific processes of care. It is not clear if these added structures eliminate decision-making heuristics based on anecdote or prejudice. METHODS: As part of a larger study across 8 VA hospitals, we taped and transcribed eight reviewers as they discussed initial quality ratings in 95 patient charts. All reviewers were board certified/board eligible in internal medicine and were trained in structured implicit methods for a study of adverse events related to inpatient laboratory abnormalities (hypokalemia, hyperkalemia, renal failure, hyponatremia and digoxin toxicity). We used qualitative grounded theory methods to build a coding structure for systematic analysis using Atlas software. RESULTS: The reliability of the ratings of quality of care and preventability of complications was between 0.2 and 0.4, which is comparable to other published studies. As demonstrated by quotes, several themes emerged among the reviewers' discussions about quality. Reviewers were averse to extreme ratings (e.g., "I don't like to say definitely to anything," or "nothing is 100% preventable"). A reviewer often claimed special expertise for a question, and their partner usually de-ferred to their opinions. The reviewers compared how they would have done in the situation depicted in the chart, and were reluctant to "second-guess" the clinicians. They commonly cited unrelated favorable events in the chart to excuse a specific instance of poor quality, such as globally good care during an episode of hypokalemia. They cited positive and negative prejudices based on identifying information in the chart, such as a certain hospital, physician, or specialty service (e.g., surgery). CONCLUSION: Our implementation of structured implicit review was comparable in design and achieved reliability to the published literature using this method. Despite the specific training to objectively focus reviewers on particular processes of care, reviews still displayed biased heuristics or emphasize irrelevant or prejudicial information when grading quality. Implicit review is a valuable tool for research, but there appears to be room for further improvement in the methods used to reduce bias and improve reliability.

LEARNING OBJECTIVES: 1. Recognize starvation as a risk for potentiating acetaminophen-induced hepatotoxicity.

2. Distinguish acetaminophen-induced hepatotoxicity from other causes of acute hepatitis. CASE PRESENTATION: A thirty-six year old woman was admitted with a four day history of fatigue, weakness, anorexia, fever, diarrhea, nausea, jaundice and glossitis. Two days prior to admission the patient developed dysparunia and abdominal pain and was prescribed acetaminophen with codeine (Tylenol #3) and TMP/ SMZ (Bactrim) by her primary physician. Her anorexia was compounded by both her glossitis, which made eating painful, and her habit of starving herself to control weight. She had a history of alcohol abuse but reportd thirty days of sobriety.

In the Emergency Department the patient was febrile at 39.2c with a pulse of 111. She was alert, extremely thin, with icteric sclera, right upper quadrant tenderness, a mildly enlarged liver, and cervical motion tenderness on pelvic exam. No ascites, petichiae or spider angiomas were noted. Labs included AST 2434, ALT 1986, alkaline phosphatase 199, total bilirubin 3.5, direct bilirubin 3.2, PT 15.0, PTT 39.5, and INR 1.3. The patient was admitted with a presumptive diagnosis of acute hepatitis. She was monitored, hepatotoxins withheld, and antibiotics started for treatment of presumed concomitant PID. Results of hepatitis panels, HIV and P24 tests, GC and Chlamydia cultures were all negative. Upon further questioning, the patient revealed that she took four Tylenol #3 each day (1.2 grams of acetaminophen) for two days prior to admission, but had also been taking four hydrocodone/acetaminophen (Vicodin) a day (2 grams of acetaminophen) for two weeks since a dental procedure. An actaminophen level was then obtained. Thirty-six hours after admission the level was two. Acetylcysteine (Mucomyst) use was discussed but not given since the patients' transaminases were normalizing with supportive care. She was discharged 10 days after admission. DISCUSSION: 1. Starvation is an important cofactor that increases a patient's susceptibility to acetaminophen-induced hepatotoxicity. Ten grams of acetaminophen in 24 hours is toxic. As little as 4 grams in 24 hours is toxic in alcoholics. Deliberate (dieting) and unintentional (pain with glossitis) starvation predisposed this patient to succumb to low dose acetaminophen-induced hepatitis. The physiologic effct of starvation on acetaminophen-induced hepatotoxicity will be discussed in further detail. 2. Even without an elevated acetaminophen level, the high AST (seen in toxic acetaminophen ingestion and viral, but not alcoholic, hepatitis) was a clue to the final diagnosis. Careful histories can decrease time-to-diagnosis and insure appropriate treatment with acetylcysteine. A review of the evidence for the treatment of unintentional acetaminophen toxicity with acetylcysteine will be presented. METHODS: Two patients a week were randomly selected from among all deaths on the medical service of two large teaching hospitals. The attending, resident, and intern who cared for the patient at the time of death were selected for study inclusion and were contacted within 2 weeks for a semi-structured interview that lasted approximately 80 minutes. The response rate was 88 percent.

The interview incorporated open-ended questions and quantitative measures to assess the physician's experience in caring for the dying patient. The following dimensions were explored: physician-patient relationship, relationship with the family, response to the patient's death, and perceived quality of care. Interview transcripts were analyzed using a grounded theory approach. Members of the research team independently read each transcript and identified themes which were discussed as they emerged. The themes were then compiled and used in coding. Atlas/ti, a qualitative analysis software, was used to assist in the coding and analysis of the transcripts. RESULTS: Of 153 physicians, 46 interns, 45 residents, 50 attendings, and 12 other physicians cared for 49 patients. Physicians' relationships with patients were brief and not considered close. Physicians reported stronger ties with fami-lies. Physicians were most likely to characterize the dying process as positive when death was anticipated and/or viewed as timely; when patients and family were regarded as rational; when patients and family goals were concordant with those of the medical team; when error or adverse events were absent; and when physicians felt connected to families or experienced their gratitude. Deaths were regarded as difficult when unexpected and untimely; when family and medical team were in conflict about decisions; when death was hastened through active interventions; and when physicians either identified with or disliked the patient or family. Deaths that were perceived as both good and bad elicited distress and selfdoubt about the quality of care delivered. CONCLUSION: Physicians have clear ideas about what constitutes "good" and "bad" dying. Experiences of discontinuity in care, lack of closeness in relationships with patients and family, and concerns about error when caring for dying patients are barriers to competent palliative care.

KM Greifer.

LEARNING OBJECTIVES: To recognize the presentation of Pancoast tumors and to review the diagnostic work-up of cervical radiculopathy CASE PRESENTATION: A 52 year old previously healthy woman presented to her primary physician with right sided neck pain radiating to her right upper extremity and hand. Physical exam was unremarkable including neurologic exam. Cervical spine XRAYs demonstrated neuroforaminal narrowing at C5-C6. Anti-inflammatory medication was prescribed. A few months later the patient was seen again with similar complaints although she stated her pain had not imroved, and had in fact worsened. Her physical exam again was normal. She was referred to orthopaedics and magnetic resonance imaging (MRI) of the cervical spine demonstrated mild disc disease with minimal encroachment at the C5-C6 level. She was sent for physical therapy. Her pain continued to worsen and she was subsequently sent to a pain specialist for cervical epidural injections. By then, the patient was on narcotic analgesics for the pain. The injections failed to relieve her pain and she began to complain of parasthesias and weakness in her right upper extremity. A neurosurgical consult was obtained but it was felt that the disease seen on her MRI was not severe enough to warrant surgery. Subsequently an electromyelogram/nerve conduction study was done. This study raised the possibility of a brachial plexopathy rather than a cervical radiculopathy. At this point the patient had overt weakness in her grip strength. An MRI was done of the brachial plexus which revealed a large tumor encasing the brachial plexus and subclavian artery arising from the apex of the right lung consistent with a Pancoast tumor. Chest/ abdomen commuted tomography showed only the large tumor and no evidence of other disease or metastasis. Biopsy subsequently showed non-small cell lung cancer. DISCUSSION: Pancoast tumors can present with neurologic findings-more commonly with a unilateral Horners syndrome. This patient demonstrated another presentation. It is important to recognize that Pancoast tumors often do not present with cough or hemoptysis or shortness of breath as do other lung tumors. In addition, this patient had pain out of proportion to her radiographic findings. Nerve conduction studies should be considered early in the diagnostic work-up where there are questions as to the correlation between radiography and symptoms. (1) the types of problems that arise in patient encounters (2) how physicians address these, and (3) the ways that they experience "trust" in the patient-physician relationship. METHODS: Qualitative methodology consisting of semistructured interviews with a convenience sample of 22 primary care physicians (19 internal medicine, 2 family practice, 1 pediatrics) in two cities. Interviews were audiotaped, transcribed and were analyzed for significant themes using grounded theory methods. RESULTS: The sample reported an average of 11.3 (sd 7.1) years in practice and saw an average of 86.2 (sd 31.8) patients per week, of whom 52.5% (sd 11.7) were in capitated managed care plans. All of the physicians reported that managed care had changed their relationships with patients in ways that ranged from increased "hassle" on non-medical tasks to a more profound mistrust that they perceive from their patients. The specific areas that physicians identified as problematic included formulary restrictions, specialty referrals, diagnostic testing, and decisions around hospitalizations. While some physicians viewed disclosure of financial arrangements as a remedy to address these conflicts, they varied in their opinions about how and when to disclose such information to patients. Almost all believed that the managed care plans themselves should be clearer in their disclosures. Physicians' in this sample largely defined "trust" as the assumption that they act on their individual patients' best interests and worried that in the current environment, the "feeling that your physician is your advocate . . . is not nearly as solid as it once was." CONCLUSION: Managed care raises troubling dilemmas for this sample of physicians in their relationships with patients in areas ranging from administrative hassles to diagnostic decisions. Many physicians see previously fulfilling relationships being replaced by mistrust in their roles, in the words of one subject, "holding both the purse strings and the prescription pad." Although disclosure of financial arrangements represents one potential remedy, further research is needed to determine whether or how these discussions should take place in the setting of the patient-physician relationship. RESULTS: Four hundred thirty one distinct comments were generated with 95% falling into one of eight broad categories: 1) Patients' description of their spirituality/religiosity 2) Church attendance 3) Mention of God/Holy Spirit 4) Prayer 5) Physician-patient communication 6) Patient preferences of whom they would prefer to discuss spirituality with 7) Context for discussion and 8) Patient recommendations to physicians. About two-thirds of patients described themselves as spiritual, half described themselves as religious, and a quarter as neither. God and prayer were often mentioned as a source of comfort, guidance, or healing. Twenty percent of comments pertained to physician-patient communication, especially psychosocial and emotional aspects such as respect for a patient's individuality and whether a physician engendered trust, comfort, and hope. Physicians with strong interpersonal skills were viewed as spiritual. Half the participants felt that physicians should talk about spiritual beliefs; this dialogue was believed to be a valuable part of the healing process. The other half felt physician-initiated conversation was not expected, but would be welcomed. Patients had different views about the context for these discussions. Some felt the appropriate setting was during serious illness; others felt discussion at this time would imply a poor prognosis. They believed the topic would be better broached during routine care. Participants recommended that physicians ask patients about their use of spirituality and sources of support, refer them to spiritual counselors, or pray with them. CONCLUSION: Patients are interested in discussing spiritual and/or religious issues with their physicians. This can be accomplished through an open and enhanced doctor-patient relationship. Future work should explore patient characteristics associated with preferences for spiritual discussion. PURPOSE: Mentoring relationships are often cited as key to the development of productive academic careers in medicine, yet the characteristics of these experiences, particularly how they fail, are poorly understood. Description of mentoring quality and variation has been hampered by the lack of survey instruments applicable to academic medical faculty. We have completed preliminary qualitative research to develop a comprehensive survey instrument. METHODS: The instrument was developed after review of the mentoring literature, existing surveys, and the analysis of a focus group. Focus group participants, chosen for spectrum of mentoring experiences, were recruited from among prior respondents to the Faculty Advancement Study Questionnaire (1995) and were asked to define and describe optimal and sub-optimal mentoring. The focus group was comprised of 7 academic faculty-4 men and 3 women, 4 clinical and 3 basic science faculty. Six participants had more than ten years' experience at their current institution. Focus group proceedings were audiotaped and analyzed by five readers who identified key words, phrases, and topics, grouping them by consensus into major themes. RESULTS: The 6 major themes were: 1) Critical mentoring situations-dealing with failure, promotion, negotiation, and publishing. 2) Critical mentoring functionsassisting the mentee in balancing professional and personal lives, aiding in networking, providing feedback, and fostering growth as an academic professional. 3) Critical mentor roles-working as a career strategist, helping the mentee develop an independent academic identity, avoiding exploitation of the mentee, holding the mentee to a high achievement standard, acting as an advocate. 4) Dynamics of the mentoring relationship-mentor learning from the mentee, producing career change for both, taking risks with traditional boundaries. 5) Effects of gender and race-including gender bias, difficulties of cross-gender mentoring, importance of minority mentors when preferred. 6) Resource recommendations-building visible, formal, mandatory, systems for mentoring, formally rewarding mentoring effort. CONCLUSION: Better understanding the elements of successful mentoring may assist in promotion of these relationships where they exist and provide resources to foster them where they do not prevail. We have built, pre-tested, and are fielding a comprehensive survey instrument to study the academic mentoring experience which we will make available to other investigators.

Education, Oakwood Hospital, Dearborn, MI PURPOSE: Applying triage pathways in the ambulatory setting can result in substantial healthcare savings. METHODS: Great variability exists in practice pattern amongst physicians in the ambulatory clinics. Patients with an acute problem may wait for an appointment 1-2 days to be seen, thus delaying therapy and potentially complicate outcome.

Some physicians treat simple problems through phone triage, others do not. Variability also exists in the choice of antibiotic and in the length of therapy. A task force was empowered to set up an acute cystitis pathway. A triage nurse phone interviews patients with suspected UTI. The goal is to establish the diagnosis and to exclude any systemic or complicating factors that might prevent treatment over the phone. The nurse discusses results with physician, then calls prescriptions to the pharmacy and instructions to the patient. Results: 43 patients were triaged, 39 were treated over the phone,. Three were asked to be seen and one had no UTI. None of the patients enrolled had any complications. Time lapse from complaint to prescription called in was one half hour to 4 hours. Total savings in this trial was $5088. Since we see an average of 25 UTI's per month, the potential savings is $42,000 per year. RESULTS: When the acute cystitis pathway is applied to low risk patients, with proper screening tools and back up, there is a very significant savings to our health care system. It also results in improved patient satisfaction, speed of therapy and lack of time off from work along with good outcome. It allows physicians to concentrate on patients with higher acuity of illness. We recommend using this model in all ambulatory settings. METHODS: Four focus groups were conducted in September, 1999. Potential participants were identified by random selection from a list of all women who received mammograms or breast biopsies at our institution during December, 1998. Patients were excluded if they were diagnosed with cancer, unable to give consent, or did not speak English. The groups included 1) women with normal mammograms, 2) women with abnormal mammograms requiring 6 month follow-up, 3) women requiring biopsies, and 4) women of lower socioeconomic status as defined by insurance status (general assistance/Title 19) at the time of the mammogram. All focus groups were run by a female professional facilitator with experience in breast cancer research. Discussions addressed anticipation of the mammogram, waiting for results, and health related behavior following the mammogram. Transcripts of the sessions were independently coded and analyzed by two of the investigators. RESULTS: A total of 45 women participated in the 4 focus groups. A wide range of opinions were expressed. Prior to having the mammogram, several women admitted to being "scared", or "concerned" about finding cancer. Worry about pain from the test was also expressed. Others described feeling "good" about having finally scheduled the procedure. Waiting for the results, many women "think about it a lot" and are quite anxious. After news of a normal mammogram, some women are more serious about self exams and health maintenance while others feel "covered" and become "lazier" about self exam. Women with abnormal results requiring 6 month follow-up had reactions ranging from "never thought about it" to substantial concerns that affected how they responded to minor irritations, how they treated those around them and their desire for knowledge about breast cancer. The most striking responses were seen in women who were advised to undergo a biopsy. These women had intense experiences that had an impact on their physical health, emotions, work, family interactions and sexual relations. The nature of an individual's response seemed to be related to both patient factors and system issues. Patient factors included faith in doctors,technology or religion; previous health experiences and other stressors at home. System issues included the doctor's attitude, nature of explanations, counseling, provider consistency and time spent waiting for results. CONCLUSION: These focus groups identified a variety of ways in which women perceive that mammograms impact their emotional and physical well being. While these findings cannot be generalized to the population at large, they can be used to guide future quantitative research into the impact of screening mammography and develop approaches to mitigate adverse effects.

LEARNING OBJECTIVES: 1. Recognize gynecomastia as a possible complication of antiretroviral therapy 2. Adopt a conservative approach to the diagnostic evaluation of a breast mass found in a male taking antiretrovirals for the treatment of HIV infection CASE PRESENTATION: A 48 year old male with HIV infection presented with a breast mass first noted 3 weeks earlier. He denied any history of trauma, nipple discharge, weight loss, or the use of non-prescription medications or recreational drugs in the past 3 years. He had been treated with zidovudine and didanosine for 2 years, but was switched to nelfinavir, stavudine, and lamivudine 3 months ago. He had a good immunologic and virologic response to the new regimen, achieving a viral load Ͻ25 copies/mL. His only other medications included fluoxetine and trimethoprim-sulfamethoxazole. On examination, he had a hard, fixed, tender mass superior and lateral to the right areola, but no adenopathy. He also had truncal obesity, wasting of the cheeks, and atrophy of the extremities. Testicles were normal in size. Because his presentation was unusual for gynecomastia and concerning for malignancy, he was evaluated by a multidisciplinary team which included an oncologist, radiologist and surgeon and subsequently underwent a right total mastectomy, without an antecedent biopsy or mammogram. The histo-pathology of the 3 ϫ 1 cm mass removed was consistent with gynecomastia. He developed a similar mass in his left breast approximately 5 weeks after the first but did not report it for 1 year. At that time, serum concentrations of free testosterone, human chorionic gonadotropin, estrone, estradiol, androstenedione, prolactin, and thyroid stimulating hormone levels were within normal limits. DISCUSSION: The syndrome of lipodystrophy is being increasingly recognized as a complication of protease inhibitors. In men, the most common manifestations include truncal obesity, buttock, extremity and facial wasting, and a dorsocervical fat pad. Breast hypertrophy, although a common complication among women, has been reported infrequently in men. This was a case of gynecomastia presenting as a unilateral, eccentric mass which was most likely due to antiretroviral treatment and resulted in a major operation. Only 5 cases of gynecomastia secondary to HIV antiretroviral treatment have been documented in men. Gynecomastia has not previously been described with the use of nelfinavir, nor does it usually present unilaterally. HIV infection alone can cause gynecomastia, however the prevalence of gynecomastia in HIV infected men has not been reported. Medical providers should identify antiretrovirals as a possible cause of a breast mass in males and choose a more conservative approach, such as needle biopsy, for the initial diagnostic evaluation. Grounded Theory methods were used to develop the theory and conclusions. RESULTS: Of the 26 subjects, 10 were women and 16 men, with a mean age 42 years. Other characteristics included 68% prior or current homelessness; 77% minority ethnicity; 13 mean years of illicit drug use; drug of choice: heroin 50%, cocaine 50%; 3 mean years since initial HIV test (range 0.1-11). At the time of HIV testing 19 ⁄ 26 were using illicit drugs daily as compared to 5 ⁄ 26 at the time of the interview. All had some periods in recovery since testing positive. For subjects in recovery at the interview, testing positive for HIV was one of a number of 'wake-up calls' that shifted their awareness only of the next high to a sense of past, present and foreshortened future. In some cases this 'wake-up call' was precipitated by developing symptoms from HIV, long hospital confinement, or incarceration. Some saw a premature death as a motivation to stop using drugs because they did not want to die an addict. All participants held the belief that using drugs would worsen their HIV status, either directly through illicit drug effects on the immune system or indirectly through poor self-care. This belief was an important motivation to decrease drug use. They were proud of their efforts to get responsible including getting off the streets, eating regularly, and keeping appointments. The longevity promised by antiretroviral therapy further motivated many to develop a stable lifestyle in order to benefit from medication. However, suicidal ideation and depression in 13 ⁄ 26 participants dampened this motivation. Subjects reported access to HIV priority housing and residential treatment programs as important resources for recovery. Subjects viewed health care providers as key members of their support system, in part, because they were frequently isolated from family. CONCLUSION: Testing positive for HIV among drug addicts appeared to interrupt a long-standing lifestyle and, paradoxically, to provide potential for a better future. Exchanging addiction for the promise of health, in combination with available social resources, helped decrease subsequent drug use and increase stable lifestyles. BACKGROUND: Adherence has been a key issue determining the success of treatment in patients with HIV/AIDS. Studies examining adherence to highly active antiretroviral therapy have focused on predicting adherence rates within demographically defined groups of patients. However, few have attempted to identify and study patients who demonstrate excellent adherence. Understanding the factors that have enabled these patients to achieve superior adherence rates may help to identify new strategies for enhancing adherence to HAART. OBJECTIVE: To examine the attitudes and behaviors of HIV/AIDS patients with excellent adherence and how they differ from those with suboptimal adherence. METHODS: A diverse group of patients with HIV/AIDS on HAART were interviewed, using a semi-structured qualitative format. We compared patients with excellent adherence to those with suboptimal adherence in terms of motivators, patient-provider relationship, substance abuse, social support and mental health. Patients categorized as having excellent adherence were those who reported taking their medication consistently 90-100% of the time or who hadn't missed a dose within 3 months of the interview. Patients who reported taking their medication less than 75% of the time were categorized as having suboptimal adherence. RESULTS: There were 43 patients: 63% were males, 37% nonwhites, 44% former substance abusers, 40% active users and 16% nonusers. 25 patients were excel-lent adherers and 18 patients were suboptimal adherers. Patients with excellent adherence rates voiced the following themes when compared to less adherent patients: 1) They believed that adherence rates needed to be 90-100% for the medication to be effective; 2) The primary care provider was regarded with a great degree of respect and trust; 3) They continued to follow their medication regimens even when actively using alcohol and/or drugs; 4) Patients were open with family and friends about their HIV status. They received substantial support from their families and attended support groups; 5) Patients cited staying healthy as their key motivator. Those with excellent adherence were not actively depressed; all had normal CESD scores. They expressed a strong desire to live and actively took steps to manage feelings of depression. In contrast 31% of suboptimal adherers were actively depressed. Of the excellent adherers, 16 were former substance abusers, 4 were current users, and 5 nonusers. Of the suboptimal adherers, 13 were current substance abusers, 3 were former users and 2 nonusers. CONCLUSION: Those with HIV/AIDS who demonstrate excellent adherence to HAART differ from their less adherent counterparts in terms of key health-related attitudes and behaviors. Former substance abuse does not appear to be an obstacle to achieving excellent adherence. Identifying and studying excellent adherers appears to provide new insights and strategies for enhancing adherence to HAART. We recruited forty-two breast cancer survivors in Cleveland who were at least one year from last treatment to participate in 8 focus groups. Women were introduced to the idea of a "process diagram" outlining their steps through the process of breast cancer diagnosis, treatment, and survivorship. We explored if, where, and why women experienced problems with information, support, or communication. All women completed an exit survey, and all transcripts were read and coded by the PI, a research assistant, and a qualitative analyst. RESULTS: The mean age of participants was 56, 85% were white, 50% had incomes over $50,000, 49% underwent mastectomy, 50% received chemotherapy, and 31% had breast reconstruction. We identified three transition points where women said their needs were not being met: 1) the transition from wellness to sickness, 2) the transition from being uninformed to making informed decisions, and 3) the transition from acute treatment to follow-up care. The transition from wellness to sickness occurred at diagnosis. Women repeatedly said that they received too much information and not enough emotional support here. The second transition occurred gradually over weeks to months as women had to make many complex decisions about surgery and adjuvant treatments. Most women described an "information learning curve" here The third transition occurred when active treatment ended. Here, all women reported being eager for information but receiving very little. No woman reported being "prepared for survivorship". CONCLUSION: We identified three transition points in the process of breast cancer care in which women's needs for information and support changed in predictable ways and were not being met. These findings and participant feelings indicate that care could be improved with relatively simple interventions that involve tailoring the delivery of information and support to the patient's stage along the diagnosis/treatment/survivorship continuum.

LEARNING OBJECTIVES: Objective: (1) Consider metabolic myopathy in the differential diagnosis of a patient with rhabdomyolysis. (2) History of exercise intolerance and exercise-induced cramping is highly suggestive of an underlying myopathy. CASE PRESENTATION: Patient is a 23-year-old, healthy, male graduate student from Thailand, who presented with generalized muscle aches. One week prior to the onset of his symptoms he initiated an intense exercise program-situps, running, and lifting weights 2 hrs per day for 3 days. Lab studies showed a total creatine kinase (CK) of 6000. He was managed conservatively with oral hydration. Ten days later he developed left arm swelling, difficulty in rising from a chair, walking upstairs, and combing his hair. He was found to be in rhabdomyolysis with total CK increased to 21,600. On exam the patient was found to have weakness, generalized muscle tenderness and edematous extremities. A doppler ultrasound study of the left brachial vein revealed a deep vein thrombosis. He was admitted to the hospital where intravenous hydration and rest were initiated. Muscle injury was evident with high CK, transaminases, Aldolase, and Myoglobinuria. Further questioning revealed exercise intolerance during sports requiring endurance and speed and longer recovery from muscle cramps. He never noticed dark urine. His father and paternal uncles had similar complaints of longer duration of muscle cramps and decreased endurance. His father had proximal hip girdle weakness elicited on exam but with normal CK. There is no documented case of myopathy in the family. The patient denied any history of drug abuse. In laboratory data-TSH, T3, T4 were normal and ANA, ANCA, Ds-DNA, and monospot were negative. A workup for hy-percoagulable state was negative, and the DVT was attributed to exertion. Eight weeks later CPK still fluctuated around 1000. Muscle biopsy was then performed which showed muscle necrosis. Biochemical analysis of his muscle biopsy revealed Phosphofructokinase (PFK) deficiency. In 4 months CPK normalized. Discussion: In this patient, starting a new exercise program precipitated an acute rhabdomyolysis. This, in conjunction with a prior history of exercise intolerance strongly suggested an underlying myopathy, leading to muscle biopsy confirming the diagnosis of PFK deficiency. PFK deficiency is a rare autosomal recessive disease, which inhibits muscle capacity to regenerate ATP through glycolysis. Hence during increased metabolic demand muscle uses fatty acids and ketones as alternative energy source. Exercise tolerance can be enhanced by a slow warm-up or brief periods of rest allowing for the start of second wind phenomenon (switching to utilization of fatty acids). A high fat diet should theoretically enhance the muscle energy, however the clinical effectiveness is only anecdotal. BACKGROUND: Adherence to Highly Active Antiretroviral Therapy (HAART) in substance abusers with HIV/AIDS is generally considered to be inadequate, and detailed studies are lacking. Since adherence to medical regimens is a complex behavior, insight into factors that influence adherence in this population may be gained through qualitative research methods. OBJECTIVES: To investigate factors and identify themes that may influence adherence among substance abusers with HIV/AIDS. MEHTODS: We interviewed 44 persons with HIV/AIDS using a semi-structured instrument. The sample was chosen to represent a broad spectrum of substance abuse patterns. Of the 44 patients, 42 were receiving HAART and 2 were receiving other antiretroviral regimens. The interview was partially based on the Health Belief Model of patient behavior. The interviews were taped, transcribed, and analyzed using ATLAS text indexing software. Transcripts were analyzed by two of three researchers. RESULTS: Four major patient beliefs emerged regarding substance abuse and adherence to HAART: 1) alcohol and drugs interfered with the actions of antiretrovirals, 2) alcohol and drugs were so detrimental to health that it was futile to take medications concurrently, 3) the side effects of HAART worsened hangovers and withdrawal symptoms, and 4) current drug or alcohol use made medication adherence a low priority. However, independent of substance abuse status, many subjects reported that their adherence improved over time. Reasons given for this included: 1) incorporating medications into their daily routine took time, 2) major life changes renewed interest in their health and led to increased adherence. CONCLUSIONS: This study generates several hypotheses regarding adherence to HAART. Substance abusers with HIV/AIDS may have specific misconceptions about HAART and drugs of abuse which, when addressed, may enhance adherence. Clinicians and researchers assessing adherence need to take into account that adherence may change over time. PURPOSE: Asian Americans (AA) are the fastest growing minority group in the United States, with a 400% increase in the last three decades. Despite this rapid growth, little is known about AA's perspectives on the components of high-quality care, especially among those who are Limited-English-Proficient (LEP). We used patient group interviews to evaluate quality of care from the view of LEP-AA patients(pts). METHODS: LEP-AA pts of Vietnamese origins were recruited from an urban community health center. We developed a moderator guide and trained bilingual group moderators to conduct interviews. We assessed quality of care using the domains of the Picker Instrument for Outpatient Care, with additional domains assessing the quality of interpretation services and patient-provider communication about the use of non-Western medicine. We interviewed 22 pts (mean age 46, 50% female,100% foreign born, 95% LEP, 58% completed high school) in 2 gender-specific groups. Video recordings and transcriptions were prepared for each group interview. One investigator reviewed these recordings and identified recurring themes about the attributes of quality health care. RESULTS: Quality of interpretive services was considered to be an important aspect of care. LEP-Vietnamese-American(VA) pts prefered using trained interpreters rather than family members. For older VA pts, gender concordance of interpreters was reported to be very important. LEP-VA pts valued providers and interpreters who show "respect" to the pts, demonstrated by having a welcoming and courteous manner. They also valued providers who show knowledge of non-Western medical practices, such as "coining" or "cupping." Many feared disclosing the use of these practices because of negative responses they have previously received from their providers, but felt that the culturally competent provider would understand, if not condone, these traditional practices. They valued providers who are conscientious in caring for pts' social and psychological needs as well as medical needs, and wanted specific treatment recommendations at the conclusion of visits. VA pts also rated highly having adequate access to their providers, adequate time with their providers, and having good continuity of care.

CONCLUSION: LEP-VA pts regard having access to professional medical interpreters as an important indicator of quality care. These pts define a culturally competent provider as one who is knowledgeable about non-Western medical practices and is not condemnatory toward the pts for using these practices. LEP-VA pts also value having adequate access to provider and good continuity of care as two important quality domains. To determine whether Objective Structured Clinical Examination (OSCE) scores of medical students reflect actual patient satisfaction in a university hospital outpatient clinic. METHODS: In 1999, an OSCE which consisted of interview, general physical examination and neurological examination stations was administered to 81 senior year medical students. 281 patients seen in our outpatient general medical clinic rated these medical students by using six items from the American Board of Internal Medicine Patient Satisfaction Questionnaire (PSQ). Students were divided in quartiles according to OSCE scores and relationships to patients' ratings were analyzed. RESULTS: OSCE items showed a good internal consistency with Cronbach's alpha (ƒ¿) of 0.89 as well as the items of PSQ (ƒ¿ ϭ 0.83), but no overall correlation between these two scores were seen. Medical students in lowest quartile, however, tended to be rated lower by patients as well (p ϭ 0.12). CONCLUSION: Medical students' clinical abilities measured by 4 station OSCE did not correlate well with patients' satisfaction as a whole, although lowermost scores might be used to identify under-performing students who require special attention. PURPOSE: Hypertension is a highly prevalent disease in African-Americans. Identifying issues hindering patients from taking their medications as prescribed is essential in order to address the issue of poorly controlled blood pressure in this patient population. The purpose of this study was to identify the barriers faced by a group of hypertensive African-American patients in taking their prescribed anti-hypertensive medications. METHODS: One hundred African-American patients with hypertension followed at an internal medicine primary care practice participated. Computerized medical records were reviewed to identify eligible patients and to obtain demographic data, comorbidity data, and blood pressure. In-depth interviews were conducted with each patient either at the time of ambulatory clinic visits or by telephone. During the interviews, patients' experiences with anti-hypertensive medications and their perceptions about the challenges they face in taking their medications were explored.

Some of the open-ended questions asked included: 1) What difficulties do you think people with high blood pressure have in taking their medications? 2) What has been your experience with taking high blood pressure medications? 3) Are there things or situations that make it difficult or easier for you to take your blood pressure medications? and 4) What do you think are the potential risks and barriers of taking your high blood pressure medications? All responses were recorded verbatim during each interview and then sorted and categorized using qualitative analysis. RESULTS: Of the 100 patients interviewed, mean age was 55 years (range 28-80), mean years of education was 13, and 60% were women. The mean systolic blood pressure was 140 mmHg, and the mean diastolic was 85 mmHg. The most important issues identified by patients that hindered them from taking their medications as prescribed were: forgetfulness (28%), side effects (21%), dislike for pills (19%), and inconvenience (13%). Others included: fear of dependence (10%), swallowing big pills (10%), absence of symptoms (7%), reluctance to take their medications in social situations (6%), cost (4%), preference for folk remedies (4%), unable to refill prescription in a timely manner (4%), and too busy to take medications (3 %). CONCLUSION: Patients cited multiple, varied issues that hindered them from taking their anti-hypertensive medications. Identifying these issues will guide the development of specific interventions aimed at helping patients overcome them and thus lead to better control of their hypertension with improved long-term outcomes. METHODS: 20 in-depth, semi-structured, open-ended interviews were conducted with patients with evidence of high-risk sexual behavior. Subjects were selected from patients from the investigators' clinical practices with self-report of an STD, multiple sexual partners, or unwanted pregnancy in the past 5 years, or with patient expressed concerns about consequences of recent unsafe sexual activity. Interviews were audio-taped, transcribed and analyzed independently by each coinvestigator. Major themes were identified by iterative consensus coding. RESULTS: In almost every instance, subjects described both safe and unsafe sexual experiences contextually in terms of their relationship with the sexual partner. Subjects did not spontaneously focus on specific sexual behaviors, nor on condom use, not even with specific questioning. Subjects frequently articulated a hierarchy of relationships, with condom use having different meanings and requiring different styles of negotiation in different types of relationships. In relationships "just for sex", talking about safer sex was perceived as more interpersonally intimate than actually having intercourse, which often occurred without verbal communication. In these situations, patients who were successful at talking to their partners described the use of rule-based ("I always use protection"), or, alternatively, more playful forms ("No party-hat, no party!") of communication, avoiding the need for intimate communication and the potential for causing offense to the partner. In more committed relationships, the notion of "trusting" the partner frequently took the place of careful risk assessment, resulting in unsafe behavior. Paradoxically, in these situations, patients described the communication about safer sex as potentially destructive to intimacy. CONCLUSION: Counseling which is focused only on behavioral aspects of sexual risk-taking may ignore the major significance of the meaning of the interpersonal interaction in which sexual behavior occurs. Helping patients to understand their own framework of interpersonal meaning and how they construct condom use within relationships, and assisting them in anticipatory communication appropriate to the type of sexual relationship in which they are engaged may be a more productive way to help patients reduce their unsafe practices. PURPOSE: Because qualitative data raise concerns about idiosyncratic interpretation, techniques to minimize and measure it must be developed. We describe one here. METHODS: In analyzing patient interviews about dying, we wanted responses to speak for themselves, as undistorted by coder bias as possible. So we designed a two-stage content analysis to minimize interpretive bias and measured residual idiosyncrasy in the results. Seeking coders likely to bring different perspectives to the analysis, we chose four coders with different genders, ethnicities, and professional disciplines. We tried to promote both a balanced perspective and an interpretive synergy by requiring at each stage two of three coders to read responses blindly, independently, and randomly and then to reach consensus about interpretation. The fourth coder decided persistent disagreements. RESULTS: In the first stage of analysis, coders A and B read all interviews and devised a thematic coding scheme from subjects' own words. The scheme was conservative: Themes suggested by one coder made the scheme only if agreed to by the other. The final saturated scheme had 81 themes. A agreed to 55% of themes suggested by B, and B agreed to 50% suggested by A. These modest agreement rates show the dependence of theme selection on individual coder perspective and, thus, the need for multiple coders and iterations in devising a coding scheme. In the second stage, coders A and C applied the coding scheme to each interview. Coder A ensured consistency of theme concept from the first stage; coder C, a fresh perspective on the interviews and theme meanings. Definitive determination of theme presence required agreement by two coders. A and C agreed on 97% of final determinations of theme presence. The other 3% required Coder D to side with either A or C. Final determinations identified any given theme in 3 to 325 responses to questions. These determinations agreed with 86% of A's initial codings and with 79% of C's, indicating substantial interpretive consistency between coders. CONCLUSION: Thus, qualitative studies can build in steps to minimize and measure idiosyncratic interpretation, and reporting such steps should become standard for these studies. example, most MAs (58%) believed the medical system controls treatment but trusted the system to serve them well (62%) and wanted no life support when terminal (50%). Most EAs (78%) believed patients-not the system-control treatment but trusted the system to serve them well (72%) and had defined for themselves unacceptable outcomes (50%). Most AAs (64%) believed the system controls treatment, few trusted the system to serve them well (36%), and most believed a patient should not express treatment wishes until very sick (50%). CONCLUSION: Thus, advance care planning must take account of both core American cultural values and specific ethnic cultural values. Core values support advance care planning in general, but such planning should be tailored to specific wthnic views about the medical system, its trustworthiness, and terminal care.

A BEDSIDE DIAGNOSIS OF PULMONARY ACTINOMYCOSIS. C Pietrantoni, M Budev, T Mekhail, W Tomford.

LEARNING OBJECTIVES: 1. The bedside diagnosis of Actinomycosis. 2. Recognize the indolent progression of this infection.

A 67 year-old alcoholic presented with a rapidly expanding chest "lump" sustained after a contusion 2 weeks prior. Examination revealed a disheveled male with numerous dental caries and a circumferentially raised 4cm, non-pulsatile, indurated, and erythematous, para-sternal lump. CT scan of the chest revealed a left upper lobe consolidation with extension of a similarly attenuating inflammatory mass into the soft tissues of the anterior mediastinum and chest wall, without osseous involvement. A prompt bedside needle aspirate of the lesion and microscopic examination by the house-staff yeilded the presence of pathognomonic "sulfur granules". Eventually, cultures confirmed A. Israelii with a few anaerobic gram-negative and gram-positive bacilli. Further surgical drainage was initially considered, however serial aspirations resulted in partial resolution of the "lump". Four weeks of metronidazole and high dose intravenous penicillin-G were given, followed by 12 months of oral penicillin. One year later, the lesion completely resolved without complications. DISCUSSION: A. Israelii, a non-acid fast, gram-positive, filamentous anaerobic bacteria, is the most common agent of pulmonary actinomycosis. It should be suspected in chronic, infiltrative, cavitary, and tumor-like mass lesions of the pulmonary paranchyma. Fundamental bedside techniques can assist to establish an immediate diagnosis, without awaiting culture assays. Chest wall lesions consist mainly of granulomatous tissue with extensive fibrosis and small abscesses, originating as a direct invasion from a subpleural lesion. Since once the organism is established locally, infection spreads contiguoulsly in a progressive manner without regard for anatomical barriers. Sinus tracts will often extend from the abscess to the skin or adjacent organs. Overlying skin may suppurate, forming tracts that spontaneously close and reform. Cervicofascial infection is most common, followed by abdominopelvic infection, while pulmonary actinomycosis only contributes to 15 percent of cases. Our patient manifested two pathognomonic features of thoracic actinomycosis, extension to the chest wall and conglomerations of filamentous rods in mineralized tissue, called "sulfur granules". An animal model of infection has not been established with Actinomycosis spp. alone, infection by copathogens supports the concept that additional microbes are important for the initiation of infection. Most mild cases are eradicated by the common use of antibiotics, thereby contributing to the overall rarity and indolent course of this infection. The diagnosis of Actinomycosis is truly the mark of a vigilant clinician. PURPOSE: Spirituality and prayer are important to most people. Two-thirds of the American public believe physicians should talk to their patients about spiritual issues and pray with those who request it. Through the use of ethnographic, in depth interviews, patients were invited to discuss their perspectives on the integration of spirituality into their current medical care. METHODS: This study was the qualitative component of a larger survey research project addressing the same purpose. 600 General Internal Medicine patients from University, Veterans, and County hospitals agreed to complete the survey. From this group, a sample of convenience of 15 inpatients agreed to in-depth interview. RESULTS: A total of 71 transcribed pages produced 16 categories. From this matrix of 113 category entries emerged a prominent theme of "patients' cautionary words about integration of spirituality into medical care." 23 cautionary statements were organized into 7 areas of concern. First, patients perceived the need to find compatibility between the physician's meaning of prayer and their own belief system-"I don't know what kind of religion they are [MDs] . . . that would matter to me." Second, before prayer is offered the physician would be expected to asked for permission. Third, for spiritual care to be meaningful "it cannot be something that is going to be written in a textbook and you follow A, B, C because every one is different, both physician and patient." The fourth and fifth areas of concern are seemingly contradictory perspectives. Introducing prayer could be interpreted as a statement of "how bad" the patients' prognosis "really is." On the other hand, when the situation is desperate and there are seemingly no medical options, prayer is welcomed and accepted even by those who might reject it otherwise. The sixth caution was for physicians not to forget that they can use prayer as a vehicle for promoting hope and "positive feelings." The last cautionary statement was the perceived need for a spiritual ambiance that includes cleanliness, religious paraphernalia (like crucifixes and paintings), and a welcoming spiritual milieu. CONCLUSION: When inviting patients for prayer, physicians are expected to be personal and meaningful not generic or emasculated from a particular faith tradition. At the same time physicians need to carefully avoid specific doctrinal statements that might be offensive to patients. This process needs to begin by a systematic informed consent that forewarns the patient. The need to communicate hope is well served by prayer and spirituality; even those who would reject a physicians' prayer may welcome prayer from a different source. In an effort to reduce very low risk hospital admissions, an interdepartmental team (Emergency Medicine, Medicine and Cardiology) developed a triage decision aid based on a previously validated prediction rule (Goldman, NEJM 1996) . Prior to beginning a 14-week prospective study of its safety and efficiency, we set as a minimum threshold 80% use of the decision aid by ED physicians. During a 17week "phase in" period in 1999, we measured use of the decision aid for all patients meeting eligibility criteria for the study. We also measured changes in number of chest pain admissions to our Observation (OBS) Unit compared with similar weeks in 1998. In weeks 1-8, ED physicians received frequent feedback and encouragement about use of the aid. In weeks 5-8, non-MD ED staff received similar interventions. In weeks 9-13, ED staff were instructed that decison aid use was required before an inpatient or OBS bed could be assigned to admitted patients. In week 14, the Hospital Director personally delivered that same message. RESULTS: The Table illustrates progressive increases in use of the triage aid after sequential interventions during the 17-week phase-in period and its sustained use during the study period.

CONCLUSION: Major institutional changes in the process of clinical care may require both collaborative persuasion and constructive coercion. Impact on outcomes merits more study. PURPOSE: Spiritual well-being has recently been recognized to be important to chronically ill individuals health-related quality of life (HRQOL). This qualitative study sought to construct a conceptual model for spiritual well-being from the perspective of older adults with chronic illness. METHODS: Eight focus groups of older individuals with chronic illness were conducted. A diversity of perspectives was sought through stratification by one or more of the following variables: age (65-74, Ͼ74 yo); race (Caucasian, African-American, Spanish-speaking Latino); chronic illness category (heart disease, lung disease, arthritis, visual impairment). Participants were asked about the meaning of spiritual health; the characteristics of spiritually healthy people; what they did to maintain spiritual health; and how aging and chronic illness impacted on spiritual health. Focus groups were audiotaped and transcribed verbatim (and translated for Spanish-speaking groups), then coded for thematic structure by two independent coders trained in content analysis. "Trustworthiness" (the qualitative equivalent of validity and reliability) was sought through checks on interrater reliability, respondent feedback to the findings, and concurrence by a panel of spirituality experts. RESULTS: Participants expressed a wide range of perspectives on spiritual health. Many participants equated spirituality with formalized religious commitment, although a variety of non-religious elements were also identified. Analysis of the focus group transcripts yielded themes in five major areas, with substantial overlap between them: 1) spiritual orientation (e.g., religious or spiritual practices, connection to God); 2) existential state (e.g., meaning and purpose in life, attitude towards death); 3) interpersonal relationships (e.g., loving and being loved, helping others); 4) outlook (e.g., "glass half full") and 5) affect (e.g., inner peace, joy). Representative quotations for each of these theme areas will be presented. CONCLUSION: Spiritual health is an important dimension of older, chronically ill individuals' HRQOL. Contrary to historical attempts to define spirituality in terms of its religious elements, our analysis suggested that a definition that includes its psycho/developmental elements may better represent patients' own models of spiritual health. Several theme areas identified in this study (e.g., affect, interpersonal relationships) correspond to previously identified HRQOL dimensions (emo-tional well-being, social well-being), although other identified areas (spiritual orientation, existential, positive outlook) are not currently represented in HRQOL assessment tools. Expanding the conceptualization of HRQOL to include spiritual well-being may be appropriate in studies of HRQOL in aging and chronic illness.

Residency; Division of General Medicine, Emory University School of Medicine, Atlanta, GA PURPOSE: Although medical decisions are made daily between physicians and their patients, we know little about the effect of culture on this process. As a first step in understanding possible cultural influences on decision-making, we explored the opinions and attitudes of African American patients in an inner-city hospital as they considered undergoing a medical procedure. METHODS: We conducted in-depth, semi-structured interviews with patients at an inner-city, hospital-based medicine clinic who were considering flexible sigmoidoscopy as a screening test for colon cancer. RESULTS: We interviewed twenty-six African American subjects. Subjects named various reasons for undergoing the flexible sigmoidoscopy, including recommendations of their physicians, fear of cancer, and the desire to detect medical problems early. Most patients wanted detailed information about the test, especially information about pain and safety, the purpose of the test, and test results. Subjects overwhelmingly named their doctor as the most trusted source of health-care information. While full disclosure was extremely important to the subjects, most wanted the decision to be made in part or full by their physician. Trust in the physician was thought to be as important or more important than knowing risks and benefits of the procedure. Most subjects stated that the main purpose of the consent document was to protect the physician in case something went wrong. All subjects stated that they would feel comfortable refusing a medical test that they did not want. CONCLUSION: In current ethical frameworks, autonomy is regarded as the primary principle of medical decision-making. Although patients in our study wanted extensive information about their medical care, most willingly deferred the decision-making to the physician. This voluntary surrender of the decision-making role raises important questions about the value placed on autonomy by African American patients in an inner-city clinic. PURPOSE: The $1.8 billion spent on direct-to-consumer advertisements by pharmaceutical companies now exceeds spending on advertising to physicians. We examined which products pharmaceutical companies are advertising to consumers and what messages are being communicated. METHODS: We performed content analysis of advertisements (excluding the FDA required PDR-like "brief summary" page) appearing in 10 widely read magazines: 4 women's magazines (Family Circle, Ladies Home Journal, Better Homes & Gardens, Family Circle), 3 men's magazines (Sports Illustrated, Men's Health, Gentlemen's Quarterly) and 3 with general readership (Time, Newsweek, People). We examined the first issue of every other month during the year July 1998-July 1999. RESULTS: Direct-to-consumer advertisements were about 3 times more common in magazines targeting women compared to those targeting men or a general readership (mean number of advertisements per issue: 5.1 (women's); 1.5 (men's) and 1.6 (general)). The products most commonly advertised were those designed to address symptoms (e.g., sneezing/runny nose, urinary incontinence, hair loss, impotence, menopausal symptoms) followed by products for treatment (e.g., Alzheimer's, HIV, diabetes, otitis media) or prevention (Lyme vaccine, bone density testing, breast cancer). Most advertisements described the benefit of a medication in vague, qualitative terms (e.g., "Zyrtec works fast") and made an emotional appeal to readers, most commonly to their desire to get back to normal (61%), to make their life better than it is (28%), or to avoid some feared outcome (21%). Half of the advertisements encouraged people to consider a medical cause for their experiences ("is it just forgetfulness, or is it Alzheimer's Disease?"). Only 14% of the advertisements described the benefit of the medication with data (e.g., "66% had visible hair growth"). In contrast to information on benefit, potential side-effects of medications were more often described with data (53%), for example, "side effects occurred in less than 2% of men". No advertisements mentioned cost. CONCLUSION: Direct-to-consumer advertisements present little information to help patients make informed decisions but may stimulate demand for products by implying benefit, generating emotional responses, and encouraging consumers to medicalize ordinary (unpleasant) experiences. A substantial proportion of advertisements also invite patients to share in disease management decisions such as the choice of antibiotic or type of insulin. Whether this approach is in the best interest of the patient or the physician-patient relationship is unknown. Week 1-4 325 50% (48-53) ϩ69%

Week 5-8 331 59% (49-68) ϩ12%

Week 9-13 381 73% (60-85) ϩ36%

Week 14-17 293 89% (83-94) ϩ125%

Week 18-31 1042 82% (58-93) ϩ79%

PURPOSE: African American men in the United States shoulder a disproportionate burden of disease and death compared to the rest of the population. Sadly, many of these men have ambulatory sensitive conditions such as hypertension, diabetes, and HIV/AIDS for which they often do not receive appropriate primary health care. This study was conducted to 1) assess African American mens' concepts of health and 2) perceived barriers to receiving health care. METHODS: Eight focus groups, with 10 participants on average, were conducted with African American men in the Woodlawn community of Chicago. Men were divided into sub-groups including adolescents, elderly men, gay/bisexual men, and HIV positive men. Focus groups were audiotaped, transcribed and reviewed by two researchers to identify dominant themes. RESULTS: Mens' concepts of health included traditional health messages regarding diet and exercise but also a recognition of factors ranging from the social environment to unemployment that affect health. They perceived a strong link between fitness of mind and of physical health. Many men described spirituality as an important determinant of their health status. The men reported a number of barriers to care including 1) lack of respect displayed by physicians and other health facility staff, 2) the clinic setting which often includes women and children or long waiting times and 3) seeking care may lead to their being seen as weak or vulnerable.

CONCLUSION: This is one of the first studies to report African American men's own perceptions of barriers to care. Analysis of the participants' insights suggests that African American men recognize basic health issues and feel health effects from their social environment and unemployment. Additionally this study points to possible remedies to systematic barriers which have been proposed by the target population.

Wynia, DS Cummins, GI Balch, EA Balch, Institute for Ethics, American Medical Association, Chicago; Balch Associates, Oak Park, IL PURPOSE: Performance measures are often reported with individual consumers as the intended audience. Yet employers make most health plan choices when they select one, or a few, plans to offer their employees-making them important potential recipients of information on plan quality. This study explored employers' use of information on health plan quality. METHODS: In December 1998 and January 1999 we conducted 6 national Computer Assisted Telephone focus groups, with 52 individuals responsible for selecting health care packages. There were 2 groups each from small (10-99 employees), medium (100-999 employees) and large (Ͼ1,000 employees) businesses. All participants were from areas with substantial competition among health plans and all offered their employees at least one plan. We asked what information on plan quality they used to select plans and where they obtained that information. RESULTS: The first decision employers face each year is whether to alter the plan(s) offered. Making a change represents a substantial burden. Even among large employers, a very small group is responsible for making this decision and selecting plans to offer. Employee complaints and cost are the primary factors considered. Few actively seek employee input (this is "looking for trouble"), but most track customer service complaints carefully and assume acceptable quality in the the absence of complaints. They rarely look at satisfaction or other traditional plan quality measures prior to making selections and they see report card information as incomparable across plans, confusing, or edited to favor a plan that presents it. Very few have heard of the most widely used quality reporting instrument, HEDIS. Experiences with employer purchasing groups are mixed. Almost all place great weight on the recommendations of an insurance broker both to suggest good plans and to "weed out" or "flag" worse quality plans. Larger employers sometimes use consultants rather than brokers for this purpose. Brokers often compile information and negotiate directly with plans, but usually also receive commissions from plans. Other trusted information sources are human resources industry peers and trade journals. CONCLUSION: Most employers display little interest in conventional quality measures as they are now presented. Instead, they use employee complaints, their peers and insurance brokers or consultants to assess health plan quality. Insurance brokers and consultants may turn out to be the most important recipients of health plan quality report cards.

HEART FAILURE, CANCER AND HEPATITIS C. SL Zickmund, DR Labrecque, R Oren, Internal Medicine, University of Iowa, Iowa City, IA PURPOSE: Interactions between patients and physicians focus on information that may provide limited insight into the impact of the disease experience. To explore alternatives, we asked patients to metaphorically describe their experience. METHODS: Patients with heart failure (CHF; n ϭ 74; 55.0 Ϯ 1.8 years; 33% women), cancer (C; n ϭ 57; 50.5 Ϯ 5.2 years; 63% women), and hepatitis C (HC; n ϭ 74; 44.9 Ϯ 1.3 years; 26% women) were randomly chosen to depict themselves as a living or inanimate object before and after their diagnosis. The answers were coded in five categories by a blinded investigator. RESULTS: A thematic analysis revealed no gender differences. However, several distinctions emerged when comparing the groups. The majority of patients with heart failure (78%) or cancer (72%) used metaphors suggesting a worsening of their lives. For example, one man described himself as a soaring eagle before the disease. Afterwards, he saw himself as an eagle caught in an oil slick. In contrast, half of the patients with hepatitis C (50%) depicted the disease as an enriching, positive influence. One woman described herself as a runaway train, out of control, before the diagnosis. She became a train on a life-saving track afterward. Metaphors related to a lack of energy or to slowing down predominated in the CHF patients (CHF:59%; C:16%; HC:22%). These included viewing oneself as a fast sports car only to become a scooter. About one third of the patients with hepatitis C depicted the disease as a chance to reflect or regain control of their lives, as expressed by the mirror which "makes me look at myself, change things I'd like to change" (CHF:0; C:4%; HC:30%). Cancer patients exhibited the least thematic consistency, articulating their experience as (1) being wounded such as an "injured dog helpless on side of road" (CHF:3%; C:19%; HC:6%) or (2) as revealing the deterioration of aesthetic dimensions, such as becoming a "tulip with no petals" (CHF:0; C:19%; HC:4%). Assessing the influence of age on the choice of metaphors demonstrated that individuals depicting a decrease in energy were older (54.8 Ϯ 1.6 years) compared with those experiencing positive effects of the illness (45.7 Ϯ 2.3 years). This was due to the younger age of patients with hepatitis C who chose metaphors indicating an increase of control as they started treatment protocols. CONCLUSION: The metaphoric description of experiences provides a powerful tool, allowing patients to highlight the impact of illness on their lives. Despite differences in age and gender, patients with chronic diseases of heart and liver embrace similar thematic conceptualizations of their illnesses. While the answers of patients with malignant diseases did not reveal a predominant theme, the majority of individuals used images related to aesthetics or existential threat.

LEARNING OBJECTIVES: 1) Evaluate a patient presenting with chronic cough. 2) Recognize the clinical manfestations of sarcoidosis, including associated medical illnesses. CASE PRESENTATION: A 51 year old metal and wood shop teacher with a past medical history notable for hypertension, gout, nephrolithiasis, and a positive tuberculin skin test presented to the clinic with a chief complaint of cough of greater than 2 months duration. His cough was non-productive and associated with moderate fatigue, such that he was unable to participate in his favorite hobby, gardening. He denied any other recent upper respiratory symptoms, including nasal congestion, rhinorrhea and postnasal drip. He had occasional sour taste in his mouth and heartburn. He denied any fever, chills, night sweats, weight loss, chest pain, wheezing, or shortness of breath. Past medical history was as noted above. He had recurrent nephrolithiasis, with calcium oxalate stones. He had mild renal insufficiency believed to be due to his nephrolithiasis. His positive tuberculin test was noted as a child. He did not recieve therapy, but was followed with serial negative chest x-rays, the last of which was less than 1 year prior to presentation. His only medication was amlodipine. He had no significant family history. He was married, sexually monogamous, and without a history of tobacco, drug or alcohol use. He had no recent travel history. On physical examination, he was well-appearing with a frequent cough, but in no respiratory distress. He was afebrile, with a blood pressure of 160 ⁄ 94 . His exam was otherwise normal. He was initially treated with a proton pump inhibitor for presumptive gastroesophageal reflux disease. However, given his history of a positive tuberculin skin test and possible toxin exposure through his work as a wood and metal shop teacher, a chest x-ray was ordered. In addition, because of his significant fatigue, a laboratory panel including complete blood count, renal function, liver function, thyroid function, calcium, and glucose were ordered. Blood tests were notable for a hematocrit of 40.2% and a creatinine to 3.3 mg/dL (up from 1.3 mg/dL 6 months prior). Other lab tests were normal. Chest x-ray revealed increased reticulonodular densities in the lower lobes. A renal ultrasound showed no evidence of hydronephrosis, but was notable for bilateral renal calculi, early medullary nephrocalcinosis, mild splenomegaly, and multiple enlarged peripancreatic and retroperitoneal lymph nodes. A thin section chest CT showed multiple nodules scattered diffusely throughout both lungs along the perivascular bundles with associated interlobular septal thickening. A bronchoscopy with transbronchial biopsy was performed revealing granulomas with multinucleated giant cells.

Cultures and special stains were negative. The patient was diagnosed with sarcoidosis. 

ago, a cerebrovascular accident (CVA) with central loss of vision in 1997, and a right lacunar infarct in the internal capsule in 1998 with residual left upper extremity paresis. She was on ASA 81mg daily since 1997.Her vitals were normal. The significant findings were gait instability, left upper extremity mild monoparesis, left pronator drift, left facial paresis and mild dysarthria. Computerized Tomographic (CT) scan of the brain showed a lacunar infarct in the anterior limb of the right internal capsule. Pertinent labs were aPTT 26.9 s,IgG anticardiolipin Ab 42.6 (normal Ͻ23 mcg/ml). The patient was anticoagulated with heparin. MRI on the following day showed an area of hypo-attenuation in the pons and medulla suggestive of an infarct. The next day, the patient developed bilateral Internuclear Ophthalmoplegia with worsening dysarthria and left facial palsy, followed within 24 hours by dysphagia and left hypoglossal palsy. On the fourth day of hospitalization, the patient developed acute respiratory distress and hypoxic respiratory failure. The patient was intubated and transferred to the MICU. A repeat CT scan of the brain did not reveal intracranial bleeding. DISCUSSION: APLS manifests clinically as arterial or venous thrombosis, recurrent fetal loss, and/or thrombocytopenia in patients with persistent elevation of ntiphospholipid antibody. Cerebral ischemia is the most common arterial thrombotic manifestation, which may be recurrent. Consideration of the combination of the pertinent clinical and laboratory findings make the diagnosis of APLS. CVA especially thrombotic in young patients, or recurrent arterial or venous thrombo-occlusive episodes in the absence of an obvious etiology should prompt an evaluation for the APLS. The tests that strongly suggest the presence of antiphospholipid antibodies-a) Lupus anticoagulants (LA) b) anticardiolipin (aCL) antibodies c) Beta-2 glycoprotein 1. Multiple tests need to be performed if the clinical index of suspicion is high since a single test will detect only 60-80% of cases. General aspects of treatment consist of anticoagulation with heparin followed by warfarin. All patients with antiphospholipid syndrome and a major thrombotic event should have lifelong anticoagulation with warfarin to maintain an INR of 3-4. Recognition of recurrent thrombotic events is an indication, which should lead a physician to think APLS in differential diagnosis. BINO is a usual manifestation of multiple sclerosis but rarely it could be of APLS.

