Statistical Models and Patient Predictors of Readmission for Acute Myocardial Infarction

BACKGROUND: Readmission after acute myocardial infarction (AMI) has been targeted for public reporting because it is a common, costly, and often preventable outcome. To assist in ongoing efforts to risk-stratify patients and profile hospitals through public reporting of performance measures, we conducted a systematic review to identify models designed to compare hospital rates of readmission or predict patients' risk of readmission after AMI and to identify studies evaluating patient characteristics associated with AMI readmission. METHODS AND RESULTS: We identified relevant English-language studies published between 1950 and 2007 by searching MEDLINE, Scopus, PsycINFO, and all 4 Ovid Evidence-Based Medicine Reviews. Eligible publications reported on readmission up to 1 year after AMI hospitalization among adults. From 751 potentially relevant articles, 35 met our predefined inclusion/exclusion criteria. Overall, none developed models to compare readmission rates among hospitals or models to predict patients' risk of readmission. All 35 examined patient characteristics associated with AMI readmission. However, studies varied in methods for case and outcome identification, used multiple types of data sources, examined differing outcomes (often either readmission alone or a composite outcome of readmission or death) over varying follow-up periods (from 30 days to 1 year), and found few patient characteristics consistently associated with readmission. CONCLUSIONS: Patient characteristics may be important predictors of AMI readmission; however, few variables were consistently identified. Thus, clinically, patient risk stratification is challenging. From a policy perspective, a validated risk-standardized model to profile hospitals using AMI readmission rates is currently unavailable in the literature.