SHM Converge Scientific Abstract ePoster Gallery-Utility of Epic Readmission Risk Score in Cancer Patients

Pdf Summary

This retrospective study evaluated whether the Epic Readmission Risk Score accurately predicts 30-day hospital readmission among cancer patients. Researchers reviewed all cancer patients discharged from one institution between September 2021 and September 2022, using Vizient and Epic EMR data, and verified readmissions and tumor types through physician chart review.<br /><br />Among 3,292 discharged cancer patients, 502 were readmitted within 30 days, for an overall readmission rate of 15.3%. Epic readmission scores were available for 3,280 patients. The score showed poor agreement with actual readmission status, with an observed agreement of 42.93% and a very low kappa coefficient of 0.022, indicating little correspondence between predicted risk and outcome.<br /><br />Performance metrics also suggested limited usefulness: sensitivity was 66.2%, specificity 38.7%, positive predictive value 16.3%, negative predictive value 86.4%, and area under the curve (AUC) was 0.5247. An AUC near 0.5 indicates performance only slightly better than chance.<br /><br />The authors conclude that the current Epic Readmission Risk Score does not reliably distinguish cancer patients at high risk for readmission. They suggest that future work should focus on developing a more accurate readmission prediction tool tailored to the cancer population.

Asset Subtitle

Patricia E. Litkowski

Meta Tag

Author List Anna McComas, Lauren Jurkowski, Ningying Wu, Olivia Aranha, Patricia E. Litkowski, Yash Shah

Category Research

Concept Epic readmission risk score

Concept 30-Day Readmission

Concept Readmission Risk Score

Concept Oncology Patient

Concept Hospital Readmission

Distinguished Non-Finalist

Presenter Organization Washington University in St. Louis

Presenting Author Patricia E. Litkowski

Track Quality Improvement

Keywords

Epic Readmission Risk Score

30-day readmission

cancer patients

retrospective study

predictive accuracy

AUC 0.5247

kappa coefficient

sensitivity and specificity

hospital readmission

EMR data

Epic readmission risk score

30-Day Readmission

Readmission Risk Score

Oncology Patient

Hospital Readmission