LEARNING OBJECTIVES: Hypercoagulable state and its work-up, presentation and management. CASE PRESENTATION: A 26 year old female with no significant past medical or surgical history, presented with a one week history of acute epigastric pain radiating to the back. The pain was continuous and associated with anorexia and night sweats. She denied nausea, vomiting or bowel complaints. On further questioning the patient noted a use of oral contraceptives for 2 years which she had stopped 2 months prior to presentation A contrast CT scan of the abdomen was obtained, which revealed portal and superior mesenteric vein thrombosis, with patent hepatic and splenic veins. In addition, a mass was seen displacing the pancreas and associated vascular structures. An exploratory laparotomy was performed which revealed the mass to be a hematoma with reaction. Histopathology revealed fibroblastic proliferation. A work up to rule out hypercoagulable state including protein C, protein S, anti-thrombin III, factor V leidin, prothrombin 20210 and homocysteine was negative. Work-up for dysfibrinogenemia, polycythemia, antiphospholipid antibody syndrome, paroxysmal nocturnal hemoglobinuria was negative. The patient was anticoagulated with heparin and later with coumadin. A repeat CT scan revealed progression of the thrombosis to involve the splenic vein. A follow up CT scan a week later did not show any further progression of the thrombosis. An esophagoduodenoscopy showed grade II esophageal and gastric varices. The patient was discharged on nadolol and coumadin. On follow up the patient has been asymptomatic. DISCUSSION: Portal vein thrombosis is idiopathic in 50% of cases. The most common presentation is gastrointestinal bleeding from gastric and esophageal varices. Acute portal vein thrombosis was described in 2 patients who presented with acute abdomen, but this is the first known case of portal vein thrombosis complicated by intraabdominal variceal bleed. Treatment is conservative with anticoagulation, although endovascular stenting and thrombolysis through a percutaneous transhepatic catheter have been reported in a few cases. Beta blockers are known to reduce incidence of variceal bleed and improve long term survival.

Hospital; Brown University School of Medicine, Providence, RI LEARNING OBJECTIVES: 1) Primary cardiac lymphomas (PCL) involve the heart and pericardium and are often high grade B-cell lymphomas. PCL is mostly found in the right side of the heart with the majority of tumors involving the right atrium. 2) Diagnosis of PCL is still a challenge. Radiologic, histologic and cytologic modalities are used to diagnose PCL. 3) We report this unusual case and show a unique surgical method used to prevent catastrophic complication of right atrial wall perforation following initiation of chemotherapy. CASE PRESENTATION: A 61 year old male with history of hypertension, presented with a three month history of cough, low grade fever, chills and shortness of breath. The patient had a twenty-year smoking history but quit seven years ago. He drank alcohol on occasions. On initial examination, the temperature was 101 F, blood pressure of 146 ⁄ 68 and pulse regular at 110/minute. Jugular venous pres-sure was 10 cm. His lungs were clear. Auscultation showed normal heart sounds and no murmurs. Chest x-ray was non-contributory. EKG showed nonspecific inferior T-wave abnormalities. A transthoracic echocardiogram showed an unusual diastolic and systolic color Doppler jet in the vicinity of the tricuspid valve. A transesophageal echocardiogram showed a 4.0 ϫ 4.5 cm right atrial mass which was obstructing flow from right atrium to right ventricle. Next day, at surgery a large tumor mass originating from the free wall of the right atrium above the inferior vena cava was found. Pathological analysis of the right atrial mass, atrial tissue surface and pericardium showed diffuse, large cell, high grade non-Hodgkin's B cell lymphoma. Five days later a large bovine pericardial patch was placed over the right atrium in order to prevent cardiac rupture from chemotherapy. One week later, the patient was admitted for chemotherapy with CHOP. A repeat transesophageal echocardiogram three months following his initial admission found no evidence of tumor in the right atrium. However, six months later he developed central nervous metastasis and ultimately died from its complication. DISCUSSION: PCL is extremely rare with an incidence of 0.25% of reported tumors of the heart of collected autopsy series. Death from PCL is typically related to direct infiltration of the heart by the tumor and not from metastasis. Surgery, chemotherapy and radiation have been tried as treatment for PCL. We describe the successful application of bovine pericardium to the heart of a patient with PCL. This surgical intervention prevented the likelihood of catastrophic cardiac rupture at the time of chemotherapy treatment for PCL.

LEARNING OBJECTIVES: 1) Recognize the association between the influenza vaccine and Guillain-Barre Syndrome, and 2) Recognize the impediment that the association may pose to attempts at widespread influenza vaccination. CASE PRESENTATION: A 62-year-old patient with a history of cervical stenosis presented with a three week history of worsening numbness in her hands, arms, feet, and legs. She also noted progressive weakness that caused her to fall numerous times. She had attributed the symptoms to worsening of her cervical stenosis and subsequently did not seek treatment until she was unable to walk. Her physical examination was notable for 4 ⁄ 5 strength of all muscles of the lower extremities. The wrist extensors also showed 4 ⁄ 5 strength. There was a loss of position sense and proprioception in her hands and feet. Deep tendon reflexes were unable to be elicited. Laboratory data revealed a normal CBC, chemistries, creatinine kinase, and sedimentation rate. Examination of the cerebrospinal fluid revealed 2-3 WBCs per high power field and a protein level of 249 mg/dl. The diagnosis of Guillain-Barre Syndrome was made. She was treated with intravenous immunoglobulin for five days and referred for rehabilitation. Of note, the patient denied any recent upper respiratory or gastrointestinal complaints. During her hospitalization, the patient's husband stated that he had searched the Internet and that he was convinced that the flu shot that his wife had received four weeks earlier was responsible for her illness. He also reported "warning" his friends of the potential dangers of the vaccine. DISCUSSION: The causal association between the influenza vaccine and Guillain-Barre Syndrome is controversial. A relative risk of 1.7 has been reported. This translates to about one additional case of Guillain-Barre per million persons vaccinated against influenza. Despite this low risk and the potential benefits of vaccination, physicians' efforts at influenza prevention may be hindered by the dissemination of anecdotal cases. These anecdotes may become more powerful with further expansion of the Internet. For the past 2 years, she had recurrent prolonged episodes of sinusitis and bronchitis. Her past medical history and family history were unremarkable. A physical examination was remarkable for bilateral tenderness of maxillary sinuses, swollen nasal mucosa and turbinates with evidence of post nasal drip. Her lung examination was remarkable for diffuse rhonchi. A chest radiography revealed diffuse bilateral interstitial infiltrates and diffuse airway thickening which was unchanged from 5 months earlier. A CT scan of the sinuses revealed pansinusitis. A CT scan of the chest revealed bilateral symmetric cylindrical bronchiectasis with scattered areas of mucus plugging. HIV antibodies were undetected. Measurement of serum immunoglobulins revealed subnormal levels of IgG Ͻ33 mg/dL (505-1364), IgM Ͻ11 mg/dL (67-354) and IgA Ͻ7 mg/dL (67-288). Three months ago patient started receiving intravenous immunoglobulin on a monthly basis and she has had no further episodes of sinusitis or bronchitis. DISCUSSION: Bronchiectasis is described as an abnormal irreversible dilation of the bronchial airways, which is usually a consequence of inflammation and destruction of the structural components of the bronchial wall. The etiology of bronchiectasis is diverse and can be divided into 2 groups. Focal bronchiectasis can be caused by endobronchial tumors, foreign body or extrinsic compression. Generalized bronchiectasis can result from viral, fungal, bacterial and mycobacterial in-fections. Less common causes include: cystic fibrosis, ciliary dysfunction disorders and immunodeficiency disorders. Generalized impairment of pulmonary defense mechanism occurs with immunoglobulin deficiency and leads to recurrent infections and generalized bronchiectasis as highlighted in this case.

LEARNING OBJECTIVES: 1) Recognize the protean manifestations of fecal impaction. 2) Review important concepts in prevention and management. CASE: A 55 year-old male was admitted with fever and change in mental status of 2 days duration. His past medical history was significant for viral encephalitis after which the patient has been demented, non-verbal and confined to bed or wheelchair. On examination, the abdomen was mildly distended and non-tender. A large, firm, oval, supra-pubic mass seemed to be arising from the pelvis. Rectal examination showed soft, brown stool negative for occult blood. A Foley catheter that had been placed in the Emergency Room had drained about 500cc of urine. A CT scan of the abdomen demonstrated a diffusely dilated colon. There was severe fecal impaction in the rectum, pushing anteriorly the decompressed, non-distended urinary bladder. The pelvic mass became completely non-palpable after the patient was successfully disimpacted with digital evacuation and multiple enemas. The patient's mental status returned to baseline and he defervesed. Fecal impaction did not recur during the hospital stay after bowel prophylaxis was instituted. DISCUSSION: Fecal impaction is a significant problem in chronically invalid patients and the elderly. Major etiologies include neurological problems such as spinal cord injury, stroke or multiple sclerosis, medications (especially narcotics and anti-hypertensives) and anorectal pathology. The physical examination can reveal altered mental status, taccypnea and fever. The abdomen may be distended with tympany to percussion. Digital rectal examination is usually diagnostic, revealing impacted stool. However, the absence of a palpable impaction does not exclude a more proximal one. Additionally, the presence of fecal incontinence should arouse suspicion for fecal impaction. Common complications of impaction include intestinal obstruction and perforation. Unusual sequalae encompass megacolon, volvulus, anorectal fistulae, stercoral uleration and lower limb ischemia. Reported urological complications in males have included urinary retention, ureteral obstruction, hydronephrosis, renal failure and urinary infection. Three reports have presented the incidental detection during nuclear imaging of bladder displacement from fecal impaction. Our report is however the first to document a normal urinary bladder presenting clinically as a large, palpable pelvic mass due to fecal impaction, despite the presence of a functioning Foley catheter. Preventive measures against impaction should be used in high-risk patients, including a daily bowel regimen of stool softeners/laxatives and maintaining hydration. For patients with established impaction a thoughtful history and examination is usually diagnostic, supplemented as needed with imaging studies. Disimpaction, enemas and laxatives are the mainstays of treatment, with endoscopic/surgical approaches for refractory cases. Metabolic, neoplastic, neurological or inflammatory etiologies should be sought where indicated. She had no prior ophthalmologic disease or complaints. Interestingly, she had been seen by her primary care physician nine days prior to this visit and had complained of a left-sided headache for two days. At this visit her exam had demonstrated tenderness over her left forehead but was otherwise normal. Her headache was thought to be consistent with either a migraine or a viral syndrome. She had no complaints of eye pain or visual loss at this time. However, during her urgent care clinic evaluation she was noted to have a decrease in the visual acuity of her left eye compared with an ophthalmologic exam six months prior. In addition, her left pupil was non-reactive, and she complained of photophobia during the fundoscopic examination. Given these findings, the patient was felt to have iritis secondary to her underlying collagen vascular disease. However, further examination revealed a few crusted lesions on her left forehead. The patient admitted to having had a rash on the left side of her forehead beginning about five days prior to this clinic visit which was now resolving. Because of the concern for herpes keratitis given this finding, fluorescein staining of the cornea was performed, revealing a classic dendritic pattern. Emergent ophthalmologic consultation confirmed herpes keratitis with secondary iritis and she was admitted for intravenous acyclovir. Despite aggressive treatment, one month later she still had significant visual loss. DISCUSSION: Iritis is often associated with underlying collagen vascular diseases. Classically, the patient with iritis presents with unilateral eye pain, visual loss, and photophobia. In advanced cases, the routine eye exam will show an unreactive pupil. Definitive diagnosis is made with slit lamp examination showing flare cells in the anterior chamber. In early disease, pain with pupillary constriction may be the only clinical feature distinguishing iritis from conjunctivitis. Herpes keratitis is a severe, rapidly progressive infection of the eye, usually occurring in patients who have trigeminal zoster. Untreated, both diseases can lead to perma-nent visual loss. As illustrated with this case, iritis may be the primary eye pathology or a secondary reaction to an underlying infection. To recognise the need for confirmation of diagnosis by lung biopsy for better management and prognosis of patients. CASE: Acute hypoxemia, especially when due to alveolar hypoventilation, represents a medical emergency and has many potential causes. We present a case of a 49-year-old diabetic lady who came to our hospital with worsening shortness of breath, cough with yellowish sputum and chest tightness, which had become severe for one week before admission. She had 30 pack year smoking history, cocaine and marijuana use which she quit 5yrs ago and multiple sexual partners. On examination, she had a temperature of 98.6, BP was 144 ⁄ 70 mm Hg, respiratory rate of 44/min and oxygen saturation was 88% on room air. She had decreased air entry bilaterally with generalized wheezing on chest auscultation. Arterial blood gas was pH 7.44, PCO2 37 mmHg, PO2 56 mmHg, oxygen saturation 88.3%. Chest x-ray showed increased bronchial markings in the right base. A CT scan of the chest done to check for pulmonary embolism showed presence of multiple areas of ground-glass appearance occupying all zones in a non-diffuse, fan shaped manner and bilateral lymphadenopathy. The abnormal laboratory data included mildly elevated alanine and aspartate aminotransferases, alkaline phosphatase and lactate dehydrogenase. Hepatitis antibody profile, HIV testing, tests for mycoplasma and legionella infections were negative.The clinical diagnosis was atypical pneumonia versus Pneumocystis carinii infection and the patient was started on intravenous bactrim, azithromycin and corticosteroids. Bronchoscopy with transbronchial lung biopsy revealed chronic interstitial pneumonia with organizing alveolitis obliterans. Intravenous antibiotics were discontinued and the patient was continued on IV corticosteroids. Nine days after admission she was discharged home on oral corticosteroids. DISCUSSION: Cryptogenic organizing pneumonitis or idiopathic BOOP (bronchiolitis obliterans with organizing pneumonia) is a relatively rare condition with an incidence of 6-7 per 100,000 hospitalizations in teaching institutes. It is characterized by excessive proliferation of granulation tissue within the small airways and alveolar ducts and associated inflammation in the surrounding alveoli. The clinical presentation often mimics community acquired pneumonia and is often preceded by flu like symptoms. Its presenting features include persistent non-productive cough, dyspnea with exertion and weight loss of more than 10 pounds. The overall prognosis is much better than conditions like idiopathic pulmonary fibrosis which is in the differential diagnosis. A high clinical suspicion and confirmation of the diagnosis of BOOP is very important because it can favorably change the management and prognosis of the patient.

LEARNING OBJECTIVES: 1. Recognise the clinical presentation of drug-induced rhabdomyolysis. 2. Demonstrate the value of carefully scrutiny of all patient medications at presentation. CASE PRESENTATION: Acute rhabdomyolysis is a clinical and laboratory syndrome that is the result of skeletal muscle injury with release of cellular contents in the plasma. The most common suspected aetiologies reported in the literature include alcohol and trauma, with increasing recognition of drug-induced rhabdomyolysis including that associated with use of lipid lowering statins. We report a case of rhabdomyolysis after mistaken ingestion of atorvastatin.

A 68 year old Caucasian male presented to the hospital emergency room with a two week history of myalgia, generalised weakness and fatigue. His past medical history included diabetes mellitus type 2. He admitted to the consumption of 5-6 units of alcohol per week and a past history of alcohol abuse. Six weeks previous to his presentation, the patient was attacked by a domestic cat (who subsequently attacked others and was later quarantined and destroyed) and had been started on a course of clavulinic acid. On examination motor weakness, more proximally than distally, was noted. Admission laboratory investigations: Creatine Kinase (CK) 52 915, Urea 22.5, Creatinine 311, AST 2 549, ALT 501, LDH 2 009. Urine was noted to be smoky-grey in colour and positive for myoglobin. The patient's CK continued to rise and peaked at 185 810. His muscle weakness progressed such that he was unable to ambulate but had no respiratory difficulties. He developed oliguria and required haemodialysis. Further consideration of the aetiology of the rhabdomyolysis and review of his medication bottles brought to the hospital revealed that the patient had incorrectly been dispensed atorvastatin 40 mg instead of Accupril 40 mg approximately four weeks previous to his presentation. The patient's renal function and muscle weakness improved and dialysis was discontinued. The patient had a slow and complicated recovery. DISCUSSION: Rhabdomyolysis is a well described adverse reaction associated with atorvastatin use. While patients are normally advised to discontinue the medication should myalgia or weakness occur, this patient received no such advice as he had not been prescribed atorvastatin and only received it inadvertently. This case illustrates an interesting presentation of a known adverse drug reaction and highlights the role for careful scrutiny of patient medications at presentation in the diagnosis of undifferentiated medical problems.

LEARNING OBJECTIVES: Recognition, early diagnosis and treatment of HZE and associated cerebral vasculitis in the elderly to reduce mortality and morbidity. CASE PRESENTATION: HZE is a rare condition with significant mortality and morbidity which can be reduced with prompt treatment. A previously healthy 80 year old woman became delirious and acutely encephalopathic several days after onset of disseminated cutaneous zoster with involvement over the right trigeminal nerve distribution. She had fever, headache, right ptosis, 6th nerve palsy, CSF lymphocytic pleocytosis with raised protein, increased signal density in the right temporal lobe on MRI suggestive of localized encephalitis but a normal EEG. Blood cultures, serum VDRL, rheumatoid factor, cryptococcal antigen and PPD testing were Ϫve and CD4, CD8 counts were normal. AFB smear, HZV DNA PCR, herpes simplex viral DNA 1 & 2 on CSF were Ϫve. Acyclovir, antibiotics and steroids for presumed vasculitis did not improve her condition. At autopsy, diffuse cerebral vasculitis, perivascular lymphocytic infiltration, trigeminal ganglionitis and neuritis was noted. DISCUSSION: Encephalitis is a rare complication of HZ associated with disseminated cutaneous disease and cranial nerve lesions. HZE is a clinical diagnosis as the viral recovery from CSF or brain tissue has been generally unsuccessful. It should be suspected in patients who present in an encephalopathic state with a recent clinical HZ infection, an abnormal brain MRI and CSF examination. Cerebral vasculitis is a serious post-zoster CNS complication (especially with disseminated zoster and cranial nerve lesions) which can present as hemisphere strokes following the trigeminal distribution of zoster due to focal internal carotid and middle cerebral artery vasculitis. In rare occasions, diffuse cerebral vasculitis is described. Early administration of acyclovir produces complete resolution of encephalitis within 72 hours with normalization of EEG, and reduced mortality. However, delayed diagnosis and treatment reduces overall effect of acyclovir. Cases of cerebral vasculitis without preceding CNS infection have been treated successfully with immunosuppressive agents. However, treatment of cerebral vasculitis due to HZ is not clear and according to some reports, there is no treatment for Zoster associated cerebral vasculitis which may be responsible for this patients' poor response to therapy. Although acyclovir did not result in improvement in our patient, it could inhibit viral replication in the CNS or decrease the antigenic stimulus that drives the CNS immunological responses. Continued treatment of undiagnosed cerebral vasculitis associated with HZE with immunosuppressive agents may lead to neurological deterioration and treatment with acyclovir is potential cure. A 25 year old Hispanic male who was taking phenytoin for two months for post traumatic seizure prophylaxis presented with fever, cough, erythematous scaly macular rash of 5 days duration over whole body except palms and soles, facial pustules, fever (101.7 F), generalized lymphadenopathy. Labs : WBC: 15,200/L, 15.6% eosinophils. Serum chemistry and liver function tests were normal. Throat and blood cultures, viral culture for EBV, CMV, rubella, rubeola and tests for HIV and syphilis were Ϫve, ruling out most other causes of this clinical picture. A history of drug exposure and clinical findings led to the diagnosis of AHS due to phenytoin. He was treated with IV hydration and corticosteroids with resolution of rash, fever and lymphadenopathy within a week and uneventful recovery. DISCUSSION: Considering high prevalence of epilepsy, clinicians must recognize potentially fatal AHS due to arene oxide-producing anticonvulsants (phenytoin, carbamazepine and phenobarbital) which presents as skin rashes, fever, lymphadenopathy, atypical lymphocytosis or eosinophilia, thrombocytopenia, hepatitis and nephritis. It occurs in 0.1%-0.01% of population,is more common in Afro-Americans, has diverse clinical features resulting in a delay of diagnosis and is fatal in 5-50% if hepatitis or toxic epidermonecrolysis occurs. AHS is caused by an inherited deficiency of epoxide hydrolase, an enzyme required for metabolism of a toxic intermediate, arene oxide, produced during metabolism of above mentioned anticonvulsants by cytochrome P-450. Inheritance is autosomal co-dominant and siblings have a 25% risk of AHS and should be counseled. Arene oxides bind to cell macromolecules resulting in cell death (cytotoxicity), teratogenecity, mutation, tumor or can act as haptenes and neoantigens resulting in hypersensitivity reactions. The lymphocyte cytotoxicity test for arene oxides is an in-vitro method for prediction and diagnosis of AHS but is not widely available. The diagnosis of AHS is made by most centers from the history of drug exposure and clinical findings. Prenatal prediction of risk of congenital malformations (Fetal Hydantoin Syndrome) is made by demonstrating low epoxide hydrolase activity in fetal amniocytes. Treatment: Discontinuation of the drug, supportive care with attention to fluids and electrolytes, admission to the burn unit depending on the skin exfoliation, and systemic steroids to reverse cutaneous eruptions. Internal manifestations do not seem to reverse with steroids. Benzodiazepines, gabapentine, lamotrigine are other anticonvulsant alternatives in these patients. Valproic acid, because of its hepatic metabolism should not be used during the acute phase. A prior electrocardiogram showed no pre-excitation or QTc prolongation. Blood pressure was 146 ⁄ 98 mmHg, heart rate was 181 beats/min and regular, and respiratory rate was 20 breaths/min. Cardiac examination revealed a regular tachycardia with normal S1 and S2; no S3, S4 or murmur was detected, and the lungs were clear to auscultation. Complete blood count, electrolytes (including calcium and magnesium), creatine kinase, and troponin I, all were within normal limits. A chest radiograph demonstrated a heart of normal size and clear lungs. A 12-lead electrocardiogram demonstrated a monomorphic regular wide QRS tachycardia with a ventricular rate of 188/min and right bundle branch block (RBBB) morphology. Valsalva maneuver, carotid sinus massage, administration of lidocaine, adenosine and procainamide did not affect the tachycardia. He received 6 synchronized shocks, 4 of which were at 400 Joules with no response. At that point a transvenous pacemaker was inserted in the right ventricle inducing a 5-beat run of ventricular tachycardia at a rate of 300 beats/min, followed by normal sinus rhythm. A transthoracic echocardiogram demonstrated normal left ventricular size; ejection fraction was 54%. An electrophysiologic study excluded the presence of an accessory pathway or dual AV node physiology and did not induce any arrhythmia. The patient was dismissed on oral verapamil, and remained asymptomatic 3 months later. DISCUSSION: According to ACLS guidelines, hemodynamically stable patients with wide QRS tachycardia of uncertain type should be sequentially treated with lidocaine, adenosine, procainamide, and bretylium. If pharmacologic treatment fails, or if hemodynamic instability occurs, synchronized cardioversion is recommended. However, if all ACLS proposed interventions fail, as in our patient, overdrive pacing or internal cardioversion could be utilized. year old man without significant past medical history presented with right-sided chest and back pain, and a dry cough of six days' duration. The pain was non-pleuritic, but was aggravated with movement. The patient denied fever or chills. He had been treated for pneumonia six months ago for similar symptoms of right-sided chest pain. He completed a full course of antibiotics with complete resolution of symptoms. He denied alcohol or tobacco use. He had no history of seizures, and denied any recent dental procedures. The patient had a desk job and was not exposed to any dust or airborne fibers. On exam, the patient did not appear acutely ill. He had a low-grade fever of 100 degrees. Other vital signs were unremarkable. Chest exam revealed dullness to percussion and diminished tactile fremitus over the right lung field. On auscultation, decreased breath sounds were heard in this region, but egophony was absent. Chest X-ray demonstrated an extensive right-sided pleural effusion suggestive of early loculation, and subsequent thoracentesis confirmed that the pleural fluid was straw-colored and turbid. The pH of 7.05 and fluid LDH of 1149 suggested the presence of a complicated parapneumonic effusion. Chest CT showed a loculated effusion with underlying right middle and lower lobe consolidation. A thoracic surgery consult was obtained, and the patient went for open thoracotomy and decortication. Intraoperative bronchoscopy revealed a 3 cm plug of aluminum foil occluding the right lower lobe bronchus. After removal of the foil decortication proceeded without incident and the patient's lung reexpanded. He did well after the surgery and was discharged home. DISCUSSION: Foreign body aspiration is more common in children, but should be considered in adults who present with complicated focal airspace disease. The majority of cases involve the right lung, and particularly the right lower lobe, due to the relatively oblique orientation of the right mainstem bronchus. Recent reviews confirm that foreign body aspiration may occur without known risk factors for aspiration, and may be diagnosed several months to years after the initial event, particularly when the foreign body is radiolucent, as in this case (Prakash and Limper, Annals of Internal Medicine, 1990; Rodenstein et al. Chest, 1999) . Such patients often present with subacute respiratory symptoms of chest pain, cough and wheezing, and may be diagnosed with recurrent pneumonias. Controversy exists whether rigid or flexible bronchoscopy is the most effective means of foreign body removal. Foreign body aspiration should be considered whenever physical exam or chest X-ray suggests the presence an occluded bronchus.

breath and right sided pleuritic chest pain. Patient had been in her usual state of health until ten days prior to admission, when she noted increasing left leg swelling after a long road trip. Five days prior to admission, patient developed sudden onset of right sided chest pain followed by a productive blood-tinged cough. She denied any fever, chills, night sweats. Her medications were notable for prednisone. Her exam was remarkable for a pulse of 101, respiratory rate of 22, temperature of 38.4, room air saturation of 93%. She had inspiratory crackles in the right middle lobe, her extremities were benign. Her WBC was 18.8, A-a gradient 50 mmHg, d-dimer 923. Her ECG revealed sinus tachycardia with LVH, her CXR showed a right middle low opactiy. Lower extremity dopplers were negative and V/ Q scan was intermediate probability. Patient was admitted with a working diagnosis of PE and started on heparin. AFB ϫ 3 was negative. Patient improved over five days, her shortness of breath and chest pain resolved. She remained afebrile and her WBC slowly decreased. On hospital day #6, the patient acutely developed worsening shortness of breath with gross hemoptysis and suffered a respiratory arrest. Due to the unexpectant clinical course an autopsy was requested. Autopsy revealed extensive bronchopneumonia secondary to Serratia marcenses. DISCUSSION: The role of clinical suspicion and V/Q scanning for diagnosis of PE has been best evaluated in the PIOPED study. This case points out the limitations of PIOPED and other proposed tests used in the diagnosis of PE. It begs to ask if pulmonary arteriography is indicated in all cases of suspected PE. This case also brings up many clinical points, namely infection in the setting of steroids, incidence of PE in end stage renal patients. Finally, this case is ironic and humbling. The decision to treat for presumed PE was in some respect to not miss a lifethreatening process. The question which still remains is should we have treated her for pneumonia from the beginning and would this have saved her life?

NOT A CLOT. J Browning, K Szauter, L Tutt, P Leonard, Internal Medicine, The University of Texas Medical Branch, Galveston, TX LEARNING OBJECTIVES: (1) Recognize the similarities between the presentation of a pulmonary embolism and that of pulmonary hypertension due to other causes.

(2) Recognize consequences of an uncorrected congenital cardiac defect in an adult. CASE PRESENTATION: A 33 year old woman presented to the emergency room with left sided chest pain and dyspnea, bilateral lower extremity pain and mild edema of the right (R) lower leg. Her past medical history was notable for menorrhagia (controlled by oral contraceptives) alopecia, and hepatitis C. Surgical history included an open cholecystectomy 12 years earlier. The patient used tobacco and had previously abused alcohol and intravenous drugs. She had a sedentary lifestyle and her work in an institutional laundry facility necessitated long periods of sitting and standing. Physical examination was notable for an uncomfortable appearing young woman in mild respiratory distress. She had diffuse alopecia. Lung exam revealed moderate air movement throughout. Cardiovascular exam was notable for a paradoxically split S2 and a soft holosystolic murmur at the left lower sternal border. Abdominal examination was unremarkable. There was moderate edema of the lower extremities bilaterally. Initial laboratory studies including electrolytes and a CBC were normal. An arterial blood gas on room air revealed a pO2 of 44. Her EKG was remarkable for an S wave in lead 1 and a Q wave and inverted T wave in lead 3. A ventilation-perfusion scan was read as low probability for a pulmonary embolism. However, because of the high clinical suspicion for a pulmonary embolism the patient was placed on oxygen therapy and started on heparin. The following morning the patient showed no clinical improvement. Lung examination revealed poor air movement diffusely. A chest radiograph showed a large heart and prominent pulmonary artery. Lower extremity doppler studies revealed no deep venous thrombosis. An echocardiogram showed (R) atrial enlargement (R) ventricular dilatation, and moderate tricuspid regurgitation. A rapid clinical decompensation necessitated her transfer to the intensive care unit. A pulmonary artery (PA) catheter was placed. Measurements demonstrated that the PA pressure exceeded systemic pressure. A second echocardiogram following the infusion of agitated saline showed significant right to left shunting at the atrial level consistent with an atrial septal defect (ASD). She was not a candidate for a heart-lung transplant and died within a few weeks of her presentation. DISCUSSION: This patient presented with Eisenmenger syndrome due to an uncorrected congenital ASD. Despite a "textbook" presentation of a pulmonary embolism her limited response to initial therapy prompted a further evaluation resulting in an unexpected and unusual diagnosis. This case provided rich learning opportunities in both pathophysiology and clinical diagnosis.

Bruns, KM Hla, Department of Medicine, University of Wisconsin, Madison, WI LEARNING OBJECTIVES: 1) Recognize hemoglobinopathy as a cause of low oxygen saturation. CASE: A 53 year-old healthy Korean woman presented to the primary care clinic for follow up of screening flexible sigmoidoscopy. During the procedure, prior to sedation, arterial oxygen saturation (SaO2) was measured at 83% on room air by pulse oximetry with a good tracing. She denied fever, cough, shortness of breath, chest pain, orthopnea, calf tenderness, tobacco use, environmental exposures or use of any medications. On physical exam, vital signs were normal, cardiopulmonary exam was normal, and there was no cyanosis or clubbing. Complete blood count, pulmonary function tests, and chest x-ray were normal. Pulse oximetry in the office was 83-84% on room air. Arterial blood gas obtained on room air re-vealed a normal paO2 of 94 mmHg while the SaO2 was only 86%. Reduced hemoglobin was elevated at 12.9%. Methemoglobin and carboxyhemoglobin were normal. Administering 50% FiO2 and 100% FiO2 caused the paO2 to rise appropriately, but the SaO2 did not correct to expected levels and remained at 91% and 96%, respectively. Mild instability of hemoglobin was identified on isopropanol testing, and 40.3% of the patients' hemoglobin was abnormal by chromatography with the remainder being largely hemoglobin A. A DNA analysis was performed and showed that the patient is heterozygous for hemoglobin Rothschild, a beta chain point mutation substituting arginine for tryptophan in the hemoglobin molecule (beta37(C3)Trp→Arg). It is inherited as an autosomal co-dominant condition. Her two children also had low oxygen saturation when screened with pulse oximetry and were subsequently shown to have the same hemoglobin abnormality. DISCUSSION: Pulse oximetry is a useful tool for noninvasive monitoring. However, its use is limited in the presence of dyshemoglobinemias, dyes and pigments (such as methylene blue), low perfusion states and anemia where SaO2 is not reflective of paO2. In our patient, the alteration in the beta chain of hemoglobin results in a decreased affinity of hemoglobin for oxygen. This leads to a shift to the right of the hemoglobin-oxygen dissociation curve and a subsequent low SaO2 due to increased concentration of reduced hemoglobin. While the patient is largely asymptomatic from this benign hemoglobinopathy, there is the potential for confusion in the event of emergency management of an injury or during a procedure where pulse oximetry might be measured. The differential diagnosis of a low pulse oximetry reading should be expanded to include hemoglobinopathy. Distinguish usual from more rare causes of otitis media. CASE: A previously healthy 24 year-old Indian accountant presented on 2/1/99 with a temerature of 103F, chills, productive cough, nausea, malaise of 4 days duration. He had returned one week prior from a three week stay in southern India. Nausea, malaise and a feverish felling commenced on the plane trip back to the U.S. On exam he appeared ill, had a temperature of 101.9F, pulse of 68 and BP of 96 ⁄ 70 . The left tympanic membrane was described as "fire red". He had frontal sinus tenderness and an erythematous pharynx without exudates. Lungs were clear. Neck was supple. He was diagnosed with otitis media and prescribed a 5 day course of azithromycin. He presented 4 days later with unremitting fever, bifrontal headache, persistent cough and nausea. On exam he had a temperature of 102.1F, BP of 100 ⁄ 80 , HR of 92 and RR of 16. He had a foul body odor but no rash. Neck was supple. Left tympanic membrane was dull and red. He had sinus tenderness on percussion. Lungs were clear. There was no hepatosplenomegaly. Chest x-ray was clear. He was switched to Amoxicillin/Clavulanate. Blood smears were negative for malaria. Stool culture was negative. CBC showed a WBC of 5.3 K/ul with 53% neutrophils, 13% monocytes, 3% eosinophils, 31% lymphocytes. Hb was 14.4 g/dl and platelet count was 225,000 K/ul. Blood cultures grew salmonella typhi and the diagnosis of typhoid fever was made. He was switched to ciprofloxacin and completed a 14 day course. He rapidly defervesced and began to feel well. DISCUSSION: Typhoid fever should be suspected in travelers to endemic areas. Because this patient had a "fire red" tympanic membrane, the physician thought he had common otitis media/sinusitis. It was only when the patient failed to respond to azithromycin that an alternative diagnosis was sought. Acute otitis media is not a common manifestation of typhoid fever but it has been secribed. This patient had several features of typhoid fever including unremitting fever, headache, putrid body odor, travel to an endemic area, pulse-temperature deficit, and normal white blood cell count.

LEARNING OBJECTIVES: Recognize the clinical presentation of adrenal insufficiency and recall that an infectious cause of Addison's Disease can be extrapulmonary tuberculosis involving the adrenal glands. CASE PRESENTATION: A previously healthy 71-year old Croatian born man presented with a two month history of persistent weakness, dizziness, anorexia, nonproductive cough, fever, night sweats, and unintentional 15 pound weight loss. During a recent hospitalization for these symptoms, he was found to have bilateral upper lobe infiltrates on chest X-ray with a strongly positive PPD (20 mm induration). A bronchoscopy with transbronchial biopsy showed chronic inflammatory changes without caseating granulomata, bronchial washings were sterile and mycobacterial cultures were negative at 3 weeks. The patient had been discharged on prophylactic Isoniazid, but his symptoms failed to improve. He then presented to the outpatient clinic with worsening weakness, dizziness and persistent fevers. The patient resided in the US for over 30 years and worked as a metal grinder. He was a former smoker with 20 pack-year history and reported drinking 2-3 alcoholic beverages daily. Physical examination was notable for a cachectic appearance, orthostatic blood pressure changes and generalized hyperpigmentation, markedly accentuated at the palmar creases and at a vaccination scar on the left arm. Hyperpigmentation of the buccal mucosa was not present. The patient was admitted and placed in respiratory isolation. Laboratory studies showed serum sodium of 133 mmol/L, potassium 6.1 mmol/L, C02 19 mmol/L, Glucose 89 mg/ dL, BUN 64 mg/dL, and Creatinine 2.1 mg/dL. A Cosyntropin stimulation test was abnormal along with a markedly elevated ACTH level. Abdominal CT scan demonstrated bilateral adrenal enlargement (LϾR) without lymphadenopathy or other abnormalities. The patient was treated with corticosteroids, fludrocortisone and was placed on a 4 drug antituberculous regimen for a presumptive diagnosis of Tuberculous Addison's Disease. A CT-guided adrenal biopsy showed acute and chronic inflammation with necrosis, most consistent with an infectious process, but acid-fast bacilli and fungal stains were negative. The patient steadily improved on therapy and was discharged home. At 6 weeks, induced sputum cultures finally grew M. tuberculosis. DISCUSSION: The clinical presentation of adrenal insufficiency can be varied, but a heightened index of suspicion is needed in patients who present with hyperpigmentation, weakness, weight loss and hypotension. Tuberculous Addison's disease, which once accounted for up to 90% of cases, is now an uncommon cause of primary adrenal insufficiency in the US. Nevertheless, it should be considered in immunocompromised patients and in immigrants from countries where TB is endemic. Four weeks prior to presentation, he had been diagnosed with costochondritis and degenerative arthritis of the spine, which were treated with a 2-week course of a NSAID. His symptoms worsened on therapy and he began to complain of severe, constant, sharp sternal and upper back pain, exacerbated by arm movements, changes in position and deep inspirations. The patient had been previously in excellent physical condition and denied any headache, trauma, unusual physical activity, recent upper respiratory tract infection, cough, fever, nausea, vomiting, diaphoresis or weight loss. He was a former 40-pack year smoker, stopping eight years ago, and reported drinking 1 ⁄ 2 to 1 case of beer per week. Physical examination was notable for extreme tenderness just below the sternal angle, without associated overlying skin inflammation or any localized tenderness at the costochondral joints. There was tenderness to palpation of the upper thoracic spine, approximately 2-3 cm to the right of the T5 vertebral body. Rectal exam showed trace guaiac positive stool with a normal prostate. A complete blood count, metabolic panel and PSA test were within normal limits. A chest x-ray demonstrated a lytic bone lesion with an associated soft tissue mass involving the posterolateral aspect of the right fifth rib. Chest CT scan showed a 2 ϫ 3-cm soft tissue mass involving the right posterolateral rib, two 5-8 mm non-calcified nodules in the periphery of the right and left lower lobes of the lungs, normal vascular structures and no adenopathy. A bone scan showed areas of increased tracer uptake in the sternum and the right posterior 5th rib. The patient underwent a sternal biopsy showing metastatic adenocarcinoma, possibly from a lung or prostate primary. Subsequent diagnostic tests including a CT scan of the abdomen/pelvis and colonoscopy failed to reveal a source of the primary cancer. The biopsy specimen did not stain for PSA, making metastatic prostatic adenocarcinoma much less likely. A pulmonary adenocarcinoma was considered the probable primary source. DISCUSSION: Sternal pain associated with other areas of bone pain should be thoroughly investigated and not simply attributed to costochondritis. Costochondrits usually involves one of the costochondral joints and can be associated with sternal joint swelling (Tietze's Syndrome). Several cancers can metastasize to the ribs, chest wall and spine including: breast cancer, prostate cancer, sarcoma and plasma cell cytoma. Metastatic adenocarcinoma of unknown primary is a rare cause of sternal pain.

LEARNING OBJECTIVES: 1) To recognize when to broaden the differential diagnosis of shoulder pain. 2) To demonstrate the need to treat the whole patient and not just the chief complaint. CASE: The patient was a 43 year old male with a chief complaint of right shoulder pain. His job required a lot of heavy lifting, but he denied any injury or strain. The pain in his shoulder had persisted for several months, necessitating at least 3 prior visits to the urgent care center where he was prescribed non-steroidals and physical therapy. Despite treatment, he had not had any relief of his pain. He said that he could no longer raise his right arm. Focused examination of the right shoulder revealed markedly diminished passive range of motion with elevation, abduction, internal rotation, and external rotation. Strength around the right shoulder joint was 4 ⁄ 5 with the limitation being primarily secondary to pain. Distal strength and left shoulder strength were normal. Reflexes were normal. Upon further questioning, the patient revealed progressive dyspnea and weight loss over several months and denied any cardiac symptoms, melena, bleeding or bruising. He reported 60 pack years of cigarette smoking. Past medical history revealed admission to the hospital one year prior for a left upper lobe pneumonia. Further examination revealed bilateral temporal muscle wasting and prominent ribs. There were hard, fixed left anterior cervical lymph nodes, the largest being 1cm. Breath sounds were markedly diminished from the left base to the left apex with dullness to percussion throughout the left side. Chest radiograph revealed opacification of the left lung field with some leftward mediastinal shift and elevation of the left hemidiaphragm. Radiographs of the right shoulder revealed a large lytic lesion in the humeral head, with some invasion into the humeral neck. There was no periosteal inflammation and no reactive bone formation. Chest computed tomography revealed a large left pleural effusion, truncation of the left mainstem bronchus about 1cm from the carina, encasement of the aorta, and poor visualization of the left pulmonary artery. The CT also confirmed the presence of a 6mm by 6mm soft tissue mass in the humeral head with bony erosion. Cytologic examination of the pleural fluid as well as brochial biopsy confirmed the diagnosis of poorly differentiated non-small cell carcinoma of the lung. DISCUSSION: Shoulder pain is a common complaint, usually benign, and mostly self-limited; even the occasional frozen shoulder has an excellent and self-limited prognosis. However, the physician must consider other diagnoses on the differential when there is no response to traditional treatment and there are other red flags in the patient's presentation. The physician should also be aware that the patient's primary complaint is not always the root of the problem. week twin gestation, presented with syncope, vomiting and a loose bowel movement. Initial systolic blood pressure was 80 mmHg, but came up to 100 mmHg after IV fluid boluses. She only complained of swelling and pain of her hands and forearms. We were called in as medical consultants after 4 hours of anuria despite 3 liters of Lactated Ringers (LR) solution. Her HCT was 54, WBC count was 30,000 with 37% bands. Pulse oximetry was normal. Her exam revealed tense non-pitting edema of her hands and forearms, 0 ⁄ 2 radial pulses, clear lungs, and an inability to actively extend her fingers. Our first impression was sepsis syndrome so we started ceftriaxone and ampicillin to cover community acquired infections and Listeria. Infected pregnant women get noncardiogenic pulmonary edema, but we had never seen compartment syndrome. We called orthopedics to check forearm compartment pressures, which were only mildly elevated. When her lips began to swell we gave corticosteroids to treat possible angioedema. Anaphylaxis was also considered, and we gave her antihistamines. Epinephrine was withheld due to adequate blood pressure and no wheezing. She miscarried during the night. By the next morning she had received 10 liters of LR to maintain adequate urine output, her compartment pressures had increased and she was taken to the OR for fasciotomies of both forearms and calves. Serum creatinine never went above 1.0 mg%. Day #4: Spontaneous diuresis began. Day #7: She still had normal brain, lung, liver, and renal function and was taken to the OR for closure of the fasciotomies. They found liquefied muscle, and performed bilateral above knee and above elbow amputations. Final lab results included: negative maternal & fetal cultures/anti-Latex IgE /ANA; normal bone marrow biopsy/serum protein electrophoresis/complement levels/C1 esterase activity. After excluding sepsis, angioedema, vasculitis, and finding literature describing profound acute hypovolemia with compartment syndromes and central organ sparing, we diagnosed her with Idiopathic Systemic Capillary Leak Syndrome (SLCS). DISCUSSION: Since 1960 fifty cases of SLCS have been reported. The hallmarks are hypovolemia, hemoconcentration and central organ sparing. Case reports describe leukocytosis, spontaneous diuresis on Day #4, compartment syndrome with fasciotomies, and death from airway edema, but never amputations. It has been associated with monoclonal gammopathy and can recur. Acute therapy is supportive. Terbutaline and theophylline may decrease capillary permeability and have been used as prophylaxis against recurrent attacks. This case was devastating to all involved. He had no history of fever, neck stiffness, visual changes, or weakness. Physical exam was negative for any neurological deficits or focal findings. Diclofenac was prescribed. Ten days later the patient returned with continued headache, worsening back pain, and a temperature of 38.4C. His history and physical were otherwise unchanged from the previous visit. A lumbar sacral spine x-ray was negative and he was prescribed ibuprofen and acetaminophen. Six days later, he was admitted with worsening headache, back pain, fever, and right sided weakness. The physical exam revealed nuchal rigidity, and inability to flex his back or hip due to pain. Motor strength was 4 ⁄ 5 in the right upper and lower extremities, 5 ⁄ 5 on the left. Lab showed Hb 11.8, Hct 37, WBC 10.6, 81% neutrophils, 8% lymphocytes, 12 monocytes; LDH 1035, HIV positive. A non-contrast MRI revealed no abnormalities. A lumber tap showed an opening pressure of 24, glucose 27, protein 983, RBC 50, WBC 680, 83% neutrophils, 13% lymphocytes, 4% monocytes. CSF gram stain: 3ϩ wbc, no organisms, KOH stain and AFB smear negative. The patient was admitted with a presumptive diagnosis of bacterial meningitis and started on cefotaxime and ampicillin. Clinically, the patient deteriorated with worsening neurological findings, a CT scan showing left basal ganglion infarction, and a MRI showing diffuse lumbar spine enhancement. A repeat LP showed continued low glucose, elevated protein and WBC values with neutrophilic predominence, and negative KOH and AFB smear. His skin tests were anergic. CSF culture for fungal, herpes, VDRL, and TB PCR were all negative. The patient was started on 4 anti-TB drugs empirically with improvement in his CSF values, but was still left with neurological deficits. It was not until one month after his hospitalization that his TB culture came back positive. DISCUSSION: Headaches are a very common disorder that present to primary care providers. Early on, TB meningitis may present as a common, non serious headache because of subacute symptoms. Fever and neurological deficits may not present for 1-2 weeks. Even when considered, diagnosis of TB is challenging because CSF AFB are seen on smear only 20% of the time. CSF culture is positive in 80% of cases, but may take weeks to months to grow out. PCR is more rapid, but only 60% sensitive. One needs to have a high index of suspicion for TB when treating populations from endemic areas, or who are HIV positive. Empiric therapy should be initiated early on because of the serious sequelae of TB meningitis. year old man with a significant cardiac history presented with 2 episodes of sudden loss of consciousness. Past medical history is remarkable for coronary artery bypass graft surgery in 1993 and wide QRS complex tachycardia which was evaluated by electrophysiologic (EP) studies in 1998. The EP study was essentially normal with no inducible ventricular tachycardia (VT). He was started on sotalol with good results. He presented with 2 episodes of unconsciousness. The first episode occurred suddenly in a sitting position with no preceding aura or prodrome such as lightheadedness, nausea, epigastric distress or palpitations. There were no convulsions, incontinence, tongue biting or postictal confusion. The second event happened abruptly while he was driving with similarly no aura or prodrome. In hospital, his cardiovascular and neurologic examinations were unremarkable. A third spell happened in hospital and this time he had a premonition of presyncope just before losing consciousness in a sitting position. The ECG showed a sinus tachycardia of 150 beats/minute. There were no convulsions, incontinence or automatisms. Further continuous ECG monitoring only showed asymptomatic runs of supraventricular tachycardia. Tilt table testing with isoproterenol was normal. An EP study was essentially normal and did not demonstrate sustained monomorphic VT. Finally, an EEG showed potentially epileptiform abnormalities occurring independently over both temporal regions during wakefulness and sleep. The patient was started on phenytoin and will be followed up in a few months. DISCUSSION: Cardiac syncope can often be diagnosed by its clinical presentation. Sudden loss of consciousness with no preceding aura or prodrome in an elderly patient with prior history of coronary artery disease and ventricular arrhythmias would strongly suggest a diagnosis of cardiac syncope. In such a patient, EP studies should be undertaken if noninvasive techniques fail to provide the diagnosis. A nondiagnostic study should prompt one to consider an alternate diagnosis. Temporal lobe epilepsy is known to mimic cardiac syncope. It is also associated with ictal tachycardia or bradycardia. With time, if phenytoin prevents further spells, a diagnosis of temporal lobe epilepsy is likely in this case.

LEARNING OBJECTIVES: *To create more awareness among general practioners about complications of PEG procedure which is getting more common with passage of time. CASE: A 79-year-old nursing home resident had recurrent diarrhea and fever. He had extensive evaluations, which were inconclusive. His past medical history was non-significant except for dementia. He had PEG placement 5 months previously, which was replaced 1 week later because it was nonfunctioning. His physical exam was significant for evidence of dementia, fever of 100.5 F, regular pulse of 104/min and peri-PEG site erythema with a foul smelling discharge through the PEG. The remarkable laboratory findings were WBC's 17,000/ul with a left shift in the differential count. He underwent a Gastrografin study via the PEG, which was diagnostic of gastrocolic fistula. The patient underwent laparatomy. Perioperatively, the PEG was found traversing transverse colon with obvious entry and exit sites before entering body of stomach. The gastrocolic fistula tract had surrounding granulation tissue. The PEG was removed and he had surgical repair of the gastrocolic fistula, with a feeding gastrostomy put in place. The subsequent course of events was unremarkable. His fever and diarrhea resolved spontaneously and he was discharged to a skilled nursing facility. DISCUSSION: Gastrocolic fistula is a rare complication of PEG although its exact incidence is unknown. We suspect that physicians are going to be faced with increasing cases of this diagnosis because many patients are undergoing this procedure. An interesting feature of our patient is that both the fever and diarrhea were intermittent and resolved spontaneously at times. The two CAT scans showed the PEG tube in stomach. So it could be inferred that the PEG migrated through the fistulous tract and hence the symptoms occurred with the PEG in the colon but spontaneously resolved when the PEG moved back to the stomach. Thus a very high index of suspicion and awareness about this complication of gastrocolic fistula is warranted in any patient who has undergone PEG.

CA Cole, Internal Medicine, Washington Hospital Center, Washington, DC LEARNING OBJECTIVES: Recognize HIV disease as a cause or cofactor in the development of severe COPD resulting in cor pulmonale. CASE PRESENTATION: A 44 year old black male presented to the emergency department with three to five days of dyspnea, paroxysmal nocturnal dyspnea, peripheral edema and abdominal and scrotal swelling. He reported being totally well prior to the development of these symptoms. He denied cough, chest pain, fever or chills. Past medical history was remarkable only for childhood seizures. There was no history of hypertension, asthma or surgical illnesses. Social history was remarkable for remote use of crack cocaine, no alcohol use, and a cigarette smoking history of 1 ⁄ 3 ppd for 25 years. Family history was notable only for hypertension and coronary artery disease. Review of systems was otherwise negative. Learn that cardiac masses are commonly metastatic in origin, 3) Understand that lung cancer is a frequent cause of cardiac tumors. CASE PRESENTATION: A 68 year old man with emphysema and an 80 pack-year history of smoking presented to his internist with sudden onset dyspnea on exertion followed by swelling of the lower extremities. He denied PND, orthopnea, dyspnea at rest, or chest pain. His physical exam revealed a change in his murmur and lower extremity pitting edema. A TTE demonstrated normal LV function, and the patient began oral diuretic therapy. The edema improved, but his dyspnea was unchanged. Over the following three weeks, the patient further noted increased lethargy, frequent chills, decreased appetite and weight loss. On re-evaluation, he was febrile and hypertensive with splinter hemorrhages. Blood cultures were positive for E. coli. At admission, the physical exam demonstrated JVD, bibasilar crackles, a 3 ⁄ 6 holosystolic murmur at the left second intercostal space, and trace lower extremity edema. The EKG revealed a new incomplete RBBB. The TEE reported a 7.2cm ϫ 4.4cm homogeneous tissue density mass seen in the apex of the right ventricle causing RV outflow obstruction and severe right atrial enlargement. A biopsy obtained by right heart catheterization identified a non-small cell, poorly differentiated carcinoma suggestive of a lung primary. Spiral CT of the chest described a 1.8cm ϫ 1.2cm spiculated nodule in the right upper lobe. DISCUSSION: Physicians often overlook the possibility of an intracavitary tumor when evaluating symptoms suggestive of right ventricular outflow obstruction of uncertain etiology. However, cardiac tumors are known to present with this constellation of symptoms. Metastatic tumors occur 20-40 times more frequently than primary cardiac malignancies. In one series of 12,485 autopsies, secondary myocardial tumors were found in 55 (0.4%) patients. The most common etiologies were cancers of the esophagus (27%), lung (24%), lymphoma (16%), and liver (5%).

Other studies also suggest breast cancer as a significant source of metastatic cardiac tumors. Patients with right intraventricular masses may present with dysp-nea on exertion, syncope, lower extremity edema, murmur, and EKG changes (arrhythmias, right axis deviation, and RBBB). Echocardiography and MRI have been used to evaluate cardiac tumors with the latter providing greater anatomic detail. Our patient had right outflow obstruction as the initial presentation of a nonsmall cell lung cancer. This case reminds us that myocardial tumors should be included in the differential diagnosis of right-sided heart failure. year-old, previously healthy female was transferred from an outside institution for evaluation of jaundice. She noticed the jaundice five days prior, and noticed clay-colored stools and dark orange urine about three days prior to admission. She denied any recent abdominal pain, nausea, or vomiting. The patient had no other medical problems and was taking no medications other than a large amount of nutritional supplements. Two months before presentation, the patient took a trip to Africa and was on Mefloquine prophylaxis, two weeks before and four weeks after her journey. She has a 41 year old sister who was diagnosed with systemic lupus erythematosis at the age of sixteen. Her physical examination was remarkable for scleral icterus, a non-tender, smooth, palpable liver edge, three centimeters below the costal margin, and generalized jaundice. Her labs at the outside hospital revealed an aspartate aminotransferase (AST) of 1236 U/L, an alanine aminotransferase (ALT) of 989 U/L, a total bilirubin of 10 mg/dL and a gamma-globulin of 2.57 g/dL. Serologies for viral hepatitis were negative. A right upper quadrant ultrasound was negative, but a computerized tomography scan revealed periportal lymphadenopathy with the possibility of an infitrative process. A liver biopsy was performed which showed severe acute hepatitis with areas of parenchymal collapse and necrosis, without atypical lymphocytes and was non-diagnostic, but suggestive of autoimmune hepatitis. Because of the family history of autoimmune disease and the increased gamma-globulin component, the diagnosis of autoimmune hepatitis was made. The patient was started on Prednisone with subsequent resolution of the jaundice, transaminitis, and hyperbilirubinemia. DISCUSSION: Autoimmune hepatitis is an idiopathic hepatocellular inflammation, characterized by histological evidence of periportal necrosis, autoantibodies against hepatocyte surface antigens in the serum, and hypergammaglobulinemia. The disease is most predominant in young and middle-aged females. Symptoms includes those of fatigue, malaise, and anorexia with physical signs of jaundice and hepatomegaly, and the picture of acute hepatitis on laboratory evaluation. Although there are no pathognomonic features of autoimmune hepatitis, the diagnosis requires exclusion of other etiologies of hepatitis, including Wilson's disease, hemochromatosis, viral hepatitis, and cholestatic liver disease. Liver biopsy usually shows piecemeal necrosis. Treatment usually includes prednisone with or without azathioprine with a very small minority of patients requiring liver transplantation.

LEARNING OBJECTIVES: Recognize that Entamoeba histolytica should be included in the differential diagnosis of abdominal pain and can closely mimic inflammatory bowel disease or appendicitis.

Recognize that immune status can affect invasiveness of amebiasis. CASE PRESENTAION: Mr. M. is a 23 y.o. white male who presents with a long history of stomach problems. He reports that he has intermittent episodes of abdominal pain associated with vomiting since he was sixteen years old. He was hospitalized with his initial episode for a presumed appendicitis. Work-up was negative and no surgery was performed. Since that time, he has had intermittent episodes of severe abdominal cramping, vomiting and diarrhea, lasting 3-4 days. The vomiting is not usually associated with nausea, but is induced by the pain. When he vomits,he does not ususally get any relief of his pain. The cramping is usually in his lower abdomen, bilaterally. It does not radiate to his back or periumbilical area. He will have diarrhea that is sometimes slightly bloody, but this did not start until a few months ago. He reports fevers, chills, and fatigue associated with these episodes. He feels well between these episodes, but this current episode has lasted several weeks and he has lost 20 pounds. He presents after seeing another physician two weeks ago, who did an upper GI with a small bowel follow through, which was negative. Lab done at that time including a CBC, comprehensive profile and amylase were all normal. He was treated with Prevacid which has given him no relief. He denies any family history of inflammatory bowel disease; he traveled to Colorado this summer, no travel out of the country except a trip to the Dominican Republic seven years ago. His PMH is otherwise negative, he has no HIV risk factors.

The only abnormalities noted on physical exam was a temperature of 99.6 and diffuse tenderness to deep palpation on abdominal exam. No rebound or guarding, hyperactive bowel sounds. A presumptive diagnosis of inflammatory bowel disease was made. Further lab showed an ESR of 5 and a lipase of 87. Stools were negative for WBC's, Giardia antigen. He was set up for a colonoscopy which was cancelled when his stool came back positive for Entamoeba histolytica. He responded to treatment with metronidazole and paromomycin, but the during the second week of treatment he came in complaining of SOB, DOE and weakness. A CBC showed a HGB of 6.3, HCT 17.8 and a WBC of 7.1 which consisted of blasts with Auer rods. DISCUSSION: Amebic dysentery can mimic other intestinal diseases. The typical presentation usually consists of sx that are of less than one month duration, although some cases will consist of a chronic nondysenteric infection in which there are episodic sx of diarrhea, abdominal pain and weight loss persisting for years. This is likely in this patient, who did not develop invasive disease until his immune system was presumably affected by his developing AML. LEARNING OBJECTIVES: 1. Manage patients with recurrent immune thrombocytopenia CASE: A 50-year-old woman underwent splenectomy for chronic immune thrombocytopenic purpura (ITP), at that time an accessory spleen was discovered and removed. After 9 years of complete remission, she developed symptomatic relapse. She responded well to medical treatment with steroids and danazol. However, she relapsed two years later with mucocutaneous bleeding and severe thrombocytopenia. Physical examination showed petechiae. The platelet count was 6,000/uL and antinuclear antibody was negative. Her peripheral blood smear showed thrombocytopenia and absence of Howell-Jolly bodies. CT scan of abdomen suggested a small accessory spleen. Radionuclide Tc 99m liver and spleen scan revealed physiologically active splenic tissue in the splenic bed. She was treated initially with steroids, danazol and Rh0 immune globulin with a good response. Subsequently she underwent exploratory laporatomy with successful removal of 4 accessory spleens measuring between 0.4 ϫ 0.3 ϫ 0.3 cm and 1.0 ϫ 0.8 ϫ 0.7 cm. Her platelet count abruptly increased to more than a million and stabilized around 600,000/mm3. She remains symptom free and in complete remission for more than two years. DISCUSSION: Recurrent severe thrombocytopenia in chronic ITP after splenectomy is a common serious problem. An uncommon cause is presence of an accessory spleen, but the number of reported cases is increasing due to utilization of radionuclide splenic imaging studies. This is suggested by the absence of Howell-Jolly bodies in the peripheral blood smear and the diagnosis can be confirmed by radionuclide scanning. Literature review indicates that accessory splenectomy can be beneficial in the majority of patients but complete remissions are uncommon. Our patient achieved a sustained complete remission with a normal platelet count and no symptoms for more than two years after the second accessory splenectomy. This supports the benefit of accessory splenectomy and that it can be a definitive treatment for recurrent ITP. During the initial splenectomy, the surgeon should search thoroughly for accessory spleens, which if found should be removed. ITP patients with recurrence of thrombocytopenia even years after splenectomy should be investigated for accessory spleen/s. An 83-year-old woman had 5q-syndrome, marked by transfusion-dependent anemia (hematocrit (Hct) as low as 12%), neutropenia (absolute neutrophil count (ANC) was 800/ul) and mild thrombocytopenia (platelet count 103,000/ul). The spleen was slightly enlarged, increasing only from 2 cm to 4 cm below the left costal margin over 7 years. Because of progressive pancytopenia and increasing transfusion requirement, she was given subcutaneous Granulocyte colony-stimulating factor (G-CSF) 300mcg daily and erythropoietin (EPO) 70,000 units weekly. Her spleen enlarged within 5 weeks to 14 cm below the left costal margin. She was subsequently hospitalized with left upper quadrant pain and abdominal distension. Her Hct was 19%, ANC 10,180/ul and platelet count 9,000/ul. Her condition deteriorated and she died with acute respiratory distress syndrome. Autopsy showed multiple splenic infarcts and hemorrhagic bronchopneumonia. Patient #2: A 54-year-old man with refractory anemia with excess blasts had a palpable spleen tip. His ANC was 900/ul, Hct 13% and platelets 81,000/ul. While awaiting a bone marrow transplant he received periodic blood transfusions. To reduce his transfusion requirement he was given G-CSF 200 mcg and EPO 20,000 Units daily. His spleen progressively enlarged to 10 cm below the left costal margin over 15 weeks on this treatment and his transfusion requirement increased. After both growth factors were discontinued his spleen size decreased to 7 cm below the left costal margin. DISCUSSION: Splenomegaly is an uncommon complication of G-CSF and EPO and has been explained by extramedullary hematopoiesis. Splenic infarction and rupture are very rare. Both patients had progressive splenic enlargement. One patient developed splenic infarction and ultimately died. The second patient also developed splenomegaly but, with cessation of growth factors, the splenomegaly reversed to a large degree. Although EPO and G-CSF have each been reported to cause and aggravate splenomegaly, life threatening splenic complication is rare. We wish to alert clinicians to this uncommon complication particularly when G-CSF and EPO are used in combination in patients with splenomegaly. Patients receiving G-CSF and/or EPO, especially those with baseline splenomegaly, need fre-quent follow-up and monitoring of spleen size. Discontinuing growth factors may reverse such splenomegaly and prevent lethal complications.

