Mining high-dimensional administrative claims data to predict early hospital readmissions

摘要

BackgroundCurrent readmission models use administrative data supplemented with clinical information. However, the majority of these result in poor predictive performance (area under the curve (AUC)<0.70).