Abstract 14960: Development of a 30-Day Angina Risk Standardization Model Following an Acute Myocardial Infarction
Introduction: The substantial economic burden and morbidity associated with angina following myocardial infarction (MI) make it an ideal target for performance measurement to expand quality improvement efforts. However, in order to report angina measures, a risk standardization model must be developed to account for case-mix so that sites are evaluated on the care provided rather than the patients they care for.
Methods: We assessed angina at baseline and 30 days post-MI in 2924 patients across 24 US hospitals using the Seattle Angina Questionnaire; angina was categorized as none vs any (SAQ AF score=100 vs <100). Hierarchical logistic regression was used to predict the presence of angina 30 days post-MI from 56 selected a priori patient variables present upon admission. The model accounted for clustering within sites, and non-linear spline terms were considered for continuous variables. Sequential removal of predictors occurred until R2 decreased by > 1%. The reduced model was externally validated in a sample of 2089 patients from 19 US hospitals. Site risk-adjusted 30-day angina incidence was calculated using the same hierarchical model framework.
Results: At 30 days post-MI, 852 patients (29%) had angina. The final model included 12 factors (Figure) and accounted for 88% of the variance of the full model. The model demonstrated acceptable discrimination (derivation c=0.71; validation c=0.69) and excellent calibration (derivation slope 1.03, r=0.985; validation slope 0.97, r=0.997). Among the 24 sites, unadjusted 30-day angina ranged from 9%-89%. Variation narrowed substantially after risk-standardization and shrinkage correction for hospital sample sizes, ranging from 21% -38%.
Conclusions: We identified a set of patient characteristics associated with angina 30 days after MI. This patient-centered hospital-level model of a meaningful outcome provides the basis to assess care quality after MI that could serve as a target for quality improvement efforts.
- © 2013 by American Heart Association, Inc.