LEARNING OBJECTIVES: 1. Recognize and treat an unusual presentation of Adult T cell Leukemia/Lymphoma (ATLL). CASE: A 71-year-old Jamaican woman presented with hypercalcemia, itching and jaundice. Physical examination showed a maculopapular skin rash in the chest wall, moderate hepatomegaly and small bilateral axillary lymphadenopathy. Laboratory tests were remarkable for: White cell count of 19 ϫ 103/uL (70% lymphocytes), normal hematocrit and platelet count, ALT 82 U/L, AST 70 U/L, LDH 1071 U/L, calcium 12.4 mg/dl and progressive elevation of serum biluribin to a peak of 23.9 mg/dl (Ͼ75% direct biluribin), alkaline phosphatase (ALP) to a peak of 952 U/L and GGT to a peak of 898 U/L. The peripheral blood smear was remarkable for lymphocytosis with 10% atypical lymphocytes and nucleated red blood cells.

Immunohistochemical studies in the bone marrow were compatible with T cell leukemia. Chromosomal analysis showed a normal karyotype. HTLV-1 antibody was positive. A bone scan showed a diffusely increased uptake. Ultrasound of the abdomen showed a moderate hepatomegaly without intra or extrahepatic biliary dilatation. A CT scan revealed a borderline enlarged spleen and mediastinal and retropertoneal lymphadenopathy. Based on these findings she was diagnosed to have an acute type of ATLL and immediately started on CHOP (cyclophosphamide, adriamycin, vincristine, prednisone) chemotherapy. After 2 cycles of CHOP the jaundice resolved completely with normalization of the liver function tests. Although she remained in a clinical complete remission in the liver she relapsed with leukemic meningitis and died before the 4th cycle of CHOP. DISCUSSION: The acute subtype of ATLL is characterized by a rapidly fatal course. Its treatment is still frustrating and the prognosis remains poor. Adriamycin based chemotherapy is the standard treatment with a median survival of 7-8 months and cures have not been reported. Our patient presented with an acute progressive obstructive jaundice most likely due to liver infiltration by the leukemic cells. This presentation is unusual as is the dramatic and complete resolution of obstructive jaundice upon combination chemotherapy. We would like to alert physicians to the mixed response of this fatal disease to combined chemotherapy. accounts for only 3% of these cases in the antibiotic era. Our case is a 19 year old previously healthy male patient who presented with a two week history of nausea, vomiting and chest pain. Review of systems was positive for dark urine, fatigue and anorexia over the same two weeks. He denied arthalgias,dysuria, uretheral discharge, or skin rash. On physical, the patient appeared acutely ill. Temperature was 102.0F, pulse 120 and respiratory rate in the 20s. Blood pressure was 100 ⁄ 70 . The PMI was hyperdynamic and a loud holosystolic murmur was heard from the LSB to the apex. The chest was clear. Pertinent laboratory results included a white blood cell count of 24,400 with a predominance of PMNs and a hemoglobin of 8.8gm. With the exception of sinus tachycardia, the ECG was normal. Chest xray revealed cardiomegaly. A transthoracic echo revealed aortic and mitral vegetations and severe insufficiency of both valves. An emergent cardiac catheterization showed a central aortic pressure of 140 ⁄ 40 and normal coronary arteries. The patient was taken urgently to the operating room where transesophageal echo revealed additionally an aortic root abscess, with extension to the right ventricle through a ventricular septal defect and tricuspid vegetation. Mechanical aortic and mitral prostheses were placed,the VSD was repaired, and the abscess and tricuspid valve were debrided. Cultures of blood and of the vegetations grew N. gonorrhoeae. The patient was placed on ceftriaxone for 1 week and discharged on oral lovafloxacin for five weeks. Post-op laboratory evaluation revealed a terminal complement deficiency which is known to be a risk factor for DGI. Although DGI is rare, it should be considered in a sexually active young patient. It can presents with tenosynovitis and uretheral discharge, although the absence of these features as in this patient doesn't rule out the diagnosis. Gonococcal endocarditis can involve normal as well as diseased valves. Typically, gonorrhea endocarditis can involve any heart valve and often requires surgical repair secondary to valvular insufficiency. With the advent of penicillin-resistant gonococcal infection,third generation cephalosporins or fluoroquinolones are an alternative treatment pending sensitivities. Our patient did not present with typical genitourinary features of Disseminated Gonococcal Infections. 2. Recognize anti-retroviral associated lactic acidosis. CASE PRESENTATION: A 45-year old man presented with a two day history of epigastric pain radiating to his back associated with nausea and vomiting. He denied alcohol use. His past medical history is pertinent for HIV infection, with the most recent CD4 count at 1,250 and an undetectable viral load, as well as hepatitis C infection. His medications were didanosine, stavudine, and nelfinavir. His physical examination was notable for mild hepatomegaly and moderate epigastric tenderness without rebound tenderness. Laboratory data revealed an elevated amylase and lipase and normal transaminases. A diagnosis of acute pancreatitis was made and his anti-retroviral medications were discontinued. Two days later, his abdominal pain had abated and his amylase and lipase returned to near normal values. He tolerated a regular diet. However, laboratory data revealed an anion gap of 20 (mEq/L) and a bicarbonate of 24 (mEq/L). His serum glucose was normal, there were no ketones in the urine, his serum and urine toxicology screen were negative, and there was no osmolar gap. His serum lactate was mildly elevated at 3 (mEq/L). The following day, he continued to improve but was found to have an increasing anion gap of 23. His serum bicarbonate was 20 and his serum lactate was 7. An abdominal CT scan demonstrated moderate pancreatitis and mild hepatomegaly with fatty changes. On the fifth hospital day, his anion gap had increased to 25 with a bicarbonate of 18. Liver tests revealed a mild transaminitis. An arterial blood gas revealed a pH of 7.30. A sugery consultation was obtained to evaluate for the possibility of mesenteric ischemia; no evidence was found to support this process. Over the next several days, the patient became progressively dyspneic and tachypneic, with increasingly diffuse abdominal pain, nausea, and vomiting. He was transferred to the medical intensive care unit where he was found to have a severe lactic acidosis, with a peak lactate of 40, and worsening tranaminitis. He suffered cardiac arrest; arterial pH was 6.85 at that time. Attempts at resuscitation were unsuccessful. DISCUSSION: This case illustrates two important points: the first is that an anion gap, especially when it is greater than 20 mEq/L, should be evaluated. Three common causes of anion gap metabolic acidosis are lactic acidosis, diabetic ketoacidosis, and uremia. The second point is the association between nucleoside analogues and lactic acidosis. The mechanism is not understood, but is thought to be related to mitochondrial toxicity. This case also serves to illustrate how well a patient can appear during the early stages of this syndrome. Treatment is supportive, while discontinuing the implicated antiretrovirals.

NK Ericsson, J Chang, Department of Medicine, Englewood Hospital and Medical Center, Englewood, NJ LEARNING OBJECTIVES: Recognize that use of diamox with thiazide diuretics may result in malignant arrhythmias. CASE: A 48-year-old woman with a 5-year history of hypertension, controlled with maxzide 75 ⁄ 50 mg p.o. daily and long-acting diltiazem 240 mg p.o. daily, was diagnosed with pseudotumor cerebri at another facility about one month prior to admission to this hospital. Her potassium level was 3.8 mEq/L. 5 days prior to admission, patient began taking diamox 500 mg p.o. B.I.D. On the day of admission, patient got out of bed to use the bathroom at about 4.30 a.m. About one hour later, patient was found unresponsive in bed. The paramedics arrived 40 minutes later and found patient in ventricular fibrillation. Patient developed a pulse after 13 minutes of resuscitative efforts. Patient was brought to the emergency department intubated and unresponsive. The patient did not use alcohol, cigarettes or illicit drugs. Physical examination revealed a moderately obese patient with blood pressure of 160 ⁄ 80 , heart rate of 130 bpm, respiratory rate of 16 per minute and temperature of 97.6o. The patient exhibited decerebrate posturing to noxious stimuli, but with reactive pupils. Corneal and oculocephalic reflexes were intact. Babinski reflexes were negative. The general physical examination was otherwise unremarkable. Her potassium level was 2.4 mEq/L. Initial EKG showed sinus tachycardia at 132 bpm, rightward axis and non-specific ST-T abnormalities. C.T. scan of brain without contrast was normal. Chest x-ray showed no cardiopulmonary abnormality. The patient was admitted as a case of severe anoxic encephalopathy resulting from cardiac arrest most likely secondary to diuretic-induced hypokalemia. Patient's diuretics were held. Diltiazem was used for blood pressure control. A total of 910 mEq of potassium chloride was used to raise and stabilize the potassium level within normal limits. Subsequent echocardiography showed no structural heart abnormality. Electroencephalography showed severe non-specific cortical neuronal dysfunction. Patient remained in persistent vegetative state throughout her hospital stay. There was no recurrence of malignant arrhythmia. DISCUSSION: Hypokalemia is prevalent in cardiac arrest. Some authors have suggested that hypokalemia may be the result rather than the cause of cardiac arrest. Ornato et al(l985) found that hypokalemia was associated with a 2.5-fold increase in relative risk of cardiac arrest.Although we would hesitate to draw conclusion from a single case, we feel that, in light of the temporal sequence of events in our patient, hypokalemia induced by the addition of diamox to maxzide is likely to have been responsible for our patient's cardiac arrest. 2. Diagnose amyloid deposition disease from its clinical and laboratory presentation. CASE PRESENTATION: A previously healthy 52 year old lawyer presented with two months of exertional dyspnea. Review of systems was negative for other car-diac or respiratory complaints. Past medical history was negative; the patient was on no medications. Initial exam was unremarkable with the exception of a mildly elevated diastolic blood pressure. EKG and chest x-ray were unrevealing. Pulmonary function testing was within normal limits. The patient's symptoms were felt to be consistent with angina and he subsequently underwent a stress test, where no significant EKG changes were noted. One month later, he presented in followup to his primary care physician. At that time, laboratory tests revealed a normal cbc and sma-7. He was scheduled for an echocardiogram, which revealed moderate concentric left ventricular hypertrophy and an ejection fraction of 60%. He was referred to cardiology for further evaluation, where it was initially felt that his symptoms were related to diastolic dysfunction. He was begun on an ACE inhibitor. One month later, he noted pronounced dyspnea and fatigue with exertion, and given increasing concern for coronary artery disease underwent cardiac catheterization, which revealed normal vessels. An extensive laboratory evaluation revealed: mild anemia (Hct 35.5), creatinine of 1.5, BUN 25, ESR 5; normal electrolytes, liver and thyroid function, urinalysis and serum and urine protein electrophoresis. Eight months after his initial presentation, the patient returned with progressive symptoms, unable to walk the short distance between the train stop and his downtown office. He had developed new lower extremity edema on exam. A repeat echocardiogram revealed new bi-ventricular dysfunction. A rightsided catheterization was performed with endocardial biopsy, which stained Congo Red and Trichrome positive, with amyloid deposition in the interstitium, endocardium, and vascular walls. He was subsequently found to have amyloid involvement in his bone marrow and kidneys, negating candidacy for a heart transplant. The patient underwent placement of an implantable defibrillator and was begun on chemotherapy, but unfortunately succumbed to a malignant arrhythmia one year after his initial presentation. DISCUSSION: Restrictive cardiomyopathies are the least common of the cardiac myopathic disorders (NEJM 1/23/97:267-274) and can be commonly overlooked in the differential diagnosis of a patient with subtle cardiac complaints. In any patient with suspected cardiomyopathy with anemia and renal insufficiency, amyloid must be considered. The absence of proteinuria and the negative serum and urine protein electrophoresis do not rule out the diagnosis. The patient is a 33 year old administrative assistant followed for her primary care in an outpatient clinic affiliated with a University medical center. She had been well generally, but over the prior six months had been seen more frequently by several providers in the clinic for allergic rhinitis and 2 to 3 episodes of sinusitis. She was referred to ENT for evaluation; complete head and neck exam and sinus CT were consistent with the diagnosis of bacterial sinusitis. On follow up with her primary care provider, the patient reported that she had felt more fatigued recently and had noticed the gradual worsening of a rash, primarily on her face. She does not smoke, drinks rarely, and has never used intravenous drugs. She is monogamous with her husband of 8 years and has a healthy 4 yearold child. On physical exam, she is 5lbs below her normal weight, has a mild facial rash consistent with seborrheic dermatitis, no oral findings, and a normal chest, cardiac and abdominal exam. Laboratory data was obtained to evaluate her recurrent sinusitis as well as her fatigue and weight loss: CBC revealed a normocytic anemia, LFT's and TSH-normal, HIV antibody-positive. The patient's PCP had sent the HIV antibody test "just to be complete" and had not prepared himself, or the patient, for the possibility that it may come back positive. She had come to the appointment alone and was shocked by the result. It had never crossed her mind that she might be at risk for HIV. Repeat HIV antibody testing was also positive, with a viral load of 500,000 and CD4 count 75. She has had an excellent response to HAART and currently has an undetectable viral load. DISCUSSION: HIV is the third leading cause of death among all women aged 25-44 in the United States, accounting for 11% of deaths. In spite of its prevalence, delayed diagnosis of women with HIV infection continues to be a major problem. Women, and clinicians, often underestimate the risk of infection with HIV, and several studies have shown that HIV screening that targets women who admit risk factors is a strategy that misses about half of infected women. It is critical that primary care physicians become familiar with these risk factors and recognize the unique manifestations of HIV infection in women. They must identify and treat women early in the course of the disease, and ideally counsel women in ways to reduce their risk of obtaining HIV infection. The primary care physician in this case was unaware of the epidemiology and clinical manifestations of HIV in women. As a consequence, the correct diagnosis was delayed, and the delivery of the bad news (the test result) was handled improperly. The spells consisted of diffuse flushing of the body, SOB, and lightheadedness leading to syncope. Some spells included chest tightness with diaphoresis. One episode was associated with a grand mal seizure. Past medical history was noncontributory. During severe spells, he typically presented hypotensive (systolic BP 40), cyanotic and with a violaceous rash. No other abnormalities were consistently noted. The patient usually was promptly resuscitated with fluids and pressors. Between spells, the physical exam was normal. Cardiac catheterization, CT scan of the head and abdomen, small bowel biopsy, EEG, tilt table, Holter monitoring, ECG and echocardiography were all normal. The CBC, electrolytes, LFTs, TSH, serum cortisol, urine metanephrines and 5HIAA were all normal. The only consistent abnormality was a prolonged partial thromboplastin time (PTT) during the spells, which resolved immediately thereafter. The serum tryptase level was slightly elevated at 1.2 ng/ml (N Ͻ1 ng/ml). BM biopsy confirmed the diagnosis of systemic mastocytosis. Systemic mastocytosis results from the overproduction of mast cells and their mediators within tissues. Heparin is a preformed secretory granule-associated proteoglycan released by mast cells. Clinically significant activity of heparin release in mastocytosis may include osteoporosis and inhibition of localized clotting. The release of heparin in mastocytosis may give rise to a transiently elevated PTT. Mastocytosis needs to be in the diagnostic repertoire of the general internist to streamline a usually invasive and unwieldy workup, as typified by this patient. A 73 year old woman was brought to the Emergency Department by her family with vomiting and lehargy. Her vital signs were normal, and physical exam was remarkable only for decreased level of consciousness. A CT of the brain showed no evidence of bleeding or infarction. A toxin screen was positive for acetaminophen. Upon questioning, the patient admitted to ingesting 50 grams of acetaminophen (100 Extra Strength Tylenol) In a suicide attempt approximately 18 hours prior to presentation. The first acetaminophen level returned at 394 mcg/ ml, which predicts a "high probability" for hepatic toxicity by the treatment nomogram. Additional relevant labs included AST 296, ALT 243, PT16.1/INR 2.1. EKG was notable for non-specific ST segment and T wave changes with a CK 500, CK MB index 6 and troponin greater than 50. N acetyl cysteine was administered at 22.5 hours post ingestion. Initial clinical impression was that the patient had early evidence of hepatotoxicity from her acetaminophen ingestion and had secondarily sustained a myocardial infarction. The patient was admitted to the MICU. On hospital day #3, AST and ALT peaked at 3000 each and INR at 6.0. Over the next 24 hours, the patients LFT's and CK's began to improve and by day #8, liver and cardiac enzymes and coagulation values returned to normal range. Despite aggressive elecirolyte repletion, several cardiac dysrhythmias were noted including runs of ventricular tachycardia up to 30 beats and sinus node re entrant tachycardia from Day #3 to Day #12. Serial EKG's revealed deep T wave inversions in all leads. An echocardiogram revealed global hypokinesis with EF of 35%. A cardiac catheterization on Day #14 found no significant coronary artery disease. The EKG abnormalities and temporal rise and fall of CK and CK MB indicate that this patient sustained subendocardial necrosis from her acetaminophen ingestion, rarely reported previously. DISCUSSION: Hepatic and renal toxicity following acetaminophen overdose are well recognized. However, case reports in the literature have also described cardiac toxicity. The exact mechanism of injury is unclear, but one hypothesis is that acetaminophen leads to the depletion of sulfhydryl groups and glutathione. In the myocardium, these SH groups are necessary for the activity of endothelin-derived relaxation factor, an endogenous vasodilator. Lack of EDRF produces coronary artery constriction and myocardial ischemia. As this case illustrates, the onset of this effect can precede serious liver toxicity. The mechanism leading to cardiac arrythmias is unclear, though reports indicate multifactorial causes including electrolyte abnormalities, direct myocardial toxicity by acetaminophen by products, and altered serum free fatty acid levels. Accordingly, an EKG and early cardiac monitoring should be performed on all patients who present following massive acetaminophen ingestion. Though treatment is largely supportive, it is critical to recognize the potential for significant myocardial damage. pack-year smoking history presented with right shoulder pain for two weeks. He denied neurologic symptoms or history of trauma or heavy exertion. Active and passive range of motion (ROM) at shoulder joint was intact but difficult. Treatment with nonsteroidal anti-inflammatory drug (NSAID) yielded no improvement. Minimal relief was obtained with a new NSAID and acetaminophen. One month later, the patient had decreased visual acuity in right eye, followed two weeks later by polydipsia, polyuria and nocturia. He had no nausea, headache, fever, weakness, numbness, head trauma, seizure, weight loss or night sweats. On examination, he appeared fatigued. His pupils were equal, round and reactive to light with visual acuity of 20 ⁄ 60 in the right eye and 20 ⁄ 25 in the left. Fundoscopic examination was normal. Cranial nerves were normal as were motor, sensory and cerebellum testing. Pain localized to the proximal humerus limited ROM of right shoulder. No abnormal cardiopulmonary, abdominal findings or lymphadenopathy found. Glucose tolerance test and prostate ultrasound were within normal limits as were chemistries, CBC, and urine. Two days later, the patient lost vision in right eye. His visual acuity was 20 ⁄ 800 in the right eye and unchanged on the left. Magnetic-resonance-imaging of the brain and orbits showed innumerable intracranial lesions, including one lesion in pituitary infundibulum. There were no lesions within the orbits. Right shoulder x-ray revealed lytic lesions over proximal humerus. Chest and abdominal computed tomography revealed an upper lobe mass in right lung, right hilar nodes and bilateral adrenal masses. Lung biopsy revealed squamous cell cancer. The patient was diagnosed with staged IV non-small cell cancer, DI due to pituitary metastasis and optic neuropathy due to meningeal compression. The patient received external beam radiation to the whole brain with no improvement. He refused chemotherapy. He was treated with desmopressin, opioids and anti-emetics. The patient died comfortably at home a few weeks later. DISCUSSION: This case illustrates metastatic malignancy as an etiology of central DI. Central DI can be familial, due to head trauma, or iatrogenic in the setting of hypophysectomy or irradiation. Other causes include infectious, vascular, neoplastic, and infiltrative disorders. The onset of DI is variable and may occur at any age. Often the patient consumes large amounts of liquids and excretes large amounts of very diluted urine. The patient is usually awakened at night by thirst, which is generally not true in psychogenic polydipsia. As long as the patient is able to replace the urinary losses, the electrolytes remain normal. with a recent increase in intensity. The pain was associated with bloating, constipation, & flatus. Physical exam was unremarkable, including a "symmetric" neurologic. He was hemoccult negative. Sedimentation rate, liver function tests, amylase, CBC & abdominal ultrasound were also negative. He was treated with a proton pump inhibitor. Three months later, his pain increased with referral to his back. He also complained of abdominal distention & difficulty voiding. A T11, T12 level on sensory neurologic exam was noted with a positive Romberg to the right. MRI demonstrated a thoracic level schwannoma which was removed. The neurologic and abdominal findings disappeared following removal of the tumor. DISCUSSION: A recent editorial in The Green Journal characterized the fundamental difference between internal medicine and family practice as one of heuristics vs. critical thinking. Essentially, heuristics or rule-based thinking can lead to the appropriate diagnosis most of the time, but the ability to think critically when the atypical presentation eventuates should be the internist's forte. Pseudovisceral syndromes (PsvS), or radicular pathology mimicking an intraabdominal process, are a true test of the critical diagnostic paradigm. Although rare, PsvS can mimic almost any abdominal disease & require appropriate diagnostic suspicionparticularly when initial workup for abdominal pain is negative. Unnecessary and expensive testing can be avoided and as such PsvS require critical thinking in the context of the history & physical exam. The words of Sir Zachary Cope may serve the internist well, "One often, if not always, learns more by analyzing the process of and detecting the fallacy in an incorrect diagnosis than by taking unction to oneself when the diagnosis proves correct." PsvS appear to serve as an appropriate paradigm for the diagnostic process of internal medicine.

LEARNING OBJECTIVES: 1) To discuss unusual sites of venous thrombosis occuring in pregnancy. 2) To illustrate ovarian hyperstimulation syndrome as a risk factor for thrombosis during pregnancy. CASE: A 36 year old woman presented to hospital complaining of left neck pain. She was eight weeks pregnant with twins, having undergone in vitro fertilization for ovulatory and severe male factor infertility. Her only past medical problem was thalassemia minor. She had been admitted six weeks prior with ovarian hyperstimulation syndrome, two weeks after her oocyte transfer. Her symptoms at that time had included nausea, vomiting, rapid weight gain and lower abdominal pain. An ultrasound had revealed ascites and bilateral enlarged multicystic ovaries consistent with ovarian hyperstimulation syndrome. She had improved with supportive care. Four days prior to presentation she noted the subacute onset of left neck pain, followed by mild dysphagia and then obvious swelling of the region. Examination revealed tender swelling of the left neck without erythema. A lateral neck xray was normal but ultrasound revealed left internal jugular vein thrombosis. She was hospitalized and treated with intravenous heparin. Testing for hypercoagulable states revealed her to be heterozygous for the Factor V Leiden mutation. She was switched to low molecular weight heparin in therapeutic doses. Her symptoms gradually resolved and the remainder of her pregnancy has been uneventful. DISCUSSION: Pregnancy is a thrombophilic state, conferring a risk of thromboembolism that is five to six times that of the non-pregnant state. The overall risk of a thrombotic event is 0.5 to 3 per 1000 pregnancies and is even higher in women with a prior thrombosis or thrombophilia. Thrombosis during pregnancy may occur at unusual sites including the pelvic veins, the ovarian veins and rarely the cerebral venous sinuses. Ovarian hyperstimulation syndrome (OHSS) results from ovulation induction with gonadotropins during the process of in vitro fertilization, with recruitment of multiple ovarian follicles leading to cystic enlargement of the ovaries. This is associated with a marked increase in vascular permeability with accumulation of ascites, pleural and pericardial fluid. Hypovolemia, hemoconcentration and hypotension may result. Venous thrombosis is common in this setting, occuring in up to ten to thirty percent of cases. The most frequent sites of thrombosis are the internal jugular and subclavian veins. Thrombosis at other unusual venous and arterial sites has also been reported. The etiology of OHSS and its complications are incompletely understood. To date there has been no systematic study of the frequency of thrombophilias in OHSS-associated thrombosis. To discuss lymphoepithelial parotid cysts as a first manifestation of HIV infection. CASE: A 30 year old Liberian woman who was thirty weeks pregnant presented to hospital with a two day history of fever, chills and right jaw pain. Her past medical history included anemia and sickle cell trait. She had received a blood transfusion in Liberia five years prior after a surgical procedure. She reported three lifetime sexual partners with no history of sexually transmitted diseases or intravenous drug use. Her current pregnancy had been uneventful until 48 hours prior to presentation when she developed fever, chills and rigors. She also noted pain and tenderness at the angle of her jaw on the right side. On examination she appeared toxic with a temperature of 100.8 degrees and a blood pressure of 81 ⁄ 44 . She was noted to have a tender swollen lump in the region of her right parotid gland and bilateral soft, mobile cervical adenopathy. Her pharynx was unremarkable and the remainder of her examination was normal. WBC was 5.7 with frequent band forms. A chest x-ray was unremarkable. An HIV test was ordered. Blood and urine cultures were drawn and the patient was started on intravenous ampicillin/sulbactam and gentamicin. An ultrasound the next day revealed a 4 cm. heterogeneous hypoechoic mass in the area of the right parotid gland, with no evidence of abscess. Cultures were negative but antibiotics were continued and her fever gradually resolved. An MRI study revealed multiple cystic masses scattered throughout the parotid glands bilaterally and extensive cervical lymphadenopathy, consistent with benign lymphoepithelial parotid cysts seen in HIV-infected persons. The HIV test was reported positive two days later. Her initial CD4 cell count was 249. She was discharged on antiretroviral therapy and antibiotics. DISCUSSION: Parotid disease in HIV-infected persons has a lengthy list of potential causes including: lymphoma, Kaposi's sarcoma, viral or bacterial parotitis and mycobacterial lymphadenitis. The parotids are the only salivary gland with integrated lymphoid tissue. Lymphoepithelial parotid cysts may occur in HIV-infected persons due to hyperplasia of this lymphoid tissue leading to ductal obstruction and secondary cyst formation. The lymphoid tissue stains positive for p24 antigen. The cysts are multiple and bilateral, leading to painless parotid enlargement. This typically occurs early in HIV infection and may be the first clinical manifestation of HIV disease. Viral and bacterial superinfection of lymphoepithelial parotid cysts has been described among non-pregnant patients but this is the first such case reported as an initial manifestation of HIV disease during pregnancy. to a tertiary teaching center with four days of fevers, myalgias, confusion and diffuse weakness after a tick bite. He had a history of type 2 diabetes mellitus, hypertension, myocardial infarction and a subcortical stroke with no neurologic sequelae. He was a binge drinker and had fallen asleep in a local park when intoxicated. Upon awakening he found a tick attached to his right shoulder. The following day he developed a headache, diffuse myalgias, and fever. Over the next three days his family noted that he became progressively weak and confused to the point that he could not get out of bed without assistance. On the day before admission a diffuse, non-pruritic erythematous rash appeared over his elbows and lower back. Upon arrival to the hospital he had a fever to 39.7 C, a HR of 128, RR of 24 and BP of 117 ⁄ 72 . Bibasilar rales we heard on lung exam. He was tachycardic but his heart sounds were normal and without murmurs. His abdominal exam was benign and no spleen tip could be appreciated. He had a diffuse, flat, blanching erythematous rash over his elbows, buttocks and back. His thighs and calves were tender. His neurologic exam was remarkable for lethargy and mild disorientation. Cranial nerves and cerebellar function were normal and muscle strength was 4 ⁄ 5 in all extremities. WBC-10.1, Hb-15.4, plt-127, PT and PTT normal, SMA-7 normal except for Cr-1.5 and glc-199. AST-83 but all other LFTs were normal. ESR-55. CSF analysis: WBC-7, RBC-20, pro-106, glc-98. UA remarkable for protein and 5-10 RBCs/hpf. CXR showed faint bibasilar opacities. The patient was admitted and administered intravenous doxycycline and ceftriaxone. Blood and urine cultures were negative, as were Lyme serologies. Initial IgM for R. rickettsii was 1:16 and IgG was 1:32. Examination of the blood smear showed toxic granulations but no WBC inclusions. A biopsy of the rash revealed a leukocyto-clastic vasculitis. The patient was persistently febrile for the first five days of admission, and then defervesced with a coincident clearing of his sensorium. The patient was discharged on the seventh hospital day with the presumptive diagnosis of Rocky Mountain Spotted Fever. Oral doxycycline was prescribed with the intent to treat the patient for a total of ten days. Upon follow-up in the internal medicine clinic three weeks later the patient complained of vertigo and had been walking with a cane. His exam at that time showed ataxia on finger-to-nose testing and a swaying to the right on gait testing. Repeat serologies were drawn. R. rickettsii IgG titer was 1:512. He was given meclizine for his symptoms, which gradually resolved over the next four months. A 54 year old male was admitted to the hospital following a three week history of fever, nausea, vomiting, and increasingly severe weakness of his lower extremities. His past medical history was significant for rheumatoid arthritis, hypertension, hyperlipdemia and prostate surgery. There was no history of smoking, alcohol or illicit drug use. There was no recent travel. His chronic medications included an antihypertensive and a cholesterol lowering agent. He had been recently started on a new arthritis medication four weeks prior to admission. He did not recall the name of the medication but stated they were 12 "little white pills" and he was to take three per week. Because of his severe arthritis pain he completed his new prescription in four days. Physical examination: Temp101.2F, BP 130 ⁄ 86 , Pulse 92. He was alert and oriented x 3. His cranial nerves were normal. Strength in his upper extremities was 3 ⁄ 5 and 2 ⁄ 5 in the lower extremities. Deep tendon reflexes in the lower extremities were 2ϩ. There were no tremors, fasiculations or sensory deficits. Plantar reflexes were normal. The remainder of the physical exam was unremarkable except for classic findings of rheumatoid arthritis. Laboratory studies: WBC 16,200, Hgb 13.9g/dl, platelets 275,000, metamyelocytes 6%, bands 11%, PMNs 69%, lymphocytes 11%, monocytes 3%, total bilirubin 0.1 mg/dl, LDH 1,169 U/L, AST 1,426 U/L, ALT 700 U/L, Alkaline Phosphatase 37 U/L, GGTP 33 U/L. Hepatitis A, B, and C studies were negative. Urinalysis: protein Ͼ300 mg, hemoglobin -4ϩ, RBC -negative, coarse granular cats were seen. CPK 15,045 U/L (99% CK-MM), BUN 35 mg/dl, creatinine 1.4 mg/dl. A toxicological screen was negative. Acetylcholine receptor AB was normal. DISCUSSION: The patient's "little white pills" were identified as methotrexate 2.5 mg. His other medications were cerivastatin 0.3 mg/day and lisinopril 5 mg/day. His progressive neurological deficits were a manifestation of methotrexate toxicity. Additionally, he developed both hepatitis and rhabdomyolysis. This case is a cautionary tale about drug interactions and patient compliance. Although the correct use of methotrexate was clearly listed on the prescription bottle and his physician had explained the purpose of the drug and its side effects, the patient self medicated himself and accordingly suffered disastrous consequences. It also underscores the potential severity of drug interactions and should emphasize that when confronted with a diagnostic dilemma, the physician must always consider the patient's medications as a potential cause of complications and morbidity. week history of intermittent pain behind both eyes associated with throbbing right sided headache radiating to the right occiput. The pain was worse at night time. She had gained 8 lb weight over the previous 4 weeks. Three years previously, in her first pregnancy she had been diagnosed with idiopathic intracranial hypertension after a similar presentation at 16 weeks gestation. She was treated in her first pregnancy with prednisone without complication, though had a premature delivery at 35 weeks. She had no other past medical history. She was taking acetaminophen with codeine as required and prenatal vitamins. She had no allergies. Her family history was of diabetes, hypertension and breast cancer. She had smoked 5-7 cigarettes daily for 8 years. She did not drink alcohol or take recreational drugs. She was recently unemployed. On examination she wighed 161 pounds, her pulse rate was 80/min, BP 118 ⁄ 70 , her cardiovascular examination was normal. Her chest was clear and abdominal examination was unremarkable other than uterine enlargement consistent with dates. Neurological examination revealed a mild left ptosis and bilateral papilledema, but was otherwise normal. A presumptive diagnosis of idiopathic intracranial hypertenison (IIH) was made and supported by normal cranial CT scan and elevated opening pressure of 50 mmHg on lumbar puncture with normal CSF. She was assessed by ophthalmology who agreed with the diagnosis and recommended acetazolamide, which was commenced at a dose of 250 mg bid. This resulted in symptomatic relief of her headaches, but resulted in side effects of nausea, mild metabolic acidosis and hypokalemia. Her acetazolamide was discontinued and she was com-menced on prednisone, 20 mg bid which has been maintained throughout her pregnancy. DISCUSSION: Idiopathic intracranial hypertension has been documented to have its onset in pregnancy, but it has recently been suggested that the association between IIH and pregnancy is coincidental due to the age and sex predilection of the disease. Onset of disease has been associated with rapid weight gain. This case illustrates onset of disease during pregnancy, with resolution postpartum, followed by recurrence in a subsequent pregnancy. year old gravida 1 woman presented at 16 weeks gestation with a few days gradually increasing peripheral oedema and shortness of breath. She admitted to poor sleep and hypersomnolence. She had a 50 pack year smoking history and a family history of diabetes and premature coronary artery disease. Initial clinical assessment and chest X ray was suggestive of biventricular failure. A blood gas showed a pH of 7.40, pO2 of 41, pCO2 of 44 and total bicarbonate of 29. A V/Q scan was low probability. Swan-Ganz catheter measurements were consistent with secondary pulmonary hypertension with a pulmonary artery pressure of 41 systolic and pulmonary vascular resistance of 20 to 42. She was treated with diuretics and improved markedly.She was maintained on 100 mg furosemide bid. Physical examination at our medical center 2 weeks later revealed a morbidly obese woman with BP 136 ⁄ 60 , pulse 98, respiratory rate 16. Her tongue was slightly large with a low lying soft palate. Her cardiac examination was unremarkable. There was 3ϩ pitting edema up to the level of her thighs. Her chest was clear. Abdominal examination showed morbid obesity. Pulmonary function tests revealed a restrictive pattern. Sleep study confirmed obstructive sleep apnea, with desaturations to 72% despite 2 liters of nasal oxygen. Echocardiogram showed normal left and right ventricular function. Doppler evaluation estimated a peak pulmonary systolic pressure of 62. She declined right heart catheterization. Her pulmonary hypertension was considered to be secondary to obstructive sleep apnea and restrictive lung disease imposed by her obesity. She was managed with bilevel positive airways pressure administration which prevented her oxygen desaturation during sleep. A multidisciplinary approach to her delivery involving maternal-fetal medicine, critical care and general medicine is planned with a slow early epidural and passive descent of the fetus. DISCUSSION: Pulmonary hypertension in pregnancy is associated with substantial maternal morbidity and mortality. Presenting features, etiologies and management of this rare but critical medical complication of pregnancy will be discussed. A 61 year old female with hypertension and chronic depression presented with a one week history of nasal congestion, cough and maxillary sinus tenderness. She was diagnosed with a bacterial sinusitis and given trimethoprim-sulfamethoxizole and dextromethorphan. Five days later, she presented with nausea and vomiting. Her physical exam was unchanged. She was told to push fluids, continue the dextromethorphan but discontinue her antibiotic. Two days later she presented with dizzyness and generalized anxiety. She was noted to be orthostatic, flushed, diaphoretic, tachycardic, hypertensive and delirious. She was admitted to the hospital. In addition to her cough suprressant, medications included venlafaxine 37.5 mg 3 po bid and enalapril 10 mg po bid. On physical, she was a tremulous anxious appearing female with flushed facies. She was afebrile, pulse was 108 and her BP 190 ⁄ 86 . On neurologic examination she was delirious with symmetric hyperreflexia. Muscle tone was normal. Her CBC, TSH, renal function and urinalysis were normal. CK was 606 with a RI 0.6. The patient's nausea, vomiting, flushing, diaphoresis, tachycardia, hypertension, delirium, and elevated CKs, in the absence of muscle rigidity and fever, were thought to be due to the SEROTONIN SYNDROME. Her venlafaxine and dextromethorphan were discontinued and she was treated with cyproheptadine with full resolution of symptoms. She was discharged to home on the fourth hospital day. DISCUSSION: The serotonin syndrome is an important, albeit rare, complication of selective serotonin reuptake inhibitors (SSRIs). It can progress to cardiac arrest, seizures and multiorgan system failure. Suspicion for the serotonin syndrome should be raised in patients with exposure to a drug, or combination of drugs, that increase CNS serotonin activity. The classic presentation includes cognitive and behavioral changes, autonomic nervous system dysfunction and neuromuscular abnormalities. There are no specific laboratory abnormalities and an elevated CK is seen in 18% of cases, making it difficult to distinguish between neuroleptic malignant syndrome. In this case, preciptation of the Serotonin Syndrome was likely secondary to the use of dextromethorphan with venlafaxine. Dextromethorphan, a potent inhibitor of serotonin uptake, can lead to increase CNS serotonin activity. Treatment includes supportive care, discontinuation of serotonergic agents and antiserotonergic drugs (cyproheptadine).

Since SSRIs are commonly administered by general internists in the treatment of depression and since dextromethorphan is a common ingredient in prescription and over-the-counter cold remedies, knowledge of the serotonin syndome is imperative.

LEARNING OBJECTIVES: To discuss a rational approach to the evaluation of unintentional weight loss in the outpatient setting. CASE: The patient is a 35 year old woman with a history of hypertension and Turner's syndrome who presented for a routine physical. She denied any complaints, including nausea, vomiting, early satiety, diarrhea, hematochezia, abdominal pain, oral ulcers, fevers, night sweats, chills, cough, dyspnea, hematuria, vaginal discharge, menorrhagia, or symptoms of depression. Medications included hydrochlorothiazide, loratidine, and oral contraceptives. Family history was negative for cancer, diabetes, or gastro-intestinal diseases. Social history included 1 ⁄ 2 pack per day of cigarette smoking without alcohol or illicit drug use. Her physical examination was notable for a 10 pound weight loss over three months, dropping from a stable weight of 124 pounds. The remainder of the exam, including rectal and pelvic exams, was normal. Laboratory data included a normal CBC, a normal TSH, normal liver injury and synthetic tests, and a random glucose of 172. A chest radiograph was within normal limits. Three months later the patient returned for further follow-up. She continued to be without complaints. Her weight had fallen another four pounds (to 110) and a fasting glucose was 87. Three months later her weight had stabilized at 110 pounds. DISCUSSION: Asymptomatic weight loss is a common problem with potentially serious implications. It can be the first sign of an underlying malignancy, point towards the development of systemic disease, or expose underlying mental illness. Weight loss can be generally classified by four pathophysiologic mechanisms: decreased intake, impaired absorption, increased loss, and excess demand. There have been three well designed studies published in the last twenty years that provide a framework for a rational clinical evaluation. Each of them examined unintentional weight loss in a predominantly elderly age group. Although they differ in methodology and setting their results were concordant. Cancer was the most frequently diagnosed cause of involuntary weight loss (accounting for 16-36%). However, the vast majority of cases were not caused by an undiagnosed neoplasm. Depression, diabetes, thyroid problems, and gastrointestinal disorders all accounted for some cases. Approximately 25% of the cases never received a diagnosis. This cohort differs from others with asymptomatic weight loss in that their mortality and morbidity do not differ from age-matched controls. All three studies showed that when an etiology was discovered it was diagnosed by a careful history and physical exam coupled with basic blood work (CBC, LFTs, TSH), a chest radiograph, and age-appropriate cancer screening. In this case, no specific cause of weight loss was determined and the patient has had an uncomplicated course.

McKenzie, Occupational Medicine, University of Pennsylvania Medical Center, Philadelphia, PA LEARNING OBJECTIVES: 1) Diagnose occupational asthma (occ asthma). 2) Recognize importance of workplace modification and manage return to work issues. CASE: A 46 yo clinical pathologist presents with complaints of increasing shortness of breath for the past 4 years. He was well until the beginning of his internal medicine residency 18 years ago when he noted an erythematous rash on his hands whenever he used latex gloves. This rash did not clear with self-prescribed steroid creams. At the beginning of his pathology residency 14 years ago he noted a non-productive cough after exposure to formalin. He remained in relatively good health except for occassional chest tightness, wheezing and coughing while at work where he was exposed to latex and formalin on a daily basis. These symptoms did not interfere with his daily activities which included vigorous exercise at least 3 times each week. Approximately 4 years ago he noted that his chest tightness, wheezing and cough became significantly and progressively worse after an exposure. He also developed progressive dyspnea on exertion such that, by the time of presentation, he was unable to walk more than a flight of stairs or a hundred yards without feeling markedly short of breath. His symptoms were worst at the end of the work week as was his self-monitored peak flow, while his symptoms lessened on weekends and during vacation and his exercise tolerance increased. He also noted heavy breathing when he "flippped" latex gloves off his hands and experienced similar symptoms when he ate fruit such as kiwi, avocado and banana. He was evaluated by an allergist a few months ago and was diagnosed with chronic sinusitis and allergy to formaldehyde and latex. He was treated with two courses of antibiotics without relief, and prescribed a steroid taper and a beta agonist inhaler to ameliorate his response to exposures at work. PE: T98.7 P84 BP130/90 Gen: obese. Lungs: distant breath sounds, no wheezing. Cardiac: nl. Abd: obese. Skin: no rashes. Neuro: AX3, nonfocal. Labs: CXR: poor inspiration. Chest CT: mild bronchial wall thickening c/w mild airway dz/bronchitis. Spirometry: FEV1 ϭ 77% predicted, FVC ϭ 81% predicted, FEV1/FVC ϭ 95% predicted. Cardiac ECHO: trace TR. Latex skin test:(ϩ). Latex RAST:(Ϫ). DISCUSSION: The prevalence of latex induced occ asthma is 2.5-6%. RAST is (Ϫ) in up to 30% of pts with latex allergy. Formaldehyde is an irritant, may exacerbate asthma and can lead to chronic bronchitis. This patient was removed from the workplace shortly after presentation as no reasonable accomodation could be made for him and he was at risk for potentially serious and fatal anaphylaxis. Clinicians should suspect occupational asthma in patients with new onset asthma or asthma symptoms that worsen during or after work. Treatment for latex allergy is cessation of exposure.

Medicine, SFGH-UCSF, San Francisco, CA LEARNING OBJECTIVES: 1) Recognize neurocysticercosis as a cause of focal seizures. 2) Review diagnosis and management of neurocysticercosis. CASE PRESENTATION: A 67 year-old homeless, white man presented to the clinic with spells of right facial twitching for two days. He denied associated loss of consciousness, headache, visual symptoms or other focal neurological symptoms. His past medical history was significant for remote tobacco and IVDU, diabetes mellitus, elevated PSA and recent negative HIV test. He denied any foreign travel. During the physical examination the patient was witnessed to have repetitive jerking movements of the right face lasting one minute. He was then noted to be dysarthric with a central 7th nerve palsy. His focal deficits resolved completely after ten minutes and the rest of his examination was normal. Work-up included normal CBC, electrolytes, RPR, urine tox and chest X-ray. A head CT with and without contrast revealed small, partially-calcified, ring-enhancing lesions. The largest in the left frontal lobe was associated with vasogenic edema. Neuroradiographically this was most consistent with neurocysticercosis but TB, metastases and toxoplasmosis could not be excluded. A PPD and serum cysticercosis immunoblot assay were negative. The patient refused lumbar puncture and refused treatment with anti-epileptics. A follow-up head MRI at 3 months revealed 5 partially calcified lesions with interval resolution of edema and no evidence of abnormal enhancement. The patient has remained seizure free for the past 8 months. DISCUSSION: Neurocysticercosis is the most common parasitic central nervous system infection worldwide and is caused by Taenia solium, the pork intestinal tapeworm. It is an important cause of adult onset epilepsy particularly in patients from endemic areas namely, Mexico, Central and South America, India and sub-Saharan Africa. The incidence in the US is not known but estimated to be over 1000 cases per year. Transmission can be from either ingestion of undercooked pork or less commonly from ingestion of T. solium eggs in fecally contaminated food or direct contact with a tapeworm carrier. Head CT and MRI are the most effective means of diagnosis. The sensitivity of antibody testing is high for multiple active cysts (99%) but much lower for calcified stage of disease and therefore, a negative serology should not exclude diagnosis. Treatment with antihelminthic drugs is controversial as most disease is self-limited and in some cases treatment may precipitate complications. Stool OϩP studies are insensitive but may be useful to screen close contacts of infected patients. This case is interesting in that it involves a patient who does not come from an endemic area but lives among many homeless people who do. A 51 year old female presented to the emergency room with total body myalgias and lethargy. Four weeks earlier, the patient's primary physician discontinued simvastatin, because of complaints of arthralgias. At that time, the patient began taking both gemfibrozil and cerivastatin. There was no history of recent trauma, excessive exercise, infection, seizure or ethanol abuse. Past history was significant for an aortobifemoral bypass with a left nephrectomy in 1992 for peripheral vascular disease and renal artery stenosis. She also had a history of hypertension, hypercholesterolemia and coronary artery disease. In addition to gemfibrozil and cerivastatin, the patient was also taking metoprolol, amlodipine/ benazepril, cyclobenzaprine and naproxen. Her baseline creatinine one month prior to this presentation was 1.6. On examination, the patient appeared acutely ill. Blood pressure was 170 ⁄ 70 . The exam was otherwise unremarkable. Laboratories revealed potassium 8.0, creatinine 7.6 and creatinine kinase (CK) 60,000. EKG exhibited PR lengthening and peaked T waves. She was diagnosed with rhabdomyolysis and acute renal failure. She was given kayexalate, sodium bicarbonate, glucose, insulin, calcium chloride, IV fluids and mannitol. Following one course of dialysis, potassium was 3.9 and creatinine 4.9. Laboratory values returned toward normal over several days, and no further dialysis was necessary. One month later her creatinine was 1.3. DISCUSSION: HMGCoA reductase (statin) medications, either alone or in combination with other medications, can cause myopathies including asymptomatic elevations of CPK, myalgias, myositis or rhabdomyolysis. Statins are metabolized through the cytochrome P450-3A4 enzyme; any medication also metabolized by this enzyme can cause an increase in statin activity and likelihood of myopathy. For both statins and gemfibrozil, the risk of myopathy increases with coexistent renal insufficiency. Approximately 5% of patients taking both gemfibrozil and a statin develop asymptomatic rises in CK that are usually reversible with discontinuation of the medications. Rhabdomyolysis is a very rare but serious complication of combination therapy, and only one case report has been published describing rhabdomyolysis in a patient taking both gemfibrozil and cerivastatin. All physicians prescribing statins should recognize that these medications, either alone or with gemfibrozil, can cause serious myopathies, especially in patients with renal insufficiency.

College of Medicine, Houston, TX LEARNING OBJECTIVES: 1) Recognize common manifestations of sarcoidosis and thyroiditis. 2) Consider the possibility of two simultaneous processes when constructing a differential diagnosis. CASE PRESENTATION: A 52 year old African-American male presented to the Houston VAMC Emergency Department (ED) complaining of dyspnea. He related a two-week history of neck swelling and pain radiating into the right ear. One day prior to presentation, he noted fever and dyspnea. He denied cough, sputum, or pleuritic pain. His past medical history included type II diabetes mellitus, depression, and lower back pain, all longstanding. He denied recent travel. His diet reflected adequate iodine intake. He quit cigarette smoking several years prior. He previously worked as a welder. He denied any family history of cancer or endocrine/autoimmune diseases. In the ED, his exam revealed a temperature of 102F, pulse 130/minute, and 24 respirations/minute. Diffuse cervical and axillary lymphadenopathy were noted. Exopthalmos or lid lag were absent. Thyroid exam revealed a visible goiter, warm to the touch and tender to palpation. Auscultation of the chest was notable for bilateral basilar rales. Evaluation by flexible nasotracheal endoscopy revealed mild to moderate extrinsic compression of the trachea, and the patient was urgently intubated. After intubation, oxygen saturation was 96% on 60% FIO2. Chest radiograph revealed hilar lymphadenopathy and a right posterior-basal infiltrate. Computed tomography revealed diffuse enlargement of the thyroid gland as well as cervical, axillary, hilar, and para-aortic lymphadenopathy. Fine needle thyroid aspirate showed nonspecific inflammation. Two separate cervical lymph node biopsies revealed non-caseating granulomas consistent with sarcoidosis, and bronchoscopy with biopsy also revealed non-caseating granulomas. TSH initially was normal, but became suppressed shortly after the CT scan (which included iodinated contrast). DISCUSSION: In our attempt to identify a parsimonious diagnosis to explain all of the patient's findings, our list of possibilities included exceedingly rare entities such as sarcoidosis of the thyroid, lymphoma with thyroid involvement, and metastatic thyroid carcinoma. Upon review of incidence and prevalence data, we realized that the coexistence of two diagnoses (thyroiditis plus acute or chronic sarcoid) is more likely. Our search for a unifying diagnosis was based on the assumption that one event is more likely to occur than two simultaneous events. In this case, however, a single diagnosis necessitates the consideration of atypical presentations of rare entities. This is in contrast to the scenario of two typical presentations of more common entities. While the concept of parsimony is helpful in constructing a differential diagnosis, we found that strict adherence to the goal of a unifying diagnosis led us to overlook the more obvious choice. He took no medications and had no known drug allergies. He smoked a 1 ⁄ 2 ppd and had 2-3 beers each weekend. He lived with his wife, who had also lost sleep from his midnight pacing. Family history was negative for mental illness and neurological conditions. On physical exam he appeared well, but had trouble sitting still. He repeatedly rubbed his legs and preferred to pace the floor. He had a BP of 145 ⁄ 95 and regular pulse of 104; he was not orthostatic. HEENT exam was notable for pale conjunctivae and non-icteric sclerae; he had no LAN; cardiac exam revealed a 2 ⁄ 6 systolic ejection murmur at the LUSB; lungs were clear; he had no hepatosplenomegaly; rectal exam revealed no masses and brown stool that was positive for occult blood; extremities showed trace edema; neurologic exam was unremarkable. Initial laboratory evaluation revealed a microcytic, hypochromic anemia, with a hematocrit of 23. The remainder of his CBC, electrolytes and liver function tests were within normal limits. His ECG revealed sinus tachycardia with evidence of LVH. His iron studies later revealed a ferritin of 8. JS was admitted to the hospital and received 2 units of PRBC. He subsequently underwent a colonoscopy which demonstrated a 3-cm malignant mass in his left colon. He underwent a partial bowel resection. His restless legs improved with iron repletion. JS has returned to driving his school bus, and both he and his wife are sleeping more soundly at night. DISCUSSION: Restless legs syndrome can be diagnosed solely on the basis of a patient's history. The salient symptom is a creeping sensation primarily on the legs, prominent at rest and relieved by movement. In many cases, the syndrome of restless legs occurs as an idiopathic condition of many years' duration; however, in a small subset of patients, iron-deficiency anemia, malignancy, uremia, and pregnancy can predispose to restless legs, and treatment of the underlying condition can lead to marked improvement in symptoms. For patients with idiopathic restless legs, therapies are tailored to the severity of the syndrome. Mild cases can be treated with benzodiazepines or opioids; moderate or severe symptoms are treated with carbidopa-levodopa, bromocriptine or pergolide.

WR Harper, Medicine, University of Chicago, Chicago, IL LEARNING OBJECTIVES: 1) Learn how excessive zinc intake can lead to anemia and neutropenia. 2) Identify people at risk for this disorder. 3) Recognize the need to be thorough in assessing supplement use. CASE PRESENTATION: The use of complementary medicine continues to increase. Zinc has been promoted for over 25 conditions, including the common cold. Excessive zinc intake impairs copper metabolism leading to reversible anemia and neutropenia. A case is presented of a 27 year old male without medical problems until 4 weeks prior to presentation when he noted progressive shortness of breath with exertion, ultimately with only one flight of stairs. He denied any fevers, chills, or weight loss. He was a non-smoker with no travel or known ill contacts. He did state that he had been taking 700-1000 mg of zinc per day for his acne. Withdrawal of zinc and administration of IV followed by oral copper normalized his zinc and copper levels. His anemia and neutropenia resolved. His diagnosis was zinc-induced hypocupremia with resultant anemia and neutropenia. DISCUSSION: Zinc promotes the expression of its intestinal intracellular binding protein, metallothionine, which has higher avidity for copper. This leads to loss of copper as the intestinal cell is shed. As little as 2.5 mg over the RDA of 15 mg may lead to a negative copper balance. Hypocupremia is also seen in patients on TPN without trace elements, patients after bowel resection, and premature infants on formula. Copper is a key component of enzymes, including those necessary for iron release from stores and those of heme synthesis. There is a sideroblastic picture on bone marrow examination often with ring forms. The neutropenia arises from delayed maturation of WBC's caused by low levels of other copper enzymes. Traditional causes of drug-induced sideroblastic anemia have been isoniazid, chloramphenicol and alcohol. Zinc should be considered as well. The use of complementary medicine continues to rise. In one study, the use of 16 alternative therapies by a random population of individuals increased from 33% in 1990 to 42% in 1997. Accordingly, we must be extraordinarily vigilant in our history taking. We must also be increasingly knowledgeable about the potential complications of these supplements.

Internal Medicine, University of Colorado Health Sciences Center, Denver, CO During the first the day of hospitalization, spiking fevers to 39 Celsius were noted. The fevers occured 3 to 4 times daily and were associated with profuse diaphoresis. A CT scan of the pelvis and abdomen revealed splenomegaly but no pelvic abnormalities. A whole body bone scan revealed clustered increased activity in the knees, ankles, shoulders, elbows and sacroiliac joints. The possibility of a bone marrow packing disorder or a systemic inflammatory arthropathy was entertained. Hematology/Oncology was consulted and a bone marrow biopsy was performed which showed a normocellular marrow with noncaseating granulomatous inflammation. The patient continued to have cyclic fevers and intense sacroiliac pain and tenderness. Further history taking revealed the patient's true travel history and a dietary history of goat milk injestion. Blood cultures drawn on the fourth day of hospitalization grew a gram-negative coccobacillus organism which was later identified as Brucella melitensis. Patient's pain and fever responded quickly to intravenous gentamycin and oral doxycycline. DISCUSSION: Brucellosis is a zoonosis which is transmitted to humans by contact with infected animals or animal products. The most common human form, Brucella melitensis, infects goats, sheep and camels. Unpasteurized milk product consumption, travel to high incidence areas and occupational exposure are important historical facts which may raise suspicion for brucellosis. Sacroilitis can occur either as a monoarticular septic arthritis or as part of a reactive asymmetric polyarthritis.

