PND19 DEVELOPING A MIGRAINE QUALITY OF CARE MEASUREMENT SET


A92 Abstracts

for 1 to 5 years in a US Medicare Health Maintenance Organi-
zation. The study population was identified using ICD-9 codes
(332;332.0) for PD and a claim for at least one filled PD pre-
scription identified using NDC codes. Medication Possession
Ratio (MPR) was estimated as proxy for adherence to PD med-
ications, with a threshold of 0.80 indicating adherence. PD
symptom progression was defined as increase in prescription
strength, addition of another PD medication, an emergency room
visit or hospitalization related to PD. Demographic, clinical, and
economic variables were extracted from the dataset. Prevalence
of non-adherence was calculated over all years and for each year
of eligibility. Logistic regression was used to assess relation
between medication adherence and PD symptom progression.
Sensitivity analyses were conducted for MPR scores less than 0.6,
0.4, and 0.2. RESULTS: The study population (N = 470 patient-
years) had mean MPR score of 0.49 (±0.38). An average of 66%
of study population was not adherent to their PD medications
(MPR < 0.8 implies non-adherence). Sensitivity analysis with
MPR scores less than 0.6, 0.4, and 0.2 indicated an average of
56%, 46% and 35% were not adherent to their PD medications
respectively. Subjects adherent to their PD medications (MPR >=
0.8) had 67% less odds of experiencing PD symptom progres-
sion (OR = 0.33; CI:0.12–0.85) compared to people not adher-
ent to PD medications. All subjects with MPR < 0.2 experienced
PD symptom progression. CONCLUSION: High prevalence of
non-adherence to PD medication and its association with PD
symptom progression, irrespective of MPR threshold chosen,
indicates it is a significant problem. There is a need for mecha-
nisms to improve medication adherence in PD, namely improved
patient understanding, simplified treatment regimens and
improved tolerability profiles.

PND18
DETERMINANTS OF EMERGENCY DEPARTMENT
UTILIZATION FOR MIGRAINE CARE
Kwong WJ
University of Georgia, Athens, GA, USA
OBJECTIVES: Over-utilization of emergency department (ED)
services for non-urgent medical conditions has been noted as a
problem for decades. Headache is the sixth most common reason
for ED visits. This study examined factors of ED use in a
migraine population. METHODS: Medical and Rx claims of
Georgia Medicaid beneficiaries who had at least one migraine
medical claim (ICD-9 of 346.xx) or one triptan/ergot claim
between Jan 2002 and Dec 2005, and were continuously eligi-
ble from 6 months before to 12 months after first migraine claim
were analyzed. Subjects who had a narcotic claim and a medical
claim for cancer (ICD-9 140–239), fractures (800–829), muscu-
loskeletal and connective tissues disease (710–739), or sickle cell
anemia (282.6x) 6 months before index date were excluded.
Likelihood of a migraine ED visit during the 12 months after
index date was estimated using logistic regression while con-
trolling for age, gender, race, metropolitan status of county of
residence, physician supply in county, butalbital and narcotic
medication use 6 months before index date. RESULTS: Data
from 43,791 subjects were analyzed. Mean age was 31 years (SD
= 17), 23% were male and 55% were Caucasians. Six percent
of subjects had used butalbitals and 28% had used narcotics 6
months before index date, and 3% subjects had at least one
migraine ED visit 12 months after index date. Females, non-Cau-
casians, residents of metro counties or counties that had lower
than average physician per 10,000 population ratio were more
likely to have an ED visit (adjusted odds ratio, AOR = 1.68, 1.51,
2.37, and 3.10 respectively, p < 0.05). Prior narcotic use signif-
icantly increased the risk of ED visit (AOR = 1.82, p < 0.001).

Age and butalbital use had no effect on ED use. CONCLUSION:
Local physician shortage and easy ED access encouraged ED use
for non-urgent medical problems. Improving access to primary
care facilities is crucial for reducing non-urgent ED use.

PND19
DEVELOPING A MIGRAINE QUALITY OF CARE
MEASUREMENT SET
Leas B1, Gagne JJ2, Goldfarb NI1, Rupnow MF2, Lofland JH1, Hopkins
MM3, Nash D1, Silberstein SD3
1Thomas Jefferson University, Philadelphia, PA, USA, 2Ortho-McNeil
Janssen Scientific Affairs, LLC, Titusville, NJ, USA, 3Thomas Jefferson
University Hospital, Philadelphia, PA, USA
OBJECTIVES: To develop a migraine quality of care measure-
ment set at the health plan level, in order to begin measuring and
improving migraine care processes and outcomes. METHODS:
The measurement set was developed through: 1) review of
migraine care guidelines; 2) literature review of quality mea-
surement for migraine care; 3) telephone interviews with thought
leaders in migraine care and quality improvement; and 4) assem-
bly of a national advisory board consisting of prominent leaders
within migraine care, quality measurement and managed care.
The advisory board reviewed collected information from tasks
1–3, discussed candidate measures, and established a consensus
on target measures to be included in the set. RESULTS: The advi-
sory board selected 19 potential measures that could be imple-
mented at the health plan level using administrative (claims)
data. These measures capture information on: migraine diagno-
sis and prevalence; use of preventive and therapeutic medica-
tions; and, primary care, specialist, emergency, diagnostic
radiologic, and inpatient service utilization. The measurement
specifications have been developed to mirror technical specifica-
tions for administrative measures in the Health Plan Employer
Data and Information Set (HEDIS). CONCLUSION: Develop-
ment of an evidence-based set of quality measures for migraine
care is an important advance in seeking to measure and improve
care for migraineurs. Despite its prevalence, and impact on direct
and indirect costs, migraine is not currently being addressed in
the national quality measurement movement. The measurement
set is now being pilot tested in health plans to assess feasibility
of data collection, properties of the measures, and correlation
among the measures.

PND20
UNDERSTANDING LIMITATIONS OF PATIENT REPORTED
CLINICAL OUTCOMES IN LUPUS
Wilke CT1, Jolly M2, Khandelwal S3, Rodby R2, Block JA3,
Pickard AS4
1University of Illinois at Chicago, Chicago, IL, USA, 2Rush University
Medical Center, Chicago, IL, USA, 3Rush Medical University Medical
Center, Chicago, IL, USA, 4College of Pharmacy, University of Illinois at
Chicago, Chicago, IL, USA
OBJECTIVES: Patient knowledge is associated with their ability
to manage their disease (self-efficacy). The objective of this study
was to examine the extent to which patients with systemic lupus
erythematosus (SLE) reported specific organs/system involvement
that was consistent with medical records. METHODS: In a cross-
sectional study, patients with SLE were asked to indicate whether
the following organ systems were affected by their lupus:
skin/hair/scalp, joints, kidneys, brain, heart/lungs, (abnormal)
blood counts, blood clots (including stroke). A retrospective chart
review was undertaken to examine organ involvement and tests
as recorded by their physician. Exact and kappa measures of
agreement between physician and patient report were calculated
for each organ system. RESULTS: The patient sample (n = 70)