key: cord-0029985-29oyl4nt
authors: Chaftari, Patrick; Lipe, Demis N.; Wattana, Monica K.; Qdaisat, Aiham; Krishnamani, Pavitra P.; Thomas, Jomol; Elsayem, Ahmed F.; Sandoval, Marcelo
title: Outcomes of Patients Placed in an Emergency Department Observation Unit of a Comprehensive Cancer Center
date: 2021-12-14
journal: JCO Oncol Pract
DOI: 10.1200/op.21.00478
sha: e3f0b542e5ad696b811cf21efc9515ed6e5724f5
doc_id: 29985
cord_uid: 29oyl4nt

Emergency department observation units (EDOUs) have been shown to decrease length of stay and improve cost effectiveness. Yet, compared with noncancer patients, patients with cancer are placed in EDOUs less often. In this study, we aimed to describe patients who were placed in a cancer center's EDOU to discern their clinical characteristics and outcomes. METHODS: We performed a retrospective observational study that included all patients age 18 years and older who presented to our emergency department (ED) and were placed in the EDOU between March 1, 2019, and February 29, 2020. The patients' electronic medical records were queried for demographics, comorbidities, diagnosis at the time of placement in the EDOU, length of stay, disposition from the EDOU, ED return within 72 hours after discharge from the EDOU, and mortality outcomes at 14 and 30 days. RESULTS: A total of 2,461 visits were eligible for analysis. Cancer-related pain was the main reason for observation in more than one quarter of the visits. The median length of stay in the EDOU was approximately 23 hours, and 69.6% of the patients were discharged. The ED return rate for unscheduled visits at 72 hours was 1.9%. The 14- and 30-day mortality rates were significantly higher for patients who were admitted than for those who were discharged (14 days: 1.7% v 0.3%, P < .001; 30 days: 5.9% v 1.8%, P < .001). CONCLUSION: Our data suggest that placing patients with cancer in EDOUs is safe, reduces admissions, and reserves hospital resources for patients who can receive the most benefit without compromising care.

Emergency department observation units (EDOUs) have been shown to increase patient satisfaction, reduce length of stay, and improve emergency department (ED) throughput efficiency and cost effectiveness 1 while providing high-quality, efficient care for patients with certain common complaints who may require short-term stays for further testing or treatment. 2 Observation care prevents inappropriate short-stay admissions, keeps the institution compliant with Centers of Medicare & Medicaid Services regulation, and avoids ED boarding. 3 Avoiding full inpatient admission for patients who improve quickly and use less resources decreases costs. Combined ED observation and fee-for-service payments are comparable with the bundled diagnosis-related group reimbursement for full admission for these types of patients, and the newly freed inpatient beds can be allotted to patients who are likely to have a larger diagnosis-related group reimbursement.

Observation units can be classified into four different types, ranging from dedicated and protocol-driven (type 1) to unstructured, in which observation care is discretionary (type 4). 4 Furthermore, type 1 observation units have been extensively studied and have been shown to reduce patient length of stay and decrease likelihood of subsequent hospitalizations compared with inpatient hospitalization or other forms of observation. 4 About 12% of patients with short-stay admissions nationwide in 2010 were eligible to be observed in type 1 units. Extrapolating from reported cost savings data by type 1 units over inpatient care, which range between 27% and 42%, the potential annual savings to the US health system was between Author affiliations and support information (if applicable) appear at the end of this article. $5.5 and $8.5 billion US dollars (USD), likely due to inherently shorter length of stays per patient and the aforementioned savings due to reduced staffing, equipment, and space needs. 4 If the two thirds of US hospitals without observation units adopted any form of observation (not only type 1), the potential average cost saving is about $4.6 million USD per hospital and $3.1 billion USD for the entire health care system according to a 2012 systematic review and simulation model. 3 

Higher hospitalization rates are observed in patients with cancer after being evaluated in the ED when compared with the general population. 5 Moreover, patients with cancer who are admitted to the hospital through the ED tend to be more ill than those who are admitted through other avenues. 6, 7 Given the increased rate of hospitalization and the wide-ranging outcomes in this population, it is important to better understand which of these patients can successfully be placed in a specialized observation unit.