LEARNING OBJECTIVES: 1) Recognize the difficulties of advance care planning in the absence of complete medical information. 2) Weigh family and patient perspectives in decisionmaking for an elderly parent. CASE PRESENTATION: An 85-year-old woman with a history of anemia, weight loss and diabetes was admitted to the hospital with diffuse abdominal pain. She had been having intermittent pain for two weeks and recently developed a fever. A CT scan revealed extensive diverticulitis and abscesses. Her only laboratory abnormality was a carcinoembryonic antigen (CEA) of 26 (nl 0-5). A colonoscopy was recommended to assess her anemia and the possibility of a malignancy. As she had previously when her anemia was discussed, she declined to have the procedure. Her husband of 57 years, son, and daughter disagreed with her refusal and tried to persuade her to have the colonoscopy. But she insisted that she "did not want to know" whether she had cancer. After discharge, a hematocrit sent from home was 21.1, down from 33.9 on the previous day. Over the phone, her daughter reported that the patient felt tired and had had one "very dark" bowel movement but no other symptoms. She refused to come to the hospital. As it was not clear that she understood what was happening, she was seen at home that evening. Upon arrival, her family said: "we've decided she needs to go to the hospital and if she doesn't want to, we'd like you to call an ambulance and make her go. Is that OK?" On exam, she was drowsy but oriented. Her pulse was 92 and her abdomen was soft. She did not believe she had anemia. Although she understood that if she were bleeding, it might lead to her death, she refused to go to the hospital. At the same time, she "wanted to be around" for her daughter. Despite her protests, the decision was made to take her to the emergency room. DISCUSSION: While models of decisionmaking and advance care planning strive for complete medical information and agreement between patient and family, these are often absent. In this case, a competent patient refused workup for a possible cancer, over the objections of her family. Subsequently, in an emergency when she was not competent, she was sent to the hospital over her objections. Could this have been foreseen? Should this patient, her family and her physician have previously come to a consensus about this type of situation? This case thus raises issues of the specificity of advance care planning, the situations when a patient's immediate wishes can be overridden and the balance of patient and family wishes in the face of urgent medical decisions. CASE: A 34 yo female was referred to our institution following a diagnosis of nephrolithiasis. Over the past 20 years, she reported 2-3 episodes of nephrolithiasis per year and was diagnosed with calcium oxalate stones. One day prior to admission, she developed severe left flank tenderness that radiated into the groin and was accompanied by nausea and chills. Prior to her transfer an IVP demonstrated a high-grade obstruction of the distal left ureter by a 5 mm stone. Upon admission, analgesia and hydration were provided; urine was strained for debris, 24 hour urine for citrate, oxalate, calcium, uric acid, and cystine was normal. She passed a stone that was analyzed and found to be composed of ephedrine. Initially she had reported only occasional use of albuterol for asthma, but upon further questioning, she reported long-term use of ephedrine and pseudoephedrine for nasal congestion up to 4 times daily. In the past year, she had also been using herbal supplements for weight loss. We advised her to discontinue all of these agents except albuterol. She has not had further episodes of nephrolithiasis after 6 months of follow-up. DISCUSSION: Ephedrine nephrolithiasis has previously been reported in patients consuming ephedrine or ephedrine-containing herbal preparations (1, 2) . Our case suggests that chronic use of ephedrine is sufficient to cause recurrent nephrolithiasis. Herbal supplements used for dietary purposes commonly contain ephedrine and we suspect our patient was likely receiving additional ephedrine through the supplement. Herbal supplements are not subject to the same review and approval process as are food products and medicines (3) . When a supplement is on the market, lack of safety needs to be demonstrated before it can be restricted. In 1997, the FDA did call into question the safety of supplements containing ephedrine (4) . However, it remains a readily available preparation and is often hidden in dietary supplements. Consumers do not usually recognize this fact. This case report add to previous findings of adverse effects related to ephedrine use. LEARNING OBJECTIVES: To recognize that other, more serious conditions can occasionally mimic routine problems such as a "simple" sinus infection. CASE: A 76-year-old female presented to an outpatient clinic with a six-month history of sinus problems. She complained of stuffiness, a full sensation in the right maxillary and frontal sinus and some post-nasal drip. She denied any significant allergy history although she frequently had sinus problems during the fall and winter. She had treated herself with over the counter medications but nothing helped. Physical exam was unremarkable except for slight tenderness over the sinuses as well as post-nasal drip. Antibiotics and decongestants were prescribed. She returned to the clinic a week later and stated she felt no better. Her medications were changed and sinus films were ordered. The films showed opacification of the sinuses. She was lost to follow-up but reappeared a month later with persistent sinus discomfort, a very slight asymmetry of her face with slightly greater fullness of the right cheek, and the story that she had coughed up a piece of "cooked turkey meat." She had the sample in a plastic bag she had brought to the office. It contained what appeared to be a dry, bloodless piece of tissue. The specimen was placed in formalin, an emergency CT scan of the sinuses and a stat referral to ENT were ordered. Studies revealed an esthesioneuroblastoma-a highly malignant tumor that metastasizes early and widely. She went to surgery but despite chemotherapy and radiation therapy, died six months later. DISCUSSION: This case illustrates that even routine clinical presentations such as sinusitis may occasionally mask or be mimicked by more serious and malignant conditions. Primary malignancies of the paranasal sinuses are rare conditions that few physicians will ever see in a primary care outpatient setting. Nonetheless, when considering the well known aphorism "When you hear hoof beats think horses and not zebras," the clinician while recognizing its wisdom, must never forget that the "zebras" are out there and that they can present at any time and sometimes in the most common of ways. CASE: A 39 year-old woman presented with 5 days of increasingly severe fatigue, weakness, and lightheadedness. Two weeks before, she had several days of nausea, vomiting, and diarrhea, was diagnosed with gastroenteritis and treated with IV fluids. Her symptoms resolved except for fatigue, which later intensified. At this presentation, she denied fever, chills, anorexia, nausea, vomiting, and abdominal pain. Over the past 6 months, she reports a 40 lb weight loss. Three months ago, the patient moved to Wisconsin from Hawaii and noticed that her tan had not faded despite minimal sun exposure. She has had salt craving for at least two years, even salting ice cubes. On examination, the patient appeared acutely ill. Vitals revealed SBP of 60 mmHg, a pulse of 98/min, and temp of 96.3F. Sclera were anicteric. Skin was bronzed and tan lines were absent. Axillary and pubic hair were sparse. Oral mucosa was dry and gingiva were darkly pigmented. Heart, lungs, abdomen and neurologic exams were normal. Laboratories revealed Na 124, K 7.0, glucose 121, BUN 62 and creatinine 2.4. CBC and LFTs were normal. EKG demonstrated peaked T waves. Cosyntropin stimulation test could not be performed. Aggressive fluid resuscitation and IV hydrocortisone were given. 3 days after admission, with only minimal fatigue and normalized labs, the patient was discharged on a hydrocortisone taper and fludrocortisone. She was instructed to double her dose of hydrocortisone during illness and has done well. DISCUSSION: Primary adrenocortical insufficiency (Addison's disease) is a potentially fatal condition resulting from deficiency of cortisol and aldosterone. Unlike secondary adrenocortical insufficiency where aldosterone production is preserved, patients with primary adrenocortical insufficiency can demonstrate marked volume depletion and hyperkalemia. Excess production of pro-opiomelanocortin in primary adrenocortical insufficiency results in the characteristic hyperpigmentation. Unfortunately, early diagnosis of adrenocortical insufficiency is often missed due to the nonspecific nature of symptoms. This can lead to patients presenting in adrenal crisis with severe hypotension and large electrolyte imbalances. High dose hydrocortisone therapy is commonly given, however dexamethasone may be used initially so diagnostic testing will not be distorted. Lifetime glucocorticoid therapy is required. Mineralcorticoid replacement is also likely to be necessary. An 18 year old white female with no significant past medical history presented with a subacute onset of right-sided pleuritic chest discomfort, with radiation to the right shoulder and breast, and slight shortness of breath. There was no associated nausea, vomiting, headache, visual changes, diaphoresis, vertigo, lightheadedness, or abdominal pain. Social history was remarkable for oral contraceptive and tobacco usage. Family history was significant for accounts of aunts having blood clots and strokes and a cousin with lupus. Physical exam revealed a pleasant young, white female in NAD. O2 saturation was 94% on room air. The lung exam was remarkable for decreased breath sounds with dullness to percussion and decreased fremitus at the right lower base. Cardiac and abdominal exams were unremarkable. No rashes were noted. Lab work revealed a platelet count of 161, 000, an INR of 0.9, and an increased PTT of 41. The ESR was 110. The RPR was positive with a negative FTA. The CXR revealed a moderate pleural effusion of the right lower lung and slight atelectasis in the left lower lung. Lower-extremity Dopplers and an ultrasound of her gallbladder were negative. A spiral CT revealed evidence of a PE in the left lower lobe. Subsequent lab work was positive for the lupus anticoagulant and anticardiolipin IgG & IgM. DISCUSSION: APS is characterized by venous and arterial thromboses, pregnancy loss, and thrombocytopenia. The antiphospholipid antibodies constitute a heterogeneous group, including the lupus anticoagulant and the anticardiolipin antibodies, and are directed against the phospholipids involved in the coagulation process. Antiphospholipid antibodies are not specific for APS. For example, the prevalence of anticardiolipin antibodies in the general population ranges from 0-14%. The mechanism of the thrombosis is considered to be multi-factorial, including involvement of the B2-glycoprotein I pathway, enhanced platelet aggregation, inhibition of the protein C pathway, and altered endothelial cell function. Though arterial thrombi are less prevalent than their venous counterparts, stroke secondary to cerebral infraction may be a prominent manifestation. Pregnancy loss, especially in the second or third-trimester, is also of concern. The diagnosis is initially a clinical one, with further confirmation by laboratory studies, such as the modified Russell viper venom time and the sensitive PTT. Treatment after a thrombotic (venous or arterial) event includes long-term warfarin therapy. In addition, various factors such as oral contraceptives, obesity, hyperlipidemia, hypertension, and smoking can have an independent, additive effect. Thus management must also include attempts at reducing these risk factors. A LP revealed no WBC and had normal glucose and protein levels. The presentation seemed consistent with NMS. The patient was treated with aggressive supportive measures and Dantrolene. His mental status slowly improved over the course of a week and he was discharged to home. DISCUSSION: NMS is characterized by fever, muscle rigidity, altered consciousness, and autonomic changes. However, patients presenting early in their course may not always have all the features. Lab work can reveal a leukocytosis and increased CK levels, the latter secondary to rhabdomyolysis resulting from muscle rigidity and breakdown. The syndrome occurs after the initiation of a neuroleptic drug or the addition of a second agent. The concomitant administration of lithium, tricyclic antidepressants, or benzodiazepines are particularly likely to evoke symptoms. Less commonly, the sudden withdrawal of levodopa, carbidopa, or amantadine has produced similar reactions. The mechanism involves the blockade of dopaminergic pathways in the basal ganglia and hypothalamus. The symptoms develop over 24-72 hours and may last as long as 5-10 days even after the medications are discontinued, given their long half-life. The overall mortality rate ranges from 10-20%, with the majority of fatalities resulting from cardiovascular collapse, renal failure, and respiratory inadequacy. Dantrolene is a muscle relaxant that reverses the muscle rigidity and has demonstrated some efficacy in reducing the duration of symptoms and associated mortality. Nevertheless, intensive supportive measures including drug withdrawal, vigorous cooling, and aggressive fluid replacement remain the cornerstones of therapy. The breast masses had progressively increased in size over a period of two months and were associated with discomfort upon raising his arms over his head. He reported night sweats without fever or weight loss. The patient underwent evaluation for presumed gynecomastia including hormonal studies which were all normal. At this time he developed mildly painful swelling of his right ankle and the left metacarpal head of left middle finger. As part of the evaluation of the presumed gynecomastia he underwent bilateral breast biopsies. The initial Gram stain, acid-fast stain, and fungal stains were negative. The histopathology revealed necrotic bone with caseating granulomas. He was then referred to the medical service and was started on antituberculosis therapy with isoniazid, rifampin, pyrazinamide, and ethambutol. Four weeks later Mycobacterium tuberculosis susceptible to all antituberculosis agents was isolated from the mycobacterial cultures of the breast tissue. He denied any exposure to tuberculosis, however a PPD test revealed an induration of 12 mm. His chest radiograph was normal and his HIV antibody test was negative. Radiographic studies of his swollen ankle and hand were consistent with tuberculous osteomyelitis. He improved on antituberculosis therapy with resolution of his night sweats and improvement of his osteomyelitis. He remains on isoniazid and rifampin and is to complete at least 6 months of therapy. DISCUSSION: This patient is an Indian, where the prevalence of TB is high. The unusual feature in this case was the gynecomastia secondary to involvement of the ribs. No other organ was involved except small joints of hand and ankle. This accounted for the delay in diagnosis even though the patient came from an area where tuberculosis is endemic. This observation demonstrates the difficult diagnostic situation presented by tuberculosis of the ribs. The key to diagnosis is to keep in mind that TB can present in multiple unusual sites, especially in immigrants from countries where TB is endemic. A positive PPD is helpful in the diagnosis. Both in their late 70s, they stated that for years they have gone to the mall for their eyeglasses and never had a problem but were urged by a relative to see a "real eye doctor". They drove together to the recommended ophthalmologist's office early one fall afternoon, having received no prior information about what procedures would be done. Immediately after check-in, they were both given mydriatic drops without either an explanation of the possible effects of the drops or inquiry into their plans for driving home. After an exam by the ophthalmologist, they mentioned to staff that they were anxious to get home because of a storm forecast and the darkening sky. They were not told to wait until the drops wore off and were not offered drops to reverse dilation. They left the clinic one hour after getting the mydriatic but discovered in the parking lot that neither of them could see well. They drove the 5 miles home at a crawl in cloudy, misting weather. Upon reaching home almost an hour later, they called their son, extremely upset at what had happened, adding that they had almost hit a parked car on the way. They chose not to tell the physician and resolved to go back to the mall in the future. DISCUSSION: This case illustrates a common, dangerous, and underrecognized problem in the care of ambulatory patients. We heard similar anecdotes from other clinic patients. Prior studies have found distance visual acuity is not usually affected by mydriasis, but patients report difficulty with glare, reading road signs, judging distance, and seeing traffic lights on the way home. Extra caution may be needed in diabetics, six percent of whom fail to meet the minimum driving visual acuity requirement after mydriasis despite meeting it beforehand. Even uniocular dilation causes similar problems. Patient management strategies requiring further study include: the use of reversal agents, patient education about mydriasis, a reminder before the clinic visit to bring a driver and sunglasses, offering disposable sunglasses to patients without them, and scheduling patients who lack a driver early in the day to allow time for the mydriatic to wear off. We were unable to find documented guidelines that could assist physicians who use mydriatic drops in the ambulatory setting, and we urge the development of specific recommendations on this issue for improvement of quality of care and patient safety. (2) Review the evidence of different treatment options using current literature. CASE PRESENTATION: A 70-year-old male with a history of severe chronic obstructive pulmonary disease (COPD), rheumatoid arthritis, and chronic bronchitis is found to have an elevated serum prostate-specific antigen (PSA) level of 13.5 ng/mL. Digital rectal examination reveals no palpable nodules. The patient is referred to a urologist and a biopsy is positive for adenocarcinoma with a Gleason score of 7. A bone scan and chest x-ray detect no metastasis. The urologist recommends either a prostatectomy or radiation therapy, but offers watchful waiting as an option. The patient and his wife return to their Primary Care Physician (PCP) with anxiety and fear over the new diagnosis. The PCP addresses each of the patients' concerns, reviews the risks and benefits for each treatment option, provides educational references, and coordinates the care plan by outlining steps that need to be taken. The patients' history of COPD, age, Gleason score, PSA level, and personal preferences are factored into the decision making. This patient eventually chooses radiation therapy as the treatment choice. DISCUSSION: The clinical evidence for the best treatment option for localized prostate cancer is not clear. Current choices fall into three major categories including a radical prostatectomy, radiation therapy, and watchful waiting. Generally, patients with a PSA level of 10 or less ng/mL, a Gleason score of 7 or lower, and a tumor size of 1.5 mL or less have been shown to have a higher curability rate. Watchful waiting, has been shown to have a 66% survivability over 15 years in patients with confined disease. This option is mostly considered if the patient has various medical problems or is concerned about quality of life after treatment. Radical prostatectomy is also a favored option for younger patients with 47-70% of patients being free from recurrence in 10 years. Risks include death, impotence, urinary incontinence, and urethral stricture. Radiation therapy has also been shown to have a 10 year recurrence free rate of 64-86% in various studies. This option also carries risks such as death, impotence, anorectal complications, and acute GI or GU symptoms. This case illustrates how the PCP, urologist, and the patient can work together towards an informed decision for treating prostate cancer. The PCP's role is to help the patient in the decision process, aid the urologist on patient's health factors, support the patient throughout the treatment, and follow up with the patient after treatment. . Chest x-ray was normal. ECG showed prolonged PR and QT intervals, U waves, diffusely flattened T waves, and ST segment depression in V2-V6, all of which were attributed to the profound hypokalemia. The patient was given IV fluid and potassium replacement. His electrolyte and ECG abnormalities rapidly corrected, and he was discharged on oral potassium supplements. After discharge, a sweat test was obtained and was abnormal. A presumptive diagnosis of cystic fibrosis (CF) was made and was confirmed by genotype analysis (mutation delta f508/R334W). Sputum culture grew pseudomonas aeruginosa and stenotrophomonas maltophilia, both organisms which typically colonize the airways of patients with CF. DISCUSSION: This patient presented with dehydration, a hypokalemic metabolic alkalosis, hyponatremia, hypochloremia, high urinary potassium, and low urinary chloride. The differentail diagnosis of hypokalemic metabolic alkalosis in a normotensive patient includes diuretic abuse, prolonged gastrintestinal fluid losses, Bartter's and Bartter's-like syndromes, and excessive chloride loss in sweat as occurs in CF. While the patient had previously been diagnosed with Bartter's syndrome, the low urinary chloride is inconsistent with this diagnosis. Cystic fibrosis usually presents in the pediatric age group with recurrent pulmonary infections, pancreatic insufficiency, and failure to thrive. The availability of sweat testing and genotype analysis allows clinicians to make the diagnosis in adults, and it should be considered when patients present with unexplained electrolyte abnormalities. year-old homeless man with suspected dementia aroused clinical concern that he could die in the streets. Over several months health care providers had noted that the patient conversed fluently while often losing the thread of conversation. He was unable to recall information from previous visits. He often smelled of urine and/or alcohol. Once while sober the patient offered what little money he had to a clinic nurse without expectation of return. A staff member had observed the patient lying in the street being beaten without resisting his assailant. He was not thought to have a coexistent psychotic or major emotional disorder. Past history was notable for chronic hepatitis C without cirrhosis, cerebral injury due to assault, occasional seizures attributed to that injury, and alcoholism. A prior CT scan had documented generalized brain atrophy as well as severe left frontotemporal encephalomalacia. The patient was subjected to involuntary hospitalization in a locked medical-behavioral unit for short-term evaluation. Medical workup was negative and CT scan confirmed previous findings. Neuropsychological evaluation found the patient to be a danger to himself on the basis of dementia and recommended placement in a supervised nursing facility. Legal review showed that the patient could be detained if he was suicidal, homicidal or at imminent risk for harm to himself. No locked facility, however, had authority to hold this patient. Regulations precluded the Department of Mental Health from holding a patient suffering from an "organic" mental condition. In the absence of tuberculosis or a life-threatening medical condition, no other facility had authority to hold this patient. The patient was released to the street. The patient returned to clinic and threatened his doctor. He was rehospitalized involuntarily and a court found the patient incompetent to make health care decisions. A guardian was appointed. Because no facility had authority to hold the patient involuntarily, he was transferred to an unlocked floor, and he eloped. He continues to reside in the streets. DISCUSSION: Clinicians should consider formal neuropsychological evaluation and legal consultation in the care of demented homeless persons. Most states permit involuntary hospitalization of persons shown to be a danger to themselves or others on the basis of a major mental illness. Some states do not provide facilities for persons at risk on the basis of dementia alone. Demented homeless persons may be at particular risk for death in the streets because of unique environmental exposures and the absence of facilities with legal authority to hold these patients. year-old man, with no significant past medical history, presented with back pain of acute onset while lifting objects at work. He denied history of direct trauma, or drug abuse, but had 50 pack years of smoking. Due to a normal neurological exam, his pain was treated symptomatically with NSAIDS. The following week his persistent pain was evaluated with an MRI which showed T12-L1 disc herniation. Since the neurological exam was unchanged, rest, corticosteroids and narcotic analgesics were recommended. Two weeks later, the family reported "recurrent disorientation and confusion", so the dosage of the narcotic analgesic was decreased. The following day, there was decreased mobility and strength in the right upper extremity. Admission exam also revealed a tender right mid-clavicular mass. A CT of the brain showed a left parietal lesion, consistent with acute ischemia, necessitating IV Heparin. Lab abnormalities on admission were platelets ϭ 41000 K/ul, creatinine 1.8 mg/dL (0.7-1.4 mg/dL), and normal WBC ϭ 9.6 K/ ul. A peripheral smear and bone marrow evaluation ruled out thrombotic thrombocytopenic purpura. Bone marrow and blood cultures grew methicillin susceptible Staphylococcus aureus (MSSA). An MRI of the spine showed discitis at T12-L1 levels. Given the sequence of events a transesophageal echocardiogram was performed, which revealed vegetations on the tricuspid and mitral valves. Subsequently, a CT revealed a right sternoclavicular abscess, splenic and right renal infarcts and bilateral pulmonary cavitations consistent with septic embolic episodes. MSSA bacteremia cleared with IV oxacillin. 10 weeks later, he remains with right hemiparesis, expressive aphasia, and respirator dependent. DISCUSSION: Occupation related back injuries are common. When warranted by a thorough physical examination, appropriate radiographic studies can support a diagnosis of musculoskeletal back injury such as disc herniation. In the early stages, this is frequently difficult to distinguish clinically from infectious discitis. When symptoms fail to respond to conservative management with NSAID's and rest, or when fever, leukocytosis, or elevated acute phase reactants are detected, further radiographic imaging is indicated to exclude infection or neoplasm. Endovascular infection with S. aureus is frequently complicated by seeding or embolization of infection to distant sites including solid organs and bones. As this case illustrates, patients may be afebrile and without leukocytosis. Also, use of narcotic analgesics may confound the significance of mental status changes. Patients with evidence of emboli sould be thoroughly investigated for endocarditis. In the most severe forms of this condition, the integrity of vasculature to the involved extremity can be jeopardized. Compartment syndrome rarely has been described in association with venous thrombosis and can be limb threatening.

Rhabdomyolysis is a well-described complication of compartment syndrome. Rhabdomyolysis can be associated with acute renal failure in less than 8% of patients. There have been no reported cases of deep venous thrombosis causing acute renal failure. We report a patient with extensive lower extremity venous thrombosis who developed acute renal failure and had evidence of rhabdomyolysis. Physicians should watch for evidence of rhabdomyolysis in patients with massive venous thrombosis. CASE: A 51-year-old white male was admitted to the hospital with a subarachnoid hemorrhage. He had been previously in good health and was taking no regular medications. He had only a vague history of hypertension in the past. Surgical clipping of a bleeding anterior communicating artery aneurysm was done on the second hospital day. In the post-operative period, he remained unresponsive, was non-verbal, and did not follow commands. One month after the onset of illness he developed deep venous thrombosis of the left lower extremity. He was not anticoagulated because of the recent subarachnoid hemorrhage and a Greenfield filter was placed. The left leg swelling resolved gradually. Four weeks after the left leg deep venous thrombosis, an extensive venous thrombosis developed in the right lower extremity. The leg became extremely tense and swollen with bluish red discoloration. Color flow ultrasonography confirmed lack of spontaneous phasic flow and incompressibility of virtually all the deep veins of the right thigh and the right calf. The common iliac vein, as well as the inferior vena cava up to the level of the Greenfield filter, also showed evidence of thrombosis. The CPK rose to 10,439 U/ L. The urine myoglobin rose to 18,740 NG/ML. The urine output dropped to 16cc/ hour for 24 hours. The creatinine rose from 0.6 MG/DL to 3.4 MG/DL. The BUN rose from 11 MG/DL to 45 MG/DL. His treatment was with standard IV unfractionated heparin, leg elevation, and forced alkaline diuresis to keep the urine pH Ͼ7.5. In a week, his CPK normalized, the urine became free of myoglobin and the creatinine returned to normal. The patient returned to normal functioning with minimal leg swelling and normal renal function. year-old female presented to an outpatient clinic with a two week history of anterior neck pain and occasional dysphagia to solids and liquids. This began shortly after a bout of an upper respiratory illness. She reported poor appetite and weight loss of 10 lb in 7 weeks. She denied trauma, upper extremity weakness, fevers, chills, palpitations, irritability and tremors, but admitted to a history of chronic constipation. On physical exam she was tachycardic with tenderness in the neck muscles bilaterally, and a palpable symmetrical and tender goiter. There was no evidence of lymphadenopathy, proptosis or exophthalmus. The cardiovascular and pulmonary exams were normal. The DTR's were normal bilaterally. She was prescribed naproxen (500 mg/day) and heat compresses for symptomatic pain relief. Given her presentation, a screening TSH was obtained. The TSH ϭ 0.02 (normal 0.4-5.5 uU/dL); subsequent thyroid indi-ces revealed a Free Thyroid Index (FTI) ϭ 19. 3 (6-11 ug/dL), T4 ϭ 13. 9 (5-11 ug/ dL), free T4 ϭ 4.2 (0.7-1.8 ng/dL). A sedimentation rate (ESR) ϭ 108. The patient's TSH-receptor antibody levels were within normal limits. One week later in follow-up, her neck pain was decreased and there was 'improvement' in her constipation. Her findings were consistent with hyperthyroidism from subacute thyroiditis. At the three week follow-up the patient denied neck pain, but complained of increasing fatigue, and had gained 11 lb. Neck palpation revealed a smaller nontender goiter. A repeat TSH was 5.83, FTI 3.2, and T4 3.3 . This suggests a hypothyroid phase, which may need treatment with thyroxine if symptoms persist. Close follow-up is therefore needed. DISCUSSION: Subacute thyroiditis is a common cause for anterior neck pain. It is a post viral inflammatory syndrome that causes release of thyroxine from follicular stores. Approximately 50% of patients present with signs and symptoms of hyperthyroidism. Symptoms may be subtle or overt with palpitations, tachycardia, diaphoresis or weight loss. Recovery may be followed by transient hypothyroidism which may require treatment. A high suspicion based on the patient's history and exam can enable an astute primary care provider to make this diagnosis on the initial visit which can later be confirmed using thyroid function tests and a sedimentation rate. An elevated ESR is so common that a normal value should question this diagnosis. Treatment involves using salicylates or NSAID's for pain; steroids are reserved for patients with severe pain. Beta-blockers can be added for patients with symptomatic tachycardia and at risk for arrhythmias. year old African American gentleman with a history of chronic obstructive lung disease and alcohol abuse who presented with a painful cyst on the right buttock and was found to have perianal fistula. He was started on oral antibiotics and asked to return in 2 days for rectal evaluation under anaesthesia or earlier if symptoms didn't improve. The patient returned a few hours later with severe left leg pain and shortness of breath. He was found to be afebrile, with pulse rate 100, blood pressure 110 ⁄ 70 , respiratory rate 20 and pulse ox 88%. His lungs were clear, heart sounds were regular and abdomen benign. Patient was unable to dorsiflex his left foot due to pain. Right buttock examination revealed mild erythema without fluctuation. Diagnostic procedures focused on causes of left leg pain including doppler ultrasound to exclude acute arterial occlusion. DVT and pulmonary embolism were ruled out by negative doppler and ventilation perfusion scan. Patient was admitted and started on intravenous antibiotics and surgery consult was ordered. A few hours later the patient was found to be unresponsive and acidotic. The patient was immediately taken for debridement and was found to have extensive Fournier's gangrene extending into both thighs, full length of sacrum and base of scrotum. During the debridement the patient went into pulseless electrical activity, shock and disseminated intravascular coagulation and died a few hours later. DISCUSSION: Fournier's gangrene is necrotizing fasciitis with extension into the perineal area and is caused by streptococcus pyogenes. It can present as unexplained fever or pain and early diagnosis is difficult. It can extend rapidly as in our patient where it extended from a small induration to extensive gangrene within a few hours leading to toxic shock syndrome.

General Internal Medicine, Denver Health, Denver, CO LEARNING OBJECTIVES: 1) Recognize that umbilical hernia perforation is a serious, potentially fatal complication of ascites. 2) Treat ascites aggressively in the presence of umbilical hernia, especially when accompanied by overlying skin changes. CASE PRESENTATION: A 46 year-old male with a history of alcoholism, hepatitis C, and hypertension presented to the clinic complaining of abdominal distention for two years. He was evaluated one year prior, and was found to have liver failure and ascites. He had been started on spironolactone and furosemide, which he had not taken for several months. Examination revealed a thin male with a distended, soft abdomen and hepatomegaly. An umbilical hernia nine centimeters in diameter was present, with brown discoloration and focal excoriation. There was no lower extremity edema. He was again started on diuretic therapy and told to return for a therapeutic paracentesis. The patient returned to clinic one month later with a flat abdomen and no lower extremity edema. He reported that two weeks prior, during a coughing paroxysm, his umbilical hernia had ruptured with the loss of his entire ascites volume within minutes, while leaning over the bathtub. His hernia had closed spontaneously, he had no recurrence of his ascites, and had no current complaints. Blood pressure was 100 ⁄ 82 and pulse was 80. He now had a flat, soft abdomen, with a seven centimeter flaccid umbilical hernia with a small healing eschar. There was no lower extremity edema, and he weighed 58 kg, a 12 kg loss from his previous visit. Electrolytes were normal and creatinine was 0.9, which was unchanged. Diuretic therapy was restarted, surgical consultation was obtained, and he underwent herniorrhaphy and peritoneovenous shunt placement six weeks later. He remained free of ascites three months later. DISCUSSION: Umbilical hernia is common in patients with ascites, occurring in more than 20% of cases. However, spontaneous rupture of an umbilical hernia is a rare and often fatal event. Since it was first described in 1901, approximately sixty cases have been reported in the English literature. Patients receiving only supportive care have had mortality rates as high as 60-90%, and even those patients undergoing urgent surgical repair experience a mortality of nearly 20%. The cause of death is usually peritonitis/sepsis, hepatic failure, renal failure, or gastrointestinal hemorrhage. In many cases, hernia rupture is heralded by skin changes in the hernia, most notably ulceration. Although this case had a favorable outcome, patients with ascites and umbilical hernia need aggressive reduction of fluid volume, usually followed by elective herniorrhaphy, to prevent the often mortal complications of spontaneous rupture. Upon identification of the CSF lymphocytosis, empiric therapy with intravenous acyclovir was begun. The team ordered a VZV PCR on the CSF, but the lab ran a herpes simplex virus PCR (which was negative). The patient's dysphagia and hoarseness started to improve, and after 2 days he was switched to oral acyclovir to complete a 10 day course. One month later he still noted mild hoarseness and dysphagia. DISCUSSION: The laryngoscopic findings of vocal cord paralysis explained his hoarseness, but the shallow lesion on the epiglottis did not seem sufficiently inflammatory to explain his severe dysphagia and odonophagia. Once extrinsic compression of the esophagus was ruled out, a polyneuritis was suspected. The diagnosis of a VZV infection causing a polyneuritis cranialis was supported by the CSF lymphocytic pleocytosis associated with slightly elevated protein and normal glucose. This patient manifested a painful reactivation of VZV in the absence of cutaneous lesions, known as zoster sine herpete. The CSF PCR can be utilized to rapidly diagnose VZV infection. IGM and IGG serology can be performed in the serum and CSF. The reactivation of VZV can affect cranial nerves 9 and 10 in addition to the more common syndromes involving cranial nerves 5 and 7. Reinforce the importance of early diagnosis and patient education. CASE: A 69-year-old woman without significant past medical history presented with 2 days of low-grade fever, malaise, myalgia, and posterior knee pain. The patient lives in an urban suburb of Boston and is an avid gardener. She denied rash, trauma or recent travel. On physical exam she was afebrile and looked well. The posterior aspect of her left knee was diffusely erythemetous and warm without scale or central clearing. No effusion, tenderness, or joint instability was found. Doppler ultrasound was significant for a 1.5 cm by 1.3 cm cystic structure in the left popliteal fossa without evidence of thrombus. She was treated with two successive courses of cephalexin 500 mg QID for 10 days, heat, and elevation for presumed cellulitis; each time she transiently improved, but worsened with cessation of antibiotics. One week after completing her second course of antibiotics she presented with a red, painful, and swollen left knee. Arthrocentesis revealed serosanguinous fluid with 475 WBC, 8225 RBC, no crystals or organisms seen on Gram stain, and subsequent cultures were negative. Lyme serology revealed positive IgG and IgM confirmed by Western blot analysis. She was treated with one month of Doxycyline 100 mg bid and clinically improved. Her Lyme studies remain positive. DISCUSSION: Lyme disease is the most common tick borne illness in the United States with over 10,000 new cases reported annually. The probability of contracting the disease after a tick bite is dependent on tick stage of development, tick attachment time of greater than 48 hours, and prevalence of ticks infected with the spirochete Borrelia burgdorferi in a given area. Presentation is variable with fever, erythema migrans, malaise, and migratory arthralgias. Early recognition is critical as later sequelae including carditis with heartblock, neurologic complications, and musculoskeletal complaints can be avoided with prompt treatment. It is important to realize that only 60% of early infections present with the typical erythema migrans rash, and fewer than 30% of patients can recall the actual bite. Primary care physicians diagnose the majority of new cases. They are often faced with requests for testing and empiric treatment. Although the chance of acquiring Lyme disease from a tick exposure is less than 10%, many physicians perform expensive serologic tests, and prescribe antibiotics regardless of actual risk. The estimated yearly economic burden of this disease approaches $2.5 million dollars. Treatment, diagnosis and vaccination strategies are widely debated and can have a large impact on healthcare practice and economics. With serologic testing and vaccination readily available, we must continue to educate both physicians and the public regarding presentation, prevalence, interpretation serologic tests, and personal protection in approaching this disease. Reinforce the importance of early diagnosis and patient education. CASE: A 69-year-old woman without significant past medical history presented with 2 days of low-grade fever, malaise, myalgia, and posterior knee pain. The patient lives in an urban suburb of Boston and is an avid gardener. She denied rash, trauma or recent travel. On physical exam she was afebrile and looked well. The posterior aspect of her left knee was diffusely erythemetous and warm without scale or central clearing. No effusion, tenderness, or joint instability was found. Doppler ultrasound was significant for a 1.5 cm by 1.3 cm cystic structure in the left popliteal fossa without evidence of thrombus. She was treated with two successive courses of cephalexin 500 mg QID for 10 days, heat, and elevation for presumed cellulitis; each time she transiently improved, but worsened with cessation of antibiotics. One week after completing her second course of antibiotics she presented with a red, painful, and swollen left knee. Arthrocentesis revealed serosanguinous fluid with 475 WBC, 8225 RBC, no crystals or organisms seen on Gram stain, and subsequent cultures were negative. Lyme serology revealed positive IgG and IgM confirmed by Western blot analysis. She was treated with one month of Doxycyline 100 mg bid and clinically improved. Her Lyme studies remain positive. DISCUSSION: Lyme disease is the most common tick borne illness in the United States with over 10,000 new cases reported annually. The probability of contracting the disease after a tick bite is dependent on tick stage of development, tick attachment time of greater than 48 hours, and prevalence of ticks infected with the spirochete Borrelia burgdorferi in a given area. Presentation is variable with fever, erythema migrans, malaise, and migratory arthralgias. Early recognition is critical as later sequelae including carditis with heartblock, neurologic complications, and musculoskeletal complaints can be avoided with prompt treatment. It is important to realize that only 60% of early infections present with the typical erythema migrans rash, and fewer than 30% of patients can recall the actual bite. Primary care physicians diagnose the majority of new cases. They are often faced with requests for testing and empiric treatment. Although the chance of acquiring Lyme disease from a tick exposure is less than 10%, many physicians perform expensive serologic tests, and prescribe antibiotics regardless of actual risk. The estimated yearly economic burden of this disease approaches $2.5 million dollars. Treatment, diagnosis and vaccination strategies are widely debated and can have a large impact on healthcare practice and economics. With serologic testing and vaccination readily available, we must continue to educate both physicians and the public regarding presentation, prevalence, interpretation serologic tests, and personal protection in approaching this disease. 1 (48-333) , and an IgA 6.93 (73-455). The patient was diagnosed with common variable immunodeficiency disease (CVID) and referred for IV immunoglobulin therapy upon discharge. One month later he reported feeling better than he had ever felt in years. He has subsequently been hired by a fire department in another state. DISCUSSION: Patients often present to primary care providers with a "common cold". Consider testing for CVID in an individual with a history of multiple respiratory infections. CVID is the most common symptomatic primary immunodeficiency syndrome and can be diagnosed with low immunoglobulin levels. It has an estimated prevalence of 1/50,000-1/200,000 and affects men and women equally. Symptoms usually occur in the second to third decade of life, but may be seen at any age. Recurrent encapsulated bacterial infections involving the respiratory tract (sinusitis, bronchitis, otitis media, and pneumonia) are the most common manifestations. If untreated, CVID patients typically develop severe bronchiectasis, frequent complications, and oxygen dependency. CVID patients have a lowered life expectancy and higher rates of lymphoma (2-7%, a 30-200 fold increase) and gastric carcinomas (a 50 fold increase). Intravenous immunoglobulin therapy is a lifelong necessity. Patients with CVID should be counseled to be evaluated for any type of infection. increasing over the past eight months. Three weeks prior to admission, her abdominal pain worsened and her visual acuity diminished markedly. Her PMD found her to be severely hypertensive and started her on Nifedipine. The patient denied a history of IV drug use or significant alcohol use. She smokes less than 1 ⁄ 2 pack of cigarettes per month. Her mother is also hypertensive, but is well controlled. The patient was employed as a gardener and horse-groomer. On physical exam, blood pressure was noted to be elevated at 180 ⁄ 110 in both arms. She had blurring of the optic disks, a systolic murmur at the apex, mild epigastric tenderness and bilateral abdominal bruits, LϾR. Examination of the skin revealed several ecchymoses on the lower extremities. EKG showed left ventricular hypertrophy, but echocardiogram was normal. A renal duplex ultrasound examination was consistent with vasculitis of the abdominal aorta, with subsequent decreased flow through the superior mesenteric artery and the renal arteries, bilaterally. A CT scan of the abdomen demonstrated concentric thickening of the upper abdominal aorta from the celiac axis to the ostia of the renal arteries. Abdominal MRA confirmed these findings. She was diagnosed with bilateral retinal detachement, secondary to hypertensive disease or Takayasu's Disease. Friability of the inflamed vessel walls prevented biopsy. The patient was started on Prednisone 1mg/kg/day and discharged with plans for angioplasty. DISCUSSION: This case is a rare variant of a rare disease. Pulses were normal and the aortic arch was normal. Retinopathy and chest pain are common in this illness. This case demonstrates that when a large vessel vasculitis is postulated, Takayasu's Disease must be high in the differential diagnosis. Recognize that atrial fibrillation may present occultly in patients with pacemakers. CASE PRESENTATION: A seventy-two year old female presented with sudden onset of apparently altered mental status and weakness. She had a past medical history of coronary artery disease and sick sinus syndrome requiring implantation of a permanent pacemaker. Initial evaluation revealed an alert woman with receptive and expressive language defects consistent with Wernicke's aphasia. The examination was otherwise unremarkable, including a normal cardiac examination with regular heart rhythm and without murmurs or extra sounds. The initial electrocardiogram showed atrial and ventricular pacemaker spikes with regular QRS complexes. Echocardiogram showed atrial enlargement (diameter 5.6 cms.); no comment was made about atrial contractions. Inhibition of the pacemaker revealed underlying atrial fibrillation. The patient's aphasia improved and she was discharged on warfarin. DISCUSSION: The prevalence of atrial fibrillation and the rate of pacemaker implantation are both increasing as the population of the U.S. ages. Atrial fibrillation can trigger rapid and irregular ventricular rhythm in patients with atrial synchronous pacemakers; however, in patients with permanent pacemakers atrial fibrillation can occur in the absence of its classic cardinal feature-an irregular rhythm. Our patient presented with new onset of CVA, a regular paced rhythm and absence of comment about atrial contractions on ECHO report. Without clinical suspicion and proper interrogation of the pacemaker, atrial fibrillation would have been missed. The possibility of occult atrial fibrillation should not be overlooked in paced patients and they should be intermittently checked for onset of atrial fibrillation. If it is discovered, anticoagulation should be strongly considered. The literature suggests that even in pacemaker clinics the diagnosis of occult atrial fibrillation is often missed, and an opportunity to prevent future strokes is lost. Head CT was normal. The patient was transferred to the ICU for plasmapheresis for possible thrombotic thrombocytopenic purpura and given antibiotics for a left lower lobe infiltrate and possible endometritis. She had a prompt response with her platelets increasing to 135K and her LDH decreasing to 844 U/L after 3 treatments. Her ICU course was complicated by a right femoral hematoma secondary to line placement with a subsequent femoral neuropathy. She was discharged on a tapering dose of prednisone with resolution of the hemolysis but presented 2 weeks later with a left internal jugular thrombosis. A thrombophilia workup revealed decreased protein S levels and an elevated homocysteine level of 17 umol/L. The patient is now being treated for postpartum depression. The femoral neuropathy has resolved. Her son is doing well at home after 6 weeks in the NICU. DISCUSSION: Women with thrombophilia are at increased risk for thrombotic events during pregnancy and the postpartum period. It is less well recognized by general internists that they are also at increased risk for obstetric complications such as severe early preeclampsia, intrauterine growth retardation and fetal loss. Women should be monitored for postpartum depression especially if they have the added stressor of a complicated pregnancy and postpartum course. Laboratory studies were significant for white blood cell count ϭ 18,000 with 78% neutrophils and 4% bands, Hgb ϭ 11, platelets ϭ 81,000, albumin ϭ 2.5 and total bilirubin ϭ 2.3. Chest x-ray revealed infiltrate at the left base with a pleural effusion. A CT of the neck showed inflammation of the soft tissues on the left with an internal jugular vein thrombosis. MRI of the left shoulder revealed no fluid collection, and CT of the chest demonstrated multiple septic emboli throughout the lungs. After cultures were obtained, the patient was placed on clindamycin, ceftriaxone and heparin. Pus was aspirated from the neck under ultrasound guidance; gram stain showed many gram negative rods and gram positive cocci. The patient remained febrile to 40C and on hospital day #3 was taken to the operating room for ligation and excision of the left internal jugular vein. Culture of the neck aspirate and thrombus grew out Streptoccus milleri, although multiple blood cultures were negative. Repeat CT showed retained thrombus, however the patient's condition improved greatly after surgery. She was discharged after 11 days of hospitalization on amoxicillin clavulanate and warfarin. DISCUSSION: (1) Lemierre's syndrome is a complication of pharyntitis with septic thrombophlebitis of the internal jugular vein and septic emboli, typically to the lungs and joints. (2) The most common organism is Fusobacterium necrophorum although other organisms such as Streptococcus, Bacteroides, and Eikenella can also be involved. (3) The mainstay of treatment is antibiotics directed at anaerobic microbes. The role of anticoagulant therapy is controversial. Patients with persistent sepsis and septic emboli despite medical therapy should be considered for surgical ligation and excision of the internal jugular vein. 2) Recognize the potential complications associated with a rare cardiac anomaly often discovered incidentally. CASE PRESENTATION: A 64 yo woman with lumbar degenerative disc disease (ddd) was admitted with one week of worsening back pain exacerbated by movement. Narcotic analgesia prescribed two days earlier provided no relief. She denied recent trauma, neurological deficit, bowel or bladder incontinence, and fever or chills. Recent history was significant for ischemic colitis managed conservatively two months earlier. Remote history was remarkable for a heart murmur diagnosed in childhood. Physical exam demonstrated T ϭ 36.5C, 2 ⁄ 6 systolic ejection murmur at the right upper sternal border, no peripheral or ocular stigmata of endocarditis, tenderness at L3/4, and a nonfocal neurologic exam. Lab demonstrated normal chemistry and urinalysis, WBC ϭ 6.9, HCT ϭ 35.1, ESR ϭ 74, CRP ϭ 14.8. EKG was normal. Lumbar MRI with gadolinium showed ddd without infection or inflammation. High dose patient controlled narcotic analgesia was necessary. A temperature of 38.1C was noted 29 hours after admission. Subsequent blood cultures were positive for enterococcus. Echocardiogram revealed a small, mobile echodensity attached to the right coronary cusp of the aortic valve, and a dilated coronary sinus with high velocity flow (4m/s). Cardiac MRI demonstrated a large serpiginous fistulous tract from the origin of the right coronary artery to the coronary sinus with a high velocity jet emptying from the coronary sinus into a dilated right atrium. Ampicillin and streptomycin treatment resulted in resolution of fever and marked improvement in back pain over the following three days. DISCUSSION: Back pain presenting days to weeks prior to the development of fever is an uncommon but well reported singular presenting symptom of bacterial endocarditis. Evidence for spinal infection or inflammation is often lacking. Hypothesized etiologies include myalgia (due to vasculitis or arterial emboli), immune complex phenomenon, minimal osteomyelitis, reactive arthritis, and glomerulonephritis. Recognition of the musculoskeletal manifestations of bacterial endocarditis is important because of the potential to mimic other processes. Coronary arteriovenous fistula is an uncommon congenital anomaly often unrecognized until adulthood. Most adult patients are asymptomatic at the time of discovery, but complications including CHF, myocardial ischemia, pulmonary hypertension, and endocarditis can result in associated symptomatology. Surgical correction or percutaneous transcatheter coil embolization has been suggested in asymptomatic patients to prevent complications. A 55 yo previously healthy woman presented to the emergency department (ED) with three days of right-sided pleuritic chest pain which worsened with cough and deep inspiration. She denied any dyspnea and reported generalized malaise but no specific fever, chills, or productive cough. The patient had a distant history of nephrolithiasis and previously undiagnosed biliary colic. Pertinent social history included 50ϩ pack years of cigarette abuse. The patient was noted to be hypoxemic on pulse oxymetry and a room air arterial blood gas revealed pH 7.44, pCO 2 36 mm Hg, and pO 2 54 mm Hg. A chest x-ray revealed only a patchy infiltrate at the right costophrenic angle. Concerned about pulmonary embolism (PE), the ED physician ordered a ventilation/perfusion (V/Q) scan which was interpreted as high likelihood for PE due to a matched defect at the right base and two areas of unmatched perfusion defects on the left. The ED record revealed no pretest suspicion for PE, and the patient was started on IV heparin and admitted to the housestaff who embarked on a hypercoagulable evaluation. After 24 hours of hospitalization, the patient's pulmonary examination changed from scant right basilar crackles to percussion dullness, egophony, and diminished breath sounds in her right lower lobe (RLL). Fever to 102.4Њ and a productive cough with blood-tinged sputum developed during hospital day (HD) #2.

Worsening but transient hypoxemia occurred on HD #2. Antibiotics for community-acquired pneumonia were started approximately 36 hours after the initial presentation to the ED. Immediately prior to the patient's decompensation, a spiral CT of the chest was performed to rule out the presence of PE but its interpretation was not readily available. The housestaff caring for the patient overnight requested resumption of IV heparin, which had been discontinued by the attending physician due to low pretest suspicion of PE and changed physical examination.

Interpretation of the spiral CT on HD #3 revealed consolidation of the entire RLL with an effusion and no evidence of PE. Blood cultures on the same day isolated Streptococcus pneumoniae. Despite the delay in diagnosis and treatment, the patient had an uneventful recovery with antimicrobial therapy and a diagnostic thoracentesis revealed an uncomplicated parapneumonic effusion She was intubated in the coronary care unit for hypoxemic respiratory failure secondary to pulmonary edema and left lower lobe pneumonia, was started on intravenous ceftriaxone, and then switched to oral cefotetan after uneventful extubation. She experienced progressive clinical improvement except for persistent nonproductive cough since admission to university hospital. Sputum samples were normal, as were screens for adenovirus, influenza A/B, parainfluenza, and RSV. She was otherwise asymptomatic without fevers or leukocytosis. On transfer to floor on day 12, paroxysmal nocturnal coughing episodes were noted. On day 14, respiratory isolation was initiated, nasopharyngeal samples were obtained, and erythromycin was started. Bordatella pertussis culture was negative; PCR analysis was positive. The coughing defervesced over 5 days, at which time she was uneventfully discharged to nursing facility. DISCUSSION: Bordatella pertussis infection is a common yet underrecognized cause of persistent cough in adults. Readily identified by its characteristic highpitched inspiratory "whoop" in children, it often passes undiagnosed by internists due to the absence of typical clinical findings (i.e., paroxsymal cough, marked lymphocytosis) and underestimation of its prevalence in adults. As illustrated, prior vaccination nor infection confers lasting immunity. Adults are the major reservoir of disease; therefore, timely recognition and treatment of infection is critical in minimizing epidemic spread. Diagnosis is limited by nonspecific symptoms and poorly sensitive lab tests. Cultures are universally negative in adults, and diagnostic PCR/DFA technology are often limited to research institutions. Nonetheless, pertussis should be suspected in adults with a persistent cough (over two weeks), particularly if associated with paroxysms, whoop, or posttussive emesis. Early administration of oral erythromycin may decrease infectivity and shorten illness duration. Improved acellular vaccines may soon lead to widespread immunization boosters in adults. A 31 year-old male with an unremarkable medical history presented with right eye redness, tearing and discharge for 3 days. He reported mild discomfort but no significant eye pain or photophobia. He denied any history of eye trauma, foreign object exposure, or ill contacts, and had not experienced conjunctivitis previously. System review revealed complaints of dysuria and a purulent urethral discharge that had started several days prior to the eye complaints. Social history was remarkable for recent episodes of unprotected sex with multiple partners. Physical exam revealed marked, diffuse erythema of the right conjunctiva with edema of surrounding orbital soft tissues. A thick and copius mucopurolent discharge was present with matting of the eyelashes. The cornea was normal in appearance. Palpation of the orbit revealed mild tenderness. Vision acuity was 20 ⁄ 20 by near-field Snellen. Lymph node examination revealed mild right preauricular lymphadenopathy. Examination of the genitalia revealed a small amount of purulent discharge at the urethral meatus. Gram stain and culture of this discharge revealed evidence of gonococcal infection. The patient was treated with intramuscular ceftriaxone and oral azithromycin, as well as opthalmic erythromycin. He was counseled to advise his partners to seek medical attention. DISCUSSION: Conjunctivitis is a common complaint encountered in the ambulatory setting. The differential diagnosis of conjunctivitis includes hyperacute, acute, and chronic bacterial conjunctivitis, as well as inclusion (chlamydial), viral, chemical, and allergic conjunctivitis. Most cases of conjunctivitis are self-limited and resolve without treatment. However, significant eye pain, copius mucopurulent discharge, photophobia, decreased visual acuity, corneal opacity or haziness, and injection of ciliary vessels are all causes for concern. This patient presented with a history and physical findings typical for hyperacute bacterial conjuctivitis. Neisseria Gonorrhoeae is the most common cause. Infection usually occurs through direct autoinoculation from genitalia and may spread quickly to the other eye. Onset occurs within two to three days of innoculation. Hyperacute gonococcal conjunctivitis requires urgent treatment, since untreated cases often result in corneal ulceration, perforation and endophthalmitis. Diagnosis may be confirmed through gram-stain of discharge or conjunctival scraping. However, establishing a culture diagnosis should not delay treatment, which should include systemic as well as local antibiotics. Evaluation should also include identification and treatment of sexually transmitted diseases in the patient and partner(s). year-old, home-bound female was referred for increasing shortness of breath, dyspnea on exertion, and chest tightness of about one week's duration. The patient also experienced coughing but denied having fever or sore throat, and no orthopnea or leg swelling. Her symptoms were precipitated by exposure to dust from window-cleaning. For many years, the patient had mild and intermittent asthma, managed with as-needed Albuterol metered-dose inhaler. She did not smoke nor drank alcohol; she did have an underlying psychiatric disorder. The patient refused to go to the doctor's office or the emergency room for care. An emergent housecall was done with a visiting nurse. In addition to basic physician's instruments, a nebulizer machine, a peak flow meter, Albuterol respiratory solution and metered-dose inhaler were also brought. The patient lived alone on the 12th floor (accessible by elevator), in a one-bedroom apartment within a naturally-occuring retirement community. On examination, the patient was alert, coherent, oriented, and in no distress. She was mildly tachypneic (‫/81ف‬min), with normal heart rate and blood pressure; her peak flow was about 300 ml/min. She tried to use her Albuterol inhaler which had expired. She was fairly nourished but poorly groomed and was able to ambulate slowly on her own. She did not use any accessory muscles of respiration. On auscultation, there was fair air entry with occasional expiratory wheezing. The rest of the examination was unremarkable. Inspection of the apartment revealed a wall-to-wall carpeting of dust and litter. The hallway and living room were poorly lit and difficult to walk through because of the stacks of magazines, newspapers, and various materials which the patient had collected through the years. The kitchen table had food scraps and bread crumbs. The sink had a stack of unwashed dishes which were several day's old. There were also mice and cockroaches in the apartment. The patient was provided with a new Albuterol metered-dose inhaler as well as a prescription for Triamcinolone metered-dose inhaler 2 puffs twice daily. The nurse was instructed to do daily home visits for the next three days to assess patient's condition and possible need for oral steroids. A referral for psychiatric evaluation was also done. The building superintindent was later involved to assist in cleaning up the apartment. DISCUSSION: There is an increasing need for physician-conducted home visits. "INHOMENESS" (standing for: Immobility, Nutrition, Housing, Others, Medication, Examination, Safety, Spirituality, Services) is an easy mnemonic for the evaluation of the patient's functional status and home environment. . Abdominal x-ray showed cholelithiasis and calcified AAA without obstruction or free air. Abdominal ultrasound was significant for AAA measeuring 5.5 ϫ 5.3 cm involving the mid aorta, mild hydronephrosis, and cholelithiasis. Emergent abdominal CT scan showed a very large, contained, ruptured infrarenal AAA with extensive clot an no extravasation. The patient went for an emergent AAA repair, which she tolerated well, and left the hospital shortly thereafter. DISCUSSION: RAAAs cause 1-2% of deaths in males over 65 years in western countries, with a rising incidence due to an aging population. Most RAAAs present with abdominal pain and tenderness, back or flank pain, and leukocytosis. Anemia and hypotension are uncommon. In one study of 23 patients, 61% were not diagnosed at presentation to internists. The mortality rate from RAAA exceeds 80%. For this reason, elective operative repair is recommended by many experts for AAAs larger than 5 cm, or that grow by more than 1 cm per year. At this point, the risk of rupture outweighs the risk of surgery. The mortality risk for elective AAA repair is 4-8%. Monitoring is recommended every 6 months by ultrasound. Screening is recommended in individuals over age 50-55 with a family history of a AAA in a first degree relative. year-old woman presented with progressive left thigh swelling and pain over 3 weeks. The pain was worse with weight-bearing. She denied trauma and was on no medications. Her past medical history was remarkable for three prior episodes of thigh swelling over 10 years, the first in 1989 during the last trimester of her only pregnancy. Each time, she was diagnosed with a DVT and treated with anticoagulation. There was no family history of clotting disorders. Physical examination revealed moderate swelling of the left thigh with mild warmth but no erythema. The distal portion of the extremity was normal. Ultrasonography of the extremity was negative for clot. Due to a high clinical suspicion, venography was performed. This revealed compression of the left iliac vein by the right iliac artery with significant clot formation, consistent with the May-Thurner syndrome. Marked venous collateralization reinforced the chronicity of this process. The patient was treated with a 24 hour catheter-directed infusion of tissue plasminogen activator followed by angioplasty, and was discharged on anticoagulation. Her symptoms, however, recurred within two weeks, necessitating endovascular stent placement. She has since done well. DISCUSSION: This case raises two important points. The first is that ultrasonography for the diagnosis of DVT, while excellent with a sensitivity of over 95% in symptomatic patients, is not the gold standard. When a clinician has an extremely high index of suspicion, venography should be pursued. The second point is that clinicians should remember that anatomical vascular anomalies may be a cause of DVT. As this case illustrates, the May-Thurner syndrome is a condition characterized by chronic compression of the iliac vein by the overlying iliac artery with resultant intimal hypertrophy and frequent clot formation. It is most common in young women and is often underdiagnosed, accounting for up to 2-5% of cases of lower extremity DVT. It is important to recognize vascular causes of DVT such as the May-Thurner syndrome, as stenting or surgical intervention may be necessary for optimal outcomes. A positive RNP & Smith antibody were consistent with MCTD as the etiology for the insulin receptor antibodies. During the hospitalization, hypokalemic, metabolic alkalemia with normotension led to further workup. High urine chloride values with magnesium wasting and hypocalciuria were consistent with GS. DISCUSSION: Extreme insulin resistance presenting in adulthood may be the result of the rare syndrome of type B insulin resistance. Measurable levels of insulin receptor antibodies are present and some of these patients have associated autoimmune diseases such as S.L.E. or Sjogrens. Also rare, and apparently unrelated to insulin resistance, is Gitelman's syndrome (GS). Previously subsumed under the umbrella of Barrter's, GS is comprised by normotensive, hypokalemic metabolic alkalemia with elevated urine chlorides, urine magnesium, renins and aldosterone, but with decreased urinary excretion of calcium. Although William of Occam's observation that simple explanations are usually the most reliable, contemporary medical sophistication may diagnose multiple zebras with today's complex hoofbeats.

