SHM Converge Scientific Abstract ePoster Gallery-Data vs. Intuition: Assessing the Added Value of Nurse Judgement to Early Warning Scores for Predicting Clinical Deterioration

Data vs. Intuition: Assessing the Added Value of Nurse Judgement to Early Warning Scores for Predicting Clinical Deterioration

Back to course

Pdf Summary

This retrospective study examined whether adding nurses’ bedside intuition to early warning scores improves prediction of clinical deterioration. The team studied adult inpatient ward patients from January 2019 to August 2021 who had both eCART scores and nursing disposition documented. eCART is a machine-learning risk model using 27 vital signs and lab values to estimate risk from 0 to 100. When patients were considered high risk, nurses completed a pathway indicating whether the patient seemed “stable/expected” or “managing instability,” which served as the intuition measure.<br /><br />A total of 14,265 patients were included. In the patient-level analysis, those labeled “managing instability” had a higher rate of deterioration during hospitalization than those labeled “stable/expected” (12.1% vs 7.0%). However, when nursing disposition was added to early warning scores, it did not improve prediction.<br /><br />In the longitudinal analysis, the models were compared for predicting ICU transfer, cardiac arrest, or death within 24 hours. eCART alone performed best among the tested tools (AUC 0.78), followed by MEWS alone (0.72). Adding nursing disposition reduced performance slightly: eCART + disposition had an AUC of 0.75, and MEWS + disposition had an AUC of 0.71. Disposition alone performed poorly (AUC 0.56). All comparisons were statistically significant (p = 0.001).<br /><br />Overall, the study found that nursing judgment, as measured here by a binary disposition label, did not add value to eCART or MEWS for detecting deterioration. The authors suggest future research should examine more nuanced or graded measures of clinical intuition.

Asset Subtitle

Leena S. Penumalee

Meta Tag

Author List Dana Edelson, Kyle Carey, Leena S. Penumalee, Matthew Churpek

Category Research

Concept Nurse Judgment

Concept Clinical Deterioration

Concept Medical Early Warning Score

Concept Electronic Cardiac Arrest Risk Triage

Concept Clinical Intuition

Distinguished Non-Finalist

Presenter Organization University of Chicago Pritzker School of Medicine

Presenting Author Leena S. Penumalee

Track Outcomes Research

Keywords

retrospective study

nursing intuition

early warning scores

clinical deterioration

eCART

MEWS

ICU transfer

cardiac arrest

hospital mortality

AUC

Nurse Judgment

Clinical Deterioration

Medical Early Warning Score

Electronic Cardiac Arrest Risk Triage

Clinical Intuition