The use of observation units among patients with active cancer has not been well described and is poorly understood. Cancer-related admissions are more expensive than are non-cancer-related admissions ($22,100 v $13,800 USD); therefore, appropriate placement is of utmost importance. 8 Two previous studies at individual cancer centers showed that observation units resulted in reduced admission rates among patients with cancer, and about one third of these patients eventually needed full admission. 9, 10 However, these studies did not evaluate comorbidities, medical complexity, or short-term mortality rates of observation in patients with cancer. Furthermore, a 2017 study showed that patients with cancer on Medicare are placed in observation units at lower rates than are their cancer-free counterparts. 11

The EDOU at our hospital is a designated medical unit in which adult patients with cancer can be placed when there is a medical necessity for ongoing short-term treatment, assessment, and reassessment before an ED provider decides to admit or discharge the patient. Our EDOU is considered a hybrid between a type 1 unit and a type 2 unit because limited protocols have been developed and most patients' care is directed by the EDOU providers. The unit is functional 24 hours a day, 7 days a week, and is staffed by an emergency physician and advanced practice providers. Most patients in the observation unit originate from the ED; however, some patients come directly from clinics or procedure areas. By reviewing the cases before placement, case managers determine whether the patients are suitable for observation placement and ensure that the patients are placed under the appropriate level of care. Generally, EDOU placement is a viable option for patients with a medical necessity for a short (less than two-night-long) hospital stay (ie, the two-midnight rule). In our institution, case managers use the Milliman Care Guidelines (MCG) for non-Medicare beneficiaries and consider both the MCG and the two-midnight rule for Medicare beneficiaries to recommend appropriate levels of care. 12 Because the MCG might contradict the two-midnight rule, we educated our ED providers to base their final admission decisions for Medicare beneficiaries on both the medical necessity of a presumed length of stay of one midnight or more and the MCG criteria. 13, 14 Appropriate reimbursable observation placement requires a clinical assessment before placement under observation. This assessment could be performed in an ED or in the clinic.

Observation patients placed outside of our EDOU are placed in observation on a regular floor. They could be discharged or admitted as an inpatient if there is a medical necessity for a hospital stay longer than one night. We developed a few observation algorithms, but because most patients with cancer present with multiple complaints and complex medical problems, most of the observation care we provide is discretionary and directed by a variety of specialists, classifying our observation unit more appropriately as a type 2 observation unit.

The purpose of this study was to discern the clinical characteristics of patients who were placed in our EDOU, their diagnosis at time of admission to the EDOU, their length of EDOU stay, the percentage of these patients who were converted to full admission, their ED recidivism rates within 72 hours, and their 14-and 30-day mortality outcomes.

This retrospective observational study included all adult patients 18 years or older who presented to the ED of our dedicated cancer hospital between March 1, 2019, and February 29, 2020, and were placed in the EDOU. Although our institution specializes in cancer, patients without cancer are not excluded from our ED and EDOU. Thus, both patients with and without cancer are included in this study. Pediatric patients (18 years or younger) were excluded, as these patients are typically admitted to the pediatrics service directly and forego EDOU placement. Patients who were sent for observation outside of the EDOU as a request from their primary oncologist or service were excluded. The majority of these excluded patients were patients with leukemia for whom the leukemia service in our institution requests that patients with leukemia be placed on observation under their service since the conditions of them are more complex and many of these patients are on study protocols. For patients with multiple visits during the study period, only the first visit was analyzed. The primary outcome of interest was disposition from the observation unit. The secondary outcomes of interest were ED return within 72 hours and 14-and 30-day mortality rates.