LEARNING OBJECTIVES: 1) Recognize that "Ecstasy" (methlyenedioxymetamphetamine or MDMA) can present with a constellation of findings suggestive of hepatotoxicity with possible gallbladder involvement. 2) Alert physicians to consider drug use when confronted with unexplained symptoms and lab abnormalities indicative of acute hepatic disease. CASE: A 27 year-old unmarried restaurant hostess presented with sudden and extreme abdominal pain. Pain was colicky, began in the epigastrium one day prior to admission, and gradually localized to the right upper quadrant. The patient described the pain as 10 ⁄ 10 without radiation. She had a history of GERD, but this pain was clearly different. She initially denied alcohol or other drug use. Her only medication was Triphasil. She denied acetaminophen use. Her restaurant served raw seafood, which she had neither ingested nor handled, and she had unprotected sex with a new partner for the previous 3 months. She reported no other risk factors for hepatitis, except, on further questioning she revealed that she had used "Ecstasy" for the first time two days prior to admission. On physical exam, she was afebrile and anicteric with stable vital signs. Cardiac and lung exams were unremarkable. The abdomen was soft, mildly distended with pronounced percussive tenderness and guarding in the right upper quadrant. No organomegaly was present. Pelvic exam was unremarkable. Relevant labwork included: WBC ϭ 5.4, Hgb ϭ 15.5, AST ϭ 943, ALT ϭ 820, Alk Phos ϭ 164, Total Bilirubin ϭ 1.3, Albumin ϭ 3.9, and INR ϭ 0.9. Amylase and lipase were normal. Hepatitis serologies and toxicology were negative. Right upper quadrant ultrasound revealed minimal perihepatic and pericholecystic fluid with a thickened gallbladder wall. Abdominal CT showed pericholecystic fluid and intrahepatic biliary dilatation. The patient was placed on IV hydration and ampicillin/sulbactam with eventual resolution of pain. Laboratory abnormalities (peak AST ϭ 1201 and ALT ϭ 1066 on day 2) gradually declined and the patient was discharged without further complications. DISCUSSION. Subacute idiosyncratic toxic hepatitis has been attributed to MDMA in nine previously documented cases. This form of hepatitis can occur after incidental or chronic ingestion, and the symptoms may develop a few days to a few weeks after ingestion. The spectrum of MDMA-associated liver disease is wide, including jaundice and anorexia, acute abdominal pain, and hepatocellular damage. Most reported cases regarding MDMA have noted severe colicky abdominal pain, which is a symptom generally not seen with hepatitis. Our patient had clinical, laboratory and radiologic findings to suggest both hepatotoxicity and acalculous cholecystitis. Clinicians should be aware that MDMA may cause a diverse constellation of hepatobiliary disorders. year-old male refugee from Somalia, who presented on 09/02/99 with a one-month history of a left anterior chest wall mass and one-year of dull pain in the same area. The pain did not have any pleuritic features and was partially relieved with common analgesics. The patient denied any previous trauma to the affected area, drug abuse, weight loss, fevers, chills, or night sweats. The review of systems was negative. His past medical history was significant for malaria, intestinal Entamoeba Histolytica, and a 15 mm PPD test with a normal chest x-ray on 09/19/96. At that time, the patient was referred to the TB Clinic, where he was diagnosed with TB Class 2 (TB infection, no disease). For 6 months, he received INH and Vitamin B6 and claimed adherence to daily treatment. On physical exam, the left anterior chest wall, between the left para-sternal border and mid-clavicular line, over the 7th rib, revealed a poorly defined, tender, non-erythematous mass measuring approximately 5 cm in diameter. The lesion had minimal surrounding edema. No associated axillary or supraclavicular lymph nodes were noted. The rest of the physical exam was normal. A CBC and a chest x-ray did not reveal any abnormalities. A CT scan of the chest revealed a 3 ϫ 4 ϫ 6 cm low attenuation soft tissue mass destroying adjacent rib and costochondral joint and several small prevascular lymph nodes. No pleural effusion or parenchymal lung involvement was noted. A fine needle aspiration biopsy revealed numerous neutrophils but no tumor cells. The aspirate's Gram and Kinyoun stain smears were negative, as well as the bacterial and fungal cultures. On 10/01/99, the laboratory reported a positive culture for a pan-sensitive Mycobacterium Tuberculosis. The patient is receiving treatment for TB Class 3 (TB infection, current disease) with Ethambutol, Pyrazinamide, Rifampin, and INH. DISCUSSION: Although tuberculous osteomyelitis of a rib is an unusual presentation of TB, it should be considered along with other infectious or neoplastic processes in patients who present with chest wall masses. Other infectious etiologies include Staphylococcus, Pseudomonas, and Actinomycosis. Sarcoma and metastatic disease are possible malignancies in this age group. Unless molecular epidemiological studies are done, it is clinically difficult to differentiate TB reactivation from re-infection. Non-compliance and drug resistance are common reasons for failure of INH prophylaxis; none of these reasons were present in this case. The patient is a 57 year-old male from Guatemala. The patient came to the clinic seeking routine health care and was found to have symptoms of depression and alcohol abuse. The physician screened the patient for previous physical and psychological trauma because of his nationality and his mental health symptoms. The patient described how his alcoholic father abandoned him during his childhood and how he had to work at a very early age to help his mother support his younger siblings. During adulthood, he decided to become a police officer for two reasons: to gain power and a steady job. The patient denied receiving any special training at the police academy, such as interrogation techniques, but he was chosen to join an elite police unit responsible for investi-gating political activists. He remembered one of his first assignments: clearing the city's downtown area of homeless people, most of whom had been displaced from rural areas by the war. It was his first opportunity to show his badge and authority. Soon, his superiors ordered him to guard political prisoners while senior officers interrogated them. He witnessed several individuals being tortured. The patient soon found in alcohol a relief for his increasing frustration. After the civil war in Guatemala, he was arrested and tortured on several of occasions for what initially appeared to be disciplinary incidents; however, he was questioned primarily about whether he knew who had ordered the arrest and torture of previous political prisoners. He went into hiding after many of his partners had disappeared or were assassinated. The patient escaped his country after receiving a phone call warning him about a plan to kill him. The review of symptoms was positive for symptoms suggestive of post-traumatic stress disorder (PTSD): flashbacks, intrusive memories, hyper-vigilance, and irritability. His past medical history was negative except for tobacco abuse. His physical exam was normal. The patient declined psychiatric and substance abuse counseling. DISCUSSION: One study found that up to 10% of all foreign-born patients seen in urban medical centers are survivors of torture and other studies found that 5-35% of all refugees have been tortured. As this patient population continues to increase, practitioners should be prepared to screen for torture sequelae. A more complex problem is when a practitioner phases a patient who was a perpetrator and a victim, raising questions of neutrality during delivery of care, criminal complicity, and safety for other refugee patients. The patient is a 28 year-old male asylum applicant from the Democratic Republic of Congo. Because a medico-psychological evaluation was needed as part of an asylum claim, his pro-bono attorney referred him to our refugee/survivor of torture program. The patient was a political activist, who was detained on 2 different occasions and tortured on multiple occasions during both detentions. Aside from symptoms consistent with post-traumatic stress disorder and depression, the patient complained of gingival bleeding for several months. The review of systems was negative. His past medical history was positive for a left forearm fracture and intestinal parasites. He denied any bleeding diathesis, as well as any tobacco, alcohol, or drug abuse. He reported having unprotected heterosexual intercourse and denied any sexually transmitted disease. On physical exam, the patient appeared well nourished. His gums revealed some degree of inflammation and bleeding with minimal blunt trauma. Few small bilateral axillary lymph nodes were noted. Skin and mucous membranes did not reveal any petechiae or ecchymosis. The rest of the physical exam was normal. The white blood cell count was 2.5. The platelet count was 120,000 per mm3 and the hematocrit was 40. While obtaining informed consent for an HIV test, the physician remembered that the diagnosis of some infectious diseases may have an adverse effect on immigration processes. It was decided to delay the test until a legal and an ethical consultation could be obtained. A moral dilemma became apparent as testing for HIV was in the patient's best interest, but positive test results could affect the patient's asylum claim, therefore, harming the patient. The consultants clarified that while some communicable diseases and welfare benefits bar individuals from gaining permanent residency, these restrictions do not apply to asylum claims. Nevertheless, if the patient is granted asylum and decides to apply for permanent residency in the future, he will need a waiver for this medical condition. DISCUSSION: The elements and the goal of informed consent are uniform for all patient populations. However, the process of obtaining informed consent is a dynamic one that varies from individual to individual. When caring for asylum applicants, clinicians must pay special attention not only to the medical facts, but also to any legal and ethical ramifications. Healthcare providers, particularly social workers, should also pay special attention when this patient population is referred for welfare benefits. Imaging studies showed mild hepatosplenomegaly with splenic infarcts, and small wedge shaped defect in left kidney suggestive of infarcts. Echocardiography was negative for vegetations. Total body gallium scan was normal except for poor uptake in splenic bed. Patient continued to spike fevers with negative cultures. Splenectomy performed two weeks later revealed necrotizing vasculitis and focal granulomatous response in medium sized vessels with splenic infarcts. Liver biopsy showed prominent perivenular sclerosis, congestion, and chronic portal inflammation. Significant clinical improvement was noted after the patient was started on therapy with steroids and cyclophosphamide. In the year prior to the onset of her illness, she had traveled to Bonaire, Cuba and Jamaica. Her past medical history was remarkable only for a remote posthysterectomy deep vein thrombosis, and hypothyroidism. She was taking levothyroxine, but she did not take any other prescription or nonprescription medications. The patient's family physician referred her to a general internist for further assessment of her symptoms. Physical examination was normal. CBC showed a mild normocytic anemia with mild eosinophilia. The sedimentation rate was high (59 mm/hr). Creatinine was elevated at 2.2 mg/dL. Urine sediment contained granular but no cellular casts. Twenty-four hour urine collection showed 0.76 g/day protein. Renal ultrasonography was normal. The patient was admitted to hospital for a renal biopsy. The biopsy showed acute tubulointerstitial nephritis (ATIN), which prompted an extensive search for infectious and immunological causes of this condition. Chest films showed no signs of sarcoidosis. Studies for filaria, strongyloides, toxoplasmosis, leptospirosis, and hepatitis were negative. Antinuclear and antineutrophilic cytoplasmic antibodies were negative. While hospitalized for her renal biopsy, the patient complained of a painful, red left eye. Ophthamological examination showed bilateral anterior uveitis. Review of the literature revealed a rare condition, called tubulointerstitial nephritis with uveitis (TINU) syndrome. Given the exclusion of other causes of ATIN, a diagnosis of TINU was made in our patient. Systemic steroids were started for treatment of ATIN, with topical steroids and atropine used to treat the uveitis. After several months of treatment, the patient's symptoms resolved. She has remained free of relapses since her initial treatment 5 years ago. DISCUSSION: TINU syndrome was first described in adolescent girls. More recently, it has been diagnosed in the adult population, with the same female preponderance seen in the original pediatric case series. Although over 50 cases have been described in the literature, it is likely that it is underdiagnosed, because many physicians are unfamiliar with the syndrome. We were successful in making the diagnosis only after discovering TINU by linking uveitis and tubulointerstitial nephritis in the medical literature. case is unique because of the relatively brief duration of Ticlopidine usage. The mechanism by which Ticlopidine causes TTP remains unclear. Ticlopidine as an anti-platelet drug has been used increasingly in different clinical settings. We recommend more judicious use of the drug. LEARNING OBJECTIVES: 1) Diagnose hemobilia and hepatic pseudoaneurysm after percutaneous liver biopsy, and 2) recognize hemobilia-induced secondary disorders such as pancreatitis as a rare consequence of the procedure. CASE PRESENTATION: A 62-year-old woman with a history of hepatitis C had an ultrasound-guided percutaneous liver biopsy for pathological evaluation of the disorder. Approximately 4 hours later, she developed transient hypotension with severe abdominal pain that resolved with fluid resuscitation and meperidine. She was discharged at the same day without further complications and was well until she presented with episodic sharp epigastric pain 7 days after the procedure. 24 hours later, she began to have hematochezia and hematemesis. Physical examination revealed epigastric tenderness, with stable vital signs, no icterus and no hepatosplenomegaly. The initial laboratory results showed strikingly elevated lipase 29,258 U/L and amylase 1,186 U/L. Liver function tests displayed AST 548 U/L, ALT 487 U/L, AP 199 U/L, Bilirubin T 1.7/D 0.9 mg/dl, Albumin 3.3 g/dl, PT 10.8 and PTT 22.9. Upper endoscopy revealed no active bleeding site. Abdominal ultrasound revealed hypoechoic materials throughout the lumen of the gallbladder, some of which were in linear strands and others adherent to its wall. Abdominal CT suggested the presence of blood in the gallbladder. Hepatic angiography was therefore performed. A pseudoaneurysm was identified in a small branch of the right hepatic artery and transcatheter embolization occluded the forward flow from the pseudoaneurysm. The patient recovered without further symptoms. At discharge, her lipase and amylase returned to 452 U/L and 62 U/L, respectively, and liver function tests also improved significantly from maximal levels: AST 600 U/L to 504 U/L, ALT 690 U/L to 305 U/L, and Bilirubin T 4.9/D 4.8 mg/dl to T 2.8/D 2.0 mg/dl. In contrast, her Hgb dropped to 9.5 g/dl from the initial 13.2 g/dl. DISCUSSION: Medline search yielded only one case of hemobilia-induced acute pancreatitis following percutaneous liver biopsy. It is important to be aware of such a potential complication of the procedure, although hemobilia itself has been well documented as an uncommon one. Angiography should be performed to confirm the diagnosis once indicated, especially if there is evidence of recurrent bleeding. These complications prompt further technical improvement to avoid such iatrogenic misfortunes. Leaming Objectives: 1. Distinguish organic diseases from anorexia nervosa.

A 17 year old female with a previous diagnosis of anorexia nervosa and H. pylori infection presented with nausea, vomiting and abdominal pain which began perimenstrually. She was admitted for dehydration. The patient had a history of being severely underweight. Despite the addition of caloric supplements to her diet she had failed to gain weight. She also had a history of severe perimenstrual abdominal pain and vomiting for 3 years, and had been admitted twice for dehydration. Treatment of H. pylori infection diagnosed by endoscopic biopsy had not improved her symptoms. The diagnosis of anorexia nervosa was made and she was assigned a counselor with whom to eat at school. The patient, however, felt strongly that she was too thin and wanted to gain weight. On admission she complained of epigastric and suprapubic abdominal pain which radiated to her back. She denied fever, dysuria, vaginal discharge, and diarrhea, but did note three to four formed stools per day. On physical examination she was cachectic and weak. Her heart rate was 96, temperature 38.2, and BP 129 ⁄ 90 . She was orthostatic by blood pressure. She had mild epigastric and suprapubic tenderness. Stool was negative for occult blood. Laboratory studies revealed a WBC 8.5, Hgb 15.5 and ESR 5. UA showed ketones, trace blood and protein. Renal function, electrolytes, LFTs, albumin, amylase and lipase were normal. Corticotropin stimulation test and TFTs were normal. Endoscopic evaluation revealed Giardia lamblia as well as colitis consistent with possible inflammatory bowel disease. Anorexia nervosa is an eating disorder affecting primarily young women. It is characterized by a refusal to maintain body weight over minimal normal weight, an intense fear of weight gain although underweight, a disturbance in the way one's body is experienced and in most females amenorrhea. Patients who meet some, but not all criteria are often diagnosed with an atypical Eating Disorder. The patient presented here would have been very unusual in that she had neither a fear of gaining weight nor an abnormal perception of her body weight. Certain medical conditions, such as adrenal insufficiency and inflammatory bowel disease, may mimic anorexia nervosa by causing weight loss and fatigue in the absence of significant other symptoms. Giardia lamblia (GL) infection can mimic anorexia nervosa as well. GL is a flagellated protozoan classically thought to cause acute traveler's diarrhea, but also causes a more chronic syndrome of intermittent soft stools, abdominal cramping, weight loss, fatigue, and anorexia. Laboratory studies often reveal anemia or signs of malabsorption, but as in inflammatory bowel disease, may be normal. These patients may have unusual eating habits or signs of depression, but usually do not exhibit all of the psychological features of anorexia nervosa. When evaluating a patient with weight loss believed to have an eating disorder, a high level of suspicion for an underlying organic cause is warranted even in the absence of significant symptoms or abnormal laboratory findings.

Learning Objectives: 1. Recognize Fitz-Hugh-Curtis Syndrome when there is no evidence of salpingitis. F.C. is a 17-year-old woman who presented with a history of right upper quadrant pain. The pain started one week prior to admission as diffuse abdominal pain accompanied by diarrhea. Two days prior to admission the pain localized in the right upper quadrant and right lower rib. It was pleuritic in nature and exacerbated by movement. The following day she developed a temperature of 101 F. Past medical history was significant for an appendectomy at age nine and a pregnancy terminated ten months ago. She took no medications except acetaminophen. Family history was significant for gallstones. She did not smoke, drink alcohol or use drugs. She reported sexual activity with two partners in her lifetime, but only one for the past year. Physical examination revealed a temperature of 37.9 centigrade, blood pressure 104 ⁄ 61 , pulse 87, and respiratory rate 20. Pulse oximetry revealed an oxygen saturation of 100% on room air. Her lung examination revealed decreased breath sounds at the right base and mild crackles. Her cardiac exam revealed a 2 ⁄ 6 systolic murmur. She had tenderness over the right lowest rib anteriorly, the right upper quadrant, and the right back. Pelvic examination was unremarkable except for scant blood from the cervix. Laboratory studies revealed a WBC of 12.6 and an ESR of 73. Further evaluation included a normal right upper quadrant ultrasound, LFTs, amylase, lipase, bone scan, lung perfusion scan, and obstruction series. Blood, urine and stool cultures were negative. Chest radiography and chest computerized tomography revealed a trivial pleural effusion. The diagnosis of postviral pleurodynia was made and viral studies were sent which later were found to be negative. On the fifth hospital day her endocervical culture became positive for Chlamydia trachomatis. She was treated with doxycycline with rapid resolution of her symptoms. Fitz-Hugh-Curtis Syndrome (FHCS) is the syndrome of perihepatitis associated with pelvic inflammatory disease originally described by Stajano in 1919 and rediscovered by Fitz-Hugh and Curtis in the 1930s. Neisseria gonorrhease (GC) was initially thought to be the sole pathogen. Now it is thought that several organisms may play a role and that Chlamydia trachomatis is the most common. Patients with FHCS usually present with pleuritic right upper quadrant pain exacerbated by movement. It is often accompanied by fever. The ESR is almost invariably elevated. Most patients have evidence of current or past salpingitis. Occasionally, as in the patient described here, there is no evidence of salpingitis. In these cases establishing the diagnosis is very difficult. The diagnosis can be definitively made only via direct visualization at laparotomy or laparoscopy. In most cases the diagnosis is made clinically, by eliminating other etiologies and by detection of a causative organism. A response to antibiotics directed against GC or Chlamydia trachomatis may also support FHCS as the etiology.

DELAYED CONSEQUENCES OF CONGENITAL RUBELLA SYNDROME. ME Pickett, Medicine, Brigham & Women's Hospital, Boston, MA LEARNING OBJECTIVES: 1) Recognize features of congenital rubella syndrome that may present in adulthood.

2) Review rubella disease incidence in recent years and recommendations for immunization CASE: A 33 year old Puerto Rican woman presented with progressive throat pain for a week and vomiting and fatigue for two days. She had been diagnosed at age 31 with insulin dependent DM and had a history of ketoacidosis. She also had a history of an elevated TSH, for which she took a low dose of thyroxine. She reported the onset of fatigue with anterior throat pain that became so severe she was unable to swallow food. Despite her poor oral intake and good compliance with her insulin regimen, her glucose levels were increased. Past medical history was notable for bilateral sensorineural hearing loss which occurred at age 2, resulting in a significant speech disability. She had a history of a patent ductus arteriosus which required ligation. Family history was notable for a febrile illness in her mother occurring during this patient's gestation. Physical examination was notable for an ill appearance and borderline tachycardia. She was slightly tachypneic. Skin and mucous membranes were dry. Her neck was tender anteriorly but had no lymphadenopathy. She had no pharyngeal erythema. Laboratories were notable for evidence of ketoacidosis with an anion gap of 23 and elevated ␤-hydroxybutyrate level. Glucose was 533. A TSH level was depressed at 0.41 despite having recently been normal with repletion therapy. Thyroid peroxidase antibodies were detected at a high concentration of 307 IU/ml. She was admitted for treatment of diabetic ketoacidosis and adjustment of her thyroid repletion therapy in the setting of a possible thyroiditis flare. Her throat pain subsided gradually over several weeks. DISCUSSION: A review of this patient's gestational history confirmed that the inutero infection was rubella. Congenital rubella syndrome (CRS) has been associated with the delayed onset of autoimmune polyendocrine failure, commonly resulting in type I diabetes mellitus and autoimmune thyroiditis. This pathology has been attributed to abnormal stimulation of the immune system during its early development. She exhibits many other features of CRS, including skeletal abnormalities, congenital heart disease, and hearing impairment which typically has its onset at toddler age. The MMR vaccine has greatly reduced the incidence of rubella infection and CRS in the United States. Nonetheless, cases continue to occur and 122 cases of CRS were reported to a national registry between 1985 and 1996. Of these cases, it is notable that less than half could be explained by foreign exposure to rubella. It is also notable that among indigenous cases, a large majority had had previous pregnancies with missed postpartum opportunities for vaccination. 44% of CRS cases in the United States occurs in our Hispanic population. There was no peripheral lymphadenopathy, or skin rash. The physical examination was normal except for tenderness in the right upper quadrant. The liver edge was felt 2 centimeters below the right costal margin. Initial laboratory tests included a normal CBC electrolytes and creatinine. Liver function tests revealed elevated serum alkaline phosphatase (382 U/L, normal Ͻ126 U/L) with normal transaminases. Blood and urine cultures were negative. A Chest X-Ray done on admission was normal. A CT scan of the abdomen showed significant retroperitoneal lymphadenopathy.Bone marrow aspirate and biopsy revealed normal cellularity and morphology. Chromosomal studies and lymphocyte flowcytometry did not reveal any evidence of a lymphoproliferative disorder. Serum alpha feto protein and beta human chorionic gonadotrophin levels were normal. The patient was anergic to PPD and control antigens. A serum ACE level was significantly elevated (244 U/L, normal Ͻ52 U/L). Laparotomy with biopsy of the liver and retroperitoneal lymphnodes revealed multiple non-caseating granulomas. Stains for acid-fast bacilli and fungal elements were negative. A diagnosis of Sarcoidosis was made. The patient was then placed on oral steroids which led to a prompt resolution of fever and other constitutional symptoms. DISCUSSION: Sarcoidosis is a multisystem granulomatous disorder, usually affecting young adults and presenting with mediastinal/hilar lymphadenopathy, and a variety of pulmonary, cutaneous and ocular manifestations. Abdominal manifestations are less frequently diagnosed, as patients with sarcoidosis do not present with abdominal symptomatology. Intra-abdominal sarcoidosis is known to involve the liver, spleen, gastrointestinal tract, kidneys and lymph nodes. All reported cases of abdominal involvement in sarcoidosis are in patients with pulmonary manifestations. Isolated retroperitoneal lymph node involvement in sarcoido-sis is rare, especially as a presenting manifestation. Our patient presented with fever and abdominal pain, with a normal chest X-ray. Review of the literature revealed only two reported cases with a similar presentation.Both patients however did have extraabdominal manifestations. Initial differentials in our patient included lymphoproliferative disorders, primary retroperitoneal malignancies and tuberculosis. This case is unique in that the disease was localized entirely to the abdomen. DISCUSSION: Despite gender predilection for women,young males with lupus are at increased risk of renal involvement and thromboembolic complications. They present less often with features such as arthritis, neuropsychiatric manifestations, skin lesions that are common in women. Therefore, the presentation may be obsure without clear defining criteria and often difficult to diagnose. However leukopenia are predictors of disease severity in males with Lupus. Most often, the primary renal manifestation is DPGN Type IV which portends higher degree of progression to ESRD. In addition thromboembolic complications were noted in higher frequency related to persistently elevated anti-cardiolipin. Therefore the ability to recognize sex-linked differences in clinical presentation and disease progression is paramount in males afflicted with Systemic Lupus. year old white female with a past medical history of hypertension, arthritis, diverticulosis, and peptic ulcer disease.She was admitted with a seven day history of swelling, pain and redness over the right flank. Her past surgeries included a laparoscopic cholecystectomy (LC) 5 months PTA for symptomatic cholelithiasis and appendectomy at age 17. On examination she had a temperature of 103ЊF. Local examination of the abdominal wall revealed an erythematous, hot, tender, indurated, non-mobile, non-fluctuant mass measuring 10 ϫ 14 inches in the right lumbar area. Her WBC count was 14,100 cells/mm3, with a differential of 32% bands. A CT scan of the abdomen and pelvis revealed an 8 ϫ 6 ϫ 6 cm multi-loculated fluid collection with pockets of air in the lateral aspect of the abdominal wall at the level of the 12th rib. A large quantity of purulent material was drained from this area. During irrigation of the wound, a marble sized gall stone was removed. Wound cultures grew alphahemolytic streptococci and klebsiella. She was treated with clindamycin and levofloxacin and was later discharged home on the same antibiotics. Review of her old records showed that at the time of the LC the gallbladder wall was accidentally punctured and several stones spilled. Despite irrigation, only one stone was retrieved at that time. DISCUSSION: Cholelithiasis is a common disorder affecting 20 million individual in the US. LC has now become the preferred method of treatment of symptomatic patients. The incidence of spillage of bile and stones is higher in LC. Removing all of the spilled stones can be difficult because of the restricted view and access. There are many anecdotal reports of complications secondary to spilled stones including ab-dominal wall abcess 2-15 months post-operatively, adhesions, umbilical sinus tract formation, erosion into sigmoid colon, cholelithoptysis and others. A review of literature emphasizes the following points: (1) The time interval between laparoscopic intervention and clinical presentation of complications can be quite long;

(2) Abscess may present at different locations unrelated to the gall bladder fossa;

(3) The small percentage of intra abdominal abscess following spilled stones does not mandate conversion of LC to open cholecystectomy simply to retrieve gall stones; (4) Gall stones may not be seen on CT because of their small size and also because they may be isodense with the surrounding fluid. Ultrasound may be more useful for the demonstration of migrated stones especially when they are small and not calcified. A 61 year old white male on chronic hemodialysis for end stage renal disease, secondary to hypertensive nephrosclerosis, presented with a painful left hip mass of three weeks duration that progressively increased. There was no history of trauma to the hip. The patient was an active smoker and there was no significant family history. Pertinent clinical findings included a temperature of 101ЊF, bibasilar rales, a left hip mass about 14 ϫ 16 cms which was warm, tender and non-pulsatile. Relevant lab data included a WBC count of 9,500/mm 3 , serum phosphate of 10.8 mg/dl, calcium of 8.5 mg/dl and albumin of 3.5 mg/dl. X-ray of the left hip revealed an extensive soft tissue calcification. Three phase total body bone and gallium scan showed abnormal intense uptake in the soft tissues of the left and right greater trochanter, proximal humerii, left ninth rib, symphisis menti and the right pubic bone indicating an active metabolic process (increased uptake of gallium) rather than an inactive calcified tumor or a mass. CT scan of the abdomen and pelvis showed dense calcifications involving muscle groups of buttocks, hips and thighs. Parathyroid scan was negative for an ade-noma. Aspiration of the left hip mass revealed cheesy material resembling pus. A gram stain did not reveal any organisms or WBC. It was also negative for AFB and malignant cells. Crystalline, non-birefringent calcium deposits were noted. Excision biopsy of the mass showed prominent blood vessels and no evidence of malignancy. Correlation of imaging results and clinico-pathological findings reveal that this patient has tumoral calcinosis. A unique feature of this case is the first time description of dystrophic calcification using both gallium and bone scanning. DISCUSSION: Tumoral calcinosis is an ectopic calcification syndrome of uncertain etiology manifested clinically as slow growing periarticular soft tissue calcifying masses. Pathology is characterised by fibrous walled cystic spaces containing structureless calcific debris and variable inflammatory reaction. Supersaturation of calcium and phosphorus in the extracellular fluid is the most important factor and more than 60 (mg/dl) 2 in the calcium-phosphorus product is usually necessary for tissue calcification. Treatment includes discontinuation of calcium and vitamin D supplements, use of low calcium dialysate and treatment of established aluminium intoxication. CASE PRESENTATION: A 71-year-old Russian woman presented with a chief complaint of itching for one week as well as right upper quadrant pain and icterus for 3 days. She had been having mild epigastric, post-prandial pain intermittently for almost two years thought to be related to gastroparesis and a paraesophageal hernia. The character of the pain had changed and appeared more localized to the right upper quadrant. She noted a 10-lbs weight loss over one month. She had a history of DM type 2, HTN, CVA, hypothyroidism, diverticulosis, NSAID associated gastritis and a paraesophageal hernia. In 1997 she had a laparoscopic cholecystectomy for cholecystitis complicated by a right subphrenic abscess that required prolonged antibiotics and percutaneous drainage. She was taking ASA, Glipizide, Metformin, Enalapril, Omeprazole and Acetaminophen with Codeine. On exam she was comfortable with a temperature of 97.8 F, blood pressure of 189 ⁄ 96 , heart rate of 76 and breathing 16 times per minute. The skin was icteric. The abdomen was soft, with guarding over the right side, no rebound, bowel sounds were present and rectal examination revealed no masses and the stools were negative for occult blood. An ultrasound examination showed dilated intrahepatic and extrahepatic bile ducts, surgical absence of the gallbladder and an area suspicious for a mass around the pancreatic head. A percutaneous transhepatic cholangiogram showed an obstruction of the extrahepatic common bile duct with proximal dilatation. A fine needle aspiration of the area of the pancreatic head was also performed while a stent was placed through the obstruction and into the duodenum obtaining good flow of bile. After 3 days of partial relief, she developed abdominal and back pain. The amylase rose to 1425 and a contrast enhanced CT of the abdomen revealed findings compatible with pancreatitis. No masses were seen around the head of the pancreas and a periaortic lymph node was thought to be enlarged. CA 19-9 was reported as 21000 and CEA as 18, while the aspirate from the head of the pancreas was reported negative for malignant cells. On day 6 her course was complicated by the development of a hospital-acquired pneumonia and on day 11 she had a sudden respiratory decompensation and ventricular arrhythmia refractory to ACLS protocols and expired. Autopsy findings: Acute pancreatitis with extensive fat necrosis; Well to moderately differentiated adenocarcinoma of the extrahepatic bile ducts with infiltration into the head of the pancreas, and metastasis to the regional lymph nodes, mesentery, peritoneum and pulmonary lymphatic-vascular invasion; Hepatolithiasis throughout the liver; Recent pulmonary emboli of small and medium vessels. Course: Pt started on 20 mcg of Calderol b.i.d. which was increased to t.i.d. Calcium intake was increased to 1800mg q day. The inital concern was that she had a marked intestinal resistance to the physiological action of 25-hydroxyvitanmin D along with secondary hyperparathyoidism. 25-hydroxyvitamin D levels were increased to the point where urinary calcium excretion increased moderatly and the secondary hyperparathyroidism was resolved. A marked improvement in bone mineral density resulted. A referral to GI for a colonoscopy was made which led to the pathological diagnosis of Celiac Sprue. DISCUSSION: Celia Sprue is a disorder in which gluten and related substances gliadin, are toxic and induce the binding of gluten to epithelial cells of the intestine with subsequent tissue damage. Usually characterized by malabsorption, abnormal small bowel structure, intolerance to gluten products, it can be found in a children and adults as well. Even in the absence of GI symptoms such as diarrhea and bloating, celiac sprue can be present and should be considered in the setting unexplained osteoporosis.

HEAD. LM Rucker, Medicine, Jacobi Hospital, Bronx, NY LEARNING OBJECTIVES: 1) Diagnose Avascular Necrosis (AVN) of the femoral head. 2) Recognize the radiographic appearance of AVN. CASE PRESENTATION: A 27 year old female presented to the walk-in clinic with one week of bilateral flank, back, and lower quadrant abdominal pain. The pain was severe and constant, worse with bowel movements and relieved by oxycodone and acetaminophen. She had completed a short course of nitrofurantoin for treatment of a bacterial cystitis one week prior to the onset of the pain. She had episodic diarrhea for two weeks, preceeding the antibiotic use. She had a past history of asthma, ectopic pregnancy, pelvic inflammatory disease, kidney stones, migraine headaches, hypertension, peptic ulcer, and bipolar disorder. Her meds included clonidine, diltiazem, oxycodone and acetaminophen, ranitidine, lithium, benztropine, perphenazine, zolpidem, and antacids. She had taken prednisone off and on for years. She used inhaled flunisolide, ipratropium, and pirbuterol. Physical examination showed an obese woman in pain. Vital signs were normal. She had expiratory wheezes but good air movement. Abdominal exam was significant for pink striae and tender lower quadrants without guarding or rebound. Pelvic exam showed right adnexal tenderness without mass and no cervical motion tenderness. Stool was guaiac negative. Subsequent exam was positive for flank pain and mild costovertebral angle tenderness bilaterally. Hip exam was significant for normal internal and external rotation. Flexion and extension were limited by, and reproduced, her flank pain. CBC, SMA, pregnancy test, abdominal and pelvic sonography, urine analysis, genprobes for gonorrhea and chlamydia, stool culture, C. difficile testing and a KUB (kidney, ureters, bladder) xray were all negative or non-diagnostic. Hip films with the note "Rule out AVN, long history of steroid use" showed increased density and flattening of the femoral heads consistant with AVN. DISCUSSION: This patient had several possible causes of her pain. Because of her steroid history it was reasonable to consider AVN of the femoral head causing pain radiating to the flank. The physical examination of the hip was consistent with AVN, in that rotation was preserved. In osteoarthritis, limitation of internal and external rotation is one of the earliest findings. In either case pain localized to the groin is more common than flank pain. Onset of pain can be more rapid in AVN than the insidious pain characteristic of osteoarthritis. Findings of AVN on xray are easy to overlook (as they were in this patient's KUB), but are essential to search for. MRI is the most sensitive diagnostic test. The earlier the diagnosis of AVN is made, the better the likelihood the patient can have surgical decompression of the joint, delaying the need for total hip replacement. Surg. 1999, 134:503-513) . The scan was done and showed two areas of increased activity; one in the region of the pancreas and the other in the right inguinal area. The JAMA series on reading the medical literature will be reviewed in the context of interpreting studies examining a diagnostic test (JAMA 1994 271(9) :703-707). On scheduled visits nine months since the onset of the pain syndrome her CEA rose to 11.5 and then 21.5. A CEA scan was performed locally and showed an increased area of activity in her right thigh. Plain films of the thigh showed an osteoblastic tumor, which was biopsied. The pathology showed welldifferentiated adenocarcinoma. A repeat bone scan was ordered and showed multiple bony metastases and shortness of breath lead to the diagnosis of pulmonary lymphangitic spread. Her family was close by her side when she died 13 months after her pain syndrome started, 14 years after her initial diagnosis of colon cancer. DISCUSSION: Recurrent colon cancer and use of emerging technologies. year old female with a history of classic migraine with visual aura presented after a tonic clonic seizure. There was mild anterograde amnesia post-ictally. She denied visual change or focal paresthesias. She took no medications, smoked one pack of cigarettes daily, drank occasional alcohol, and used no illicit drugs. Physical examination revealed a tongue laceration, subtle dysarthria, and a left palmomental reflex. The remainder of the examination, including detailed neurologic evaluation, was normal. CBC, chemistries, CXR and EKG were normal. Urine drug screen, blood cultures, and lumbar puncture results were negative. MRI of the brain revealed a lesion in the right frontal lobe, as well as multiple punctate lesions in the right and left centrum semiovale, consistent with ischemic infarcts. Evaluation for etiology of ischemic stroke, including carotid duplex examination, echocardiogram, and hypercoagulable states, was negative. Homocysteine and cholesterol levels were normal. Serologies for ANA, HIV, ANCA, VDRL, and antiphospholipid antibodies were negative. Cerebral angiogram revealed stenosis of the distal portions of the internal carotid arteries (ICA), and a network of dilated capillaries at the base of the skull. These findings are pathognomonic for Moyamoya disease, a rare disorder of unknown etiology in which progressive narrowing occurs in the terminal ICA and proximal portions of the anterior and middle cerebral arteries, followed by extensive collateralization of lenticulostriate arteries. "Moyamoya," a Japanese term for "hazy puff of smoke," describes the angiographic appearance of these vessels. Clinical manifestations result from subcortical cerebral ischemia, or from hemorrhage of dilated collateral arteries. Medical therapies include antiplatelet agents and steroids. Surgical revascularization, in which stenotic portions of the middle cerebral artery are bypassed, is an increasingly promising therapeutic option. She also complained of nausea and decreased oral intake. She had been paraplegic since age 9 secondary to a vertebral fracture, had a chronic neurogenic bladder, and had had multiple previous urinary tract infections and episodes of pyelonephritis. 7 days prior to this visit, she was given levofloxacin for a presumed recurrent urinary tract infection and did not received any relief. Her exam revealed temperature 36.6, pulse 101, respiratory rate 16, blood pressure 125 ⁄ 99 , left costovertebral angle tenderness, and mild suprapubic tenderness. White blood cell count was 4.9 with 47% neutrophils. Her urinalysis showed no nitrite, 2ϩ leukoesterase, 10-25 WBC, 10-25 RBC, and numerous bacteria. She was admitted for a recurrent urinary tract infection recalcitrant to levofloxacin with a possibility of pyelonephritis. Urine and blood cultures were obtained, and her antibiotics were switched to ampicillin and gentamicin. Her symptoms resolved overnight, and the patient was discharged on the second hospital day. The urine culture grew more than three organisms, and the blood cultures were negative. DISCUSSION: Multiple studies have shown that blood cultures rarely contribute to the management of pyelonephritis and, even when positive, seldom vary from the urine culture results. A recent study of 338 inpatients with pyelonephritis showed that, although 91% had blood cultures obtained, only 1 grew a pathogenic organism not found in the urine culture-with no impact on clinical management and a cost of about $18,000. Two other studies showed that the blood cultures from over 200 patients with pyelonephritis had no impact on treatment. The yield of blood cultures in the face of potential pyelonephritis is very low, and the potential harm of false positives is significant. In addition to the potential side effects from antibiotics, studies have shown that treatment based on false positive blood cultures increased length of stay and cost more than $3000 per patient treated. Historically, the rims of felt hats were impregnated with mercury (Hg) and chronic cutaneous exposure to Hg could result in a progressive dementia-hence the phrase "Mad as a Hatter". Hg, also known as "quicksilver," has been injected by boxers as a perceived means of enhancing quickness. Hg injections for a "high" by drug users have also been described. We present an unusual case of parenteral, elemental Hg poisoning. A 34 yr old male presented to the emergency room with renal colic, chills, and a cough productive of green, blood-streaked sputum. He was admitted for fluids and pain control. Admission CT of the abdomen revealed a small nephrolith in the proximal left ureter and mild hydronephrosis. Admission CXR was negative. On hospital day two the patient complained of increasing shortness of breath and a repeat CXR was obtained. It revealed diffuse bilateral branching opacities greatest in the lower lobes thought to be cosistent with a heavy metal exposure. CT revealed high density foci in multiple areas including the lungs, the anterior right ventricle, apex of the left ventricle, liver, and kidney. CT of the head was negative. Overwhelming evidence suggests that the route of Hg exposure was through intravenous injection. The patient denied knowledge of an intravenous injection. He, however, adamantly insisted that a broken sphygomanometer sprayed Hg onto his head when the cuff was inflated. The preponderance of Hg in the pulmonary vasculature and the distinct lack of it in his upper airways and bronchial tree make inhalation exposure extremely unlikely. Most notably, the CXR had the appearance of a pulmonary angiogram. A piece of glass from the broken sphygmomanometer was found in the patient's drawer with a tourniquet. The X-ray of the patient's left hand at the site of an IV infiltration revealed small hyperdense foci consistent with traces of Hg in the skin. Initial blood level was 530 ug/dl. The first urine Hg level obtained after chelation had been implemented was 1193 ug/24 hours. When asked he could describe the implications of his disease and what the therapy did. Within a few weeks the patient began calling for frequent headaches accompanied by nausea and vomiting. Despite being encouraged to go to ED during these episodes he chose not. He stated the symptoms usually went away after the phone call. Within one month he was admitted for labile HTN, intractable nausea and vomiting. When his symptoms resolved, though his HTN was still labile, he demanded to be released from hospital. It was explained to him, that in his current condition his BP could lead to a stroke. He stated he understood and agreed to remain over night. This sequence repeated itself every day. Psychiatry determined that the patient did not appear to have any comprehension of his disease and that this behavior might be a form of a passive suicide attempt. He was determined incompetent and a restraining order was put in place. Shortly after, he walked out of hospital despite the restraining order. The police were contacted. The police chief informed the PCP that the patient was a known offender who had repeatedly gotten out of serving time by pleading his illness. After repeated calls, his wife called back to ask if his medications could be filled over the phone. She stated she did not know his location. At this time the VNA arrived at the his home. She reported that she found him hypertensive, without any medicine and smoking marijuana. Shortly after the patient called asking if the restraining order could be rescinded since the police were after him. It was arranged for him to be seen at his local ED to be reevaluated. The psychiatrist there determined the patient to be competent. The patient did not return to clinic. Two months later he was admitted for a hypertensive emergency and died of a hemorrhagic stroke. DISCUSSION: Competency has several components. The ability to communicate a choice, to have factual understanding of the situation, to appreciate the implications of the decision and be able to rationally manipulate the relevant information.

In conversation, this patient appeared to be competent on each of these levels yet his actions were irrational with regard to his health. Competence does not necessarily imply rationality and actions can speak louder than words. The patient needed reminders to bathe but was still quite functional.She did not have any behavioral or sleep problems, nor any falls or gastrointestinal complaints. The patient had a history of atrial fibrillation on warfarin and osteoarthritis on acetaminophen. Her physical examination was remarkable for an irreg-ularly irregular heart rate of 80, marked short term memory deficits, word finding and naming difficulties, apraxia without parkinsonism or gait abnormalities or lateralizing neurological signs. The patient's daughter, after perusing the internet and visiting her local health food store,had purchased and started her mother on ginkgo biloba, lecithin, and coenzyme Q10. She wondered why her motherhad not shown any improvement in cognition or physical functioning. DISCUSSION: Alternative therapy use is increasing rapidly. Little is known about the use of alternative therapies by the elderly. The elderly are particularly at risk when using alternative therapies because they have limited incomes, many chronic conditions and medications, and caregivers who are demographically the highest users of alternative therapies. Ginkgo biloba is an extract of the maidenhair tree with antioxidant and anticoagulant properties. Three RCTs in the English language literature, evaluating 569 patients for an average of 29 weeks reported modest cognitive improvement. However, dropout rates were high often due to gastrointestinal side effects. Ginkgo biloba is a potent inhibitor of plateletactivating factor essential for platelet aggregation. Interactions with aspirin and warfarin resulted in a spontaneous hyphema and an intracerebral hemorrhage. Lecithin, a cholinergic precursor found in egg yolk,liver and soybeans was examined in at least five RCTs. No improvement in cognition or function was demonstrated. Inositol and choline, other cholinergic precursors, were each studied in one small RCT of short duration without benefit. Coenzyme Q10, a lipid soluble antioxidant, has not been studied in the elderly with dementia. Bovine phosphatidylserine, a membrane stabilizer, was studied in four RCT but cognitive improvement was not consistently demonstrated. Zinc, essential to antioxidant enzymes in the hippocampus, was not beneficial in one small study. The administration at high doses of vitamin E, an antioxidant, delayed death, nursing home placement, loss of function or worsening dementia but only after statistical data manipulation. Given the evidence and the patient's use of warfarin, she was advised to stop all the alternative therapies. bleeding, she felt she had reversed her menopausal transition. Upon the request of her primary care physician, she stopped using the herbal estrogens. Her hot flashes returned. An endometrial biopsy revealed a proliferative endometrium, and a pelvic ultrasound revealed several fibroids. Her hot flashes were controlled with combination hormone replacement therapy, but she ultimately required a hysterectomy for heavily bleeding fibroids. DISCUSSION: In the US, use of herbal remedies and supplements has risen 380% in the last decade, with an estimated 12.5% of American adults using herbal medicines in 1997. Women ages 35-49 use alternative medicines more than any other demographic group. In order to counsel their female patients, physicians must be knowledgeable about possible beneficial and harmful effects of those herbs with estrogenic effects. Although several herbs have estrogenic activity in vitro, clinically relevant estrogenic effects of herbs remain to be proven. At least one randomized trial of single herb treatment for hot flashes has shown no significant effect. However, many herbs are sold as combination tablets using several different herbs. Case reports such as presented here suggest that in high doses and in combination, estrogenic herbs may indeed exert a clinically relevant estrogenic effect. Patients taking such herbs may be taking active estrogen supplements, unaware of the potential risks and benefits of unopposed estrogen. Potential benefits include reduction of hot flashes and vaginal dryness, while potential harmful effects include stimulation of breast cancer cells, endometrial hyperplasia and endometrial carcinoma, hypercoagulability and hepatic injury. American women in their perimenopausal years are turning to herbal medicines in record numbers. The use of certain herbs, especially in combination, may indeed expose women to meaningful doses of exogenous estrogens. Therefore, physicians must be knowlegeable about these herbs and their potential actions in order to provide effective counsel to their patients. Further study is needed to ascertain the safety and effectiveness of higher doses and combination herbal therapies. Hodgkin's Lymphoma A subtype of T-cell angioimmunoblastic T-cell lymphoma occurs in older adults in whom it presents with generalized lymphadenopathy, rash and hepatosplenomegaly. It is locally aggressive and has a high recurrence rate after therapy. We report a case of angiocentric T/NK cell lymphoma presenting as a non-healing gangrenous foot ulcer, which is a very unusual presentation. CASE: An 86 year old woman presented with right leg cellulitis, deep vein thrombosis and femoral arterial stenosis. She reported right leg pain, swelling and redness for 2 weeks The physical examination was unremarkable except for right leg cellulits with erythema, edema of the second and fifth toes with laceration and drainage of serosanginious discharge. CBC and serum chemistry results were within normal limits. The patient developed gangrene of the toes, secondary to peripheral vascular disease. A femoral popliteal bypass and transmetatarsal amputation of the toes was performed. Biopsy (transmetatarsal amputation) showed coagulative and liquefactive necrosis. The skin at the margin showed an atypical, predominantly perivascular lymphoid infiltrate an also revealed blood vessels surrounded and infiltrated by a collar of atypical lymphoid cells associated with brisk mitotic activity. Small angulated cells with dark chromatin pattern and larger cells with open chromatin/prominent nucleolus predominated. Immunostains showed that the neoplasm had an unusual phenotype (CD2ϩ, CD3Ϫ/ϩ, CD5ϩ, CD4ϩ, CD8Ϫ, betaF1Ϫ, CD56Ϫ, TIA1Ϫ), best fitting a CD4ϩ Gamma/delta T-cell lymphoma. This type does not fit an NK cell neoplasm, which would be CD4Ϫ, CD8Ϫ, CF56ϩ, TIA1ϩ, nor the usual gamma/delta T cell lymphoma which would be CD4Ϫ, CD8Ϫ, CD5Ϫ and strongly TIA1ϩ. DISCUSSION: T Lymphocyte receptor has two hetrodimer subunits, alpha/beta and gamma/delta. In the peripheral blood and lymph node gamma/delta receptor contribute 5% in the adult. Most gamma/delta T-cell lymphocytes are negative for CD4 and CD8. But a CD4ϩ and CD8Ϫ, gamma/delta T lymphocyte clone, which is a small population shows lytic activity against the Daudi lymphoma cell line and stimulates, Daudi cells to secrete more gamma interferon and GM-CSF when compared with CD4Ϫ, CD8Ϫ, gamma/delta lymphocyte. There has been only one previously reported case with this phenotype which presented with a cervical mass (Hum Patho 27:1370 ,1996 . Here we are reporting the second case of this disease which also had a unique presentation. A 67 year old woman was brought to her physician by her sister for "confusion." Her history was notable for hypercholesterolemia, domestic violence, and depression. Medications were atorvastatin and aspirin. On physical exam, the patient was disheveled and tachycardic to 115, but the remainder of her exam was unremarkable. On mental status exam, she was alert, hypervigilant, and claimed to be "an African queen." Thought processes were tangential, and speech was pressured. Labs showed Ca 11.5 and albumin 3.8 with normal renal function. Urine toxicology was negative, and TSH was normal. Following psychiatric hospitalization and resolution of psychotic symptoms, the patient returned for followup. Review of her personal and family history failed to reveal nephrolithiasis or hypercalcemia. Repeat Ca was 11.2 and PTH was 7.2 pmol/L (1.3-7.6). 24 hr urinary Ca was 166 mg and creatinine clearance was 68 ml/min. DEXA scan is pending. The patient refuses a surgical procedure to correct her hypercalcemia. DISCUSSION: We found mild hypercalcemia in this patient during evaluation for psychosis. We felt that the patient's elevated calcium warranted investigation but did not contribute to her psychosis. The profile of hypercalcemia and a normal, but inappropriately high, PTH in this patient is compatible with primary hyperparathyroidism or familial hypocalciuric hypercalcemia (FHH). Her 24 hr urinary Ca rules out FHH. Management of primary hyperparathyroidism is controversial.

In 1990, an NIH Consensus Conference stated that all patients with primary hyperparathyroidism should be considered as surgical candidates, but that a subset of patients with asymptomatic primary hyperparathyroidism may be candidates for "surveillance." A recent observational study of individuals with primary hyperparathyroidism, most of whom were asymptomatic, demonstrated no change in serum Ca concentration, urinary Ca excretion, or bone mineral density after 10 years; 27% of patients developed symptomatic kidney stones. Observational data suggest calcium intake can and should be liberalized to 1,000 mg/d if the 1,25-dihydroxyvitamin D level is not elevated. Estrogen replacement therapy, demonstrated in a randomized, controlled trial to suppress bone turnover, reduce urinary Ca excretion, and increase bone mineral density in postmenopausal women with primary hyperparathyroidism, should be considered in appropriate candidates. The role of bisphosphonates in long term medical management of primary hyperparathyroidism remains to be studied. A 72 year old woman with multiple medical problems including irritable bowel syndrome presented with worsening abdominal pain. She initially visited a community physician with complaints of increased abdominal cramping for two weeks. She also described occasional loose stools, nausea, and a loss of appetite. She denied vomiting, fevers, melena, or bloody stools. Her exam was reportedly normal, and she was sent home with loperamide. One week later she presented to our acute care clinic with no imporvement in her symptoms. Stool studies were sent, including bacterial cultures, ova and parasites. No therapy was administered, and these studies subsequently returned normal. She followed up a few days later in the general medicine practice. Upon careful history taking, it was discovered that she had been taking a Chinese medication, a "stomach tonic", for over a month. Laboratory data at this time revealed a white cell count of 12,000 and normal liver function tests. Stool examination for Clostridium Difficile (C.Diff.) toxin was positive. She was asked to bring in the Chinese medicine, and one of the ingredients was identified as "SMZ" (sulfamethoxazole). The medication was discontinued, and she initiated a course of oral metronidazole. She returned to clinic two weeks later and reported resolution of her diarrhea and abdominal pain. DISCUSSION: Patients with chronic abdominal pain are often difficult to evaluate. New symptoms or worsening of symptoms should prompt investigations into new medicines including nonprescription drugs and herbal remedies. Chinese medicines are frequently contaminated with antibiotics, steroids, analgesics, and other often-toxic substances. In this case the adulterant was declared in the packaging, however many herbal medicines come unlabelled. C. Diff. infection is associated with prolonged antibiotic use. While most inpatient cases of C. Diff. present fulminantly, outpatient infections may have a more indolent course, making recognition difficult. Symptoms may include chronic diarrhea, nonspecific abdominal pain, anorexia, and weight loss. Diagnosis is established with an assay for the toxin produced by the organism. Discontinuation of the offending antibiotic and administration of oral metronidazole is the treatment of choice. The renal function eventually stabilized with the addition of cyclophosphamide. The post admission course was complicated by pulmonary embolism. DISCUSSION: SLE affects women nine times more than men, except during childhood and over age 50 where the incidence is nearly equal. This case illustrates that SLE should be considered in young Latino men who present with arthralgias and renal impairment. SLE presents with higher frequencies among Latinos and other selected ethnic groups (African Americans and Asians) and earlier consideration of this diagnosis is warranted in these populations. This highlights the importance of ethnic consideration in the presentation of a not uncommon illness.

This patient had many features for SLE, but the diagnosis was not strongly considered early on because of the patient's age and gender. Early diagnosis is important because initiation of therapy is critical in preventing the disabling and life threatening complications of lupus. The patient's male gender is associated with worse prognosis. Prognosis may also be related to socioeconomic factors such as access to health care. Lungs were noted to be clear to auscultation. The remainder of the examination was unremarkable. Her white blood cell count was 11.6/mm 3 . Chest x-ray revealed bilateral upper lobe infiltrates. Sputum obtained for Gram's stain, acid-fast bacillus stain, and silver stain showed no organisms. The patient was treated with intravenous cefuroxime and initially defervesced with resolution of many of her symptoms. However, she soon began to again manifest fever despite antimicrobial therapy. Suspecting atypical pneumonia, the inpatient physicians added high dose erythromycin to her regimen. She again defervesced after 24 hours. She was discharged to home on oral azithromycin. She was seen in the outpatient clinic one week later for follow up. At that time, the patient complained of continued fever and cough. Her temperature was 39 degrees Celsius, pulse was 120/minute, respirations 24/minute, and blood pressure was 175 ⁄ 92 mmHg. Examination notably revealed decreased breath sounds in the posterior upper lung fields and amphoric breath sounds in the apical regions. Review of the sputum culture data from the hospitalization revealed growth of Mycobacterium tuberculosis on day 12 of culture. Because of her respiratory distress and continued cough she was readmitted to the acute care hospital and placed in pulmonary isolation. She was started on a four-drug antituberculous regimen and gradually improved. On further questioning, the patient related a history of tuberculosis exposure as a child: her sister had died of pulmonary tuberculosis. Skin testing during childhood was reportedly negative and she did not receive prophylactic antituberculous therapy. Tuberculin skin testing was not performed prior to the initiation of corticosteroid therapy when the patient was diagnosed with SLE four months previously. DISCUSSION: Patients with SLE are at increased risk for active tuberculosis, whether or not they are treated with immunosuppresive therapy. SLE patients tend to more often manifest with extrapulmonary tuberculosis or atypical presentations of pulmonary tuberculosis.

Hopkins Medical Services Corporation, Baltimore, MD LEARNING OBJECTIVES: 1) Review the differential diagnosis of fever and rash in a patient with a history of intravenous drug use; 2) Recognize the rash assoicated with disseminated gonococcal infection; 3) Discuss the host and bacterial characteristics of disseminated gonococcal infection. CASE: A 25-year old woman presented to the emergency department after a suicide attempt in which she lacerated both wrists. Her urine toxicology screen was positive for cocaine, opiates, cannabinoids, salicylates and acetaminophen. The wounds were sutured, and tetanus-diphtheria vaccination was given. She was admitted to the psychiatry service for detoxification and treatment of depression. Two days after admission, she became febrile with a temperature of 38.5 degrees Celsius, and painful, red and violaceous nodules were noted on her skin. She denied other constitutional symptoms, headache, chest pain, shortness of breath, abdominal pain, dysuria, or vaginal discharge. Her past medical history was significant for hepatitis C infection, and prior hepatitis B infection. She had been treated for gonorrhea and chlamydia cervicitis in the past, and was noted to be HIV negative one month prior to admission. Medications at the time of transfer were oxazepam and buprenorphine. Her social history was significant for the heavy use of both tobacco and alcohol, as well as intravenous heroin and cocaine use. She has supported herself through prostitution for the past five years. On examination, she was a thin, disheveled young woman. Temperature was 38.7 degrees Celsius, pulse was 90/minute, and blood pressure was 130 ⁄ 61 mmHg. Head and neck examination revealed poor dentition and a normal funduscopic examination. She had no lymphadenopathy. Pulmonary, cardiac and abdominal examination were unremarkable. Examination of her extremities revealed sutured superficial skin wounds on both wrists, and nodules which appeared violaceous on an erythematous base without pustules or ulcers. There were nine lesions counted, which were scatted on the palms of her hands, arms, and legs. There were no warm, tender or swollen joints. Pelvic examination revealed no cervical motion tenderness or adnexal masses. On laboratory evaluation, the white blood cell count was 9.3/mm 3 , with normal hematocrit and platelet count. Serum chemistries were within normal limits. The urinalysis was normal, and a chest x-ray was unremarkable. Blood cultures were obtained, and empiric antibiotics were administered. Serum rapid plasma reagin and HIV tests were obtained. A transthoracic echocardiogram was obtained and was normal. One day after transfer to the internal medicine service, the blood cultures became positive for gram-negative diplococci, subsequently identified as Neisseria gonorrhoeae. The antibiotic regimen was changed to cover for disseminated gonorrhea and the patient rapidly improved. CASE: A 46-year old African American female with end stage renal disease and obesity presented with tender subcutaneous nodule in the abdomen. She had been on dialysis for 12 years, and was on anticoagulants for recurrent fistula graft thrombosis. She was symptomatically treated intially in the out-patient clinic. In the next few weeks, she developed multiple painful nodules in the abdomen and breast. Lab values showed Calcium 10.5-11.5 mg/dL, Phosphorus 4.5-5.5 mg/ dL, Intact PTH 433-651 pg/mL, Albumin 3-3.5 g/dL, BUN 40-50 mg/dL, and Creatinine 7-9 mg/dL. ANA and Anti-DNA were negative. Biopsy of nodule showed a medium sized artery with medial calcification and intimal thickening due to fibrosis. It also showed occlusion of an arterial branch by an organizing thrombus. This confirmed the diagnosis of calciphylaxis. The nodules were removed surgically. In the post-operative period, wound healing was poor and the patient developed secondary infections of the wound. This was treated with appropriate antibiotics. A subtotal parathyroidectomy was performed to control hyperparathyroidism. The PTH level decreased to 70 pg/mL post-operatively, and increased to 355 pg/ mL in a few months. The patient has continued to develop multiple skin and subcutaneous lesions despite the treatment. Her condition is being managed with local wound care, analgesics and antibiotics, and, surgery when necessary. Her phosphate, calcium, and PTH levels are monitored on a regular basis. DISCUSSION: Calciphylaxis is a rare, painful, and debilitating disorder commonly seen in patients with end stage renal disease who are on dialysis or who have recently received a renal transplant. Early diagnosis and treatment may interrupt the progression of the disease process. Hyperparathyroidism, hyperphosphatemia, normal or increased plasma calcium concentration, and Vitamin D supplementation are some of the factors commonly seen in these patients. Epidemiological factors include white race, morbid obesity, administration of warfarin, and low serum albumin. The usual presentation is ischemic necrosis that affects the dermis, subcutaneous fat and, less often, muscle. The diagnosis is confirmed by skin biopsy which shows arterial occlusion and calcification of the arterial wall. 2) To recognize socioeconomic factors that may delay the presentation of a congenital disorder into adulthood CASE PRESENTATION: A 26 year old male from southern Mexico presented with a three month history of progressive dyspnea on exertion associated with right upper quandrant abdominal pain. He reported that 6 months ago he felt strong and could work endlessly. His exercise tolerance had since diminished progressively, and lately he could only walk 1-2 blocks. His RUQ pain was described as a pulsatile, pressure-like pain. Other symptoms included cough, two pillow orthopnea, PND, and intermittent nausea. He did report a doctor telling him when he was fourteen that he had some kind of heart problem. His sister and his aunt have heart problems. He was seen two months earlier for his symptoms, was found hypertensive and given an antihypertensive medication. On presentation, he was afebrile, tachycardic at 125, hypertensive 151 ⁄ 85 , and saturating at 95% on room air. His JVP 13 cm, carotids 2ϩ, had a laterally displaced apical impulse, a left parasternal heave, a holosystolic blowing murmur at the apex that radiated to the axilla, and a similar murmur at the left lower sternal border but that enhanced with inspiration. He had rales half way up the lung fields. His abdominal exam revealed a 10 cm liver by percussion, a tender RUQ, but no signs of ascites. His radial pulses were 2ϩ bilaterally, but distal pedal and femoral pulses were unpalpable, though measurable by Doppler examination. He had 2ϩ ankle edema. His EKG revealed sinus tachycardia of 120, a right axis deviation of 140 degrees, left ventricular hypertrophy, left atrial enlargement, and nonspecific ST and T wave changes. His chest radiograph revealed a markedly enlarged heart with congestive heart failure. His laboratory results were significant for a creatinine of 0.9, AST 158, ALT 191, total bilirubin 1.6, direct bilirubin 0. There was no history of joint swelling, redness, or warmth. While her pain limited her activity, she had no actual weakness. Her past medical history was significant for depression, anorexia nervosa, and menometrorrhagia. During her childhood she had been severely verbally and emotionally, but not sexually or physically, abused. She had no history of thyroid disease, diabetes, seizures, renal dysfunction, photosensitivity, oral ulcers, or rash. Her medications included buproprion, oral contraceptives, acetaminophen, and vitamins. She denied the use of alcohol or illicit drugs. On physical examination, she appeared comfortable, but depressed and teary-eyed. Her skin revealed no rashes. None of her joints had any swelling, erythema or warmth, although her metacarpophalangeal and proximal interphalangeal joints were mildly tender. Thirteen of eighteen tender points for fibromyalgia were positive. Her neurologic examination was normal with no weakness in any muscle group. The Schirmer test was normal. The patient was presumptively diagnosed with fibromyalgia, but treatment withheld until her work up completed. She was scheduled for a follow up visit which she will have in the near future. Laboratory test results : CBC, electrolyte panel, and liver tests all normal. Calcium 9.8, creatine kinase 62, aldolase 3.7. glucose 86, ESR 9 mm/hr, rheumatoid factor negative, TSH normal. Radiographs of her chest, hands, feet, pelvis and cervical spine were not suggestive of any inflammatory or degenerative process. DISCUSSION: Fibromyalgia is a diagnosis of exclusion supported by the absence of the following: history of myopathic drugs; objective weakness and arthritis on physical examination; evidence of systemic inflammatory processes or endocrinopathies on laboratory investigation. While most cases of fibromyalgia are idiopathic, the disorder can be associated with both connective tissue diseases and depression. Fibromyalgia is less commonly associated with a history of abuse than are somatoform disorders. This case, however, illustrates that emotional trauma, depression, and fibromyalgia seem to be a continuum of presentations that interplay with a history of depression, emotional trauma, and somatic symptoms consistent with fibromyalgia. The management of fibromyalgia will be discussed in more detail, but includes patient education, stress management, counseling, pharmacotherapy, exercise, and biofeedback. A 68-year-old man was admitted to the hospital for substernal chest pain. The patient described progressive shortness of breath and a non-radiating, sub-sternal chest pain beginning twelve hours prior to admission. His past medical history was notable for COPD and long-standing tobacco use, but there was no history of exertional angina or coronary artery disease. With the exception of diffuse end-expiratory wheezes, the physical examination performed in the emergency department was normal. All laboratory values including a troponin, CXR, and EKG were normal. The patient was admitted to a telemetry floor to exclude myocardial ischemia. Twelve hours after admission, the patient had a syncopal episode during micturition, and a medical code was initiated. The patient had fully recovered by the time the physician had arrived. A repeat physical examination, however, revealed a persistently split S2 with a loud P2. A spiral CT revealed a large saddle pulmonary embolus. Thrombolytics were administered and the patient transferred to the ICU. Over the next two days, the shortness of breath and chest pain resolved. There were no additional syncopal episodes and no complications from thrombolytic therapy. DISCUSSION: In many situations the physical exam can provide live-saving clues in the diagnosis and treatment of a medical crisis. Because the syncope occurred during micturition, it would have been reasonable to ascribe the event to neurogenic syncope. The repeat physical examination, however, provided evidence of an acute elevation in pulmonary artery pressure, prompting evaluation for pulmonary embolism. Aside from tobacco use, this patient had no risk factors for pulmonary embolism, and with the exception of the chest pain there was no historical data to suggest PE. The history of present illness often provides clues as to what the physical examination will show. This case, however, illustrates that while the history can be useful in guiding the exam, the physical exam remains an independent evaluative measure, and should be carefully performed. Fidelity to the diagnostic method of considering all possibilities and using the physical examination to assess each possibility was life-saving in this case.

