Abstract 1060: Readmission Rates Among Patients With Non-ST-Segment Elevation Myocardial Infarction: Results in 38,131 Patients From the CRUSADE Quality Improvement Initiative
Background: Hospital readmission after acute MI is being considered as a potential outcomes-based hospital performance measure. Yet, factors influencing acute and long-term risk for readmission have not been well characterized.
Methods: We linked Medicare claims data to detailed clinical data from 38,131 NSTEMI patients >=65 years of age discharged alive from 448 CRUSADE hospitals between 2003 and 2006 to examine longitudinal outcomes. Our primary outcome was all-cause 1-year readmission, but we also considered 30- and 90-day readmissions. Cox proportional hazards modeling was used to examine factors associated with readmission, adjusting for baseline patient characteristics.
Results: Overall, 21%, 35%, and 58% of NSTEMI patients were readmitted by 30 days, 90 days, and 1 year, respectively; median time to readmission was 134 days (IQR 29 – 409) from the index discharge. Of those readmitted, 66% were cardiac readmissions (17% recurrent MI, 25% heart failure). Differences in baseline and in-hospital variables between patients with and without readmissions are shown in the table⇓. After multivariate modeling, factors significantly associated with 1-year readmission included older age, diabetes, prior heart failure, peripheral arterial disease, dialysis, current tobacco use, increased admission heart rate, signs of heart failure on admission, lower baseline hematocrit, and increasing baseline creatinine. Similar factors were associated with 30- and 90-day readmissions; c-statistics for models were modest (0.60 –0.61).
Conclusions: The majority of NSTEMI patients aged >=65 years are readmitted within 1 year of their index NSTEMI event. Risk for readmission is only moderately determined by baseline clinical factors. Thus, other factors such as provider quality of care, transitional care planning, and longitudinal patient compliance with therapies likely affect readmission risks and could be modifiable targets for reducing this risk.