Abstract 12126: Derivation and Validation of a 30-Day Congestive Heart Failure Readmission Model
Introduction: In 2006, there were over 1 million hospital admissions for heart failure (HF), and the estimated cost to the U.S. in 2009 was over $37.2 billion. Better models to target aggressive therapy to patients at the highest risk for readmission are clearly needed.
Methods: We studied 3631 consecutive admissions for HF based on discharge diagnosis codes between October 2007 and August 2011 from a single academic center. We randomly generated derivation and validation sets in a 3:1 ratio. We used generalized estimating equations to develop and test our models, accounting for repeated hospitalizations.
Results: The 30-day readmission rate in both sets was 25%. Among 25 candidate variables, the best fitting model included creatinine level and hyponatremia at discharge; race; zipcode of residence; suspicion of ischemia; discharge hour; and number of hospitalizations in the previous year. Insignificant variables included IV diuretic use on day of discharge, discharge service, diabetes, atrial fibrillation, age, and sex. Among the subset of patients with measured BNP levels, change in BNP during admission also tended to be associated with readmission. Men, but not women, who were married had a lower risk of readmission. The predicted / observed risks of readmission across ascending tertiles of predicted risk were 14%/13%, 21%/24%, 33%/37% in the derivation set and 14%/18%, 22%/29%, and 35%/38% in the validation set. The area under the ROC curve for the model was 0.67 in the derivation set and 0.65 in the validation set.
Conclusions: We derived and validated a simple model relating discharge-specific characteristics with risk of 30-day readmission. The model was well-calibrated and yielded better discrimination than most models in the literature. Application of this approach may facilitate targeted intervention to reduce the burden of rehospitalization among patients with HF.
- © 2012 by American Heart Association, Inc.