We abstracted study data by querying the electronic medical record for patient demographics, comorbidities, treatment history, diagnosis at the time of placement in the EDOU, reason for observation, length of stay, disposition from the EDOU, ED return within 72 hours after discharge from the EDOU, and 14-and 30-day mortality outcomes. The reason for observation was assigned on the basis of the primary indication, as stated by the initiating physician. Patients undergoing active chemotherapy and radiotherapy were defined as patients who received chemotherapy or radiotherapy within 30 days before hospitalization, and patients undergoing active immunotherapy were defined as patients who received immunotherapy within the past year. Because side effects of immunotherapy can occur months after the last immunotherapy dose, 15 we choose one year as the threshold for immunotherapy. For patients with multiple cancer types, the most active cancer was considered to be the primary cancer. Hypertension, diabetes, renal failure, heart failure, chronic obstructive pulmonary disease, and myocardial infarction were chosen for abstraction, as these are some of the most commonly cited comorbidities in the literature for patients with cancer. 16, 17 Descriptive statistics were used to summarize the sociodemographic characteristics of the study population. Continuous variables were reported as medians and interquartile ranges or means and standard deviations, and categorical valuables were analyzed as counts and percentages. Clinical characteristics and 14-and 30-day mortality rates were compared between discharged and admitted patients.

All statistical analyses were performed using R software (version 3.6.2, The R Foundation). 18 The institutional review board of MD Anderson approved this study and granted waivers of informed consent.

Of the 28,358 visits to our center's ED during the study period, 3,334 visits (11.8%) resulted in patients being placed under observation; 2,461 (85.7%) unique first visits were eligible for analysis once the exclusion criteria were applied (Fig 1) . The majority (2,372 [96.4%]) of the included patients had at least one cancer type. The median patient age was 63 years, and the majority of patients (67.4%) were White and/or Caucasian. Breast, lung, and colorectal cancers were the most frequent cancer types observed (Table 1) , and 89 patients (3.6%) had more than one cancer type. Most patients (72.6%) had an urgent acuity level (level 3) on the Emergency Severity Index (ESI).

Pain due to neoplastic disease was reported as the main reason for observation in 25.8% of the visits (Table 1) .

Other frequent reasons were electrolyte and/or metabolic disturbance (20.7%), cardiac problems (9.4%), infection (9.2%), GI symptoms (7.8%), anemia (5.3%), and pulmonary problems (4.9%). Of the patients with cancer who presented to the EDOU, 55.2% were undergoing active cancer treatment. Only 32.7%, 26.9%, and 13.9% of the visits were for patients who were receiving active chemotherapy, radiotherapy, or immunotherapy, respectively. Hypertension, diabetes, and renal failure were the main patient comorbidities. The median length of stay was 23 hours, with an admission rate of 30.4%. The 14-and 30day mortality rates were 0.7% and 3.0%, respectively. Forty-seven patients (1.9%) from the whole cohort (including 46 [2.7%] from the ones who got discharged) had returned to the ED within 72 hours of their initial visit ( Table 2 ). Of these, seven patients had scheduled return visits for neutropenic fever follow-up. Comparing the percentage of patients who returned to the ED within 72 hours of their visit after ED discharge without observation during the same study period with those discharged from observation, we found the former group had a significantly higher return rate (8.2% [812 of 9,886 patients] and 2.7% [46 of 1,713 patients], respectively, P , .001). Appendix Table  A1 (online only) shows patient management and outcomes, excluding noncancer patients.

Demographics were not significantly different among patients who were admitted versus those were who discharged, but significant differences were observed in these patients' comorbidities, cancer types, and mortality outcomes (Table 3) . Compared with patients who were discharged, patients who were admitted had significantly higher rates of hypertension (53.9% v 46.0%; P , .001), diabetes (28.6% v 23.6%; P 5 .010), heart failure (11.5% v 6.8%; P , .001), chronic obstructive pulmonary disease (10.2% v 6.0%; P , .001), and renal failure (29.0% v 19.0%; P , .001). The 14-and 30-day mortality rates were also significantly higher for patients who were admitted than for those who were discharged (14 days, 1.7% v 0.3%, P , .001; 30 days, 5.9% v 1.8%, P , .001). Other significant differences associated with disposition from the EDOU were chemotherapy within 30 days (discharged patients, 34.5% v admitted patients, 28.5%; P 5 .004) and consultations requested during the EDOU stay (discharged patients, 29.7% v admitted patients, 56.8%; P , .001). Similar results were observed once noncancer patients were excluded (Appendix Table A2 , online only).

