Readmission Prediction Using Trajectory-Based Deep Learning Approach

Abstract

Hospital readmission refers to the situation where a patient is re-hospitalized with the same primary diagnosis after discharge. It causes $26 billion preventable expense to the U.S. health systems annually and may indicate suboptimal care for patients. Predicting readmission risk is essential to alleviate such financial and medical consequences. Yet such prediction is challenging due to the dynamic and complex nature of the hospitalization trajectory. The state-of-the-art studies apply statistical models with unified parameters for all patients and use static predictors in a period, failing to consider patients’ heterogeneous illness trajectories. Our approach – TADEL (Trajectory-BAsed DEep Learning) – addresses the present challenge and captures various illness trajectories. We evaluate TADEL on a unique five-year national Medicare claims dataset, reaching a precision of 0.780, a recall of 0.985, and an F1-score of 0.870. This study contributes to IS literature and methodology by formulating the readmission prediction problem and developing a novel personalized readmission risk prediction framework. This framework provides direct implications for health providers to assess patients’ readmission risk and take early interventions to avoid potential negative consequences.