Related Experiment Video
Updated: Jun 12, 2025
07:31
Implementation of a Real-Time Psychosis Risk Detection and Alerting System Based on Electronic Health Records using CogStack
Published on: May 15, 2020
7.0K
Evaluating Equity in Usage and Effectiveness of the CONCERN Early Warning System
Rachel Lee1, Kenrick Cato2,3,4, Patricia Dykes5,6
1Department of Biomedical Informatics, Columbia University, New York, New York, United States.
Applied Clinical Informatics
|June 10, 2025
Summary
The CONCERN Early Warning System (CONCERN EWS) showed equitable usage and reduced mortality risk across diverse patient groups. Non-English primary language patients saw greater mortality reduction, highlighting AI-CDSS equity in healthcare.
Area of Science:
- Healthcare technology assessment
- Clinical decision support systems
- Health equity research
Background:
- The CONCERN Early Warning System (CONCERN EWS) is an AI-CDSS predicting patient deterioration using nursing documentation.
- Previous trials showed CONCERN EWS reduces inpatient mortality and length of stay.
- Evaluating implementation outcomes is crucial for ensuring equitable AI-CDSS results across patient populations.
Purpose of the Study:
- To assess if CONCERN EWS usage varies by patient demographics (sex, race, ethnicity, primary language).
- To evaluate if CONCERN EWS effectiveness in reducing in-hospital mortality differs across demographic groups.
Main Methods:
- Retrospective analysis of EHR data from a multi-site pragmatic cluster-randomized controlled trial.
- Assessed equity in CONCERN Detailed Prediction Screen launches across demographic groups.
- Examined mortality risk reduction using Cox proportional hazards models, adjusting for patient characteristics.
Main Results:
- CONCERN EWS usage was equitable across patient demographic characteristics.
- CONCERN EWS significantly reduced in-hospital mortality risk (aHR=0.644, p<0.0001).
- Effectiveness was consistent across most groups, with a greater mortality risk reduction observed for non-English primary language patients (aHR=0.419, p=0.0082).
Conclusions:
- This study contributes to the literature on evaluating AI-CDSS equity in usage and effectiveness.
- Findings suggest equitable engagement and effectiveness of CONCERN EWS.
- Ongoing evaluations are necessary to understand variability and ensure responsible AI-CDSS implementation.