Four different settings can be described for observation care units. Type 1 observation care is usually protocoldriven and has a dedicated location in the hospital. Type 2 observation care has a dedicated location in the hospital, but the care is more discretionary and directed by a variety of specialists. 4 The care in type 1 and type 2 units is usually provided by ED providers. Type 3 observation care is protocol-driven, but a type 3 observation unit does not have a dedicated location. Type 4 observation care is discretionary, and the patient could be in any bed on any floor in the hospital. Type 1 and type 2 observation units are more favorable because the care is delivered in a controlled environment by a specific service in charge. Type 1 observation care carries the highest evidence for a favorable outcome. 4 Since the establishment of the EDOU pilot at our institution in 2012, the number of patients with cancer placed in the EDOU has consistently and rapidly increased. From 2012 to 2014, about 30 patients per month were placed on observation, almost 1.5% of our monthly ED census at the time. 9 Over our current 1-year study period, approximately 10% of the patients presenting to the ED were placed in the EDOU, which is also much higher than the 2% previously reported in general observation units. 19, 20 As most patients presenting to our ED are patients with cancer, 5 our finding was more consistent with the 9% observation rate shown in a 2015 study that studied patients with cancer. 10 Patients with cancer presenting to the ED are believed to have more comorbidities and medical conditions, requiring complex decision making and management that ultimately leads to higher hospitalization rates than those for patients without cancer. 5 Because patients with cancer are often older and more medically complex, they often require more time and resources than are available in an ED; thus, the EDOU should be considered for these patients. 21 Our patients were older and had more comorbidities than did conventional patients in EDOUs, 19 ,20 yet we believe that observation medicine can be a viable alternative in patients with cancer. A recent study of a national geriatric patient cohort showed that observation status was becoming more accepted in older adults, even at the cost of higher admission rates from observation units than the traditional ideal 20% admission rate. 22, 23 Although cancer status was not taken into consideration in this study, it is reasonable to assume that the same principle applies to the subgroup of geriatric patients with cancer. In fact, our EDOU patients were admitted 30.4% of the time, which is similar to the findings of existing reports on patients with cancer and older adults. 10, 24, 25 Despite the available data for older adults, a 2017 study matching Medicare patients with cancer to control patients without cancer found that patients with cancer were placed on observation status at lower rates than were their counterparts without cancer. Additionally, patients who are older than 75 years, have a higher comorbidity score, and have been hospitalized previously are even less likely to be placed on observation status. 11 Our institution uses an adapted version of the ESI classification process. 26 In our center, the ESI includes consideration of systolic blood pressure and temperature to better differentiate high-risk scenarios commonly seen in patients with cancer. Furthermore, Adler et al 27 showed that the ESI is predictive of ED disposition in patients with cancer. Most of our patients had an urgent acuity level on ESI classification.

The median length of stay in the EDOU was approximately 23 hours, and only 46 patients (2.7%) who were discharged from the EDOU returned to the ED within 72 hours after discharge. Interestingly, this was significantly lower than the 72-hour return rate (8.2%) for the patients who were discharged directly from the ED during the same study period. Although the study was not designed to compare these groups, this finding may highlight another potential benefit of the observation unit in the management of patients with cancer, which may potentially lead to lower short-term revisits if some discharged ED patients are . We believe that this novel finding is a surrogate for the safety of the decision-making process in addition to ED revisit rates.