Medicine, Univ. of California at San Francisco, San Francisco, CA LEARNING OBJECTIVES: To recognize the importance of monitoring for drug toxicity from non-HIV medications upon institution of therapy for HIV. CASE PRESENTATION: A 58-year-old HIV positive man presented with one week of progressive weakness. He had been seen in the clinic three days earlier, at which time normal CBC and electrolyte panels were obtained. There were no gastrointestinal, neurologic or cardiovascular complaints. The patient's past medical history was notable for hypertension and hypercholesterolemia for which he had been taking atenolol and lovastatin. There had been no recent alteration in the doses of atenolol and lovastatin. There was no history of opportunistic infections and a CD4 count one month prior was 275. A viral load at that time was 30,000 copies, prompting the addition of indinavir to his lamivudine-stavudine regimen. Physical examination revealed normal vital signs. There was objective proximal muscle weakness without atrophy or fasciculation. The remainder of the examination was normal. Repeat laboratory studies revealed an elevation in the creatinine (3.5 mg/dl), AST/SGOT (350 U/L), ALT/SGPT (180 U/L) and creatinine kinase (CK Ͼ50,000 U/L). Rhabdomyolysis was diagnosed, and the patient was started on aggressive hydration with alkalinization of the urine. Lovastatin was discontinued. Over one week, the patient's laboratory abnormalities and physical examination returned to normal. DISCUSSION: Rhabdomyolysis due to HMG-CoA reductase inhibitors (statins) is not uncommon, occurring in 0.7% of all patients receiving the therapy. This case illustrates that statin myopathy can occur due to elevated statin levels because of a drug interaction. The addition of the protease inhibitor decreased the clearance of lovastatin, presumably by inhibiting the hepatic P-450 CYP3A4 enzyme that is responsible for metabolism of both drugs. Similar drug interactions have been described with the combination of HMG-CoA reductase inhibitors with drugs that inhibit the CYP3A4 enzyme (gemfibrozil, cyclosporin, ketoconazole, nefazodone, and niacin). As the life expectancy of the HIV patient is extended, management of other medical disease such as hypercholesterolemia will become increasingly important. Because the medications used in HIV therapy commonly interact with the metabolism of other medications, physicians must exercise special vigilance in monitoring for side effects of all of the patient's medications. Where possible, medications that do not share the same method of metabolism and clearance as HIV medications should be chosen. A 60 year old man with a history of diabetes mellitus presented to his primary care provider's office with the complaint of impotence. The patient had a history of normal erections until two months prior to presentation. He reported a moderate loss of libido, but denied depression or anxiety. The patient denied smoking or alcohol intake. Physical examination was unremarkable including normal genital, prostate, neurological, and peripheral vascular examinations. Serum glucose was 158 with a glycosylated hemoglobin of 9.9 mg/dl; electrolytes and renal function were normal. Cholesterol panel and thyroid function tests were normal. Serum testosterone was found to be 81 ng/dL (normal ϭ 350-720 ng/dL). Additional laboratory evaluation revealed a prolactin level of 650 ng/mL (normal ϭ 17-25 ng/mL). Magnetic resonance imaging of the head revealed an intrasellar pituitary mass lesion consistent with a pituitary macroadenoma. The patient was started on bromocriptine, however discontinued therapy due to intolerable gastrointestinal side effects, and was subsequently treated with cabergoline. DISCUSSION: Erectile dysfunction is a common condition which affects an estimated 10 million men in the United States and is a common complaint encountered by primary care clinicians. Since the prevalence of hyperprolactinemia in patients with erectile dysfunction is low (1-5%), routine determination of prolactin is not justified. Screening for hyperprolactinemia should be limited to men with low sexual desire, gynecomastia, galactorrhea, or low testosterone. Applying this diagnostic strategy proves to be a cost effective means of identifying cases of erectile dysfunction due to hyperprolactinemia. There was no evidence of Parkinson's disease. She was rehydrated and nortriptyline was discontinued. She improved although orthostatic hypotension persisted to a lesser degree. Esophageal studies were normal. Bladder dysfunction disappeared. In the following days the family confided that she had recently been on 17 medications and had been diagnosed with complications related to polypharmacy. Her medications were a compromise between herself and her physicians. She refused to stop several medicines because she was convinced of their efficacy. She also had been stockpiling medications since 1962 and self-medicating. These medications included flurazepam and OTC sleep aids. DISCUSSION: This case illustrates several points. First, it underscores the problem of drug interactions and polypharmacy. Her use of tricyclic antidepressants and antihypertensives all contributed to her orthostatic hypotension. Second, selfenforced bed rest and deconditioning magnified the problem. Third, since an asyptomatic orthostatic drop in blood pressure occurs in a large number of normal elders, age-related physiologic changes can be postulated as a contributing factor separate from any pathologic changes. Fourth, surreptitious use of medications, sedatives with long half-lives, and unknown over-the-counter medications all contributed to her problem and made diagnosis and treatment difficult. Lastly, this case underscores the multifactorial nature of illnesses that while less common in the young are so common in the elderly. 1980, 1984, 1986, and 1990 . Women reported other known and suspected ovarian cancer risk factors on biennial mailed questionnaires from 1976-1996. Newly reported ovarian cancer diagnoses were documented by review of medical records. During 16 years of dietary follow-up (1980-1996) Skim and low-fat milk were the largest contributors to dietary lactose. Women who consumed one or more servings of skim or low-fat milk daily had a 32% higher risk of any ovarian cancer (RR 1.32, 95% CI 0.96 to 1.80; p for trend 0.06) and a 66% higher risk of serous cancers (RR 1.66, 95% CI 1.10 to 2.51; p for trend 0.008) compared to women consuming 3 or fewer servings monthly. We observed an increased risk of serous cancers in women consuming 5 or more servings of yogurt per week (RR 2.35, 95% CI 1.09 to 5.07; p for trend 0.04), but no association with all tumor subtypes together. We found no evidence of confounding by fat or calcium intake and no effect modification by oral contraceptive use. CONCLUSIONS: Our findings support the hypothesis that lactose intake from skim or low-fat milk increases risk of epithelial ovarian cancer. The excess risk was most apparent for the serous subtype of ovarian cancer, and was observed at consumption levels often recommended for women to reduce fracture risk. A 65 year old female with history of recurrent gastrointestinal bleed requiring 26 units of blood transfusion over six months was admitted with anemia and extreme weakness. She had undergone multiple endoscopies and cauterizations of bleeding points but no specific diagnosis. On the fifth endoscopy, it was noticed that she had gastric vascular ectasia. She then had Roux-en-Y gastrojejunostomy. She was also found to have a cirrhotic liver nodule. DISCUSSION: Gastric antral vascular ectasia or watermelon stomach is a clinicopathological condition characterized by severe gastrointestinal blood loss, iron deficiency anemia, endoscopic appearances of longitudinally prominent erythematous mucosal ridges converging on the pylorus and histological appearance of ectatic vascular channels, intravascular fibrin thrombi, and fibromuscular hyperplasia. Etiologic factors include hypergastrinemia, slower gastric emptying, increased levels of prostaglandin E2, and increased proliferation of local neuroendocrine cells containing 5 hydroxytryptamine and vasointestinal peptide. Diseases which gastric antral vascular ectasia is associated with include portal hypertensive states resulting from hepatic cirrhosis or venoocclusive diseases, cirrhosis without portal hypertension, autoimmune diseases, chemotherapy and growth factor use in bone marrow transplant patients. Diagnosis of gastric antral vascular ectasia is delayed for an average of 5-6 years as both endoscopic and histologic findings can often be missed or misinterpreted. This presentation focuses on such a case with pitfalls in diagnosis and treatment. Therapy consists of endoscopic laser ablation, oestrogens, progesterone or calcitonin. Antrectomy offers cure in patients who fail medical management. Awareness of gastric antral vascular ectasia can lead to early diagnosis by limiting the number of investigations, reducing morbidity associated with chronic blood loss and repeated blood transfusions. This review emphasizes the importance of being aware of the condition gastric antral vascular ectasia so that it can be diagnosed and appropriately treated. DISCUSSION: Timely recognition of acute progression of renal artery stenosis helped to restore the kidney functions. Delay in treatment could end in end stage renal disease. The postulated mechanism of pulmonary edema is due to direct effects of angiotensin II in myocardium, reduced pressure natriuresis and its resultant volume expansion. In our patient it was an acute renal failure rather than chronic and it was due to progression of stenosis itself rather than any precipitants. The patient's symptoms were improving at the time the hematuria began. There was no history of trauma, strenuous exercise, heavy alcohol intake, IV drug abuse, or other possible causes of rhabdomyolysis. He had continued taking the Oseltamivir 75 mg twice a day even after the flue symptoms resolved. Muscle discomfort began as tenderness in the buttocks and progressed to the lower extremities three days after his flu symptoms had improved. Physical findings were positive for muscle tenderness in the gluteal areas and lower extremities. There was no rash. The patient's urinalysis was positive for hemoglobin although there were no red blood cells present on microscopic exam. His CPK was 109,817 and the myoglobin was found to be 2540. No electrolyte disturbances were found. The patient was treated agressively with fluids and forced alkaline diuresis. He responded well. Renal function remained within normal limits. RESULTS: Rhabdomyolysis is an established although less frequent complication of Infuenza. It is most commonly seen in young children. The pathologic mechanism remains unknown but may be caused by a virus generated muscle-specific toxin. Influenza virus has also been isolated by direct hemagglutination and electron microscopy from the muscle specimen. In the current case the fact that the muscle tenderness started after the improvement in the flu symptoms and in the absence of other demonstrable causes of rhabdomyolysis raises the possibility of a drug induced etiology. Oseltamivir, a new antiviral which inhibits the viral surface structure protein neuraminidase, has not been associated with rhabdomyolysis. Although a single case report is not enough to document a causal relationship, this patient's illness raises the distinct possibility that it may have been drug related or drug induced. Further postmarketing studies and further observations are in order. Our patient presented with a picture strongly suggesting myocardial infarction. However, a sudden onset of focal cardiac abnormalities together with a history of a recent respiratory tract infection, is also the typical initial presentation of myocarditis in young men. This diagnosis was bolstered by findings of non-significant coronary artery disease and a rapid clinical deterioration, marked by arrhythmias, cardiac arrests and the development of global cardiac hypokinesis. The isolation of infective organisms from serology or cultures may give some clues as to the etiology but the isolated organisms may not be related to the myocarditis. Endomyocardial biopsy is considered the gold standard for diagnosis; however, it lacks sensitivity or specificity depending on the criteria used. Myocarditis is thus a clinical diagnosis. Treatment options beyond the management of heart failure and shock are controversial. This is probably due to a poor understanding of the disease process. For certain etiologies, specific therapy may treat the predisposing disease, but may not relieve the myocarditic process. The use of steroids, immunosuppressants and immunoglobulins have been studied with equivocal outcomes. What is known however, is that myocarditis is potentially self-limiting and that aggressive supportive measures must be accorded patients. If all measures fail, then cardiac transplantation can be a life-saving option. (1) To recognize the work-up for hypercalcemia in the setting of previous malignancy. (2) To diagnose and manage hypercalcemia. CASE: The patient was a 75 year old white female with a past medical history of non-small cell (squamous cell) lung cancer diagnosed one year ago, status postright middle lobectomy. She presented to her physician's office with 6 weeks history of loss of appetite and constipation. The patient also had a history of frequent falls with the most recent fall causing an undisplaced radial fracture four days prior to admission. In the office before admission, serum calcium was 15 mg/dl. and she was admitted for treatment of severe hypercalcemia. Physical examination revealed a thin frail lady Pulse-80, BP 130 ⁄ 80 . Significant findings were dry oral mucosa, kyphosis and spinal tenderness from T 12 to L 1 -. Other lab data on admission showed phosphate 1.5 mg/dl, albumin 3.5 gm/dl and calcium oxalate crystals in urine. Chest x-ray showed cardiomegaly and right pleural thickening. Initial treatment included fluids and diuretics, which decreased Calcium to 14.

Because of her past history of squamous cell cancer, the patient was worked up for recurrence of malignancy with CT scan of the chest/abdomen/brain, which were unrevealing. Initially the patient responded poorly to fluids and diuretics and required bisphosphonates. Because the patient was not responding to the above treatment, hypercalcemia secondary to primary hyperparathyroidism was considered, despite a normal calcium level one year ago. Her PTH levels and PTH-related protein levels were 924 (normal limit 12 to 72 pg/ml) and 0.5 gm/dl (normal limit 1.3 or less) respectively. Because of elevated PTH levels and calcium oxalate crystals in urine, she was considered a surgical candidate and underwent a thyroid Sestamibi scan which showed intense uptake in left parathyroid gland. Surgical exploration of her neck revealed a 3gm parathyroid adenoma. After removal she was discharged to home with a normal calcium level. DISCUSSION: Hypercalcemia is a relatively frequent medical problem. The commonest causes are primary hyperparathyroidism and cancer. This patient with a history of lung cancer had severe hypercalcemia, which was thought to be caused by recurrent cancer, and much less likely by hyperparathyroidism. As compared to other types of lung cancers, squamous cell lung cancer has the highest associ-ation with hypercalcemia. When determining the etiology for hypercalcemia in such group of patients one must not forget to consider hyperparathyroidism, as the diagnostic work up and management will be considerably different for these patients. years (with diabetic retinopathy and sensory neuropathy), hypertension, chronic renal insufficiency, and fibromyalgia presented to the emergency department. At home, she became confused, dizzy, sweaty, and tremulous; then she lost consciousness. The finger stick glucose checked by her husband was 20 mg/dl. She regained consciousness after 50% dextrose was given intravenously by emergency services. Her medication list included: insulin, quinapril, amlodipine, furosemide, amitryptiline (for painful neuropathy), and cyclobenzaprine (started only recently for fibromyalgia). By the time she reached our hospital, her blood sugar level was 68 mg/dl; blood urea nitrogen was 46 mg/dl, creatinine was 1.9 mg/dl. The remaining work-up was unremarkable. DISCUSSION: Prescribing multiple medications to patients with several medical problems is always a bit of a minefield. This problem is particularly acute in the elderly. We present a case of complex drug interactions that led to hypoglycemia. Case analysis suggested the following sequence of events: (1) cyclobenzaprine, together with amitryptiline, deteriorated an already existing baseline diabetic cystopathy (the patient had significant urinary retention in the past); (2) worsening urinary retention created an acute exacerbation of the already existing chronic renal failure; (3) declining renal function reduced insulin excretion and requirements, so that the unchanged dose of insulin became excessive and led to hypoglycemia. CONCLUSION: A diabetic cystopathy has to be assumed in every patient with diabetic neuropathy, and should be considered when starting any new medications, to avoid potentially life-threatening complications of therapy. (1) To be able to diagnose polyarteritis nodosa. (2) To consider polyarteritis nodosa in the differential diagnosis of pancreatitis.

CASE: A 41 year-old man with history of generalized weakness, intermittent fever, loss of weight, arthralgia, wrist drop (EMG showed axonal pattem neuropathy) and intermittent testicular pain (biopsy showed non-specific atrophy) presented with severe abrupt abdominal pain. Work up for polyarteritis nodosa (PAN) was done. Repeated mesenteric and celiac angiograms were normal. The patient was treated with low dose steroids, for presumptive diagnosis of PAN, any infective or other inflammatory etiology having been ruled out. CT scan of the abdomen revealed pancreatitis and a peripancreatic necrotic fluid collection. Laparotomy with pancreatic ebridement was performed. The biopsied tissue was consistent with PAN. The patient refused immunosuppressive therapy hence was maintained on steroids, analgetics, and antibiotics. He deteriorated gradually, developed polymicrobial sepsis and died few weeks later. DISCUSSION: Polyarteritis Nodosa is a multisystem vasculitis that may affect any organ. Most frequently it presents with acute abdomen, occasionally due to pancreatitis. In order to be diagnosed The 1990 American College of Rheumatology requires 3 of the following 10 criteria to be present: (1) weight loss of more than 4 kg (2) livedo reticularis (3) testicular pains (4) mono-or polyneuropathy (5) myalgias (6) elevated blood urea nitrogen or creatinine (7) diastolic hypertension Ͼ90 mmHg (8) hepatitis B infection (9) visceral angiograms with aneurysms and occlusions (10) pathognomonic small or medium-sized artery biopsy specimen. (9) and/or (10) remain the gold standard of diagnosis. Polyarteritis survival at 5 years is about 55%, and is unaffected by adding cytotoxic agents to corticosteroid treatment. CASE: A thirty-one year old woman, mother of four, with a past medical history of SLE (diagnosed ten years ago, treated at that time with steroids), was admitted for new onset worsening angina with anterolateral ST depressions noted on the electrocardiogram. Two years ago she underwent a balloon angioplasty for a 100% blockage of the Left Anterior Descending coronary artery. Since then she remained asymptomatic. The patient had neither family history of CAD nor any other risk factors for CAD. Her home medications included aspirin and metoprolol. Incidentally she was found to be pregnant (8 weeks). The patient refused an angiogram because of the risk of radiation exposure to the child. She opted to be managed medically. DISCUSSION: SLE is a significant risk factor for the development of CAD. The risk of myocardial infarction (MI) is increased up to 50-fold and the risk of fatal MI is increased three times over that in the general population. Atherosclerosis is the most common mechanism for CAD in SLE. Immune complex deposition causes the initial intimal damage, which is followed by accelerated development of atherosclerosis. Patients with advanced CAD also have a higher prevalence of pericarditis and valvular disease, which suggests that an immune factor is the cause of the CAD. The development of coronary atherosclerosis may be related to steroid use and prolongation of life (which allows more time for the development of atherosclerosis), or it could be related to the exacerbation by steroids of hypertension and hyperlipidemia. The treatment of CAD is the same regardless of the SLE status. If coronary bypass grafting is considered, arterial grafts rather than venous grafts should be used, because of a high thrombosis risk. After atherosclerosis, arteritis is the second most common type of CAD in patients with SLE. Pathologic examination is the only definitive method of distinguishing arteritis from atherosclerosis. Rapidly developing stenoses or restenoses suggest the diagnosis of arteritis. Clinically, the distinction is very important because to treat arteritis one has to increase the corticosteroid dose, while on the other hand increasing steroids might worsen risk factors in a patient with atherosclerotic disease. Finally, a high anticardiolipin antibody level, regardless if SLE is present or not, is an isolated risk factor for CAD. LEARNING OBJECTIVES: 1) To consider sarcoidosis in the differential diagnosis of ischemia. 2) Recognize the association between periaortitis, retroperitoneal fibrosis and sarcoidosis. CASE: A 59-year-old woman with advanced sarcoidosis, retroperitoneal fibrosis, ureteral obstruction and bilateral ureteral stents was admitted to the urology service reporting difficult urinating. Ureterolysis was planned. Soon, she developed chest pain and pulmonary edema; she ruled out for myocardial infarction, but a nuclear stress test revealed reversible inferior wall ischemia. Furthermore, she complained of recurring abdominal pains, particularly after large meals. Her stools were positive for occult blood. There was a high clinical suspicion of mesenteric ischemia. A CT of the chest and abdomen, performed to evaluate the etiology of her pains, revealed extensive periaortic fibrosis (involving the entire aorta including aortic root and origins of coronary arteries, as well as the mesenteric arteries) and retroperitoneal fibrosis with ureteral obstruction and bilateral hydronephrosis. In addition there was lung fibrosis and hilar lymph node calcification seen. The patient refused invasive work-up, and is currently managed conservatively. DISCUSSION: Sarcoidosis is a multisystem granulomatous disease, most frequently affecting the lungs. The exact etiology is unknown. It is more common in young, black women and within families. This is the third reported case (Snow 1977 , Godin 1980 describing the triad of sarcoidosis, retroperitoneal fibrosis and periaortitis. In this setting, vascular involvement and ischemia (cardiac, mesenteric, and peripheral) have been reported in several cases. In our patient with this triad, the cardiac and mesenteric ischemia were aggravated, if not caused, by extensive (peri)vascular fibrosis and compression of involved blood vessels. Sarcoidosis should therefore be considered as a rare cause of vascular insufficiency. Keeping in mind that both sarcoidosis and retroperitoneal fibrosis are relatively common, this "triad" is probably underreported. 

FINDING TO DATE: Both residents and medical students were able to identify more than half the simulations correctly. Medical students performed just as well as the residents on these three not uncommon CV simulation exercises.

KEY LESSONS LEARNED: 1. We need to improve physical examination skills at all levels of training. 2. Unlike the residents the medical students had spent some time with "HARVEY" prior to the test. This might have contributed to the lack of difference between the two groups and probably demonstrates the teaching effectiveness of simulation exercises.A formal teaching curriculum utilizing a CPS might enhance learning experience and improve upon our current abilities. . As with most academic medical practices that residents are exposed to, the patient population at CIMA is older and has multiple medical problems. In addition to providing much needed medical services to Hunter College, we also felt that this would be a unique opportunity for junior and senior medical residents (mostly primary care and a few categorical) to participate in the care of a younger and relatively healthy population. CIMA staffs the Hunter College Health Services two half sessons per week. Each session is precepted by an attending physician who supervises two residents. For continuity, the attending physician is consistent for each session and the same two residents attend each session on their ambulatory medicine block. In addition to usual medical exam, the residents and attending are strong advocates for preventative health care. FINDING TO DATE: Since the Hunter College Health Services formally opened with medical services on February 17, 1999, we have seen narly 2,000 studenets. Survey feedback from students has been extremely positive. Resident feedback for the experience has also been extremely positive. KEY LESSONS LEARNED: As medical services at Hunter College expand, we will strive to continue to provide excellent health care to the student body. By exposing our residents to the college health service, we provide not only needed medical care, but also young positive role models for the students and a fulfilling experience for the medical residents. BACKGROUND: The mini-CEX is an efficient, effectice evaluation method that promotes assessment of resident's clinical skills, attitudes and behaviors that are essential in providing high quality patient care. In conjunction with the ABIM's earlier work in developing the mini-CEX 1 , the Board has recently expanded that activity toward the goal of national implementation among the 402 currently accredited internal medicine residency programs training over 23,000 residents. Twenty (20) residency programs located in the northeast and over 500 PGY-1s are now involved in an implementation pilot (July 1999 -June 2000 . PURPOSE: The mini-CEX focuses on the core skills that residents demonstrate in patient encounters. It is a 15-20 minute snapshot of a resident/patient interaction that provides a valid, reliable measure of clinical performance based on multiple encounters (optimally four per year per resident) by different evaluators. GOAL: The pilot project is designed to promote the use of the mini-CEX as a routine, seamless evaluation for residents during clinical rotations. It can be easily implemented by attending physicians in any setting. FORMS AND RATING SCALE: Convenient pocket-size, duplicate forms were developed by participating program directors for use by attending physicians with the goal of immediate feedback to the resident and documentation for the program file. A nine-point scale is used; the rating of 4 is defined as "marginal" and conveys the expectation that with remediation the resident will meet the standards for Board certification. CONCLUSIONS: The mini-CEX assesses residents in a much broader range of clinical situations than the traditional CEX, has better reproducibility, and offers residents greater opportunity for observation and feedback by more than one faculty member and with more than one patient. The ABIM encourages the use of this method in conjunction with or as an alternative to the traditional CEX. The focus is on using statistics to assess the applicability of studies to individual patients. During the 3rd year medicine clerkship, students are required to address a clinical question raised by a case seen during their ambulatory medicine rotation using a recent medical study. This study is then appraised using the questions from the "User's Guides to the Medical Literature" during a presentation to a faculty member and a small group of fellow students. Skills stressed are critical appraisal, application of research to clinical care and the dissemination of knowledge to colleagues. During the 4th year, students may choose to take a one month selective, "Critical Appraisal of Influential Medical Literature." This is an intensive course aimed at improving a student's ability to critically read studies and apply them to clinical practice while emphasizing that current medical practice is grounded in the medical literature. Associated with the above curriculum is a faculty development effort aimed at improving the skills of all medical and surgical faculty in teaching evidence based medicine. FINDING TO DATE: The effort has been positively received by students. Students have exhibited an adequate knowledge base entering the fourth year course.

Those who complete the senior course excel on a post course knowledge assessment. KEY LESSONS LEARNED: It is possible to institute a vertically integrated curriculum in the practice of evidence based medicine. Among the most challenging aspects is involving multiple specialties in the curriculum during the clinical years. year residents participated. The 9 SP stations included four pediatric/adolescent cases: 1) a child with chronic abdominal pain (school refusal); 2) telling parents bad news; 3) adolescent counseling; and 4) newborn hospital discharge instructions. There were five adult SP cases: 1) domestic violence; 2) cardiac risk factor reduction counseling; 3) HIV risk factor reduction counseling; 4) chest pain; and 5) adult survivor of childhood sexual abuse. The four non-SP stations included three EKGs, ten X-rays, twelve dermatological and ten laboratory slides. Criterionbased evaluation was used to determine excellent, adequate and unsatisfactory performances for each station. Two of the SP stations were videotaped and analyzed for specific verbal and nonverbal communication skills. Feedback was given to the resident individually and to the group for each station and for specific skills (interviewing, counseling, physical examination). FINDING TO DATE: As a group, residents performed poorly on HIV counseling suggesting a need for curricular revision. Residents easily recognized depression in one SP but few inquired into the possible history of childhood sexual abuse. Many residents could not formulate a safety plan for the SP experiencing domestic violence. KEY LESSONS LEARNED: To our knowledge, this is the first performance-based evaluation of residents in a combined med-peds residency. Our twenty minute stations are longer and address higher-order clinical skills than have been reported using the objective structured clinical examination (OSCE) format. The CPE has served as a useful tool for assessment of the individual residents and of the residency program. 2) Feasibility planning with 6 resident attendees-Faculty and residents worked together to find a time, place, and format acceptable to all for providing feedback.

Residents identified several issues of importance from their perspective and helped brainstorm suitable solutions.

3) Integration-Faculty reviewed skills, brainstormed recent feedback sessions to determine what could have gone better, and developed a formal process for providing feedback on inpatient rotations. FINDING TO DATE: Six months after the workshops were given, residents completed a telephone survey (response rate 30%). Three resident representatives who had attended the feasibility session attended a 4th follow-up workshop, presented the results of the survey, and discussed ways to further improve. Residents reported faculty to be slightly better to much better at providing excellent feedback. They noted: "feedback is more frequent, faculty is showing more interest, and mid-month evaluation is very good." When asked what was useful about the feedback they were given, residents replied: "feedback is useful when it is specific to a situation, when attendings give coping strategies and other process feedback instead of just content feedback, and when attendings take 5-10 minutes to discuss ways to improve." Faculty unanimously enjoyed the workshops, felt they had learned from them, and felt they had changed their practices in giving feedback. KEY LESSONS LEARNED: Residents and faculty felt the collaborative workshops were a successful and valuable method to improve skills in giving excellent feedback on inpatient rotations. Both groups are now participating in a similar process to improve feedback in the outpatient setting. , and yet, many formative lessons about being a medical professional are learned during clinical training. We created these "Ward Ethics" sessions in part based on student concerns that there are relatively few opportunities to meet and talk with colleagues about issues relating to professionalism and being a medical student on the wards. OBJECTIVES OF PROGRAM/INTERVENTION: During these sessions, students will be able to:

-discuss with peers challenging experiences occurring in clinical rotations; -outline how to address difficult issues with superiors and manage risks, mistakes, and failures on future rotations; -identify strategies to maintain integrity throughout medical training and practice; -improve awareness of personal values and beliefs, strengths and weaknesses, interests and aversions. DISCUSSION OF PROGRAM/INTERVENTION: Ward Ethics sessions are primarily student-run, with an experienced clinician-teacher on hand to facilitate. Students are responsible for bringing the cases for discussion and reflection. The facilitator insures that the group works together to discuss alternative solutions, with a focus on building concrete skills in anticipation of future ethical or professional dilemmas.

The sessions occur during the first week of the Medicine Clerkship and repeat every 6 weeks with new students. By placing the discussions early in the Clerkship, we intend to focus on experiences from the prior clinical rotation and limit the pressures students often feel regarding evaluation with current faculty. As students have completed a variety of clinical rotations, the discussions are mostly anonymous and always confidential. Students and faculty-facilitator complete an evaluation form at the end of each session. FINDING TO DATE: 1) Students reported overwhelmingly positive ratings for the sessions. 2) Students readily shared strategies for coping with difficult situations. They reported that it was helpful to hear that other students had similar experiences. 3) Faculty presence was appreciated by the students, particularly in creating the climate for students to openly discuss difficult situations and affirming their ideas for coping strategies. KEY LESSONS LEARNED: 1) Facilitator training is essential. We have developed a tool-kit for facilitators and held brief workshops to equip them with teaching techniques as they shift roles from lecturer to facilitator. 2) Feedback from faculty and student evaluations drive modifications for future sessions. We plan to expand the program to include all required clerkships, with sessions occuring at multiple sites.

Providing feedback to residents in ambulatory care is an essential component of training programs, but it does not always occur in a frequent and timely manner. We hypothesized that a feedback form would encourage attendings to give feedback to residents, because it does not require excessive attending time. The investigators created a feedback form (check-off sheet) summarizing all written feedback given to Internal Medicine residents in the ambulatory clinic during the months of January and February 1997. The form contained three subheadings: major problems, minor problems, and improvement noted/good work. If a major problem was checked, the resident was required to respond. For major problems only, a copy of the form was placed in the resident's file and reviewed during ambulatory care grading sessions. We collected feedback forms used during January and February 1998 and compared them to written feedback given prior to introduction of the form. In May 1998, attendings and residents completed anonymous surveys regarding the feedback form. Prior to introduction of the form, feedback was given 137 times compared to only 32 times after the form. None of the feedback given during either time period was positive. Of attendings surveyed, 75% felt the form was a convenient, effective way to document problems and was helpful in filling out resident evaluations. However, only 50% used the form regularly. Reasons for non-use included the form not readily available, no time for the paperwork, and the perception that the residents viewed the form as punative. Of residents surveyed, 80% felt the information given in the form was helpful. Because the feedback was usually negative, 56% of residents agreed or were unsure if the form made them angry. Sixty-seven percent of residents agreed that they would like to receive more feedback however, they were equally divided over their preference for verbal or written comments. KEY LESSONS LEARNED: We conclude this feedback form was not associated with an increase in the frequency of feedback given to residents in our ambulatory clinic. Although we designed the form as a check-off sheet, many attendings still wrote notes on the form thereby not saving time. Furthermore, one attending expressed concern that those who used the form would have a disproportionate influence on the resident's evaluations. Additionally, although the residency program director appreciated having written documentation of resident behavior, the residents viewed this unfavorably. Most importantly, the form was never used to give positive feedback, strengthening the resident's negative reaction to the form. Future attempts at improving the frequency of feedback should focus on resolving these barriers and providing a better balance of positive and negative feedback. OBJECTIVES OF PROGRAM/INTERVENTION: It has been recognized for some time that residents play a central role in the education of interns and medical students, but receive little in the way of formal training to prepare them for this function. DESCRIPTION OF PROGRAM/INTERVENTION:We sought to develop an educational module that would serve as an instructional paradigm for training medical residents to become more effective teachers. Our model utilized a multi-media ap-proach with both didactic and role playing components, with the notion that the opportunity to practice new skills with immediate feedback improves learning and the acquisition of these skills. FINDING TO DATE: Our program spanned a four week time period. Twelve senior residents participated, as well as ten students from the medical school who were rotating on the Internal Medicine service. The sessions were divided into three main modules, each focusing on an essential element of the teaching process: giving effective feedback, doctor-patient communication, and bedside teaching. Each session began with a brief didactic component, followed by an interactive set of video clips derived from popular culture as well as educational tapes. The second half of each meeting was devoted to role playing in smaller groups, utilizing patient-actors and pre-arranged scripts. In the final session, the students and residents were separated into three role-playing "stations" which required the application of skills learned from the prior lectures. Each resident was videotaped during the session, allowed to view the videotape alone, and returned to the group to receive feedback with faculty supervision. KEY LESSONS LEARNED: Verbal and written evaluations revealed that the program was well-received by both residents and medical students, who agreed that it was a needed addition to the curriculum. Suggestions included having the sessions during the daytime rather than in the evening, and increasing the proportion of time devoted to role-playing exercises. STATEMENT OF PROBLEM: The learner as teacher program is an effective tool to teach residents much-needed skills that are often overlooked in the formal curriculum. The use of an active participation model further strengthens the training exercise. The positive response from participants has prompted plans to incorporate similar modules into a longitudinal program to enhance the current residency curriculum at our institution. They are required to keep a journal in which they note how care is given and received, observe contextual factors which influence the clinical dynamic, and reflect upon their own responses. Our goal in using this pedagogical strategy is to promote self-reflective practice. We believe that the cultivation of these observational and reflective skills is an essential ingredient in the promotion of professionalism and ethical practice. Course objectives include identifying psychosocial and contextual factors that influence care, ethical and legal issues at the end of life, and principles of pain and symptom management. Students are expected to demonstrate the ability to apply ethical norms to patient care, describe methods of pain and symptom management, communicate in an effective and humanistic manner, and articulate models of patient advocacy. Attitudinal goals include fostering professionalism in the care of patients, appreciating the importance of respecting cultural diversity, and assuming the responsibility for developing competency in these areas. To develop their technical skills in pain and symptom management, students participate in palliative care rounds at Memorial Sloan-Kettering Cancer Center, attend case management problem solving seminars in palliative care, and make a site visit to observe hospice care. Students attend additional seminar sessions to discuss their observations in patient care with a multidisciplinary faculty with expertise in palliative care and clinical ethics. These instructors are from the departments of medicine, neurology, psychiatry, pediatrics, and pastoral care. Students are evaluated on the basis of their seminar participation, written journal, and a 5-7 page paper in which they identify an ethical problem and formulate a solution which promotes patient advocacy. Preliminary evaluation of this clerkship shows that students report increased confidence in their palliative care skills and a heightened awareness of their ethical obligation to provide humane end-of-life care.

TEACHING PRIMARY CARE SKILLS. MA Fischer, EJ Malcolm, Internal Medicine, Stanford University, Stanford, CA BACKGROUND: Increasingly medical practice is shifting from inpatient hospitalization to ambulatory treatment. Generalists are called upon to perform more procedures, and operate with greater speed. Though more training in outpatient sites is offered in some medical residency programs, categorical residents in many programs still spend the bulk of their training time on hospital based rotations. There is little time for exposure to, and proficiency in, skills required for success in outpatient primary care. Some primary care skills are procedure-based requiring interactive instruction. GOAL: To develop primary care skills training sessions to address the educational needs of our residents and the diverse requirements of the subject matter, while integrating easily into our teaching program. METHODS: The curriculum was developed through several methods: subject matter was generated from practicing clinicians in General Internal Medicine, Internal

Medicine subspecialties and other specialties; resident suggestion; published curricula. Final topics of two types were chosen: procedures or examination skills, and knowledge. Teachers were identified who could present the topics in small groups emphasizing hands on experience with models or volunteers, use of equipment, and open question-answer format. Sessions were scheduled for upper level residents (R2 and R3) once weekly. This limitation allowed smaller groups, and a narrower baseline of participant experience and skills. RESULTS: Response to the program from both house staff and educators has been positive. We will discuss feasibility, attendance, teacher cooperation and selection, and program evaluation. Additionally we will offer details regarding the development of this curriculum that has become imbedded in our teaching program. Traditional hand-written patient sign-out systems can be cumbersome and can lead to medical errors. To assess improvement in physician satisfaction and perceptions of traditional and online sign-out systems as used by housestaff at a large academic teaching hospital.

A dynamic database-driven electronic alternative to the sign-out sheet used by housestaff throughout the hospital will be implemented. This electronic clinical reference serves as an on-line resource for physicians as they sign-out patients in their care. It is a fully functional, editable, on-line sign-out sheet. After receiving a user name and password, a physician may use the sign-out sheet by logging onto the site. They may then edit their patient list, change existing data, add, and delete information as necessary. Because the information can be accessed via the Internet, it is continuously available. Paper surveys were distributed to all housestaff in a university based training program in order to assess perceptions of the paper sign-out system. Survey questionnaires prior to the implementation of the online sign-out system were completed by 104 out of 173 (60%) department of medicine residents. These included 46 (68%) PGY 1; 33 (59%) PGY 2; and 25 (51%) PGY 3. Surveys showed that during their last clinical rotation: 1) 71% of all housestaff and only 58% of interns were satisfied with the quantity of information they received in a sign-out; 2) 60% of all housestaff and 57% of interns felt a poor sign-out affected patient comfort; 3) 38% of all housestaff and 30% of interns felt a poor sign-out affected patient safety. KEY LESSONS LEARNED: Paper sign-out systems result in poor satisfaction with the quantity of data received and are perceived by housestaff to have a negative impact on patient comfort and safety. OBJECTIVES OF PROGRAM/INTERVENTION: Cross-cultural training is now recognized as a critical component of medical education, especially given the current attention to racial/ethnic disparities in health care. Initiatives in this area have not been widely implemented and have employed educational methods that are more theoretical than practical given the challenges of caring for diverse patient populations in busy clinical settings. DESCRIPTION OF PROGRAM/INTERVENTION: We developed and implemented a cross-cultural curriculum for medicine residents with the overall goal of improving their ability to understand, communicate with, and care for patients from diverse backgrounds. Our specific objectives were to: Raise awareness of racial/ethnic disparities and sociocultural factors in health; Teach an individual patient based (rather than culture specific) approach to care; Emphasize practical, efficient application of skills in the clinical setting; Employ innovative, experiential teaching techniques; Use open-ended surveying methodology for continuous quality improvement. FINDING TO DATE: The curriculum is composed of five modules (basic concepts, core cross-cultural issues, the meaning of the illness, the social context, and negotiation) taught in four weekly two-hour sessions. The following innovations facilitate discussion and feedback, and help residents put cross-cultural skills into practice: Cases for problem based learning are used to develop awareness of the issues; Videotapes of simulated cross-cultural doctor-patient encounters are shown to demonstrate the concepts and skills in action and to generate discussion; Actors for cross-cultural medical interviewing are employed to allow residents to put the skills to use and create a "living laboratory" for learning. Through these simulated cross-cultural encounters residents utilize the techniques by directly interviewing the actor/patient, by observing and critiquing each other, and through feedback from the actor and the instructors. KEY LESSONS LEARNED: Qualitative data obtained from the residents through open-ended surveying techniques has been used for continuous quality improvement. This has led to the innovations described above and some of the key lessons learned below. STATEMENT OF PROBLEM: Cross-cultural curricula can be feasibly integrated in standard medical training. Interactive case-based learning was greatly preferred over didactic presentation. Use of actors for simulated patient encounters was well-received and useful. Succinct introductory discussion of residents' own cultural backgrounds and attitudes towards health within their families was helpful.

Medical literature and media coverage on racial/ethnic disparities in health was effective in increasing awareness of importance of "cultural competence."

Occupational Medicine, University of Pennsylvania Medical Center, Philadelphia, PA STATEMENT OF PROBLEM: Many general internists move into the field of OM without formal training. This can limit their progression in the field. They may be excellent clinicians but lack in-depth training and experience in population based preventive medicine, risk assessment and communication, knowledge of workplace disability and work placement, and organizational management. In contrast, OM physicians with 1yr clinical residency training may lack the clinical expertise to evaluate complex causation and employment issues. Overall, there is a shortage of physicians with a combination of both in-depth clinical and OM training. Barriers that may deter otherwise motivated internists from specialty training in OM include the need to leave current employment, relocation and disruption of the family. Overcoming these barriers will allow better medical care for underserved workers and a better physician workforce in OM. STATEMENT OF PROBLEM: The stresses of residency can be challenging for relationships. Anecdotally, we had observed that most discussions or seminars on resident stress and relationships were somber and tedious affairs, with limited participation from either residents or spouses, and what participation that occurred was so unrelentingly serious that we doubt any stress was alleviated. OBJECTIVE OF PROGRAM: To remedy this, as part of two two-day resident retreats, we devised a "Newlywed Game" format to enliven discussion of the effects of resident stress on relationships. DESCRIPTION: On the first day of the resident retreats, we recruited five resident couples to participate in the Newlywed Game, with the other approximately 30 residents as the audience. Although focused on married residents, it was stressed that the themes our game were exploring applied to all resident relationships, including family or any significant other (including same sex relationships). As in the Newlywed Game, initially husbands would leave the room while their wives would answer five questions, with the husbands returning and points received if the husband's answers matched their wives. Next, the wives would leave the room, with husbands answering questions and then wives returning to answer the same questions, to receive points for matched answers. Questions were devised to bring up major issues regarding resident stress in a relaxed, often humorous atmosphere, to engender and enliven discussion. Sample questions included: If you could add a medicine to your husband's food, what would it be? a) Prozac; b)Valium; c) Viagra; d) Ritalin; e) Insulin. Other questions asked, for example, what cartoon animal is your spouse most like when stressed? (most common answer was Eeyore); or what song best describes your love life? a) U Can't Touch This; b) You Shook Me All Night Long; c) Minute Waltz; d) All by Myself (most common answer was Minute Waltz). Full discussion of the issues (and conflicts) which may arise in relationships with resident stress occurred after the "game". The second day of the retreat focused on stress-reduction techniques, informed by the discussions. FINDINGS: The Newlywed Game proved to be wildly popular with the residents, and engendered lively, interactive discussion of resident stress and relationships (including a great deal of discussion of the stresses on other relationships besides marriage). Spouses were much more involved in this format than in our usual small group discussions. STATEMENT OF PROBLEM: There is a need to develop educational strategies, which will 1)improve a physician's understanding that the relationship between doctor and patient is integral to medical care, and 2)improve a physician's ability to create meaningful relationships with patients. Advances in communication skills training for physician's have provided necessary but not sufficient solutions to these problems. OBJECTIVES OF PROGRAM/INTERVENTION: 1) To enable learners to increase their own sense of imperative to establish meaningful relationships with patients in order to succeed in providing excellent medical care. 2) To enable learners to improve their ability to establish meaningful relationships with patients. 3) To develop a laboratory for the study of empathy as an independent pathway to meaningful physician-patient relationship. DESCRIPTION OF PROGRAM/INTERVENTION: Interpersonal Rounds consists of an hour-long learning experience with a medical service ward team, two faculty facilitators, and a patient as participants. Interpersonal Rounds occurs in three phases over the course of the hour. 1) The team is assembled on their inpatient unit in order to form the agenda for the rounds. Facilitators provide the general guideline of learners getting to know one of their patients better, as a person. The team selects a patient for whom this activity might clarify or address obstacles the team is facing. The team selects someone to lead a 15-minute bed-side encounter. The team then reassembles in the conference room to debrief. The debriefing is facilitated such that learners are most likely to reflect on their experience of exploring the patient as a person, and what new perspectives they may have acquired. Facilitators attempt to support the natural teaching relationships that form on the team in the dynamics for the discussion. FINDING TO DATE: Thus far, our findings are based upon facilitator observations, and feedback from learners and chief residents (solicited and unsolicited). Learners enjoy the rounds and state that these conversations make a real difference for them and for patients. Learners report that they come to understand the patient's perspective on his/her life, experience of the illness, treatment, and hospitalization overall. Learners report that they want to and are more confident that they can have similar conversations with patients routinely. Participants reported that they had experienced making time-efficient, meaningful connections to patients. KEY LESSONS LEARNED: Empathy can be observed and experienced in this practical and palatable educational format. Empathy holds promise as an accessible, independent pathway to meaningful physician-patient relationship. Learning experiences which occur in the context of patient care and permit learners to arrive at and address salient barriers to care are perceived as enjoyable, useful, and transfortmative by participants. STATEMENT OF PROBLEM: Problem-based learning (PBL) is a modality of teaching that is employed with increasing frequency in undergraduate and graduate education throughout North America. Primary care physicians make unusually good facilitators because of their tolerance for ambiguity and experience in physicianpatient communication. A problem, however, is that most current faculty recruited to be facilitators have not experienced PBL during their own education. Further, extensive training in facilitation is often not feasible, given faculty time and practice commitments. OBJECTIVES OF PROGRAM/INTERVENTION: Improve the quality of facilitation in PBL groups and enhance learning outcomes. DESCRIPTION OF PROGRAM/INTERVENTION: We have developed an innovative process to provide facilitators with immediate, formative feedback. Our process, while based upon the experience of expert trainers, employs the facilitators themselves to observe one another and provide feedback. In this process, each facilitator participates in two sessions per semester, once to observe and once to be observed. These sessions provide an opportunity for facilitator's personal growth by focusing them on the process (rather than the content) of teaching/facilitating. Since facilitators may have varying levels of experience and sophistication in PBL, we developed a qualitative/quantitative instrument that guides the peer observer through a series of observational tasks focused on teaching behaviors. Using our instrument, we train facilitators in observation and feedback at the beginning of each semester. FINDINGS TO DATE: Our initial experience in pilot testing this program has shown it to have two advantages over traditional forms of feedback. First, the feedback that is generated is immediate and is grounded in changeable teaching behaviors rather than in summative outcomes. Second, it requires a minimum of resources, since the facilitators themselves act as the observers. KEY LESSONS LEARNED: This process has potential for faculty development beyond PBL, in that it can be adapted to a variety of small-group learning situations including attending rounds, journal clubs, and others. Our exhibit will feature descriptions of the development and use of our process. We will also have our observation instrument available and a training video that will allow participants to experience our guided observation process. We will be on hand to discuss our process and our plans for testing its effects on feedback, teaching behaviors, and learning outcomes. STATEMENT OF PROBLEM: Residents need to be trained for effective practice in the managed care environment. However, programs have been slow to incorporate new curricula, in part because faculty themselves are unfamiliar with the content and lack easy-to-use instructional materials. Objectives of Program/Intervention: As an original participant in the Partnerships for Quality Education initiative, Tufts Managed Care Institute (TMCI)-a nonprofit educational collaboration of Tufts Medical School and Tufts Health Planproposed to develop faculty-friendly, modular courseware that would provide residents with quality instruction in managed care principles and practices. TMCI would conduct faculty development programs focused on the content and methods to facilitate implementation. DESCRIPTION OF PROGRAM/INTERVENTION: TMCI developed courseware, Preparing Residents to Succeed in Managed Care (PRS), comprised of four exportable modules in three formats: a self-paced CD ROM presenting the essentials of managed care; a rotation comprised of onsite learning experiences in community practices and managed care organizations; and structured classroom courses in patient-centered care and clinical practice guidelines. The CD ROM is an interactive learning tool that places the resident within a group practice; through a series of case-based scenarios the resident learns basic tenets, tools and techniques of managed care. The rotation module recommends flexible structured sessions with physicians, practice staff, hospital administrators, health plan managers, etc., designed to teach residents about the healthcare environment and keys for successful practice. The classroom modules incorporate didactic and role playing components including discussions of the ethics of managed care practice, negotiating treatment plans with patients, and how to find, assess and apply practice guidelines. TMCI conducts faculty development programs to train faculty to use these modules most effectively. FINDING TO DATE: TMCI has trained faculty from 75 residency programs and distributed more than 500 copies of the CD ROM. Faculty have rated the faculty development programs highly (overall average score of 4.7 out of 5), and report that residents have responded well to the substance and instructional designs of the modules. Tests of resident attitudes towards managed care concepts before and after exposure to PRS modules are available with the courseware; seven programs have used the pre-and post-surveys, and results are being analyzed. KEY LESSONS LEARNED: Flexible, pre-packaged teaching materials can help faculty more easily incorporate the principles and practices of managed care into residency training. Faculty themselves benefit from instruction in the content as well as in the use of these materials. STATEMENT OF PROBLEM:In working with medical students, developing the evaluative and teaching skills of housestaff is a dynamic process. Faculty development initiatives should be tailored to the unique needs of housestaff, including the frequent change in services and their level of experience. OBJECTIVES OF PROGRAM/INTERVENTION: The formal evaluation sessions held during the internal medicine clerkship at the Uniformed Services University (USU) provide a unifying forum of "real-time", "cased-based" faculty development for housestaff at multiple teaching sites, while also yielding credible student evaluation and formative feedback. DESCRIPTION OF PROGRAM/INTERVENTION: Formal evaluation sessions are planned meetings, held monthly at all 7 USU clerkship sites, at which the clerkship director, or on-site clerkship director, meets with all of the instructors who are working with students currently on the clerkship. Clerkship goal and expectations are reinforced (Reporter-Interpreter-Manager/Educator; Acad. Med. 1999; 74:1203-7) . Fifteen minutes is spent discussing each student. The clerkship director makes notes of the instructor's comments. Individual feedback is provided to the students the following day during meetings with the clerkship, or onsite, director. At the evaluation sessions, we model and teach the core components of the Stanford Faculty Development Program. FINDING TO DATE: An intern will spend 8-10 hours in formal evaluation sessions each academic year; a resident will spend 4-5 hours. This allows us to focus our attention on the least experienced instructors. Through the use of formal evaluation sessions, a synthetic evaluation framework, as well as pre-clerkship orientation meetings and post-clerkship student feedback, we have achieved a reliability of evaluation similar to quantifiable examinations (Assoc. for Med Educ in Europe Conference, Sept. 1997), have improved the identification of marginally performing students (Acad Med 1997; 72:641-3; Acad Med 2000; In Press), and identified students at risk of poor performance during internship (Acad Med 1998; 73:998-1002) . This reflects, in part, the power of the evaluation sessions in helping housestaff mature in their role as evaluators by modeling and teaching effective skills for use in their interactions with students. KEY LESSONS LEARNED: Making formal evaluation sessions work as a faculty development initiative entails: a realistic commitment of time and resources from the Department Chairperson; a clerkship (or program) director invested in the process; the identification and on-going training of onsite directors; avoiding schedule conflicts which might preclude housestaff from attending. Nevertheless, these sessions play a critical role in the development of the evaluative and teaching skills of some of our most heavily relied upon instructors-our housestaff. STATEMENT OF PROBLEM: Most all medical students enter residency training programs where one of their many responsibilities is teaching medical students. As they progress through the residency program, they also teach junior residents. Some residency training programs have developed Residents as Teachers programs. However, this is not yet universal. One purpose for getting senior medical students involved in teaching is to prepare them for residency. Another is to promote improved clinical skills. Also they serve as a resource for clinical faculty. OBJECTIVE OF PROGRAM/INTERVENTION: The four presenters will share various experiences of their medical schools including successes and failures, barriers and pitfalls. The benefit of such training programs will also be explored. DESCRIPTION OF PROGRAM/INTERVENTION: Each institution believes providing fourth year medical students with the skills to be better teachers is an important endeavor, but has approached it somewhat differently. We will discuss commonalties and differences in approach. Each panelist will take 10 minutes to describe his/her institution's approach. Specific examples of senior medical students demonstrating teaching techniques will also be displayed by videotape. The session will be continuously open for questions, answers, and sharing. A copy of all slides and summaries of the curriculum from each presenting institution will be provided to seminar attendees. FINDING TO DATE: Survey data is currently being tabulated, to be summarized at a later date. KEY LESSONS LEARNED: Utilization of MS 4's is an effective teaching tool. If properly taught, MS 4's are effectively utilized as teachers. STATEMENT OF PROBLEM: As we educate physicians for the 21st century, a thorough understanding of the behavioral, psychosocial, and multicultural contexts of medical illness will be increasingly important in order to successfully manage chronic diseases of an aging population. OBJECTIVES OF PROGRAM/INTERVENTION: To address these issues, the UC Berkeley-UCSF Joint Medical Program is developing an undergraduate medical curriculum in "contextual medicine" focused on the lives of individual patients in the contexts of their families, communities, and society at large. The objectives of this curriculum are: to create an educational setting in which medical students develop a greater understanding of the societal, ethical, and human "context" of medical care; to create a generalizable set of models for developing a contextuallybased curriculum; and to study the effect of contextualization and evaluate outcomes. DESCRIPTION OF PROGRAM/INTERVENTION: The curriculum in contextual medicine will feature a fully integrated, patient-centered curriculum. All core medical education courses will be focused upon a series of patient histories and the illumination of illness from behavioral, community, societal, and ethical perspectives integrated with those of human molecular biology and pathophysiology. Specific components of this curriculum are currently being developed by collaborative groups of research scientists, clinical teaching staff, practicing community physicians, and students. They include: integration of basic and clinical sciences courses around a core group of richly complex patient cases taught in cooperative, small group learning settings; an electronic curriculum database; a student-organized free clinic for the homeless; training projects and student research opportunities in community medicine and public health; year-long clinical experiences in which individual students follow and assist a single patient in the midst of a major life transition; and long-term student-faculty mentoring relationships. FINDING TO DATE: The contextual curriculum is a work-in-progress. Faculty, administration, and students have given unanimous support for the development of this program and curricular changes are being well received. Our exhibit will describe the research design for the contextual curriculum, highlight faculty and curriculum development models, outline student evaluation techniques, and discuss applicability to other programs. KEY LESSONS LEARNED:The process of increasing faculty awareness and knowledge of the entire curriculum, and engaging active faculty participation and cooperation in continuing curricular change is critical, as is "top-down" political support. We believe a contextually-based curriculum in medical education will better prepare physicians to deal more effectively with the behavioral and psychosocial aspects of medical care and improve the quality of patient care. [193] [194] [195] [196] [197] Teams from internal medicine teaching programs are selected to attend one of three national conferences based upon a submitted plan for faculty development addressing ambulatory teaching. Teams include an administrative team leader, a community-based teacher, and other key faculty members. Participants attend facilitated team meetings, plenary sessions, and 4 of 17 workshops related to educational skills and program implementation. Sixty-two teams (259 persons) were selected to attend the first conference in December, 1999. Response rate for the evaluation was 91%. Overall, 1% would not recommend the conference to a colleague, 9% would recommend as a satisfactory experience, 42% as a good experience, and 48% as an outstanding experience. Mean ratings for individual workshops ranged from satisfactory to outstanding. Workshops with the highest attendance were "Teaching in the outpatient setting: precepting skills" and "How Doctors Learn: Tips for Teaching Faculty to Teach." The "precepting skills" workshop received the highest rating (70% outstanding). Participants who identified themselves as being primarily community-based (CB) (N ϭ 85) rated the conference similarly to hospital-based teachers (HB) (N ϭ 130) (p ϭ 0.4). On a scale listing prior training in 24 teaching skills, on average, CB teachers were trained in 4.5 fewer teaching skill areas, compared with the HB teachers (p Ͻ 0.001). CB teachers were less likely to have protected time for their faculty development project compared with HB teachers (odds ratio 0.49, 95% CI 0.24-0.98), but supervisor support and salary support rates were similar. Participants rated the likelihood that their group will accomplish the objectives of their plan with a mean ϭ 4.2 (SD 0.73) on a likert scale from 1 ϭ Not Likely to 5 ϭ Extremely Likely. KEY LESSONS LEARNED: This collaboratively implemented national faculty development project on ambulatory teaching succeeded in soliciting acceptable plans for faculty development from teams composed of hospital and communitybased teachers from 62 teaching institutions. The first conference was highly rated. Information gathered will be used to improve subsequent conferences. We will assess the accomplishments of individual teams through prospective follow-up. We developed a case-based, single-session exercise to help participants identify different approaches to curriculum development, improve their knowledge of the elements of curricular design, and practice skills of collaborative curriculum reform. Since this is a curriculum about developing curricula, we call our session "metacurricular development." FINDING TO DATE: Participants are given a written case scenario and one-page assessment form that we created based on our prior experience in curriculum development and study of the literature on curriculum design (copies will be made available at SGIM). Participants are told they are newly hired faculty members at an academic medical institution and have one month to revise an existing curriculum for teaching Evidence Based Medicine (EBM). First, each participant writes down answers the following questions: Who is the target audience? Whom, if anyone, are they going to seek out for help? What concepts will guide their revision? What do they think a three-session EBM course should contain? How would they assess their curriculum? Participants then form small (3-4 person) groups and negotiate one "optimal" curriculum plan. Finally, the groups come together in a plenary session to present their thought processes and results. We will demonstrate the analysis grid we developed to compare these group data using concepts from literature on curriculum design. KEY LESSONS LEARNED: This session was pilot tested among internal medicine and pediatrics fellows at Harvard Medical School in Fall 1999, with favorable informal feedback from most participants. Specifically, they reported that the session improved awareness of the potential for flexibility in curricular design and provided the opportunity to discuss the roles of experts and end-users (student learners) in the curriculum development process. STATEMENT OF PROBLEM: This session on curriculum development using a realistic junior faculty scenario was well received by internal medicine and pediatrics fellows. In 1 1 ⁄ 2 hours, important concepts in curricular design and both intrafaculty and faculty-student collaboration can be addressed using this interactive case. We believe this session meets an important training need for future academic medical faculty. DESCRIPTION OF PROGRAM/INTERVENTION: The program consists of physician led sessions and uses the UMedic multimedia computer curriculum that includes 10 patient-centered cardiology programs. The programs are linked to "Harvey," the Cardiology Patient Simulator, and include the history, bedside examination, laboratory data and treatment. Two sets of multimedia computer test items, equivalent in content and empirical performance, are used as pre-and posttests. (Reliability coefficients KR 20:0.81, 0.84, respectively.) Items assess auscultatory and non-auscultatory skills including the identification, interpretation and correlation of findings with cardiovascular diseases. The scope of bedside skills was framed by a multicenter consortium of cardiologists and educators. Eight residents participate in five 2-hour sessions during a one-month ward teaching rotation that was created to provide structured learning time separate from their dayto-day patient responsibilities. Residents take the pretest at the beginning of the rotation and participate in instructor-taught sessions, using "Harvey" and UMedic, that focus on bedside skills. Residents take the posttest at the end of the rotation and complete an evaluation form designed to measure their satisfaction with the program and to quantify their patient care responsibilities during the rotation. FINDING TO DATE: After the first 3 rotations, 25 out of a total of 90 second-and third-year residents completed pretests, 5 remediation sessions and posttests.

Overall pre-and posttest means, standard deviations and ranges were 44.86, 14.86 and 20-64 and 70.43, 11.8 and 52-88, respectively (p Ͻ 0.001). Similar improvement was demonstrated in all areas of auscultatory and non-auscultatory findings. Residents rated all items regarding the program structure and content highly. Each house officer was responsible for an average of 36 inpatients and 18 outpatients during the 4-week rotation. KEY LESSONS LEARNED: We have successfully developed a structured program in cardiac bedside skills using a simulator/computer-based standardized method to teach these skills to internal medicine residents that includes outcome measures. Residents enjoyed and appreciated the program and indicated that they wished to have similar experiences in other medical disciplines. During residency, young physicians are frequently confronted with suffering and death. House officers report feeling increased isolation after the death of one of their patients (Ferris T et al., 1998) . There is little time to engage in an in-depth discussion of these issues in traditional educational forums. The goal of the group is to provide a safe place for medical residents to discuss experiences with suffering and dying. Our objectives are to improve patient care at the end of life as well as to foster the development of each resident by acknowledging and exploring the complicated interface of the personal and professional in our lives as physicians. The group meets in the evenings for two hours twice monthly facilitated by two internists and one psychiatrist who share an interest in end of life issues and group process. The group begins with the sharing of a meal followed by a short silent writing exercise in transition to the time spent sharing thoughts and feelings about patients, family, and friends. The evening ends with a moment of silence, or a reading of poetry or prose. In the first year of the group, 66% of the medical house staff attended this voluntary evening group at least once. Twelve of 22 participants responded to a survey about the group. Mean values were calculated for the 5-point Likert scales. Residents felt the group provided a safe place to grieve (4.3) and that the group helped them explore their role as a doctor and as a fellow human being in caring for dying patients (4.5) . Residents reported the group improved their care of patients at the end of life (4.0). KEY LESSONS LEARNED: Residents will attend and value a group that provides a consistent respectful environment in which to discuss personal and professional experiences with suffering and death. Residents feel this type of group improves the care they give patients at the end of life. Facilitators face many challenges in offering an open-ended group with frequently changing busy participants.

