Abstract 6: Preventable Readmission Rates among Medicare Beneficiaries with Ischemic Stroke

Background: Hospital readmissions are being used to assess quality of care for several conditions. Recent studies, however, find that only a small proportion of readmissions after MI are preventable, casting doubt on the merits of using all-cause readmissions to compare hospital performance. Data related to preventable readmissions after stroke are limited. We determined rates of all-cause and potentially preventable readmissions within 30 days of discharge after ischemic stroke using national Medicare data, and identified patient-level factors associated with preventable readmissions. Methods: The cohort included all fee-for-service Medicare beneficiaries ≥65 yrs discharged alive with a primary diagnosis of ischemic stroke (ICD-9 433, 434, 436) in 2006. Only the first readmission was used if there were multiple readmissions within 30 days. Preventable readmissions were identified using Agency for Healthcare Research and Quality methodology. Overall, hospital, and regional preventable readmission rates were assessed. A random effect Cox proportional hazards model was used to identify patient-level factors associated with preventable readmissions. Results: Among 302,565 ischemic stroke hospital discharges, 39,057 (14.8%) patients were readmitted within 30 days. Of these, 13.6% (n=5,322) were potentially preventable. The median hospital-level rate was 14% (Inter quartile range [IQR] 11.6-16.8%) for all-cause and 1.6% (IQR 0.7-2.3%) for preventable readmissions. Preventable readmission rates were higher in the Southeast and Mid-Atlantic and lower in the Mountain and Pacific regions. In risk-adjusted analysis, older age, previous MI, CHF, and diabetes were strong predictors of preventable readmissions. Conclusion: Almost 15% of elderly stroke patients were readmitted within 30 days, but only 2% with a readmission that was potentially preventable. National programs aimed at reducing 30-day post-stroke readmissions may have limited impact unless they focus on the small subset of potentially preventable readmissions. Identifying patients at greatest risk is an important step for developing post-discharge interventions to minimize preventable readmissions.