Interestingly, we found that the patients who received chemotherapy within 30 days before admission to the EDOU were more likely to be discharged than were those who did not. No association was seen with radiotherapy or immunotherapy. Although we did not specifically examine the characteristics of this patient subgroup, we speculate that this difference was due to chemotherapy side effects generally being more transient than radiotherapy or immunotherapy side effects and also more quickly treatable.

Comorbidities that were statistically significantly associated with admission over discharge from the EDOU were hypertension, heart failure, chronic obstructive pulmonary disease, diabetes, and renal failure. A previous study found that increased comorbidities in patients with cancer were predictive of admission over observation status from the ED. Unlike our study, it did not specifically look at whether cancer observation patients with comorbidities were more likely to be admitted compared with those without comorbidities after the period of observation. 11 Comorbidities are associated with decreased long-term survival duration in patients with cancer because of alteration in cancer treatment, increase in treatment toxicity, and increased adverse drug interactions. 28, 29 Therefore, comorbidities can translate into longer hospital stays. 30, 31 Several studies have shown that comorbidities contribute not only to the length of stay but also to in-hospital mortality. 32 Still this does not mean that some patients with cancer and comorbidities cannot be safely placed on observation. More research is needed to see how the presence of comorbidities in patients with cancer should steer observation placement decision making.

Cancer-associated pain was the main reason for placement in the EDOU, with nearly 26% of our patients having this as the primary diagnosis. Patients with cancer frequently visit the ED for cancer-related pain. However, management of cancer-related pain continues to be challenging for ED clinicians, and many patients leave with continued pain. 33 For this reason, the EDOU should be considered for patients with cancer whose pain complaints are not fully addressed in the ED. Pain management is one of the most common reasons for EDOU placement, not only in our cohort but also as previously reported. 34 Adequately addressing pain will lead to not only decreased mortality but also decreased risk of 30-day hospital readmission and increased patient satisfaction. 33 Patients admitted to the hospital from the EDOU had more consultation requests during their observation stay than did those who were discharged, likely because of the complexity of their medical problems. This highlights that these patients were sicker, requiring further hospital care.

Whether the need for consultation should be a predictive factor for deciding observation initiation for patients with cancer is also something that should be further explored.

Our study has several important limitations. First, this was a retrospective chart review in a single cancer-specific hospital. Second, the variables collected were limited to what was available from the ED encounter. At our institution, the ED clinicians typically do not document staging or performance status. This is typically documented by the oncologist and may no longer be accurate by the time the patient arrives in the ED. Third, our results may not be the most generalizable. Most patients with leukemia were placed on observation under the leukemia service and not the EDOU. Therefore, they were not included in our study. Additionally, we included 92 patients without a cancer diagnosis, which could affect how these results are applied to the cancer population. However, because of the lack of currently published observation data for patients with cancer, we believe that our findings may affect the number of patients with cancer placed on observation status. Fourth, we limited our study to six comorbidities (diabetes mellitus, heart failure, hypertension, renal failure, chronic obstructive pulmonary disease, and myocardial infarction). Other comorbidities that were not accounted for could have contributed to the increased rate of admissions in our patients. Similarly, we limited the chemotherapy-and radiotherapy-related data collection to 30 days and immunotherapy-related data collection to 1 year before the EDOU visit. Despite these time constraints, we believe that our data are representative of our patients since major acute side effects related to chemotherapy and radiation treatment are usually seen within 30 days of such therapy. Finally, the knowledge of the clinicians in the EDOU of a cancer center might differ from that of general emergency medicine practitioners. Previously, we have shown that the majority (93.4%) of the patients seen by the physicians in the ED at MD Anderson Cancer Center consists primarily of patients with cancer. 5 The extensive experience in cancer care that this ED's physicians have may differ from the experience of a general emergency physician or provider, potentially leading to better performance on the metrics measured at the ED studied versus a general ED. We acknowledge that this study was conducted in a tertiary cancer center. Thus, because of our unique setting, our results may not be generalized to other health care organizations given that our data and our medical practice may not reflect those of other noncancer institutions.