HOMELESS. RI Jahiel, PJ Ellis, Yale Primary Care Residency Program, New Haven and Waterbury, CT PURPOSE: To understand and overcome barriers to physicians in the delivery of quality health care in homeless settings. METHOD: Qualitative assessments of internal and external barriers were done before and/or during service of 30 general internal medicine residents at a homeless shelter clinic. RESULTS: Internal barriers: 1) inability to comprehend the devastating effects of lack of a home; 2) spotty knowledge of homelessness and homeless populations; 3) stereotypes of and psychological distance from homeless people; 4) unrealistic view of physician roles in caring for homeless patients; 5) little preparation for work in low technology environments, 6) little advocacy training. External barriers: 1) time constraints; 2) incompatibility between circumstances of homelessness and medical management; 3) obstacles to follow up of patients in the medical care system; and 4) lack of needed resources. CURRICULUM TO ADDRESS THESE BARRIERS: A) 90 minute seminar on: a) 10 functions of a home; b) diversity of homeless people and causes of homelessness; c) hazards of homeless life; d) psychological distancing, stereotyping, and unrealistic clinical expectations. B) 16-hour supervised clinical experience at the shelter with a) 40 minutes per patient encounter; b) check list to elicit information on homelessness his-tory and obstacles to a stable home; c) help in decisions on a medically optimal management versus a less-than-optimal one that is more feasible in homeless life; d) guided reflection on perceptions of homeless people; e) guidance to bypass obstacles to health care by adaptation of care, physician role modification, or team work. Outcome shows satisfaction of the medical residents. CONCLUSION: Efforts to improve the performance of physicians in the care of people who are homeless need to address internal and external barriers. It is feasible to package interventions targeting these barriers in a short time period in a community setting. In other stations, residents evaluated risk of coronary artery disease and osteoporosis or dealt with issues of domestic violence and abuse. Some of the stations were videotaped. Feedback was provided to residents by standardized patients and faculty using predetermined checklists. Workshops were followed by an interactive, didactic session to reinforce key learning points. Videotapes were individually reviewed with each resident. FINDING TO DATE: 1) Important deficits exist in resident evaluation of cardiac risk in women patients. 2) Many residents failed to uncover a history of childhood sexual abuse in a depressed patient. 3) Videotaping of selected workshops uncovered important problems in communication in some residents, which were previously undetected by faculty. KEY LESSONS LEARNED: Clinical Skills Workshops are a valuable venue for resident teaching and may reveal deficiencies in resident skills that are not otherwise apparent. Limitations include cost, faculty time, and resident availability. Commitment by program directors to ensure attendance is important to maximize effectiveness of this intervention. Further evaluation is needed to determine if these workshops have a positive effect on patient outcomes. DESCRIPTION OF PROGRAM/INTERVENTION: Our goal was to develop a brief, dynamic seminar to: 1) provide residents with a basic set of teaching skills; 2) improve teaching of interns and medical students; and 3) facilitate positive team interactions. FINDING TO DATE: We conducted two, half-day workshops for senior residents, using a curriculum adapted from one previously described at Stanford. Topics included an introduction to educational frameworks, learning climate, communication of goals, understanding and retention, and feedback. Participants completed a post-seminar questionnaire (using a 5-point Likert scale) addressing overall usefulness of the program; its impact on attitudes; and a retrospective self-assessment of teaching performance.

KEY LESSONS LEARNED: Overall, residents found the program useful (mean score 4.11, standard deviation (SD) 0.68), despite varying expectations beforehand (mean score 3.33, SD 1). They felt their teaching performance improved in all four topic areas (see table below ). The program impacted their attitudes, knowledge, and teaching behaviors, and enhanced cooperative interactions among residents. They formulated a variety of personal goals, and identified barriers to their implementation, the most frequent of which was time. STATEMENT OF PROBLEM: Senior residents in this community-based program derived benefit from a short, interactive program aimed at improving teaching skills and behaviors, despite variable expectations beforehand. Time constraints were felt to be the most important barrier to positive behaviors. Hospital, Ann Arbor, Mich., which includes both teaching and non-teaching members. The survey addressed demographic features of respondents, attitudes toward teaching residents, and perceived benefits of or barriers to teaching. Univariate analyses on aggregate data were performed using SPSS software. FINDING TO DATE: There was a 47% response rate. Most respondents indicated that they enjoyed working with residents (91%) and preferred settings which involved residents (88%). 63% agreed that they had made teaching a priority when selecting their current position. They identified five benefits of teaching: continuing medical education, keeping up-to-date with the current literature, personal and job satisfaction, and mentoring opportunities. They targeted four barriers: office time considerations, financial concerns, managed care, and administrative duties. Male and fellowship-trained physicians were more likely to have made teaching a priority (LR 7.41, p ϭ 0.006 and LR 9.7, p ϭ 0.002, respectively). Younger physicians were more likely to perceive faculty development and mentoring as benefits (LR 11, p ϭ 0.011 and LR 10, p ϭ 0.018). Physicians with more teaching experience were more likly to see different patient populations and compensation issues as barriers (LR 7.5, p ϭ 0.023 and LR 5.9, p ϭ 0.05), while those with less experience found faculty development to be of benefit (LR 7.5, p ϭ 0.02). Those with more than 20 contact hours per week were more likely to identify time considerations as a barrier (LR 6.74, p ϭ 0.009). Fellowship training increased the likelihood of identifying the four barriers of managed care (LR 7.5, p ϭ 0.006), office time considerations (LR 3.9, p ϭ 0.047), compensation issues (LR 14, p ϭ 0.001), and administrative duties (LR 4.5, p ϭ 0.035). KEY LESSONS LEARNED: Community-based physicians in this study identified internally-oriented benefits and business-based barriers which did not preclude teaching. Several physician-specific characteristics influenced attitudes toward teaching. Understanding these influences may enable graduate medical programs to more effectively interact with community faculty members. STATEMENT OF PROBLEM: Quality control across multiple clinical teaching sites is a major concern for medical educators. Tuft's Internal Medicine clerkship presents major challenges due to 8 geographically dispersed hospital sites, each using multiple ambulatory sites to provide 33-50% ambulatory experience during the 12-week rotation. OBJECTIVES OF PROGARM/INTERVENTION: Tufts sought an evaluation system that would provide meaningful feedback for faculty to improve their clerkships, foster sharing of ideas among clerkship directors, allow comparisons across sites, adapt to needs of other core disciplines, and be administratively sustainable. DESCRIPTION OF PROGARM/INTERVENTION: Our standardized evaluation system for the Internal Medicine clerkship provides detailed feedback to every site. Web-based evaluation instruments administered to students yield a 97% response rate; general questions allow comparisons across all core disciplines, specific questions allow focused queries within a discipline. Quarterly reports of results (comparative bar graphs and extensive comments) are sent to clerkship directors, chairs, and directors of medical education, with established feedback loops among the dean's office, faculty, and sites. Every three years the Curriculum Committee performs in-depth evaluation of clerkships, with annual reports tracking the status of action plans. Targeted discussion of successes and problems occurs during regularly scheduled clerkship directors meetings. To maximize clerkship improvements, refinement of evaluation instruments and the feedback process is ongoing. An identical evaluation system is used for all 5 other core clerkships at over 100 departmental sites. In designing a nine-month longitudinal ambulatory care experience, we faced the difficulty of communicating with medical students and preceptors in diverse community and university settings. 1) Develop an interactive web site which allows students and preceptors in remote continuity clinic sites access to curricular materials and course information.

2) Insure frequent and timely feedback by making student evaluations available to preceptors on the web site.

3) Provide an on-line forum for students to discuss patient issues in between monthly didactic sessions. Stanford Medical School has 4 years of experience providing comprehensive support for 41 medical school courses including: on-line searchable syllabi, streaming video, and interactive web based teaching programs. We have leveraged this experience to produce a web site for our longitudinal ambulatory clerkship for medical students that provides: (1) access to the web based clerkship syllabus; (2) online information about the clinic, searchable maps and didactic schedules; (3) electronic journals that can be accessed by students and preceptors in remote sites; (4) web-based feedback and evaluation forms; and (5) links to primary care resources on the web. Finally, in order to establish a base of institutional memory, we routinely update the web site with useful teaching points from the monthly student presentations, relevant communication from e-mail correspondence between students and preceptors, as well as frequently asked questions. Students have found that the web site provides easy access to didactic materials and references. Students have spent an average of 7 minutes per page of didactic material. They have also found it useful for obtaining logistical clerkship information. The most commonly visited pages on the site (excluding lecture topics) included: "Goals and Objectives", "Course Structure and Policies", and "Purpose". We have not yet been able to evaluate preceptor use of the site as we have only recently acquired user access for the preceptors. KEY LESSONS LEARNED: 1) The web site is a useful adjunct to traditional means of distributing course materials, particularly in a longitudinal ambulatory clerkship.

2) Preceptor use of the web site requires specific attention to education of the faculty regarding access and use of the web site.

3) The web site may provide a secondary incentive for community providers as it provides access to the medical library, librarian assisted searches and full text articles. STATEMENT OF PROBLEM: Increasing evidence gathered during the 1990s has indicated that errors in physical diagnosis may be quite common among physicians-in-training. Cardiac auscultation, a time-honored art and long the centerpiece of physical examination, seems to have particularly suffered. Whether this mounting evidence of trainee's deficits and errors might have caused a change in teaching practices and attitudes towards cardiac auscultation is unknown. OBJECTIVES OF PROGRAM/INTERVENTION: 1) To assess the time and importance currently dedicated to the teaching of cardiac auscultation during internal medicine residency; 2) To compare 1999 data to those previously gathered by us in a similar survey conducted in 1990. DESCRIPTION OF PROGRAM/INTERVENTION: We surveyed all accredited U.S. residencies in Internal Medicine. Data were collected by mailed questionnaire between January and June 1999. All non-respondents were sent a second copy of the survey one month after the initial letter. The survey instrument consisted of a one-page questionnaire identical to the one previously used in 1990. Time and importance dedicated to cardiac auscultation were then compared between 1999 and 1990 data. FINDING TO DATE: There were 258 ⁄ 434 responses in 1999 compared to 328 ⁄ 426 in 1990. On a six-step scale, with six being the highest, 1999 respondents attributed significantly less clinical value than their 1990 counterparts to 7 ⁄ 11 cardiac auscultatory findings. And yet, 1999 respondents gave significantly greater importance to cardiac auscultation (5.6 Ϯ .6 vs 5.3 Ϯ .8, P Ͻ .001), and expressed greater desire for more auscultatory teaching during residency (5.2 Ϯ .9 vs 4.9 Ϯ 1, P Ͻ .001). This desire translated into a significant increase in the number of programs offering structured teaching of cardiac auscultation (48% compared to 27.1% for 1999 and 1990 respectively, P Ͻ .001). In contrast to the 1990 survey there was no inverse correlation between respondent's age and their wish for greater teaching of cardiac auscultation. There was, however, a persistently negative association between presence of structured auscultatory teaching and university affiliation of the residency program. This association was weak for 1990 (30.7% of nonuniversity affiliated programs offered teaching compared to 21.9% of universityaffiliated, P ϭ 0.07) and stronger for 1999 (53.7% vs 40.2%, P ϭ .04) KEY LESSONS LEARNED: These data indicate a significant increase in the number of programs currently offering formal teaching of cardiac auscultation. This increase is significantly greater among non university-affiliated residencies. Whether this greater teaching emphasis will translate into greater clinical proficiency of primary care trainees needs to be demonstrated. The evaluation program developed emphasized formative as well as summative evaluation. Components, tools, time lines, responsibility and outcome processes were included. Tools were developed to evaluate all components of the private office experience. These included evaluation of the experience by the residents and the physician preceptors, site evaluaton, including resident patient panels, resident and preceptor performance and resident self assessment. The quantitative tools consisted of Likert Scale measures with open ended items and patient panel computerized tracking. Qualitative methodology included interviews and focus groups. Reliability and validity of the tools were established utilizing content and construct validity and for reliability internal consistency and concurent review as appropriate. FINDING TO DATE: The evaluation plan developed has yielded significant outcome data to date. The findings have supported that private office sites provide improved experiences in managed care, patient counseling and patient continuity without impacting on patient diversity. Important feedback from the physician preceptors indicated that mentoring did not interfere with patient relationships however areas of concern were staffing and productivity. Stipends for the mentors were also identified as important as was continued faculty development. KEY LESSONS LEARNED: The findings from the evaluation have enabled refinement and future planning for a high quality educaitona experience for residents. STATEMENT OF PROBLEM: Recorded encounters with real and standardized patients are invaluable for teaching physicians, residents, and medical students to communicate with patients and for conducting research on the medical encounter. However, most researchers and educators use outdated audio technology to produce analog recordings that are cumbersome to work with and difficult to analyze. OBJECTIVES OF PROGRAM/INTERVENTION: To develop a computer-assisted system that simplifies archiving and analyzing audiotape data, facilitates time measurement and improves sound quality. DESCRIPTION OF PROGRAM/INTERVENTION: We adapted hardware and sofware initially developed for the music industry to meet our research and education needs. A PC computer equipped with a multi-media player and a 32-bit sound card converted our analog recordings to digital ".wav" (sound) files. Individual encounters were processed using Sound Forge XP filtering tools to enhance sound quality. The processed encounters were archived on compact discs (CD) as digital data files using Adaptec Easy CD Creator and a HP CD-R writer. Once archived, Sound Forge XP and Microsoft Access were used simultaneously on a split screen to analyze individual encounters. Sound Forge XP plays the audio file, provides a visual image of the sound frequencies, and enables researchers to time stamp significant events for easy retrieval and to time specified regions of talk to 0.001 seconds. While the recording plays, raters can complete the Access data-entry form that we programmed to automatically perform calculations and prepare the data for entry into SAS. FINDINGS TO DATE: Digital file processing greatly enhances the sound quality of poorly audible analog recordings. Visualizing the sound frequencies enhances analysis. Events within encounters can be easily marked, timed and retrieved. Digital files may be sent via the internet, and data is easily incorporated into research and education presentations. KEY LESSONS LEARNED: This user-friendly system makes audiotape analysis as easy as a mouse-click. Processing software makes noisy clinic backgrounds and background hiss disappear. Automated data processing minimizes errors, and digital archiving makes the data portable. We are currently using the system to evaluate an educational intervention designed to improve resident communication skills. We expect higher inter-rater reliability than previously possible. We will also use this technology to export data to distant collaborators for joint analysis and to share existing and future databases with other researchers. Furthermore, digital technology is a rapidly growing field, and we look forward to incorporating advances that will expand our capabilities and the system's efficiency including digital recorders and high-compression formats. Departments of Medicine and Medical Schools rely on residents to teach interns and medical students. Most residents, however, have little formal training for this role. In 1996 Montefiore developed the Resident As Teacher (RAT) Project. RAT is a mandatory full day of interactive seminars designed to help PGY 2's become better teachers, evaluators, and feedback givers. This survey was conducted to evaluate the impact of RAT on resident's beliefs, skills, and attitudes towards teaching. The instrument consisted of two surveys-a Pre-test and Post-test-administered and analyzed in a blinded manner. The Pre-test was given before RAT. The Posttest was given after the residents had completed at least one rotation where they were responsible for teaching interns/medical students, an interval of 2-4 months after the RAT intervention. Thirty-six PGY 2's participated in the study, 30 returned Pre-tests (83%) and 34 returned Post-tests (94%). Before the intervention 93% of the participants reported they would like to become better teachers. However, prior to RAT, only 43% reported they knew how to achieve this goal. Prior to RAT, 43% described their previous teaching experience as limited (i.e.: "a topic presented on rounds") and 50% as moderate (i.e.: "math tutor in high school"). None described their prior experience as substantial. Of note, 90% reported they were unsure of their current role as teacher and reported little or no previous help developing their teaching style. After the RAT intervention, residents demonstrated improvement in three of the questions posed (on a Likert scale of 1-5, where 1 ϭ strongly agree, 3 ϭ neutral, 5 ϭ strongly disagree): 1) "I consider myself a teacher" (p ϭ 0.02), 2) "I feel I know how to become a better teacher" (p ϭ 0.04) 3) "I feel comfortable giving feedback and evaluation to medical students and interns" (p Ϫ 0.05). There was a trend towards improvement after RAT for five questions with p Ͻ 0.1, though they did not reach statistical significance: 1) "The Department of Medicine relys on me to teach" 2) "The Medical School relys on me to teach" 3) "Giving feedback is an integral part of teaching" 4) "I would like to be a better teacher" 5) "I know how to achieve the characteristics which I feel make excellent teachers". After RAT 89% reported they had learned useful tools at the workshop to develop their teaching style, 73% reported making changes in their approach to teaching, and 73% reported feeling more prepared for their role as a teacher. KEY LESSONS LEARNED: The RAT intervention has a positive impact on resident's beliefs, skills, and attitudes towards their role as a teacher. There is a general lack of understanding of what the costs are to train residents in community-based ambulatory settings. It is important to address this especially as patient care has shifted into ambulatory settings. In our general medicine clinic 6 faculty members see patients independently most of the week,but 4 faculty also spend one or more half days supervising 1-4 residents. Residents see both resident and faculty patients. Faculty supervise under HCFA primary care exception.This study compares practice income and expenditures when the faculty practice is compared with the resident practice.

To determine the added costs or savings attributable to a resident continuity practice in a combined faculty resident practice. Fulltime equivalent providers were defined and salaries distributed according to percent effort in the faculty or resident portion of the practice. Faculty and resident visits were collected from the hospital billing system, as were billed clinic revenues and payor mix. Total practice overhead and support staff costs were taken from hospital cost reports. These costs were assigned proportionally by percentages of patients seen in each portion of the practice. Data was analyzed for a 4 month period. The monthly revenue or cost was determined for the faculty only practice and the resident practice. Costs to the resident practice included the portion of faculty salary used in supervision. Faculty practice total patients seen per month were 1117 compared to 132 for the resident practice. The monthly revenue for the faculty practice is $52,251 compared to $25,736 for the resident practice. Faculty salary is $51,826/mo at 3.98 Full time equivalents versus $10,917 for the resident practice at 1.6 FTE. Total monthly costs including overhead and support staff are $113,887 for the faculty practice and $35,664 for the resident practice. The total monthly loss for the faculty practice is calculated to be $15,487. The monthly loss for the resident practice is $6,205. KEY LESSONS LEARNED: In this community staff model clinic for resident training the resident portion of the practice was significantly less costly. The lower salary of residents with faculty supervision offset their lower patient volumes; also the support staff and overhead costs are significantly less for the resident portion of the practice. These variables contributed to the overall lower cost of the resident practice.

Medicine, St Vincents Hospital and Medical Center, New York, NY STATEMENT OF PROBLEM: As managed care becomes the predominant health care delivery system in the United States, there is a pressing need for physicians to be trained for practice in this new environment. This is especially true for many urban internal medicine programs where the majority of residents receive their training in traditional hospital based clinics. When New York State initiated a plan to implement mandatory managed care enrollment for all Medicaid recipients it became imperative that we appropriately prepare our residents for this new practice environment.

OBJECTIVES OF PROGRAM/INTERVENTION: In the summer of 1998 we developed and implemented a curriculum in "practice management skills for the managed care era" for our third year internal medicine residents. We wanted our residents to be familiar with the language of managed care, to understand the dynamics that led to its evolution, to recognize the various methods of assessing quality and controlling cost in a managed care environment, and to recognize ethical issues a physician might encounter in a managed care setting. DISCUSSION OF PROGRAM/INTERVENTION: The curriculum consists of six units: 1) Health care financing and managed care, 2) Computers in clinical practice, 3) Cost effective clinical practice, 4) Quality and managed care, 5) Coding and billing and 6) Ethical issues and managed care. The didactic portion of the curriculum is taught during the ambulatory care block. The practical aspect involved a reorganization of our clinic practice site, specifically the development of attending and resident practice teams. Residents were given a pretest prior to the didactic sessions to assess their baseline knowledge. They were given the posttest an average of 6 months after completion of the didactic curriculum. Residents evaluated each of the units they completed. FINDING TO DATE: Over the past year a total of 20 third year residents have participated in the didactic portion of the curriculum. General knowledge of managed care terminology and procedures improved from a mean score of 49% to a mean score of 62%. Response to the curriculum was overwhelmingly positive, with 60% of residents stating "strong agreement" and 40% stating "agreement" that the curriculum should be continued. The only suggestion for improving the curriculum was that it be offered earlier in their training and that more time be dedicated to it. KEY LESSONS LEARNED: Internal medicine residents see managed care training as a much-needed part of their education. They are eager and receptive to interventions that will improve their knowledge of the key competencies needed for managed care practice. It is possible in a traditional hospital based clinic to implement a didactic curriculum that will improve resident's knowledge of managed care terminology and procedures.

evidence-based conferences while on subspecialty rotations. For those residents who are interested in more in-depth training, a two-week, intensive EBM elective is offered for second and third year residents. To provide an easily accessible representation of resident work and to measure their acceptance of EBM, we created an Internet based bank of Critically Appraised Topics (CATS), a short formatted easily read style for educational prescriptions. Residents submit these one-page summaries of articles that are evaluated by two faculty members who provide specific structured feedback. Once revised, the CAT's are published on the World Wide Web. To promote interest in this, we have created a weekly contest for the best resident CAT and award the resident with the most weekly prizes with a trip to the McMaster EBM course. FINDING TO DATE: In the first twelve months, we have received 105 CAT submissions from 23 residents. The program has been immensely popular with residents. KEY LESSONS LEARNED: The use and acceptance of the CAT bank provides initial evidence that residents accept the application of EBM in daily clinical practice.

STATEMENT OF PROBLEM: How to make a modular curriculum in obstetric medicine, designed for use in medical residency programs, readily available to educators across the country. DESCRIPTION OF PROGRAM: Using an extensive review of medical consultations performed at a busy obstetrics and gynecology hospital, a brief practical curriculum in medical problems in pregnancy was developed for medical residents. This curriculum included resident handouts, educator teaching scripts, bibliographies and cases for discussion for each of 13 specific topics. The curriculum was designed to be readily adaptable into a wide variety of educational forums within residency programs. The curriculum has previously been presented as a precourse at the national meeting of the Society of General Internal Medicine in 1998 and has subsequently been sold to over 50 programs across the country.

Having completed the curriculum development project, we became concerned with how to more effectively disseminate the materials. Therefore, in association with the Society of Obstetric Medicine, we have placed our entire curriculum on the web. Our handouts, teaching scripts, cases and bibliographies can now be readily accessed and utilized by residency programs across the country. Slides can also be downloaded from the website and adapted to local needs. The website also provides access to an obstetric medicine examination which educators can employ to stimulate and document resident learning. We wish to present our website and make available to others the lessons we have learned thus far through its development and utilization.

STATEMENT OF PROBLEM: The decision of a primary care physician to refer a patient to a specialist requires formulation of a question concerning diagnosis or management. In an era of managed care it is essential to create best practice strategies so that appropriate and cost effective referrals to medical specialties will be requested. OBJECTIVES OF PROGRAM/INTERVENTION: In residency training it is important to have house staff (HS) recognize the value not only of framing a proper question for a consultant, but also of examining the pre-probability of benefit to the patient, and providing the appropriate history, physical findings and lab data expected of an internist. DESCRIPTION OF PROGRAM/INTERVENTION: Four educational interventions were created. These varied from 1) a "consult" resident assignment to review requests of other HS to a specialist; 2) assignment of selected consults from HS to a resident for critique with other residents; 3) faculty presentation of incomplete consults; and 4) teaching a need to seek evidence to answer the posed question and determine benefit to the patient of the referral. FINDING TO DATE: Review of consult requests by HS to five medical subspecialties revealed inappropriate and cost ineffective referrals. Preliminary examination of HS appreciation of the different interventions found general rejection of the first, but interest in selected consult review, faculty presentations and evidence gathering approaches. Fifty per cent of HS responding to an initial questionnaire agreed that phrasing a question, providing proper data and knowing established indications for recommending a procedure were all important in residency training. However, attempts to quantify measures of appropriateness of consults by attendings and specialists were hampered by lack of significant interrater agreement (kappa value Ͻ.6). KEY LESSONS LEARNED: HS referrals to medical subspecialties arise from uncertainties in knowledge, diagnosis or management. Consult review provides a good base for examining educational needs and teaching appropriate communication to subspecialists, especially in an era of managed care restrictions.

STATEMENT OF PROBLEM: Peer-and self-assessment are important for continued quality improvement. Physicians-in-training are often resistant to peer-evaluation and may not have the skills necessary for self-assessment. One solution may be to address these issues throughout the medical curriculum. We describe a multifaceted approach to introduce peer-and self-evaluation to first-year medical students. OBJECTIVES OF PROGRAM/INTERVENTION: First-year students will: 1. Understand the principles of giving and receiving feedback. 2. Practice giving verbal and written feedback to peers. 3. Assess one's own performance by discussion and by written evaluation form. DESCRIPTION OF PROGRAM/INTERVENTION: The program takes place during our interviewing course. An introductory lecture is presented addressing the basic principles of giving and receiving feedback. To apply this information, students view a trigger tape of a medical interview and provide written feedback on a transcript of the interview. The class discusses student's responses to the trigger tape. Our interviewing course is conducted in a small group setting with individual students interviewing standardized patients in front of their group. After the interview, students are encouraged to critique themselves and their peers in a constructive and supportive fashion. Using the feedback skills practiced in the introductory lecture, the students comment upon their peer's skills. At midcourse, a student's interview with a patient is videotaped. After watching the videos, students complete self-and anonymous peer evaluation forms using rating scales of knowledge of the structure of the interview and interviewing style and techniques as well as providing open-ended comments. Evaluations are not used in grading. Students receive a written summary of their evaluations. FINDING TO DATE: The written evaluations from the videotaped interviews have been evaluated. Peers generally rated their colleagues higher than the students did themselves or did the faculty, and provided very few negative comments. Peers scores on the numeric scales show little variation between categories (low interdimentional variance) or between students (low discriminability). Students were more critical on their own performances and were more likely to give themselves negative rather than positive comments. KEY LESSONS LEARNED: First-year medical students were willing to provide feedback to their peers, but were reluctant to assign "grades" to their colleagues. Peer evaluations using an open-ended format for feedback may be more beneficial than quantitative assessment. Students had difficulties providing positive selffeedback. Interventions are needed to teach students how to have a more balanced perspective of their performance. In the future, students will be surveyed as to how useful they found conducting the peer and self-evaluation as well as receiving the written feedback from their classmates. STATEMENT OF PROBLEM: Introducing professionalism into the pre-clinical curriculum is challenging in that many medical students immersed in studying basic sciences may not see the relevance of psychosocial issues. Beginning student's motivation for learning is external, i.e., passing examinations. Professional development occurs through internal motivation and life-long reflective learning. How might this discrepancy in student intellectual development and motivation and the goal to teach professionalism be resolved? OBJECTIVES OF PROGRAM/INTERVENTION: The objective of our program is to encourage pre-clinical students to appreciate the importance of professionalism by allowing the students to become aware of their own perceptions, followed by observing role model physicians, self-reflection, and discussion. DESCRIPTION OF PROGRAM/INTERVENTION: Our professional development curriculum begins during the application process. Applicants are asked to answer the following question on their secondary application "What are the competencies and qualities you feel a physician should possess for practice in the 21st century?" The composite results were presented to the students during a professionalism seminar in our Introduction to the Medical Profession Course. Students were challenged to strive to meet these expectations. During an early communitybased clinical experience students were asked to observe the professional qualities of their physician preceptor and write an essay on how they plan to practice medicine in the future. Following the early clinical experience small groups of students met with a faculty member to discuss their experience. FINDING TO DATE: The top 5 competencies were: technology (keeping up with technological advances) 63%; business (management, finance, cost) 62%, patient advocacy (helping, counseling, demonstrating concern ) 53%, lifelong learning (keeping up with developments and research) 48%; communication (talking to patients and families, listening) 47 The top 5 qualities were: compassion (give aid or support) 49%, intelligence (ability to gain and apply knowledge, grasp of field) 45%; adaptability (flexible, open-minded, willing to change) 34%, honesty and integrity (sincerity, trustworthiness) 32%; dedication (dependable, hardworking, diligent) 29%. These qualities are consistent with expectations published in the literature. KEY LESSONS LEARNED: Medical school matriculants are very insightful into understanding the professional attributes they will need to practice medicine in the 21st century. The above approach to introducing professionalism allows students to define necessary competencies and qualities of physicians and reflect on the importance of those attributes through observation and experience. Such experiences are designed to help students assimilate professional behaviors. (RTJC) is a multi-professional, collaborative, learner-centered model emphasizing evidence relevant to patient care. It evolved from a facilitator-centered session, that emphasized critical appraisal of journal articles that didn't have direct relevance to patient care. A resident coordinator, general internist, medical librarian, and clinical pharmacologist conduct this once monthly one-hour session. The preceding month, the resident selects a clinical question of interest to him/her. Search techniques are pursued with the librarian. Abstracts and at most two articles are selected for the group to review in "real time" at the journal club session. Groups of three learners seek to answer the clinical question posed by the resident coordinator, and site the best evidence provided. Worksheets from the Information Mastery Working Group are utilized to enhance the critical appraisal process. Findings are summarized and implementation to patient care discussed. FINDING TO DATE: Outcomes of RTJC are the following: Increased learner participation and enthusiasm for journal club with improved voluntary attendance at each session. Open dialogue, and improved communication between internal medicine residents and multi-professional staff on patient care issues. Observation of resident behaviors,in the ambulatory internal medicine clinic, demonstrates self-directed generation of clinical queries, search for best evidence relevant to patient care, appraisal of what was found and discussion of the applicable literature with the supervising attending in the clinic. Preparation workload responsibiltiy shared among participants decreases the amount of work for one person. KEY LESSONS LEARNED: Changing the format from a facilitator-centered journal club, to a learner-centered, multi-professional collaborative model provides an environment for community based internal medicine residents to learn skills of evidence based medicine, which facilitate information management with application of relevant evidence to patient care. This program, designed to give students early clinical experience, requires students spend at least one half day per month in a generalist's clinic during the first two years of medical school. Currently, there are 175 volunteer physician preceptors throughout Southern Wisconsin who devote countless hours and provide rich educational experiences for 288 first and second year students. We have attracted and retained a network of practicing community-based generalists to serve as preceptors using a variety of materials and methods. These include A brochure briefly outlining the advantages of becoming a partner in the program. A handbook detailing the purpose of the program and responsibilities of preceptors and students. A pocket guide that summarizes suggested clinic activities and tips on involving students in busy practices. Free internet access including e-mail and electronic clinical resources. Eligibility for volunteer clinical faculty appointment at the UW Medical School. Use of instructional and recreational facilities including all libraries.

Recognition of program participants in the UW Medical School Alumni Magazine. UW Medical School GPP lapel pin. University of Wisconsin bookstore gift certificate. FINDING TO DATE: Our materials have served to 1) enhance visibility and recognition of our program, 2) increase physician interest in our program, and 3) improve retention of our generalist preceptors. KEY LESSONS LEARNED: Recruiting and retaining a network of generalist physician volunteers is the key to the success of community based educational programs. A multitude of materials and rewards may help accomplish this goal. More studies are needed to determine the effectiveness of various materials and rewards. Pairs of housestaff answer three consecutive questions. These pairs rotate so that all team members are able to participate. The final question is similar to final jeopardy and is a team effort. FINDING TO DATE: Informal feedback has been quite positive. We are in the process of evaluating the educational value of this program as it relates to several important issues: use as a medical teaching tool, promotion of teamwork, interest in computers and the internet, as well as its entertainment value. Results of this evaluation will be available in early February. KEY LESSONS LEARNED: The I.M. Quizitor conference is a well attended conference at the Oregon Health Sciences University. Novel conference formats should be included in residency curricula as they increase resident's interst in attending conferences and likely enhance learning. STATEMENT OF PROBLEM: Professionalism in medicine is facing serious challenges in this era of managed care, increasing cost containment, and medical economic turmoil, from both practice and medical education standpoints. Many postgraduate medical training institutions have found the task of developing a curriculum in professionalism that is both useful and meaningful equally challenging. In our own attempt to do this, we piloted a curriculum designed to address some of the many issues surrounding the teaching, learning, and instillation of professionalism in residency training. OBJECTIVES OF PROGRAM/INTERVENTION: 1) Survey our interns about their beliefs and attitudes regarding professionalism in training; 2) Provide a realistic forum within which interns could define and identify different aspects of professionalism, share their experiences with professional and unprofessional behavior, and explore their responses to different clinical vignettes; and 3) Set the stage for collecting future data about the effectiveness of this approach. DESCTIPRION OF PROGRAM/INTERVENTION: A mandatory, 8-hour retreat was held off campus for our interns only. They were relieved of all clinical duties that day and overnight. Pre-and Posttest questionaires [modified 4-point Likert scale, Strongly Agree (SA) to Strongly Disagree (SD)] were administered at the beginning and end of the day, respectively. A brief didactic preceded small group breakouts which were then used to discuss 12 clinical vignettes, with summaries being presented to the large group, followed by a moderated group discussion. Lastly, the interns drafted a Code of Honor for their class, capturing the key elements of professionalism. FINDING TO DATE: 100% A/SA that: professionalism was equally important as medical knowledge or clinical judgment/skills; they could improve their degree of professionalism. 95% A/SA that: professionalism could be taught and learned; their colleague's professionalism could be improved; they have a heightened sense of the behaviors which allow for a greater degree of professionalism; the retreat was useful/valuable. 80 to 85% A/SA that: professionalism is not stressed enough in medical education; they have a better understanding of altruism, excellence, duty, honor and integrity; they have a better appreciation of how power, arrogance, and greed can erode professionalism. 65% A/SA that more than just the retreat was needed to fully grasp the concepts discussed. 80% SD/D that learning about professionalism in other settings would be more valuable. KEY LESSONS LEARNED: We felt this retreat was a success, and is a useful forum for our professionalism curriculum. Further retreats and other fora need to be developed, studied, and refined to improve this. Our residents (and faculty) enjoyed the retreat, and most are extremely willing to help plan and moderate future sessions. In response to the need for enhanced residency training in geriatrics, eight academic medical centers (including Baylor College of Medicine; Harvard University; Johns Hopkins University; Stanford University; University of California, Los Angeles; University of Chicago; University of Connecticut; and University of Rochester) and the American Association of Family Physicians developed educational programs and implemented them locally during a 3-year initiative, funded by the John A. Hartford Foundation. In 1998, the sites formed a consortium with the purpose of creating exportable resources that would assist residency programs to strengthen the quality of geriatric education at their institutions. The resources include curriculum manuals, packaged methods for teaching geriatric skills, stand alone teaching aids, faculty development programs, and consultation programs. They are disseminated through the non-profit, Stanford University Geriatric Education Resource Center. Between April 1998 and October 1999, the resources were disseminated to approximately 30% of internal medicine and family medicine residency programs. Evaluation of the resources by users has been very positive. Evaluation results are used to enhance the resources. Two key lessons have been learned: (1) a noncompetitive, collaborative venture involving diverse academic medical centers is possible, and (2) funding is critical to bringing locally developed and tested resources to a national audience. We developed a repeating curriculum for our interns during their ambulatory time. We have new interns in clinic every 2 weeks. Each intern spends two weeks between July and December and two weeks between January and June on ambulatory block. Each Thursday, the interns have didactic sessions. At noon, they have case-based conferences on telphone medicine and clinical time management taught by the general medicine faculty using materials developed by our curriculum committee. In the afternoon, they have conferences on primary care topics during the first half of the year and critical evaluation of the medical literature during the second half of the year. A further exciting aspect of this program is the use of our primary care senior residents as teachers. The residents initially observe a faculty talk, then develop talks for the interns. During their talks, they are either videotaped or observed, so that their faculty mentor can provide feedback on their teaching skills. KEY LESSONS LEARNED: 1. Improvement in intern confidence and independence in the most common problems and issues seen in our clinic. 2. Development of a standard primary care curriculum across the intern year. 3. Novel teaching opportunities for primary care senior residents in basic ambulatory topics with feedback from faculty and residents. Smoking Cessation, An Approach to the Management of Patients with Chronic Pain, Introduction to Alternative Therapies-Acupuncture, An Approach to Cross-Cultural Medicine, and Cost-Effectiveness in Ambulatory Medicine. FINDINGS TO DATE: Residents find the seminars a valuable experience. The methods of case-based interactive discussions, role-plays, and emulations (e.g., simulating physical therapy exercises), along with minimization of lecture-based presentations, are key factors to motivate resident interest and learning. KEY LESSONS LEARNED: Resident practice behaviors can change immediately as a result of this educational intervention. A major obstacle to continuing the program is providing adequate compensation to faculty presenters in the form of academic credit and financial support. 1) The Fellowship Forum: A dinner session with presentations by third year residents on how and where to apply for fellowship and discussion by faculty members from different subspecialties. 2) Interview Skills: An interactive presentation providing strategies on preparation for the time prior to, during, and after the interview. 3) Practice Choices: Which position is right for me?-A panel discussion of representatives from managed care, private, academic, and other types of practices. 4) Practice Management: A workshop on how to join or set up a practice. 5) Negotiating Your Contract: A seminar conducted by health care lawyers detailing specific issues concerning contracts and finding a lawyer. Materials developed for the program include the "Fellowship Guide for Dummies" and the Career Pathways Handbook "How to Get the Job that You Want." FINDING TO DATE: Written evaluations are conducted after each session. Resident satisfaction with the programs has been high. Revisions to programs are made based on resident feedback. The evaluations have revealed that 1) the guidebooks have been found to be useful resources, 2) the panel discussion and session on legal issues are the most informative programs, and 3) location of venue and timing influence attendance. KEY LESSONS LEARNED: Given faculty support, trainees can develop successful programs in areas where there is a perceived need. Both residents and faculty should share a role in assisting trainees in their career search. A "Career Pathways" Program with a series of workshops and seminars for trainees can lead to increased competence and comfort in making career decisions. attitude, communication, knowledge, organization, time management, physical examination, medical decision making, and interpersonal skills. A numerical score is given for each of the seven topic areas, and the area of greatest deficiency is addressed. Students with low attitudinal scores undergo personal/professional goal setting and interest assessments. Interpersonal skills are augmented using an adaptation of Carnegie and Covey's techniques of conflict resolution. Using interactive video, students are shown situations requiring interpersonal skills and are asked to make decisions at critical moments in each scenario. Written communication is augmented using an interactive computer program. Oral communication is strengthened with digital subtraction video, in which student presentations are digitally edited and re-arranged to emphasize prioritization of data. Organization of thought is improved by showing students segments of a patient interview, and asking them to state their hypothesis at different parts of the presentation. Deductive reasoning exercises help students re-formulate their thought process at each point in the history/physical. The student's method of recording patient data is reviewed. Johnson's One-Minute Manager techniques are combined with exercises in prioritization to improve time management. Medical decision making is improved with exercises that teach the use of pre-and post-test probabilities. A pre-test is used to assess the student's knowledge base. Faculty provide a list of reading/multi-media resources that are specific to the student's needs. The student's home library and reading strategy is reviewed. FINDING TO DATE: This diagnostic strategy has assisted eight students and two interns. The areas of greatest deficiency have been communication skills (3 students/1 intern), interpersonal skills (2 students/1 intern), organization/time management (1 student), attitude (1 student) and knowledge (1 student). Every student successfully completed the medical clerkship, and both interns are currently re-engaged in a successful training program. KEY LESSONS LEARNED: 1) Each struggling student is unique. 2) It is rare that clinical failure is due to global inadequacy. 3) A program tailored to the student's primary deficit can improve student progress. 4) Innovative technology can assist in timely correction of difficult problems. STATEMENT OF PROBLEM: Housestaff play a significant role in student education, but receive little formal education in teaching methods. In order to be effective educators, residents need practical teaching strategies and practice in developing these strategies. OBJECTIVES OF PROGRAM/INTERVENTION: 1) Provide specific teaching strategies for three teaching venues: board teaching, bedside teaching, and teaching medical decision-making. 2) Illustrate teaching principles through exercises in which residents actively teach students. 3) Objectively evaluate this teaching curriculum with analysis of pre-and post-curriculum videotape of each resident's teaching. DESCRIPTION OF PROGRAM/INTERVENTION: Before participating in the curriculum, each resident is videotaped teaching students. They are then given a packet of selected readings and asked to screen five selected movies that illustrate important principles of education. The teaching curriculum consists of 4 two-hour sessions in which residents learn education principles through active teaching of medical students. Didactic discussion is used only to consolidate lessons learned in the resident's teaching activities. Session 1: Residents learn how to inspire learners, promote independent learning and teach methods of problem solving. Session 2: Residents are videotaped while teaching students a 15-minute board presentation. Each resident's teaching is analyzed by the group and board teaching strategies are discussed. Session 3: Each resident teaches the group a clinical topic at the patient's bedside. Through the bedside teaching sessions, residents learn to engage and educate the patient, to teach the physical examination while respecting the patient, and to involve the patient in clinical decision making. Session 4: Each resident guides a student through the decision-making process of a complicated case. The group critiques each resident's performance, and additional methods of teaching medical thinking are reviewed. At the conclusion of the curriculum, residents are videotaped teaching students. The pre-and post-curriculum videotapes are evaluated by a blinded, independent panel of observers. FINDING TO DATE: 1) Resident teaching ability is improved by observed practice with immediate feedback. 2) Instructing residents how to teach at the bedside increases the frequency and quality of bedside instruction.

3) The prevalence of resident teaching increases as self-perceived teaching ability improves. KEY LESSONS LEARNED: 1) Videotape is an effective way to review and critique resident teaching. 2) Venue-specific teaching strategies enable residents to use generic teaching principles in the context of their ward responsibilities. 3) Involving residents in active teaching exercises improves interest in the curriculum and refines teaching skills. STATEMENT OF PROBLEM: Ensuring that residents achieve the educational goals of a rotation is ideal, but difficult. It can be especially difficult in an ambulatory setting if there are multiple faculty and multiple residents with varied schedules. We faced such a situation in our general medicine and women's health portions of our two-month ambulatory block rotation. OBJECTIVES OF PROGRAM/INTERVENTION: Our two main objectives were: (1) To standardize the content of the main educational topics in the rotation. (2) To develop a system to track resident's exposure to each educational topic in the curriculum for general medicine and women's health in the ambulatory block. DESCRIPTION OF PROGRAM/INTERVENTION: We first developed a list of the core topics in the medicine and women's health curricula that we felt were most important to cover during this rotation. A set of reference articles and teaching materials were compiled for each topic. We then developed a combination orientation, evaluation and educational goals checklist. On a single sheet of paper for each resident we list the core topics with check boxes. Also included are empty lines for personal educational goals, mid and end-of-term evaluation feedback and future goals. On the same page is a checklist with all the administrative and orientation requirements of the rotation. This sheet is reviewed with each resident on the first day of the rotation, and used again at mid and end-of-term evaluations. Two other sheets are used as well. One lists the women's health and one lists the medicine core topics. Each list is in table form with each resident's name heading a column. These are taped to the wall of the charting rooms. As a faculty member reviews a topic he or she initials it under each resident present. This serves as a visual reminder to the faculty to make sure that each resident receives teaching in every topic. FINDING TO DATE: Since we began using this system we have noted several improvements. The orientation to the clinic is definitely more complete for each resident. The residents are more likely to receive mid and end-of-term evaluations. Use of the check sheets has allowed us to focus residents to certain types of patients and experiences so that their personal educational goals can be met. We feel more confident that each resident receives a more complete education and a more satisfying rotation. KEY LESSONS LEARNED: Simple check lists in an ambulatory setting can help ensure that rotational goals are met.

A PRIMARY CARE APPROACH TO TEACHING WOMEN'S HEALTH. LA Williams, Internal Medicine, St. Joseph Mercy Hospital, Ann Arbor, MI STATEMENT OF PROBLEM: The residents graduating from our Internal Medicine residency program did not feel prepared to handle gynecological and women's health problems with confidence. Our program formerly used two private ob/gyn practices for their gynecology eperience, but the residents were not given adequate opportunity to perform exams and were in more of an observer role in that setting. Our other out patient teaching sites did not serve populations who presented commonly with problems that represented key aspects of the core curriculum in women's health recommended by the ABIM. Examples of lacking elements included sexually transmitted diseases, abnormal uterine bleeding, and contraceptive methods among others. OBJECTIVES OF PROGRAM/INTERVENTION: The objectives we had were threefold: (1) To teach residents the principles of women's health through a curriculum based upon the ABIM's core curriculum in women's health. (2) To teach in a setting that would provide the type of patients needed, and that would allow for residents to have hands-on experience in patient interviews, exams and procedures.

(3) To use internists and other primary care providers in lieu of specialists to teach or residents women's health. DESCRIPTION OF PROGRAM/INTERVENTION: Because we could not achieve our goals effectively within our current system of out patient teaching, we looked outside of it. In 1998 a collaboration was formed with a local university health center. The center is made up of a separate women's clinic in addition to a general medicine clinic. That spring, during a trial period, one faculty and one to two residents each month saw patients in the women's clinic two half days a week. By July of that year both the women's clinic and the general medicine clinic were incorporated into our two month ambulatory block rotation taken by seniors FINDING TO DATE: The response by the residents has been incredibly positive. The residents are the primary providers performing all initial interviews, exams and procedures. The population is very different from any of our other out patient sites. They are young, and generally healthy with no or few chronic medical problems. The residents get plenty of exposure to most aspects of the women's health curriculum. The pace is slow enough that a lot of case-based and didactic teaching can be done. The residents have voiced a much greater degree of confidence in practicing women's health independently. The general medicine clinic has also proven to be a unique experience as well in that it provides exposure to acute problems (such as URI's and injuries) that are not common in their continuity clinics. KEY LESSONS LEARNED: Instead of struggling to change our current out patient structure to meet the educational needs of our residents we looked out side of it and formed a collaboration that not only met our original needs, but has become a resource for multiple educational programs. STATEMENT OF PROBLEM: The importance of the teaching role of residents is undisputed. However, much of the teaching that third and fourth year medical students receive is provided by residents who have had little or no instruction on how to teach. OBJECTIVES OF PROGRAM/INTERVENTION: We described and cataloged the characteristics of the questions and statements that residents used in teaching medical students both before and after a course designed to develop their teaching skills. From this we characterized the teaching strategies that residents used and the influence that the course had on these strategies. We then looked at the strategies for common themes that could help us understand the conceptual models in use when residents teach, and if the course influenced these models. DESCRIPTION OF PROGRAM/INTERVENTION: Thirty five first year residents elected to attend a longitudinal teaching skills development course. Before and again after the course, most of these residents were videotaped during three teaching interactions with standardized medical students. A sample of these tapes were reviewed. The characteristics of each question or statement made by a resident were identified and coded into categories using a coding system that has been previously developed and validated. The codes were tabulated and converted into frequencies. The independent samples T-test was used to measure the difference in the frequency with which each code appeared before and after the course. Codes which changed significantly in frequency were grouped and common themes in these groups were sought. Independent coding of notes written by the residents after each videotaping will be used to triangulate these themes. FINDING TO DATE: The preliminary review showed that first year residents frequently asked factual and convergent questions, gave direction on case presentation, and tended to analyze the case for the student. After attending a teaching skills development course, the same residents asked more process-focused questions, gave fewer directions on case presentation, and more frequently encouraged students to analyze the case themselves. Comparison with the coding of notes written by residents after the videotaping is pending. Coding of additional videotapes is underway. KEY LESSONS LEARNED: A teaching skills development course appears to influence the teaching strategies that residents use from content-focused and teachercentered toward process-focused and student-centered. Further study is underway to better understand these changes. is an approach to primary care that integrates clinical medicine, epidemiology, social sciences, and health service research in a complementary fashion to develop programs to meet the health needs of a community. A COPC curriculum development team was formed at our institution, and based on targeted needs assessments from a piloted curriculum, a detailed curriculum was developed. The goal of this curriculum is to cultivate in general internal medicine residents the knowledge, attitudes and skills necessary to incorporate a COPC approach in their future practices. OBJECTIVES OF PROGRAM/INTERVENTION: Specific learner objectives include:

(1) Knowledge-Identify populations with which residents work; describe methods to assess the health needs of a community; describe a method to evaluate the impact of a COPC intervention. (2) Attitude-Identify own role as a primary care physician who addresses needs of community as well as individuals; rate as valuable the involvement of non-physicians and community representatives in COPC projects. (3) Skills-Define a community and its characteristics using databases; use results from community health assessments to identify and prioritize health problems; work in a group setting to design an intervention based on identified community health needs. Program level goals include incorporation of COPC in resident's future practices and community leader recognition of resident participation in community health issues. DESCRIPTION OF PROGRAM/INTERVENTION: A longitudinal approach integrating didactic instruction with hands-on involvement in COPC projects was developed. The designed COPC curriculum is a two-year longitudinal course that consists of 8 sequential modules. Each module consists of one didactic and facilitated small group session and one flexibly scheduled project work session, incorporating selected readings and specific project assignments. Modules are led by GIM faculty, but include speakers from the community and other disciplines. FINDING TO DATE: A pilot COPC curriculum was implemented in academic year 1998-1999. Residents undertook diverse projects such as preventing tobacco use through middle school student education; supporting caregivers of the homebound elderly; and assisting medical clinic patients having difficulty obtaining prescription medications. Preliminary evaluation results show that residents have great enthusiasm for COPC, citing rekindled interest in volunteerism important to many of them prior to residency. Retained knowledge of the principles and practices of COPC was high. KEY LESSONS LEARNED: (1) Residents value structured opportunities to learn about and participate in interventions in community health.

(2) A longitudinal, project-based learning approach appears to be a successful model for incorporating COPC training in a medical residency curriculum. STATEMENT OF PROBLEM: By 2020, one-third of North American women will be menopausal. The complex health needs of these women are poorly met by our current health care system. OBJECTIVES OF PROGRAM/INTERVENTION: To improve the quality of care provided to menopausal women by developing a personalized decision support aid (DSA). DESCRIPTION OF PROGRAM/INTERVENTION: We developed a DSA for menopausal women and their clinicians that consists of computer-generated personalized health reports. Women complete a brief written questionnaire about their lifestyle, family, and medical history. This information is entered into a software program that generates a comprehensive personalized report for the patient (35-45 pages) and a 2-page summary report for her clinician. Both reports include information about the woman's menopausal symptoms, her personal risks for common chronic diseases and the impact of hormone replacement therapy on these risks (based on published risk models), her utilization of recommended screening modalities, and general strategies for risk reduction and treatment. The DSA also screens women for depression, alcohol abuse, sexual dysfunction, and domestic violence, providing relevant information about diagnosis and treatment to clinicians and patients when needed. 10 clinicians and 13 patients have evaluated the DSA to date. FINDINGS TO DATE: Most physicians rated the DSA very favorably (60% excellent, 30% good) and 70% would like to use it with their menopausal patients. None felt that it was scientifically inaccurate, and 90% reported that all of the topics they normally discuss with patients were included in the DSA. All felt that it increased their knowledge about their patient and their patient's knowledge about menopause. Most felt that it would improve their ability to counsel women about menopause, the efficiency of the consultation, and the quality of care. Half felt it would improve their interaction with the patient. Most patients felt the DSA would improve communication with their clinician (69%), help them identify questions to ask their clinician (77%), and improve the quality of care they receive (54%). KEY LESSONS LEARNED: The menopausal DSA appears to be acceptable to clinicians and patients, may facilitate counseling, and may improve the quality of care provided to menopausal women. Knowledge about pt 100 0 0 OBJECTIVES OF PROGRAM/INTEREVNTION: Pharmacy industry sponsored patient assistance programs are increasingly available and hold the potential to offset drug costs for indigent patients. However, many providers are hesitant to use these programs due to perceived barriers. DESCRIPTION OF PROGRAM/INTEREVNTION: In order to evaluate the perceived usefulness and ease, we surveyed all providers at our urban community health center through a questionnaire given between June and August of 1999 (response rate 60%, 31 ⁄ 51 ). FINDING TO DATE: The questionnaire was used to help develop a systematic approach to enroll patients into the patient assistance programs. Data from the provider yielded information on the proportion of providers using the programs, the number of times the provider applied to a program, and the success rate for actually receiving medications. Data from the application yielded information on the method of enrollment, barriers specific to that medication or program, the amount of time it took to receive that medication and the overall satisfaction with the patient assistance program. KEY LESSONS LEARNED: Fifty-eight percent (18/31) of responders had applied for medications through a patient assistance program at least once, and 55.5% (10/18) had applied four or more times. These applications represented 19 different medications. With most applications (14/39), providers became aware of the patient assistance programs through the clinic pharmacist. Written applications or letters to the pharmaceutical company were used to enroll patients most often (versus phone contact or sales representative). The most common barriers were receiving the medications and reapplying to the programs compared to the initial contact with the company. Medication was usually received in 1 to 6 weeks. The majority of providers (90%, 28 ⁄ 31 ) felt that medications should be obtainable in Ͻ4 weeks and the best method of initial contact with the pharmaceutical company was through phone contact from the provider or provider representative. The providers (96%, 27 ⁄ 28 ) felt the best reminder to reorder medication was a notice from the pharmaceutical company or pharmacist rather than a notice to the patient. Eighty-two percent (29/35) of providers were satisfied or very satisfied with the programs. STATEMENT OF PROBLEM: The patient assistance programs were perceived as useful and fairly easy to use in this setting. We have designed multiple systems allowing providers easy assess to program applications and follow up, which has allowed us to overcome barriers perceived by the providers. Our systematic approach towards enrollment resulted in providers using these programs more often. The response from the pharmaceutical companies towards the increased use of these programs may be a factor in how available these programs are in the future. Program was started in 1994 as a partnership between medical school faculty and a regional integrated health system. Project funding is limited to $25,000; proposals are solicited from all employees and medical staff. We use five simple guidelines to identify desirable projects: (1) clear-cut application to an important health care problem; (2) evidence that the innovation is likely to improve the value of healthcare; (3) high probability that benefits can be demonstrated in 18 months; (4) potential for applicability to multiple sites within the health system; and (5) rigorous evaluation of innovations (most funded studies have been randomized clinical trials). FINDING TO DATE: We have funded 26 projects. The fifteen projects completed to date have demonstrated clear improvements in diverse outcomes including patient satisfaction, length of mechanical ventilation, and length of stay. Annual cost savings from completed projects (total funding of $236,795) is estimated at $2.2-3.1 million. Examples of projects which have demonstrated cost-beneficial clinical improvements include: the elimination of daily suction catheter changes in patients on mechanical ventilation; institution of protocol-driven weaning to reduce time on mechanical ventilation; substitution of tincture of iodine for povidone-iodine as a skin preparation to prevent contamination of blood cultures with skin flora; the substitution of normovolemic hemodilution for autologous blood donation in patients undergoing prostatectomy; and early ambulation for patients with community aquired pneumonia. KEY LESSONS LEARNED: Most project proposals have come from within the academic medical center, while efforts to encourage research projects at community hospitals have met limited success. System-wide implementation of successful projects is difficult to achieve in a heterogeneous health care system. Despite these challenges, the Innovations in Health Care Program appears to be a viable model for encouraging changes in health care delivery.

PREDICTING MISSED OR CANCELLED APPOINTMENTS. RI Gianani, TD MacKenzie, General Internal Medicine, University of Colorado Health Sciences Center; Community Health Services, Denver Health Authority, Denver, CO STATEMENT OF PROBLEM: The overall Westside Adult Clinic missed appointment rate is over 30%, which has resulted in long waits for new and follow-up appointments and uneven workloads for providers. OBJECTIVES OF PROGRAM/INTERVENTION: To identify factors predicting missed or cancelled appointments in an urban adult medicine clinic. DESCRIPTION OF OBJECTIVES/INTERVENTION: We analyzed 25122 clinic appointments made to the Internal Medicine clinic at Westside Family Health Center in Denver, Colorado during a 15 month period. The patient specific data include age, gender, race, language preference, primary care provider, and income. Provider-specific factors include training level and Spanish language ability. Visitspecific factors include prospective provider for the scheduled visit, whether or not prospective provider was the PCP, day of the week, time from appointment scheduling to visit, and type of visit. FINDING TO DATE: Factors strongly associated with missing or cancelling an appointment are shown below.

Factors that were strongly associated with missing or cancelling an appointment in the univariate analysis remain so using multiple logistic regression. Factors which were not strongly associated with missing or cancelling an appointment include language preference, language concordance between provider and patient, and income. KEY LESSONS LEARNED: Patient-, provider-, and visit-specific factors strongly influence the likelihood of missing or cancelling an appointment.

Physicians practicing at a General Internal Medicine (GIM) clinic at a University teaching hospital were at risk of being terminated from a health plan due to consistently higher expenses than revenue. This project was undertaken to determine if the University GIM clinic had higher expenses than the other primary care clinics due to adverse selection. Claims data for 1998 was obtained from 1 health plan. A case mix analysis was performed which utilized the first visit of the year for each patient. The top 30% of diagnoses were identified for each clinic and 2 physicians rated the diagnoses on a scale of 1 to 5 based on typical resource utilization expected over 1 year. The physicians chosen to rate the diagnoses included the Medical Director of the health plan mentioned as well as a practicing physician who is also the Utilization Management Director for the University. The number of visits per diagnosis were then multiplied by the average score for that diagnosis and an average was calculated for the clinic. Cohen's Kappa (Cohen, 1960) was used to assess physician agreement in rating the 21 diagnoses regarding resource utilization with an a priori Kappa of .40 set as acceptable based on Fleiss's (1981) criteria. The diagnoses were divided into tertiles of low, medium, and high resource utilization. Chisquare analyses with Bonferroni-adjusted planned comparisons were used to determine if there were significant differences among the clinics and their respective panel of patients based on resource utilization, and if so, whether the GIM clinic had a higher resource utilization patient panel than its 5 counterpart clinics. Case mix analysis demonstrated the GIM clinic to have the highest mean score indicating this clinic has patients with higher resource utilization. There was good agreement among the physician raters for the ratings of diagnoses for expected resource utilization (Kappa ϭ .48). There was a significant difference among the six clinics regarding resource utilization expected in their panel of patients (Chi-Square ϭ 112.58, p Ͻ .001). When compared to its counterparts, the GIM clinic was also seeing a different panel of patients (Chi-Square ϭ 57.69, p Ͻ .001), specifically more patients with higher resource utilization and fewer with moderate or low resource utilization (Chi-Square ϭ 40.31, p Ͻ .001; Chi-Square ϭ 40.93, p Ͻ .001, respectively) compared to the other 5 clinics. KEY LESSONS LEARNED: According to the case mix analysis, the University GIM clinic has patients who would be expected to have higher resource use than the patients in the other 5 clinics. These data were accepted by health plan administrators as an explanation for the higher utilization experienced by the GIM clinic. This methodology could be used in other teaching clinics where health plans do not have case mix and severity adjustment tools available to evaluate for adverse selection. maximixe profit (or minimize loss) for the MCO by gaining access to corporate rebate programs, often to the detriment of provider units that, increasingly, bear full risk for pharmacy costs. As such, MCO formularies lack credibility among the provider community. OBJECTIVES OF PROGARM/INTERVENTION: To develop and implement a single-source, physician-generated unified formulary founded on the complementary principles of best efficacy and value to patients at least cost to provider groups.

