Prediction using a randomized evaluation of data collection integrated through connected technologies (PREDICT): Design and rationale of a randomized trial of patients discharged from the hospital to home

BACKGROUND: Hospital readmission prediction models often perform poorly. A critical limitation is that they use data collected up until the time of discharge but do not leverage information on patient behaviors at home after discharge.

METHODS: PREDICT is a two-arm, randomized trial comparing ways to use remotely-monitored patient activity levels after hospital discharge to improve hospital readmission prediction models. Patients are randomly assigned to use a wearable device or smartphone application to track physical activity data. The study collects also validated assessments on patient characteristics as well as disparate data on credit scores and medication adherence. Patients are followed for 6 months. We evaluate whether these data sources can improve prediction compared to standard modelling approaches.

CONCLUSION: The PREDICT Trial tests a novel method of remotely-monitoring patient behaviors after hospital discharge. Findings from the trial could inform new ways to improve the identification of patients at high-risk for hospital readmission.

Pennsylvania Department of Health: Commonwealth Universal Research Enhancement (CURE) Program
University of Pennsylvania Health System: Penn Medicine Nudge Unit