In conclusion, our data suggest that placing patients with cancer in a type 2 observation unit is safe, as evidenced by our low ED return rates within 72 hours and low 14-and 30day mortality rates, although a higher percentage of these patients were admitted than were patients in the general observation units. Even so, observation in cancer has the potential to avoid admissions and reserve inpatient hospital resources for patients who can receive the most benefit without compromising care, as it has been shown to do in noncancer populations. Further study is needed to elucidate predictive factors that may further maximize efficiency of observation status by reducing observation to admission rates. Abbreviations: COPD, chronic obstructive pulmonary disease; ED, emergency department; IQR, interquartile range.

Short-stay units and observation medicine: A systematic review

Emergency department observation units: A clinical and financial benefit for hospitals

Making greater use of dedicated hospital observation units for many short-stay patients could save $3.1 billion a year

Protocol-driven emergency department observation units offer savings, shorter stays, and reduced admissions

Oncologic emergencies in a cancer center emergency department and in general emergency departments countywide and nationwide

Association of emergency department admission and early inpatient palliative care consultation with hospital mortality in a comprehensive cancer center

Presenting symptoms in the emergency department as predictors of intensive care unit admissions and hospital mortality in a comprehensive cancer center

Statistical Brief #270. in Healthcare Cost and Utilization Project (HCUP) Statistical Briefs

Utilizing TQM and CQI prinicples in establishing a type 1 observation unit in a comprehensive cancer center

Impact of observation status on hospital use for patients with cancer

A population-based assessment of emergency department observation status for older adults with cancer

Changes to inpatient versus outpatient hospitalization: Medicare's 2-midnight rule

Observation and inpatient status: Clinical impact of the 2-midnight rule

Diagnosis and management of immune-related adverse effects of immune checkpoint therapy in the emergency department

Annual report to the nation on the status of cancer, 1975-2010, featuring prevalence of comorbidity and impact on survival among persons with lung, colorectal, breast, or prostate cancer

Patient-reported comorbidity and survival in older adults with cancer

Use of observation care in US emergency departments

National study of emergency department observation services

Observation Medicine: Principles and Protocols. Cambridge, United Kingdom

Key operational characteristics in emergency department observation units: A comparative study between sites in the United States and Asia

History, principles, and policies of observation medicine

Care of special populations in an observation unit: Pediatrics and geriatrics

A national dataset analysis of older adults in emergency department observation units

Emergency Severity Index: A Triage Tool for Emergency Department Care

Validation of the Emergency Severity Index (version 4) for the triage of adult emergency department patients with active cancer

The impact of comorbidity on cancer survival: A review

The impact of comorbidity on cancer and its treatment

Factors associated with head and neck cancer hospitalization cost and length of stay-A national study

Prolonged hospital stay and readmission rate in an enhanced recovery after surgery cohort undergoing colorectal cancer surgery

Risk factors for in-hospital mortality and prolonged length of stay in older patients with solid tumor malignancies

Cancer pain management in the emergency department: A multicenter prospective observational trial of the comprehensive oncologic emergencies research network (CONCERN)

Pain management, including musculoskeletal and low back pain

We thank Ann Sutton, Scientific Editor, and Ashli Nguyen-Villarreal, Associate Scientific Editor, in the Research Medical Library at The University of Texas MD Anderson Cancer Center, for editing this article.

The following represents disclosure information provided by authors of this manuscript. All relationships are considered compensated unless otherwise noted. Relationships are self-held unless noted. I 5 Immediate Family Member, Inst 5 My Institution. Relationships may not relate to the subject matter of this manuscript. For more information about ASCO's conflict of interest policy, please refer to www.asco.org/rwc or ascopubs.org/op/authors/author-center.Open Payments is a public database containing information reported by companies about payments made to US-licensed physicians (Open Payments).No potential conflicts of interest were reported.