Committee of the Provider Services Network of Caregroup includes practicing network primary care clinicians and clinical pharmacists. This 22-member core committee, with input from 10 specialty subcommittees, has developed and implemented a unified formulary. Specific medication classes are reviewed on a monthly basis. The unified formulary has been distributed to the 2,800 physicians in this large, multi-hospital, multi-practice network in a number of formats:

(1) multi-fold pocket cards (2) wall charts (3) binders (4) electronic versions on hospital intranets and the internet, and (5) downloadable electronic files for palm computing devices. Average wholesale prices of individual agents are featured prominently. The committee has counselled network physicians to use the unified formulary exclusively and to ignore formularies from individual MCOs. Pharmacy compliance is assessed against the PSN formulary alone. FINDING TO DATE: Compliance with formulary recommendations has increased within every therapeutic class addressed in the document, yielding first year savings in excess of $1.4 million. In certain drug categories, market share of PSN "preferred" drug products has risen by 30-40%. The unified formulary has become the cornerstone of other PSN pharmacy efforts within individual provider units, such as therapeutic interchange initiatives and educational efforts. KEY LESSONS LEARNED: This physician-driven unified formulary has had tremendous support among network provider groups and enjoys substantial credibility as the work product of prescribing clinician's colleagues. It is a user friendly, single-source, easily accessible reference, and has become a widely used teaching tool, and the foundation of pharmacy interventions at both the network and local provider unit level. STATEMENT OF PROBLEM: There is no experience in academic health center (AHC) development including information management planning (IMP) from the very first beginning of its operation in Argentina. It is important to mention that the market does not offer any support for health care information management planning. OBJECTIVES OF PROGRAM/INTERVENTION: We conducted since 1998 an IMP process and designed the managerial and clinical information right into the "genetic code" of a new emerging AHC. The objectives were: a) to put the information management at the core of the future university hospital (UH)and integrated delivery system (IDS), b) to include paperless clinical information system along with administrative information management, through an electronic medical record (EMR)as a commanding system, c) develop the basis of a database for research, education, and health care provision and management. DESCRIPTION OF PROGRAM/INTERVENTION: A program leaded by the medical and managerial groups was designed to: 1) describe an IDS process oriented planning as a template for the requirements to software and hardware potential providersof the hospital information management systems (HIS), 2) select a system available closest to the stated objectives stated, 3) use that providers system to develop the remaining objectives (such as the EMR, among others), 4) introduce system wide diagnostic and procedure coding, and 5) implementation of the previously described objectives. FINDING TO DATE: 1)No similar experiences were found in Argentina up to date, 2) no EMR nor computerized medical record (CPR) were found in the market, 4) the best proposal of system provider met 60% of the requirements, 4) we worked with the provider in the development of the 40% of the remaining objectives not met and the EMR, 5) in order to use CPT coding and to be able to bill we had to develop an conversion table to the national practice coding system. KEY LESSONS LEARNED: 1) Introducing IMP in the "genetic code" of an emerging AHC is in fact a significant advantage, 2) substantial effort in the planning phase is required, 3) there is organizational, medical and market inmaturity in Argentina for the scope of the project, 4) organizational stress to adaptation of providers and physicians is widespread, 5) to upgrade a HIS into a clinical information system through an EMR was required, 6) there is a lack in procedure coding systems, 7) there is marketing and contractual risk of the coding decision performed, 8) difficult recruiting of trained human resources, 9) costs of the project. The Bellevue Hospital Medical Service, 110 beds with 40,000 annual inpatient days, had a part-time faculty from the NYU School of Medicine (NYU) to supervise clinical care provided by house staff. Increasing oversight requirements and pro-longed lengths of stay could not be addressed by the house staff-dominated system with only two hour daily rounds with attendings. 1) To create a full-time hospitalist attending program to supervise clinical care without disenfranchising the voluntary NYU Faculty nor ending its contribution to clinical care and teaching; 2) To reorganize the medical wards to: improve house staff oversight and attending accountability; meet Residency Review Commission requirements for distinct clinical and management attending rounds; improve discharge planning and outpatient followup; and decrease length of stay. Since July 1, 1999, one hospitalist and two teaching attendings (voluntary) serve with two house staff teams per ward. The hospitalist oversees all patient care as the attending of record. Teaching attendings, with the hospitalist, conduct daily house staff teaching rounds. 16 hospitalists, chosen from the Primary Care Clinic for clinical and teaching skills, serve one month in every four. Five Physician Assistants (PA) were hired to follow hospitalist's clinic panels while on wards. PAs attend daily ward discharge planning rounds to facilitate outpatient followup. Hospitalists on wards still precept house staff continuity clinics. 1) Initial responses by attendings and house staff are positive as per multidisciplinary committees on Hospitalist and Residency Programs. End of year house staff and attending surveys will be compared to baseline data. 2) Surveys on all four wards two months in a row documented no house staff sense of decreased autonomy. 3) All hospital accreditation requirements were met. 4) Both Primary Care Clinic visits and medicine inpatient admissions increased 10% in July- December, 1999 , compared to July-December, 1998 . 5) Initial data indicates a 5% decrease in length of stay. 6) The program was funded at an incremental cost of less than $400,000, including four PAs hired with a New York State Community Health Partnership grant. Hospitalist positions were created from existing primary care physician lines, with an 8% salary increase to compensate for increased duties. KEY LESSONS LEARNED: A universally implemented, full-time hospitalist attending program, utilizing primary care physicians, can increase supervision of a house staff-centered medicine service without: decreasing house staff sense of autonomy; disenfranchising prior ward attendings; decreasing inpatient or outpatient productivity; or incurring great financial costs. Additionally, continuity of care to an outpatient setting can be facilitated. STATEMENT OF PROBLEM: Beta blockers have been shown to reduce morbidity and mortality when administered to high risk patients undergoing major non-cardiac surgery. Clinical practice guidelines (CPGs) are often used to enhance the adoption of effective treatment strategies but before implementing any CPG it is important to estimate its potential clinical and financial impact. OBJECTIVES OF PROGRAM/INTERVENTION: We sought to determine how frequently peri-operative beta blockers were being used at our institution and to estimate the potential clinical and financial impact of a CPG designed to increase their use. Desrciption of Program/Intervention: We developed a CPG for identifying high risk patients undergoing major non-cardiac surgery and for facilitating and standardizing beta blocker usage in the peri-operative period. We conducted a retrospective analysis by chart review of patients undergoing major non-cardiac surgery at our institution during a 1 month period in 1999. FINDING TO DATE: 1132 consecutive surgical patients were evaluated, 212 met inclusion criteria, and the charts of 158 patients were available for review. "Ideal candidates" were predefined as having established CAD or 2 or more major CAD risk factors and no initial contraindications to beta blocker use. 67 ⁄ 158 (42%) patients appeared to be "ideal candidates" for treatment with peri-operative beta blockers, yet of these only 31 ⁄ 67 (46%) received a beta blocker at any time during the course of their hospitalization and nearly all of these patients had been taking beta blockers chronically. Annual costs to the institution associated with additional drug and monitored bed use were estimated at $32261 or $60 per patient.

Using the results of a large RCT to estimate the clinical benefit at our institution, an additional 60 lives might be saved each year at a cost of $424 per life-year saved if all ideal candidates received beta blocker therapy. KEY LESSONS LEARNED: A large opportunity exists to improve the quality of care at our institution at relatively modest cost. A clinical practice guideline may be one method to achieve this goal.

QUALITY IMPROVEMENT FOR ADVANCED HEART AND LUNG DISEASE. J Lynn, MW Schall, KM Nolan, A Kabcenell, CC Milne, Center to Improve Care of the Dying, The George Washington University, Washington, DC; The Institute for Healthcare Improvement, Boston, MA; Associates in Process Improvement, Silver Spring, MD STATEMENT OF PROBLEM: Only a few generations past, most people died suddenly. Now, most will die after months or years of living with slowly progressive chronic illness. Care routinely mismatches needs, perhaps especially with advanced heart (CHF) and lung disease (COPD). Discontinuity and prognostic uncertainty make for aggressive medical treatment with inadequate advance planning, symptom management, and reliability.

B & C screening, and, if deemed necessary by the physician, a prescription for sterile syringes. Follow-up appointments and referrals for substance abuse treatment, mental health and social services are scheduled as needed. FINDINGS TO DATE: To date, we have enrolled 87 participants, and more data will be presented at the meeting. Of the first 35 participants, 57% are male, 74% are white, 57% homeless, and the average age is 41. On average, participants inject 4.4 times a day. The drug of choice is primarily heroin (83%). 57% have shared a syringe in the previous 30 days and, on average, participants reuse each syringe 16 times. Although 69% of participants use the Rhode Island syringe exchange program, of these, 63% were using it on a less than weekly basis. 51% of participants do not have a regular doctor or medical insurance. 86% have been in drug treatment. 17% reported being HIV positive and of the HIV negative participants, 54% believe that they are at risk for contracting HIV. KEY LESSONS LEARNED: This is the first study to explore syringe prescription as a strategy to reduce the transmission of HIV and other blood borne pathogens in IDUs. Thus far, physician-patient interactions have gone well, illustrating that this is a feasible practice. In addition to providing legal access to sterile syringes, syringe prescription can increase access to medical care and substance abuse treatment services. Other physicians are encouraged to consider the prescription of syringes to IDUs with inadequate access to sterile syringes. Patients perceive the quality of care to be high in academic-based practices, but they also report poor access to visits with providers of their choice. Traditional scheduling systems often defer current patient needs far into the future. With fully booked schedules, physicians are often unavailable to provide service to their own patients when they most need care for acute issues. Poor access may also drive patients to suboptimal care sites, such as urgent care centers and emergency rooms, and engender extra visits. Attraction and retention of new patients is correlated with how well a practice provides access. OBJECTIVES OF PROGARM/INTERVENTION: An "open access" system is predicated on the principle that a patient is "able to seek and receive care from a provider of choice at the time the patient chooses". Operationally, this implies that a practice must match, quantitatively, the demand for service with the capacity of each provider. Many non-academic practices have successfully implemented such systems of "doing today's work today". We have embarked on an ambitious plan to implement open access in a large academic-based primary care practice. DESCRIPTION OF PROGARM/INTERVENTION: Because many of our primary care faculty do not practice daily, we have modified the basic open access model:

(1) establishing mini-care teams comprised of a faculty member, 4-5 comanaging residents and 1-3 nurse practitioners; and (2) expanding the time response to support the concept of "doing this week's work this week". Each team is responsible for servicing all its patient's needs within 7 days, with an expectation that most will be met on the day requested. FINDING TO DATE: We have captured data establishing functional panel size and phone demand for visits with each faculty member. Those faculty who are implementing the new model have worked down backlogs of scheduled patients, through a number of interventions, such that the proportion of "open" schedule slots in the first and second months has risen from 0 and 6% to 34% and 58%, respectively. Eight to ten faculty members and their associated scheduling teams anticipate full implementation of open access by 5/00. KEY LESSONS LEARNED: Implementation of open access requires continued monitoring of demand for visits, simplified schedule templates with a shift away from categorizing visits as "urgent" versus "routine", and aggressive management of available visit slots. Most importantly, open access necessitates a commitment to servicing the needs of patients as they arise. Change comes hard to physicians anxious about nominally unfilled schedules. Those patients and physicians who have already experienced improved access report improved satisfaction with no adverse effect on productivity. STATEMENT OF PROBLEMAlthough previous studies document poor physician compliance with practice guidelines for cholesterol management, the reasons that providers fail to achieve lipid goals are not well understood. OBJECTIVES OF PROGRAM/INTERVENTION: We evaluated the relative importance of three potential areas of inadequacy facing the primary care provider: i) failure to screen, ii) failure to initiate therapy, and iii) failure to titrate therapy to optimal dose (lipid goal not achieved and patient receiving less than half of maximal statin dose). DESCRIPTION OF PROGRAMS/INTERVENTION: Medical records from selected patients receiving primary care from 24 general internists and 5 nurse practitioners at an urban Veterans Affairs Medical Center were audited. Approximately 150 patients with at least one previous visit with the provider were selected randomly from each provider's panel over a nine month period beginning 1 ⁄ 99 . The lipid goals were derived from National Cholesterol Education Program guidelines and defined as follows: CHD present-LDL Ͻ115, HDL Ͼ30, TG Ͻ250 mg/dl; CHD not present-LDL Ͻ160, TG Ͻ400.

FINDING TO DATE: During the nine months, 4757 patient records were reviewed from 29 providers. Among all records audited, 82% (provider range 64-97%) showed evidence of timely lipid screening (yearly for patients with cardiovascular disease, otherwise every three years). Among those screened, 52% had abnormal lipids requiring drug therapy. Among these patients with a lipid disorder requiring drug therapy, 46% (provider range 25-70%) had not achieved their lipid goal. Among these patients, 42% (provider range 17-73%) did not receive lipid lowering therapy, and 40% (provider range 14-66%) received suboptimal doses of therapy. Therefore, optimal therapy could potentially have improved lipid responses in 82% of patients. Only 18% of patients who did not achieve lipid goals were receiving optimal therapy. KEY LESSONS LEARNED: We conclude that failure of providers to initiate and optimize lipid lowering therapy were important contributors to most treatment failures. Teaching programs and/or reminder systems designed to overcome these deficiencies are likely to significantly improve provider performance in hypercholesterolemia management.

Breudigam, CC Williams, E Pomiecko, Regional Medical Practice, The Cleveland Clinic, Strongsville, OH; Regional Health Affairs, The Cleveland Clinic, Cleveland, OH STATEMENT OF PROBLEM: Cleveland Clinic Strongsville (CCS) is a satellite office of the Cleveland Clinic Foundation (CCF) that provides primary and surgical care. Eight primary care physicians (PCP) practice at this facility. Three primary care RNs answer phone calls that deal with prescription refills, referral authorizations, labs/x-ray notifications, and triage. Because of an increasing volume of phone calls and complexity of triage, patient complaints regarding the office's phone process are also increasing. OBJECTIVES OF PROGRAM/INTERVENTION: CCS instituted an automated phone menu system on November 1, 1999. Its purpose is to improve the efficiency in managing patient phone calls by: 1.) reducing the length of time a patient was put on hold (hold time), 2.) decreasing the number of patients who hang up while waiting (abandonment rate), and 3.) facilitating patient access to same-day (urgent slots) or next-day appointments. DESCRIPTION OF PROGRAM/INTERVENTION: The automated phone system allows patients to choose from the following selections: prescription refills; referral authorization; lab/x-ray results; non-urgent messages; or remain on hold if experiencing an urgent medical problem. Remaining on hold automatically connected the patients to more experienced triage nurses at CCF's Nurse On Call (NOC), a major component of CCF's call center. NOC has access to each PCP's schedule and can add patients to their urgent slots (typically 4/day per PCP). The three primary care RNs continue to handle all other calls. FINDING TO DATE: Total calls to the primary care RNs decreased by approximately 50% (see Table below ). Consequently, the average hold time and abandonment rate decreased, and continue to improve into the second month of the project. NOC triages all other phone calls and has been successful in facilitating same-or next-day appointments.

KEY LESSONS LEARNED: Implementing an automated phone system menu and using existing resources (NOC) improved the office's efficiency in managing patient calls dealing with prescription refills, referral authorizations, lab/x-ray notifications, and triage. tients with diabetes and Group 3 had the EMR template and an updated patient education card to give out if desired. The primary care providers were educated on how to update this template themselves. Documented compliance with five evidence-based HPIs (BP, Hb1Ac and foot exams every 4 months; annual dilated fundoscopy and albuminuria screening) were used as the outcomes. These were measured at baseline, six months and 1 year. The main outcome for analysis was a composite of the five HPIs.

FINDING TO DATE: In October 1998, 304 consecutive patients of participating PCPs were randomly assigned to each group. PCPs were stratified by location within the building. There were 7 PCPs and 92 patients in Group 1, 7 PCPs and 118 patients in Group 2, and 7 PCPs and 94 patients in Group 3. One year follow up was over 99%. At baseline, Group 1 patients had significantly lower HPI adherence than Group 2 or 3 (38% vs 54% and 57%, respectively, see Table) . At 6 months, only Group 3 patients had significantly better composite HPI than baseline, but this improvement was lost at one year. In fact, at study conclusion the only significant difference occured in Group 1, possibly reflecting a Hawthorne effect or regression to the mean. KEY LESSONS LEARNED: Although the combination of an electronic reminder template and a patient education card might transiently increase adherence to HPI guidelines in patients with diabetes, this improvement is not sustainable at one year. Ongoing EMR template updating and/or feedback might be necessary to achieve this. is an academic, closed panel, staff model managed care organization that delivers primary care to 19,000 adults. After 28 years of operation using processes that had once served it well, YHP was no longer meeting the needs of its patients for timely access to their clinician of choice by appointment or telephone. OBJECTIVES OF PROGRAM/INTERVENTION: Our goals were: 1) to improve significantly access to appointments of all types by creating a simplified and flexible scheduling system, 2) to increase face-to-face time with the clinician by maximizing efficiency of patient flow, 3) to improve the ability of patients to see their own clinician, and 4) to create a system of care which promotes population-based patient management. DESCRIPTION OF PROGRAM/INTERVENTION: Based on a planning process involving every member of the department working in multidisciplinary teams supported by senior administration and an external consulting firm, the organization of care delivery was redesigned to incorporate the following key changes: 1) An open access appointment system was developed with 50% of appointments held for same or next-day availability. 2) All patients were linked with a primary clinician, with panels benchmarked and balanced. 3) Multidisciplinary clinical teams were formed and clinical and support roles were re-defined to facilitate patient flow and cross-coverage. 4) Flexible clinical schedules were developed to expand appointment hours and optimize the use of constrained clinical space. 5) Clinical information capabilities were enhanced to track clinical performance and outcomes. FINDING TO DATE: The proportion of patients with an identified primary clinician has increased from 65% to 100%. A pilot team was developed to test and implement the new design with full department implementation planned for February 2000. For the pilot team, clinical "face-time" as a proportion of each visit has increased, on average, by 20%. Third available appointment time has improved from 8.7 days to Ͻ2 days, and for physicals from 43.6 days to Ͻ3 days. Other key measures including visit and message cycle time and patient satisfaction are being monitored. KEY LESSONS LEARNED: Patient demand for improved access creates opportunities to develop state-of-the-art care delivery systems. Cultural change cannot be underestimated as a challenge to systems change. Key elements of success include building commitment for change within the leadership team; engagement of all levels of staff in the process; creating a high level of tension for change; accounting for the teaching mission of the academic practice; and maintaining constant attention to timely, clear and repeated communication. Utilizing the entire team to "do today's work today" in addressing patient demands for timely access is possible. The cost in treatment of allergic rhinitis ranks third in overall pharmacy costs in our network over the past 3 years exceeding costs used to treat chronic conditions such as hypertension and congestive heart failure. The control chart on the use of non-sedating antihistamines over the past 18 months demonstrates the fluctuations in costs concordant with seasonal variation. The chart also demonstrates the most recent costs exceed the statistically significant upper limit indicating a special cause variation. The network pharmacy improvement committee applied the FOCUS-PDCA improvement model to the cost-effective management of allergic rhinitis. The committee followed the FOCUS-PDCA process, which involved biweekly group meetings. The team utilized brainstorming to clarify the current knowledge of the problem. Special cause variation was understood using the pharmacy utilization data collection from our PHO, a flow-chart of the office prescription refill process and a chart review of 70 patients treated for allergic rhinitis in August 1999. Using the 20 ⁄ 80 rule the committee selected two variables to improve: 1) Standardize the phone refill policy in physician offices and 2) disseminate allergic rhinitis practice guidelines to all providers. The committee developed a force field analysis to obtain the desired state of physician compliance with practice guidelines and office implementation of standardized refill policies. KEY LESSONS LEARNED: The program demonstrated the importance of group process techniques necessary to identify variation in the process of care in ambulatory settings. Using the FOCUS-PDCA process the committee is able to quickly identify key areas for improvement and to identify the forces which will drive and restrain the improvement process. to use a non-paper-based educational tool to educate patients about HIV and to assess their acceptance of this model. The clinical nurse specialist in the hospitalbased primary care HIV clinic designed an interactive, hands-on tool that teaches adults about HIV. She collaborated with the Village of Arts and Humanities, a community-based agency that develops innovative health promotion programs in Philadelphia, Pennsylvania, to create this tool. The tool is a life-sized soft sculpture of the human body which is able to become infected with HIV by allowing virus to attach to it. The nurse can demonstrate viral replication and CD 4 cell decline using virus and CD 4 cell attachments to the model. The tool can also teach the concepts of viral mutation and drug resistance by using color-coded virus and medication attachments. DESCRIPTION OF PROGRAM/INTERVENTION: Random patients attending the HIV Clinic at Temple University within the Section of General Internal Medicine were invited to attend individual and group educational sessions with this visual tool when their primary care physician identified them as potentially having low literacy based on having less than a high school diploma. The nurse specialist used the model in sessions with identified patients lasting from 30-60 minutes each. Patients were given a verbal pre-session and post-session test which contained basic questions about the role of CD 4 cells and viral replication in HIV disease. Information about the highest level of education reached was obtained. FINDING TO DATE: Ten patients so far have used the tool in patient education sessions with the clinical nurse specialist. All patients tested to date have found the model easy to use and helpful in understanding HIV disease. Pre-session and post-session testing results with the model will be analyzed. KEY LESSONS LEARNED: This visual teaching aid is useful in teaching low literacy patients about HIV. Future areas of research will be the application of the teaching tool to larger numbers of patients and correlating its use with patient care clinical endpoints. cal findings of malignancy in the database. 61.5% were found as a direct referral from the program. Approximate FS cost per patient without biopsy was $142. Estimated additional community costs for biopsy alone was $150 per patient. This charge was incurred at our biopsy rate of 16% of patients per year, therefore the average cost of FS with biopsy was $166 per patient. KEY LESSONS LEARNED: The program was successful in detecting 3.4 malignancies per 1000 patients, 61.5% of all colorectal malignancies found. Because none of the biopsies submited by PC detected malignancies and referrals to GI were primarily made because of the FS exam itself, the value and cost effectiveness in biopsing an average risk patient in our screening program is in question. Evaluation by a GI specialist following FS without biopsy may be a more cost effective strategy. Further research is needed to minimize the cancellation and no-show rates for the screening clinic, thereby maximizing practice efficiency. STATEMENT OF PROBLEM: Inadequate reading literacy is a major barrier to better educating patients. Despite its high prevalence, practical solutions for detecting and overcoming low literacy in a busy clinical setting remain elusive. OBJECTIVES OF PROGRAM/INTERVENTION: In exploring the potential role for the multimedia computer in improving office-based patient education, we com-pared the accuracy of information captured from audio-computer interviewing of patients with that obtained from subsequent verbal questioning. DESCRIPTION OF PROGRAM/INTERVENTION: A multimedia computer presentation that used audio-computer interviewing with localized imagery and voices to elicit responses to four questions on prior computer use and cancer risk perceptions. Based on the findings below, we have developed an array of computer-based tools for routine educational tasks in the office setting. FINDING TO DATE: Three patients refused or were unable to interact with the computer at all, and three patients required restarting the presentation from the beginning but ultimately completed the computerized survey. Of the 51 evaluable patients (72.5% African-American, 66.7% female, mean age 47.5 (ϩ18.1)), the mean time in the computer presentation was significantly longer with older age and with no prior computer use but did not differ by gender or race. Despite a high proportion of no prior computer use (60.8%), there was a high rate of agreement (88.7% overall) between audio-computer interviewing and subsequent verbal questioning. KEY LESSONS LEARNED: Audio-computer interviewing is feasible in this urban community health center. The computer offers a partial solution for overcoming literacy barriers inherent in written patient education materials and provides an efficient means of data collection that can be used to better target patients educational needs.

Ephedrine nephrolithiasis associated with chronic ephedrine abuse

Ma-huang strikes again: ephedrine nephrolithiasis

Dietary supplement Health and Education of 1994, from U.S. FDA, Center for Food Safety and Applied Nutrition

FDA Guide to Dietary Supplements. FDA Consumer

KS STATEMENT OF PROBLEM: Primary care residents lack exposure to principles of geriatric medicine and are unfamiliar with care processes across the continuum of community-based sites available for frail older adults. OBJECTIVES OF PROGRAM/INTERVENTION: To develop and evaluate a yearlong program to deliver geriatric content to primary care residents using computer-based content linked to focused clinical experiences

GYN) were pilot program participants. 2) The one-year program consisted of 10 sessions, held once monthly for 3 hours. 3) Independent review of geriatric content on one specified topic each month from: (a) new computer-based learning modules developed by the authors, and (b) selected learning modules on CD-ROM from The Computer-Based Self-Instruction Modules in Geriatric Medicine

wound care; geriatric assessment; gait/falls; care processes in long-term care, assisted living, and home care settings.) 5) Pre-and Post-testing of knowledge and attitudes using previously validated tools, and self-reported confidence about common geriatric syndromes. A satisfaction survey was conducted at program completion. FINDING TO DATE: The completed one-year pilot project was well-received. Mean attendance was 60% of sessions. On a 5-point Likert scale (1 ϭ poor, 5 ϭ excellent), mean ratings were 4.3 for the overall program, 4.7 for the faculty-lead group clinical exposures, and 3.3 for computer-based materials. Residents felt this content was not redundant with other training. All six found the out-of-hospital clinical exposures valuable and felt they should be viewed as an integral component of the program. Applicability to future practice was rated more highly by IM and FP than by GYN residents. Confidence improved in all residents, and attitude im

Residents appreciated the opportunity to learn geriatric content, and valued the faculty-lead, small group clinical experiences highly. Sources of dissatisfaction included computer-based learning methods, difficulty in finding time for independent study prior to the sessions, and logistical challenges in getting away from the hospital for outside experiences

FINDING TO DATE: Student ratings of the Medicine clerkship have remained strong despite major increases in ambulatory sites. The system stimulates friendly competition among sites during discussions that also identify areas for deeper investigation through focus groups. Examples of impact include improving a problematic site (multiple systems problems were revealed, then corrected); providing focus for faculty development initiatives (workshops and discussions improved ambulatory teaching); and monitoring mid-and end-of-clerkship feedback to students (clerkship directors are improving compliance rates). KEY LESSONS LEARNED: Tuft's Evaluation System provides a dynamic model for quality control and improvement across multiple clerkship sites. Capturing both student and faculty perspectives fosters critical discussion of content and teaching. Web-based evaluation instruments greatly increase administrative efficiency to convey essential results quickly to faculty and the Curriculum Committee.

INTERNAL MEDICINE RESIDENCY. BR Leslie, L Adams, L Cyran, S Kick, J Rifkin, Division of General Internal Medicine, University of Colorado Health Sciences Center, Denver, CO PURPOSE: To describe the development and implementation of a behavioral science education program for primary care internal medicine residents. METHODS: Psychosocial problems account for up to 45% of patient visits in a primary care setting. Physicians who can address these problems produce better patient outcomes and higher patient satisfaction. Recognizing the compelling need for behavioral science training, our primary care residency program decided to hire a full time psychologist in September 1999. Goals for this position included: 1) development and implementation of a three-year behavioral science curriculum; 2) provision of individual precepting for residents; 3) consultation with faculty and residents on specific patient problems; and 4) role modeling collaboration between physicians and behavior scientists in primary care settings. Needs assessment activities included individual meetings with faculty and residents, observation of residency clinic operations, and review of existing curriculum to identify opportunities for integrating behavioral sciences into traditional medical teaching. Existing curricula from family medicine programs and ABSAME (Association for the Behavioral Sciences and Medical Education) were reviewed. RESULTS: The initiation of a more comprehensive behavioral science program has been well received. Feedback from faculty and residents has been positive and has helped prioritize the following program activities. 1) Curriculum development and implementation: The curriculum is delivered in multiple formats including longitudinal three year block sessions for complex topics (i.e., communication skills, behavioral change) and noon seminars on specific topics (i.e., somatic disorders, pain management). 2) Individual precepting: All residents have regularly scheduled one-hour behavioral science appointments. This time is used for case presentation or conjoint patient interviews. Videotapes of patient/physician interactions are also reviewed. 3) Precepting: Residents have continuity clinics at three sites. The behavioral scientist is regularly scheduled at these sites to consult with the residents and faculty around specific patient presentations. 4) Role modeling: As a result of consultation on specific patients, the behavioral scientist sees patients for assessment and psychotherapy. Care is delivered in collaboration with the resident or faculty member. CONCLUSION: The need for behavioral science training in primary care residency is well documented. Implementation of a behavioral science curriculum involves identifying existing resources, designing a longitudinal curriculum, and responding to resident and faculty needs as they care for patients.

AJ Lipman, RM Sade, MF Marshell, A Qlotzbach, C Lancaster, Medical University of South Carolina, Charleston, SC PURPOSE: In medical training, computer-based methods of instruction offer the possibility of improvements in teaching clinical skills and professionalism. Without rigorous documentation of its pedagogic advantages, belief in the utility of internet-based teaching is not solidly grounded. We have carried out a prospective, randomized study of educational outcomes, comparing a traditional classroom course in clinical ethics with the same course supplemented by internet-based discussion. METHODS: Introduction to Clinical Ethics is a sophomore medical school course that teaches a specific method for analyzing clinical ethical problems. Our sophomore class was randomly assigned to either classroom teaching alone (traditional group; n ϭ 65) or classroom teaching supplemented with internet-based discussion of cases illustrating ethical issues (cyberethics group; n ϭ 62). A final case analysis comprehensively evaluated student's understanding of the analytic method taught in the course. We compared the grades assigned by external reviewers to the case analyses of the two groups. RESULTS: The student's understanding of ethical analysis, as measured by grades on the final paper, was significantly higher in the cyberethics than in the traditional group (3.0 Ϯ 0.6 and 2.6 Ϯ 0.7, respectively; p Ͻ 0.005). CONCLUSION:: Our study provides objective data documenting the incremental value of internet-based teaching of clinical ethics to sophomore medical students. STATEMENT OF PROBLEM: Community-based interdisciplinary curricula comprise one approach to teaching medical students how to work with health care teams and access community resources to improve patient and community health. The relative novelty of this approach to medical education requires a systematic mechanism to (i) monitor curricular processes and outcomes, and (ii) facilitate continuous improvement of these curricula. OBJECTIVES OF PROGRAM/INTERVENTION: The purpose of the communitybased Interdisciplinary Rural Health Training Program (IRHTP) is to prepare students in medicine, health education, nursing, nurse practitioner programs, nutrition, pharmacy, physician assistant programs, and social work for interdisciplinary practice in rural settings and improve the health of the community. DESCRIPTION OF PROGRAM/INTERVENTION: Consistent with the tenets of continuous improvement (CI), the Plan, Do, Study, and Act (PDSA) cycle was used to assist course planning, administration, and evaluation. The "Plan" phase occurred when an interdisciplinary faculty team revised IRHTP content and evaluation to enhance the experience for all students. The "Do" phase was characterized by implementation of planned revisions to the course. The "Study" phase involved examination of data obtained from students and faculty before, during, and at the end of, the revised course via questionnaires, journals, case conferences, webbased threaded discussions, and meeting notes. Further modifications were implemented during the "Act" phase. FINDING TO DATE: The results indicated (i) there was a significant increase in student knowledge of the education and roles of other members of the health care team (p Ͻ 0.005) (ii) 83% of students could describe the inter-relatedness of social, political, legal, and economic factors in terms of their influence on community health, and (iii) 70% could list recommendations for improving community compliance with current standards of care. Other results, however, prompted several curricular changes. Structured opportunities for interactive discussion of interdisciplinary care among students and community members and for peer teaching were increased, time allocated toward lectures and independent reading was decreased, paperwork was reduced by integrating some assessments into required assignments and by adding banked cases to the case conference schedule, and the criteria for attaining outcome objectives were modified. KEY LESSONS LEARNED: The PDSA cycle provided a systematic approach to improving an interdisciplinary rural health curriculum. Elements consistent with CI and central to the PDSA cycle included, 1) an intensive planning phase, 2) implementation of the revised curriculum and collection of relevant data, and 3) data review and utilization to guide further changes to curricular content and assessment.

RESIDENTS. TD MacKenzie, Community Health Services, Denver Health Authority; General Internal Medicine, University of Colorado Health Sciences Center, Denver, CO STATEMENT OF PROBLEM: Prior to 1997, we were using multiple distinct forms for internal medicine residents to evaluate their experiences at 5 hospitals and 7 continuity clinic sites. Lack of standardization limited objective comparisons across sites. In addition, filing and distributing the forms was labor intensive. OBJECTIVES OF PROGRAM/INTERVENTION: The primary objective of this project was to standardize resident's rotation and attending evaluations across five hospital sites and seven continuity clinic sites. Secondary objectives were to reduce the administrative expense of distributing, copying, and entering data from paper evaluations, to transition all 135 internal medicine residents to electronic mail (email) communication, to improve computer literacy of the residents, and to provide residents with greater assurances that their comments about rotations and attendings are confidential. DESCRIPTION OF PROGRAM/INTERVENTION: In July, 1997, we required that all internal medicine residents in our program sign up for an email account. We then developed three standard evaluation forms that were approved by the participating University-affiliated institutions. These include a monthly rotation evaluation form, an attending evaluation form, and a continuity clinic site evaluation form. Using Microsoft FrontPage TM , a web page development program, the author converted the evaluation forms into web pages accessible to anyone on the Internet. In 1998, it became clear that a more sophisticated server application was needed to maintain the complex relational database and to allow for easy updates to the system. In 1999, the forms and database were converted to Cold Fusion, a web server application with links to Microsoft Access. With this conversion, the completed evaluations are automatically sent by email to the program director, the chairman of Medicine, the attending of record, and the current Chief Residents. Also, residents and the supervising faculty are better able to track delinquent rotations evaluations. FINDING TO DATE: Since the new pages went live in February, 1999 we have ostensibly required clinic evaluations every 6 months and rotation evaluations every month. We have received 103 ⁄ 270 continuity clinic evaluations and 335 ⁄ 1620 monthly rotation evaluations. We have also received 131 optional attending evaluations. The response rate is much less for senior residents than for interns. Despite the low response rate, we have been able to make objective comparisons across rotation sites for the first time. Feedback from faculty and residents is mixed. KEY LESSONS LEARNED: In the eyes of those who look at evaluation forms, the program has been very successful. The challenges that face us are 1) acceptance by rotation directors so that paper forms can be discarded, and 2) building incentives for residents to complete the web pages.STATEMENT OF PROBLEM: Despite the growing number of refugees, asylees, and survivors of torture, healthcare providers learn little during their training about caring for these individuals and other international human rights issues. In addition, providers working primarily in urban areas recognize they are ill prepared for this task. OBJECTIVES OF PROGRAM/INTERVENTION: The project's primary objective is to devise an educational tool accessible to any healthcare provider or trainee who is interested in learning about victims of human rights violations. The project has 4 educational objectives: 1. To understand the scope and magnitude of human rights violations. 2. To recognize the symptoms and signs that victims of human rights violations present with. 3. To learn how to approach these patients. 4 . To inform about other resources available for these individuals. DESCRIPTION OF PROGRAM/INTERVENTION: We devised an introductory Internet-based course covering 8 areas: definitions, history, refugee status and asylum process, epidemiology, symptoms and signs, clinical approach, education and prevention, and resources. The project has a pilot phase during which the course's content and format will be tested and an implementation phase during which changes from the pilot phase will be incorporated into the project and the final course will be placed on the Internet. FINDING TO DATE: The first phase of the project is complete and the first part of the second phase is underway. The course was introduced to faculty members, community-based healthcare providers, residents, and medical and public health students during a series of academic conferences. Overall the audiences agreed that such a course would fulfill knowledge needed in today's urban healthcare environment. Participants suggested expanding the following areas: education and prevention, definitions, and asylum process. Although designed to be an introductory course, we noted the importance of gaining feedback to further develop certain areas, such as epidemiology, education, and prevention for public health providers or signs, symptoms, and clinical approach for medical practitioners. KEY LESSONS LEARNED: Teaching human rights through an Internet-based course demonstrates the flexibility and the power of this tool as an educational media. By facilitating access to course materials in human rights, levels of awareness and advocacy in healthcare providers will increasingly benefit a significant number of individuals whose human rights have been violated. (JAMA 1987; 257:1629 -1631 , literary classics were applied to ethics. Well-known affinities exist between medicine and literature providing an alternative approach to traditional didactics, bedside discussions and forums. Yet, recent comprehensive bibliographies addressing areas wherein literature may lend to education-ethics (reports of the American Society for Bioethics and Humanities) and Project Professionalism (from the ABIM)-literature is absent.

OBJECTIVES OF PROGRAM/INTERVENTION: To introduce a literary context to graduate medical education with an emphasis on core education in professionalism. DESCRIPTION OF PROGRAM/INTERVENTION: The "Great Books Seminar" offered by Evanston Northwestern Healthcare began in 1981. The course has evolved, but originally was intended to be a forum in which literature serves as a springboard for exploring issues in clinical medical ethics. FINDING TO DATE: Currently core bibliographies present issues related to professionalism in isolation. For the resident, such an approach risks being didactic, presenting professionalism as another process to be mastered. Such an approach confounds the idea that professionalism is dynamic. Recent sample readings and the area they engage include: Face of Stone and The Use of Force by William Carlos Williams-portrayals of compassionate but overworked doctors confronting personal reactions to patients. The Student's Wife and Blackbird Pie concern academics, who while deeply committed to their work, ignore people close to them (written by Raymond Carver). Addressing the duty to treat during the AIDS epidemic, The Plague by Albert Camus and My Own Country by Abraham Verghese are foundations for discussion. Physician arrogance was discussed via the novel Frankenstein. Regarding physician impairment, the story entitled Old Doc Rivers by William Carlos Williams is contained in a larger volume entitled The Doctor's Stories. It is approximately 30 pages long and tells the story of a well-liked physician with a substance abuse problem. Although Doc Rivers got there-"anywhere, anytime, for anybody-no distinctions"-he began to take "the dope." His loyal clientel said, "he's a dope, but when he's right, you can't beat him." The bitter irony of his addiction was that "he became a legend and indulged himself the more." KEY LESSONS LEARNED: As the "Great Books" seminar appears to be an appropriate venue for education in the core issues of professionalism. Retreat is held at a remote location on a weekday in June. The morning session consists of seven small groups discussing different aspects of the resident's job. The afternoon session is aimed at boosting morale, teaching professionalism and team building. Our exhibit will describe the goals, development and execution of a retreat. We will display the packet given to each participant, color photographs, and results of the written survey evaluating the educational impact of this exercise. We will also give concrete guidelines and instructions to those programs seeking to develop a similar retreat at their institution. FINDING TO DATE: Improved intern to resident transition,decreased intern anxiety,and increased intern morale. KEY LESSONS LEARNED: A retreat for interns prior to residency will smooth the transition from intern to resident. The faculty can teach upcoming residents what is expected of them and tools with which to accomplish it. The retreat is extremely well received by the interns increasing their confidence and decreasing their anxiety. and access to care, the "Introduction to Patient Care" elective course was developed to train 1 st and 2 nd year medical and pharmacy students to provide social support and advocacy for outpatients with end-stage CHF, COPD, or cancer in the General Medicine Practice at UCSF. Participating patients were those enrolled in a clinical research project entitled the Comprehensive Care Team (CCT)an interdisciplinary team of professionals providing comprehensive health care and family caregiver support. DISCUSSION OF PROGRAM/INTERVENTION: Groups of 2-3 students were assigned to serve as patient advocates to individual CCT patients. Students provided weekly patient contact in person or by telephone, primarily listening to patient stories and communicating needs back to the CCT. Students were required to attend weekly, one-hour classroom sessions. Sessions provided didactic instruction on the following topics: patient advocacy; working with an interdisciplinary team; the patient's experience of illness; caring for patients with CHF, COPD, and cancer; supporting family caregivers; dealing with death and dying; and advance care planning. Additionally, 50% of classroom time was reserved for students to share stories of their experiences with patients. FINDING TO DATE: Of the 32 student participants, 8 (25%) were medical students and 24 (75%) were pharmacy students. At the end of the course, students completed anonymous, self-administered course evaluations (response rate ϭ 63%).

Based on a 1-5 Likert scale (1 ϭ strongly disagree; 5 ϭ strongly agree), students agreed that the quality of the course was high (mean ϭ 4.2) and that the course stimulated their interest in patient advocacy (mean ϭ 4.5). Students felt comfortable speaking with their patients (mean ϭ 4.0) and sharing their experiences with the class (mean ϭ 4.2). While students felt that this course was more challenging than other electives (mean ϭ 3.9), they felt that it was more valuable (mean ϭ 4.5). All 20 respondents (100%) would recommend the course to a classmate. CCT patients are surveyed semi-annually about their experiences with the CCT project. All 26 CCT patients (100%) working with student advocates reported that their contact with their advocate was one of the main benefits of the CCT intervention and one of the most satisfying aspects of their health care overall. KEY LESSON LEARNED: A course promoting interdisciplinary care and patient advocacy for 1 st and 2 nd year medical and pharmacy students was well-received, considered important, and mutually satisfying to students and patients. To follow patients over time is a major goal for continuity clinic training. Despite advantages to having residents train in community practices, residents risk losing continuity when they share patients with community preceptors. 1. To develop a tool to measure continuity simply, reliably, and completely. 2. To develop a resident activity report that includes continuity as a regularly reported variable 3. To compare patient continuity in resident clinics in the hospital vs. the community. An electronic data base tracks resident-patient encounters using the billing/encounter form as the data collection tool. Patient demographics, visit type, diagnoses, procedures, and level of faculty supervision are recorded for each encounter. To determine continuity residents indicate with a check mark the number of times they have seen the patient previously. If the resident has seen the patient before, it is a familiar patient. Another check indicates the patient's primary care provider. The resident marks "Myself" to indicate a personal patient. These indicators of continuity were tallied for three months of weekly clinics for 50 PGY 2 and PGY 3 internal medicine residents. Results are for the first three months of academic year 1999-2000. At the hospital clinic, 43 upper level residents were supervised by rotating faculty who did not see their own patients in this all resident practice. At two community sites, 7 upper level residents were supervised by faculty who had their own patients in those practices. The table summarizes the differences in 'familiar patients' and 'personal patients' at the three sites.

KEY LESSONS LEARNED: 1. Residents at community sites saw significantly fewer familiar or personal patients in their primary care clinics compared to residents at the hospital clinic. 2. The range of continuity differed greatly from resident to resident, regardless of clinic setting. 3. The resident data base provides a simple tool for measuring continuity, as well as many other aspects of resident experience. 4. Program leadership hopes to use this information to optimize continuity in primary care training. We have developed an EBM curriculum that involves all major aspects of our residency program. During the first year, residents are introduced to the concepts of EBM and learn the skills of asking four part questions encompassing the patient group, the clinical and comparison intervention and the outcomes of interest (PICO's questions). This is reinforced throughout the year in Intern Morning Report where questions are identified by the group and answered during the next session. In the second and third years of residency, EBM skills are sharpened in an advanced skills block and morning report. In these settings, residents complete Educational Prescriptions by developing PICO's questions, searching the medical literature, critically appraising the evidence, and considering the application to their patients. Educational Prescriptions are presented daily in morning report. The process is reinforced in the outpatient clinics where residents are encouraged to identify knowledge deficits in the form of a PICO's question and relate the answers in the form of an Educational Prescription. During their third year, residents are required to give Residency programs have been increasingly emphasizing the teaching of evidence-based medicine (EBM) to their residents. The most common method currently employed is the traditional journal club where one or a few articles are critically appraised within the context of a group. While this has been shown to modestly increase resident's critical appraisal skills, it focuses almost solely on these skills alone, leaving the remainder of the EBM process not formally taught. For the past year we have been attempting to teach the whole process of EBM to our residents through formal feedback on their creation of critically appraised topics (CATs). In November 1998, we began a system within which residents were encouraged to answer their clinical questions using the process of EBM and formalize the answer in the form of a CAT. Their questions are drawn from the ambulatory clinic, subspecialty electives, intensive care unit, floor experiences, and our own EBM elective. The various attendings on the faculty would help them formulate their question and search, and they were directed toward a librarian facile in searching techniques as well. The evidence is summarized using the one-page CAT format and submitted to 2 faculty reviewers. The reviewers independently scored the CATs on explicitly stated criteria: how common, important, and useful is the information, how readable is their format, how well their comments help us apply the evidence, and how valid is the evidence their CAT is based on. We evolved this into a contest where the resident with the highest score on any CAT submitted that week won "CAT of the week. FINDINGS TO DATE: Resident education will require attendance at four IRB meetings (3rd Friday every month). The HBO "made for TV movie" Miss Evers Boysaddressing the Tuskegee Study-serves as an introduction to the protection of human subjects. A 90-minute session follows the movie during which residents design a study with attention to informed consent, safety monitoring, the ethics of placebo-control and the reporting of adverse events. The first IRB meeting will consist of observation, during which time the resident will focus on the ethical scrutiny applied to protocols. In advance of the following three sessions, the resident will be provided with copies of two protocols to be reviewed at the meetings. The resident's role will be to highlight areas related to human subject protection. The resident would then attend the meeting at which the protocols are discussed, participate, and summarize the CONCLUSION. After each of the meetings, the resident and the IRB director (B.A.) review the resident's evaluative performance in the context of each protocol. KEY LESSONS LEARNED: The IRB provides a fertile environment for IRB ethics education. It should become an integral portion of ethics and research training for internal medicine residencies. This equation accounted for slightly less of the daily variation in patient volume, 75.2%, but also had a p value of Ͻ0.01. The validation set of data showed that the calculated patient volumes, using the equation including weather variables, averaged within 7.6% of actual. When the equation without the weather variables was used, the predictions were within 8.1% of actual patient volume. In 71% of cases the predicted number of patients seen was less than the actual number of patients seen. This is likely due to an 8-9% increase in patient volume since 2/99. FINDING TO DATE: Using the prediction equation including only calendar variables, and adding a correction factor of 8% for the increase in volume, in 9/99 the clinic was accurately staffed on 70% of the days, overstaffed on 20% of the days, and understaffed on 10% of the days. Due to this, the number of patient complaints decreased by more than 50% and the number of patients who left without being seen, which had been increasing during the first half of 1999, decreased by 14% in September. KEY LESSONS LEARNED: Accurate prediction of volume allows for improved staffing. Adequate staffing allows for a decrease in patient complaints and increased patient access. 

study to compare patient satisfaction at one of the RCC's in our program with that of the PCC at the same site run by the same staff members. DESCRIPTION OF PROGRAM/INTERVENTION: A cross-section of patients seen in the previous 5 months at both the RCC and PCC received a mailed questionnaire covering 20 aspects of patient satisfaction in addition to overall clinic satisfaction and demographic data. Returned surveys were sent to an independent data entry firm for tabulation, and final analysis was done by our group using computerized statistical software packages. Pearson chi-square analysis was done for each question. FINDING TO DATE: Of the 1028 surveys sent out to PCC patients, 335 were returned. 85 of the 350 RCC surveys were returned, for an overall response rate of 30.5%. This rate is consistent with most mailed surveys. Patients in the RCC group felt that routine appointments (p ϭ 0.022) and phone call returns were not as convenient (p ϭ 0.002), and it was more difficult for them to see the doctor of their choice (p ϭ 0.015). In addition, RCC patient felt it was more difficult to have their questions answered during clinic hours (p ϭ 0.024), and they were not as well informed of test results as those in the PCC clinic (p ϭ 0.032). However, overall patient satisfaction did not statistically differ between the two groups (p ϭ 0.268). KEY LESSONS LEARNED: While it was heartening to find that patients in the RCC were just as satisfied with their medical care as those in the PCC, this study certainly identifies important areas of routine ambulatory patient care which would benefit from directed improvement in the RCC. Some of these areas speak to the decreased availability of residents (one half-day per week) compared to most faculty members, and as such are unlkely to be areas for possible improvement. However, there clearly is a disparity with regards to communication with patients, as seen by the lower ratings by RCC patients for questions such as phone call returns and communication of test results. These areas should be noted by the RCC directors for intervention development in order to improve the care and satisfaction received by patients who are generous enough to enroll in these clinics which help to further medical education.OBJECTIVES OF PROGRAM/INTERVENTION: Working for seven months with 34 provider organizations, we aimed to learn how to improve care for patients with advanced CHF or COPD. Participating organizations included VA medical centers, hospices, integrated health systems, and one eldercare organization. Descriptition of Program/Intervention: The Breakthrough Series Collaboratives, a program of the Institute for Healthcare Improvement, recruits provider teams to learn methods to efficiently adapt improvements to their programs. In the Collaborative on Advanced Heart and Lung Disease, co-sponsored by the Center to Improve Care of the Dying and the Department of Veterans Affairs, 34 teams joined in three Learning Sessions, internet and telephone communication, site visits, monthly reports, and especially on-site work in quality improvement. Most teams were from the DVA or from hospices. FINDING TO DATE: * Patients to target for advanced illness interventions are often best identified by asking their physician, "Is this patient sick enough that you would not be surprised if he/she died in the next few months?" * Reliability over time is essential. The Collaborative developed a conception of good care that builds on making promises to patients: that medical treatment will be correct, that symptoms will never be overwhelming, that no gaps will affect care, that planning will encompass expectable emergencies, that care will be customized to patient/family preferences, that family issues will be important, and that caregivers will help to make every day as good as it can be. * Good care systems turned out to need medication guidelines implementation, creative adaptation of medication schedules and costs, rapid response to serious symptoms, home medication kits, 24/7 on-call by caregiving team, patient/family self-management, and advance care planning. * Care systems regularly cut emergency room and hospital use to about half of these patient's past utilization, and patients and families were much more confident and satisfied. KEY LESSONS LEARNED: Clinical teams can substantially improve care for patients with advanced heart and lung disease, usually within a few months. * Medicare financing will pose barriers to replicating this work in fee-for-service and many managed care settings. * Patients and family involvement in self-management is welcome and effective. We performed a retrospective cohort study, comparing the care stroke patients received before and after implementation of a stroke pathway, ISIP. The ISIP consisted of admission pre-written order sheets, multidisciplinary rounding with a case manager and social worker, and a diagnostic algorithm based on local expert opinion by the Neurology department. Patients admitted by the neurology service for an ischemic stroke during September 1998 to February 1999 were in the pre-ISIP group. The ISIP was introduced March 1999. Post-ISIP patients were admitted from September 1999 to February 2000. Data was obtained from medical record review and a hospital database. We measured antiplatelet agent or anticoagulant use, deep vein thrombosis (DVT) prophylaxis, physiatry consultations, brain imaging tests, electrocardiograms (ECG), in-hospital mortality, the National Institutes of Health Stroke Scale (NIHSS), the Glasgow Outcomes Stroke (GOS) scores and hospital length of stay. FINDING TO DATE: Seventy-three patients, 46 from the pre-ISIP and 27 patients from the post-ISIP cohorts, were evaluated. The groups had similar admission NIHSS (6.9 vs. 7.8) and GOS (2.2 vs. 1.9) scores. Among patients in the post-ISIP group there was a higher rate of either antiplatelet agent use or anticoagulation on admission (100% vs. 89%, p Ͻ .002), ordering physiatry consults on admission (100% vs. 48%, p Ͻ .002), and obtaining admission ECG (100% vs. 78%, p Ͻ .02). The rate of brain imaging tests (100% post vs. 98% pre) and DVT prophylaxis (75% post vs. 68% pre) was not significantly different between the groups. There were no differences in length of stay, in-hospital mortality, NIHSS or GOS scores between the groups at discharge. KEY LESSONS LEARNED: Patients cared for using the ISIP had higher rates of therapies, tests and consultations that have been associated with better outcomes. Although the ISIP changed process of care we do not know its impact on outcomes. Since our study did not have the power to evaluate outcomes we suggest a larger prospective trial to evaluate the effect of a stroke critical pathway on patient care and outcomes.

STATEMENT OF PROBLEM: Nearly 90% of influenza-related deaths occur in persons over the age of 65. Yearly immunizations can significantly decrease the morbidity and mortality related to influenza. While national recommendations call for immunization rates of 60% in these individuals, many medical centers fall short of this goal. OBJECTIVES OF PROGRAM/INTERVENTION: We examined vaccination rates of patients over the age of 65 at a major university-affiliated general medicine prac-tice. The results may be used to target our efforts to increase immunization of high risk populations. DESCRIPTION OF PROGRAM/INTERVENTION: Our study dates were from August 1, 1998 to July 31, 1999. Using CLIMACS, a medical information and tracking system, we were able to identify all patients over the age of 65 who had an appointment made within two years of our study dates. We then generated a list of all patients who had received an influenza vaccine documented in CLIMACS's electronic flowsheet. By comparing these lists we were able to obtain an estimate of our overall vaccination rate. In addition, a second list of patients over the age of 65 who had an appointment between September, 1998 and January, 1999 was also obtained. A random chart review of 100 of these patients provided an estimate of vaccination rates among those who actually saw a physician during this period. We tabulated the number of vaccines given that were either documented in the flowsheet or recorded within the text of the progress note. FINDING TO DATE: 5048 patients over the age of 65 had an appointment made between August, 1998 and July, 1999 . 1443 of these were recorded on our CLI-MACS system as having received a flu vaccine, corresponding to a 28.6% vaccination rate. 2530 patients had an appointment made from September, 1998 through January, 1999. Random chart review revealed that 91% kept their appointment. Of these, more than 2 ⁄ 3 (68%) received a flu vaccine. 27 patients did not, of which 14 refused, and 3 received a flu shot at another location. KEY LESSONS LEARNED: Our overall vaccination rate was lower than national goals. However, of those patients who made and kept their appointments during the fall, a substantial proportion were vaccinated, exceeding national guidelines. These results imply the need to especially target those patients who are not seen by a physician in the months preceding the flu season. The differences in measured vaccination rates found by the two techniques of chart review reflect problems of compliance with flowsheet documentation within the electronic medical record. Potential strategies for improving compliance and record keeping include a computer generated letter reminder system, electronic chart reminders for physicians, and scheduling of annual exams during the fall months. The techniques outlined here can be applied to other areas of health care maintenance, such as pneumonia vaccination, to tailor effors at improving preventive care to individual practices. English as a second language. Observations made during patient care suggest that patients who received integrated care suffer less psychological re-traumatization compared to patients who received segmented care. On average, patients have struggled for more than 4 years to complete their transition into society. KEY LESSONS LEARNED: A comprehensive approach to the complex needs of these individuals is essential to facilitate their reintegration into society. Preliminary observations suggest that our model of care appears to decrease the incidence of psychological re-traumatization in this vulnerable population; however, studies are needed to confirm these anecdotal observations. ing the 5 boroughs of New York City were randomly selected from a primary care practice and tested for antibody to hepatitis A,B, and C. In addition to obtaining demographic and past medical history information, patients were also screened for thyroid disease, abnormalities of liver function, evidence of blood dyscrasias, exposure to syphilis, and HIV status. Predictors were screened by univariate and mutltivariate analysis to identify risk factors associated with being hepatitis A antibody positive. FINDING TO DATE: Patients were screened for evidence of antibody to hepatitis A (anti-HAV)with 321 (64.2%) being positive (p Ͻ 0.01). The overall average age was 49yo; the average age among anti-HAV (ϩ) patients being 54yo; anti-HAV (Ϫ) being 41yo. Of those screened, 59% were male. Significant (p Ͻ 0.10) predictors on univariate screening included: age Ͼ48yo, being an emigre from outside North America (USA and Canada), race other than Caucasian, and having a prior exposure to hepatitis B. However, multiple regression analysis revealed statistically significant predictors (p Ͻ 0.05) to be age greater than 48yo and being from outside North America. If the patient was from North America and less than 48yo, 83% were anti-HAV negative. However, if the patient was from outside North America and less than 48yo, only 22% were anti-HAV negative. Among patients older than 48yo and from North America, 40% were anti-HAV negative. Patients older than 48yo but from outside North America were 97.9% anti-HAV positive (p Ͻ 0.001). KEY LESSONS LEARNED: If the patient is younger than 48yo and from North America, they should be vaccinated and not tested for prior exposure to hepatitis A. If patients are older than 48yo and from North America, they should be tested before vaccination. Patients older than 48yo and not from North America should neither be screened nor vaccinated. Patients not from North America and younger than 48yo are likely to have been exposed to heaptitis A but should probably be tested before ruling out vaccination. Non-North American patients from Asia, the Caribbean, Central and South America were more likely to have anti-HAV antibody than Western Europeans who they outnumbered 5:1. In screening patients from Western Europe, this model may not be as reliable as for other more represented groups. STATEMENT OF PROBLEM: Running a resident continuity clinic is a challenge for the program leadership and clinic management. Whether the clinic is in the hospital or in a community practice, managers need to make decisions regarding scheduling, faculty supervision, and clinic support staffing. Program leadership needs to monitor resident patient panels, patient continuity and quality of faculty supervision. At our institution, good data on resident experience was limited. Attempts to track resident experience and faculty involvement via questionnaire or log books had serious limitations. OBJECTIVES OF PROGRAM/INTERVENTION: To develop and implement a simple but complete data base for use by clinic management and residency program leadership. DESCRIPTION OF PROGRAM/INTERVENTION: The patient billing/encounter form was modified to include information on patient continuity and supervision level by the precepting faculty. Residents and faculty must complete the form prior to patient check out. Information from 16 data fields are entered daily into a Microsoft Access Database program. Reports can be generated for any date interval required and reviewed with the house officer, supervising faculty, or other authorized user. FINDING TO DATE: At 3 Internal Medicine clinics, 2 Pediatric clinics and 1 combined Med-Peds clinic, over 7000 patient encounters have been entered in 6 months. Three standard reports are being generated. A 'Resident Activity Report' summarizes for each individual resident the number of encounters, ages of patients, distribution of visit types, list of diagnoses (with frequencies), list of procedures, continuity indicators, and faculty supervision levels. The 'Resident Comparitive Report' summarizes key indicators and compares the resident with his peers. The 'Faculty Report' summarizes the numbers of patients seen by each of his or her residents, the faculty involvement level, and the total RVU's earned by the resident. This report also monitors compliance with HCFA supervision requirements by flagging encounters which call for a level of supervision greater than that documented. KEY LESSONS LEARNED: 1. Prompt feedback increases resident willingness to ensure that all items are filled in. 2. With feedback, some faculty pay more attention to charge entry and supervision requirements. 3. Program directors appreciate the comprehensive listing of diagnoses and procedures. 4. Several faculty members are requesting access to the data base to make inquiries about patient panels, continuity, and faculty supervision. 5. Since the hospital accounting system only assesses cost based on faculty billing, this data base will allow a more accurate analysis of the cost of resident training.