(1266) Widespread Adoption of a Clinical Deterioration Tool Across a Large Health System

Introduction: Delays in recognition of clinical deterioration lead to increased mortality and morbidity. Our health system used several clinical deterioration identification tools over the last decade with variable but limited adoption. Electronic Cardiac Arrest Risk Triage (eCART) is an artificial intelligence (AI) early warning score that uses a machine learning model incorporating 97 variables (including vital signs, labs, and nursing assessment data) to generate risk prediction of ICU transfer or death. The decision to implement eCART was made after a large, retrospective analysis of our health system patients comparing eCART to previously implemented tools (Rothman Index and Epic Deterioration Index). The analysis demonstrated that eCART had significantly higher precision in predicting all-cause hospital deterioration. The goal was to incorporate eCART into our daily workflow to better identify patients at risk of clinical deterioration.

Methods: A multidisciplinary group of nurses and providers convened to develop a workflow for patients at risk of clinical deterioration. The group: 1) created “Care Signature” clinical pathways to address a uniform reassessment of at-risk patients; 2) provided educational tools for front-line staff; 3) hosted weekly team meetings with bidirectional sharing of best practices, process metrics, and patient stories; 4) developed support from nursing and physician bedside and senior leaders; and 5) encouraged unit-based champions to incorporate eCART into the daily workflow.

Results: Process metrics collected since go-live in January 2024 show widespread adoption. Median time to recognize an elevated score is 0.7 hours (goal: ≤1 hour) and compliance with completing the pathway is 62% (goal: ≥ 55%). In addition, the mean number of rapid response team (RRT) activations per month increased from 209 pre-intervention to 243 post-intervention (p=0.0067). A key advantage of the eCART approach is the integration of nursing and physician workflows (clinical pathways) to assist in implementation of standardized approaches to clinical deterioration.

Conclusions: We observed widespread adoption of an empowering bedside tool to detect deterioration. Future analysis will include number and timeliness of ICU transfer, floor level cardiac arrest, and all-cause mortality.