Truace
Subscribe

Validation of an Electronic Health Record Model for Unplanned Readmission in Patients With Heart Failure

Risk scores aim to predict 30-day heart failure (HF) readmissions. Current models, including Epic's Risk of Unplanned Readmission for a 30-Day Event, demonstrate modest to moderate discriminatory ability (area under the curve: 0.59-0.72). The goals of this study were to create an electronic health record (EHR)-automatable 30-day HF readmission risk calculator; to test the accuracy of Epic's model to predict 30-day HF readmission risk; and to compare the 2 models. To derive our EHR-automatable model using machine…

By the Truvace desk2 min read1 sourcePeer-reviewedTRV-2026-1960· cite & verify
Validation of an Electronic Health Record Model for Unplanned Readmission in Patients With Heart Failure
Medicare participating heart bypass center demonstration : executive summary : final report by Cromwell, Jerry United States. Health Care Financing Administration Health Economics Research, Inc. Public domain

In brief

Researchers built an electronic health record model that uses machine learning to estimate which heart failure patients will be readmitted within 30 days. They based its design on a systematic review of earlier 30-day heart failure risk prediction models, then tested it against Epic's Risk of Unplanned Readmission model.

The result matters because risk scores help direct care resources to the patients who need them most. Both tools showed only moderate discrimination, so neither is a precise predictor, and the source says the new model mainly offers a base for later improvements.

Main points

  1. Prior readmission models, including Epic's, had reported area under the curve values of 0.59 to 0.72.
  2. Epic's model scored a C-statistic of 0.74 in this evaluation.
  3. The newly derived EHR-automatable score reached a C-statistic of 0.76, with a P value of 0.41 for the difference.
  4. The authors suggest adding electrocardiogram-based artificial intelligence diagnostics in future versions.

The gain

A machine-learning risk calculator built from automatable EHR data predicted 30-day heart failure readmission about as well as Epic's commercial model, giving hospitals an independently derived alternative.

The rundown

The study had three goals: to create an EHR-automatable 30-day heart failure readmission risk calculator, to test the accuracy of Epic's model, and to compare the two. The new model was derived with machine learning after a systematic review of existing prediction models.

The comparison found 0.74 for Epic's model and 0.76 for the new one, a difference that was not statistically significant. The authors read this as comparable performance, and both models were described as having moderate discrimination.

The authors frame risk scores as a means of personalizing care and accelerating resources for those who need them most. They describe the independently derived model as a platform for enhancements such as ECG-based AI diagnostics.

Sources

  1. Peer-reviewedJACC: Case Reports2026-10-16

The debate