Abstract 16794: Is 30-Day Readmission Preventable after Acute Myocardial Infarction? A New MI Patient-Specific Derivation of the Intermountain Risk Score for 30-Day Readmission

Abstract

Background: Composed of the complete blood count (CBC), basic metabolic profile (BMP), and age, the Intermountain Risk Score (IMRS) is a sex-specific clinical decision tool derived to predict mortality. IMRS also predicts incident myocardial infarction (MI), but it is unknown if it predicts 30-day readmission (30dR). A new Medicare rule will assign penalties to hospitals with worse-than-expected 30dR for acute MI patients. This study evaluated if IMRS or a new derivation, IMRS-MI, predicts 30dR among acute MI patients.

Methods: Patients presenting with acute MI (1993-2012) were enrolled in the Intermountain Heart Collaborative Study. Electronic medical records provided age, sex, CBC, BMP, 31 other risk factor, diagnostic, and treatment variables, 30dR, and 30-day mortality. Mortality data were enhanced by Utah death certificates and Social Security death records. Sex-specific Cox regression beta-coefficients were used in 70% of patients to assign scalar risk values for IMRS-MI and the other 30% were used for validation.

Results: Average age was 67.7±12.9 (N=1251) and 68.5±12.9 years (N=546) for females in the derivation and validation groups (p=0.22), respectively, and 61.7±12.3 (N=2850) and 61.4±12.0 years (N=1373) for males (p=0.53). Original IMRS coefficients did not predict 30dR (p>0.05) for females (ROC area under the curve [AUC]=0.517) or males (AUC=0.518), but predicted 30-day mortality (p<0.001: females AUC=0.728, males AUC=0.765). IMRS-MI, derived using all 51 variables, predicted 30dR in females (validation: HR=1.14 per +1 score, 95% CI=1.06, 1.23; p<0.001) and males (validation: HR=1.20 per +1 score, CI=1.13, 1.28; p<0.001). For IMRS-MI categories, proportions free from 30dR were, for females: 83%, 81%, 74%, 77%, 58%, 57%, 63% (HRs=1.17-2.76) and for males: 86%, 73%, 72%, 69%, 67%, 60%, and 62% (HRs=2.13-3.81). IMRS-MI discriminated 30dR (all p<0.001) for females (AUC=0.639, 0.614 for derivation and validation, respectively) and males (AUC=0.624, 0.590).

Conclusions: IMRS-MI strongly stratified 30dR in acute MI patients, with differences of 3-4 fold in relative risk and >25% in absolute risk. IMRS-MI uses data commonly available to hospitals and may be useful in guiding patient care to prevent 30dR and avoid Medicare penalties.