Written by Peter Liu
Spoon Feed
Implementation of an early warning system (EWS) utilizing nursing documentation to predict inpatient deterioration resulted in 36% lower mortality, 11% shorter length of stay.
Machine learning unlocks a treasure trove of insights from nursing documentation
As a machine learning enthusiast, I think I’ve found the pinnacle of its application in today’s study. Researchers have developed an automated way to use information from nursing documentation to predict patient deterioration in the hospital in a system called the “COmmunicating Narrative Concerns Entered by RNs” (CONCERN) early warning system (EWS). This system utilizes the content and metadata in nursing documentation tasks to categorize patients into low (green), medium (yellow), and high (red) risk categories for patient deterioration and display to clinicians in an easy-to-interpret interface. The upshot? It works, and saves lives.
This pragmatic cluster-randomized controlled trial assessed whether informing care teams with the CONCERN EWS improved patient outcomes. Across 60,893 hospital encounters, the intervention group experienced a 35.6% reduction in instantaneous in-hospital mortality risk (AHR 0.64; 95%CI 0.53–0.78; P < 0.0001) and an 11.2% shorter length of stay (IRR 0.91; 95%CI 0.90–0.93; P < 0.0001) compared to usual care. Secondary analyses showed reduced sepsis risk (AHR 0.96; 95%CI 0.86–0.99; P = 0.03) and increased unanticipated ICU transfers (AHR 1.25; 95%CI, 1.09–1.43; P = 0.001).
If these findings generalize to real-world practice settings, it would be hard to find criticism of what this EWS does or the value that it adds to patients and physicians seeking ways to incorporate nursing insights into their decisions.
How does this change my practice?
The effectiveness of the CONCERN EWS dramatically increases my existing veneration of nursing work and nursing insights in the inpatient setting. There is clearly rich prognostic information encoded in the documentation of their bedside evaluations. While I excitedly hope to utilize a similar EWS in the future, prior to implementation, it would be prudent to first increase ICU capacity in hospitals.
Source
Real-time surveillance system for patient deterioration: a pragmatic cluster-randomized controlled trial. Nat Med. 2025 Apr 2. doi: 10.1038/s41591-025-03609-7. Epub ahead of print. PMID: 40175738
