Abstract 12685: Hospital Variation in Adherence to Secondary Prevention Medications After Acute Myocardial Infarction

Abstract

Background: Hospital transition of care practices may influence patient adherence to secondary prevention medications post-discharge. We assessed inter-hospital variation in 90-day post-MI medication adherence and its association with downstream patient outcomes. Methods: We linked 19,704 post-acute MI patients >65 years discharged alive from 459 hospitals in ACTION Registry-GWTG to Medicare data. Adherence was defined as ≥80% proportion of days covered using Part D prescription fill data. Based on the distribution of odds ratios, hospitals were divided into high (>75 th percentile), medium, or low (<25% percentile) adherence. We used Cox regression to compare 2-year MACE outcomes (death, MI, stroke, or non-staged revascularization) after adjusting for hospital case mix differences. Results: By 90 days post-MI, 68% of patients were adherent to discharge-prescribed beta-blocker therapy, 63% to statins, 64% to ACEI/ARBs, and 72% to thienopyridines. There was significant inter-hospital variation in 90-day adherence (p<0.001), the probability of which varied the least for ACE-I/ARBs, and most for beta-blockers (Figure). Hospitals with medium medication adherence had lower unadjusted 2-year MACE risk than hospitals with low adherence, but this did not reach statistical significance after adjusting for differences in patient case mix (unadjusted 33.4% vs. 37.3%, adjusted HR 0.94, 95% CI 0.86-1.02); however, high medication adherence hospitals had significantly lower unadjusted (29.9%) and adjusted risk of MACE (HR 0.90, 95% CI 0.82-0.98). Conclusion: Patient adherence to secondary prevention medications varied markedly across U.S. hospitals. Discharge from hospitals with higher adherence rates was associated with lower risks of MACE. Our results suggest that the development of collaborative initiatives between healthcare providers and patients to improve post-discharge medication adherence may enhance longitudinal patient outcomes.