(1237) Predicting Patient Survival Using the St. Marianna Modified Early Warning Score With the 24-6 Rule

Introduction: Early Warning Score (EWS) systems, especially the National Early Warning Score (NEWS), mandate regular vital sign checks and scoring, a substantial burden for ward nurses. We developed a system that automatically calculates a score based on the NEWS rules using five parameters entered into the electronic medical record by ward nurses as their routine patient care tasks: heart rate, systolic blood pressure, respiratory rate, SpO2, and body temperature, the St. Marianna modified early warning score (SMEWS). To calculate the SMEWS every hour for each patient while addressing the variable intervals between vital sign measurements, The "24-6 rule" is adopted in this system, where the most recent vital sign within the last 24 hours and the remaining four parameters within its' preceding 6 hours are used to calculate the SMEWS on an hourly basis for each patient. We evaluate the predictive performance of SMEWS using the "24-6 rule" in predicting unexpected deaths.

Methods: From the SMEWS system implemented in 13 Japanese hospitals, we created an hourly SMEWS dataset for the 24 hours before discharge from every ward for each patients admitted and discharged from every hospital between January 1, and December 31, 2019. To examine the association between unexpected patient deaths and the SMEWS, we used the following criteria: (1) SMEWS score of 0-4 with no red score (RS) that indicates a score of 3 by a single vital sign, 24 hours before discharge, and (2) no missing values in the vital data at the time of discharge. We created Kaplan-Meier survival curves from each point above 4 is evaluated in each score.

Results: A total of 21,322 cases were included. Cases with a score of 5 or higher or the presence of an RS showed a significantly lower survival rate compared to non-occurrence cases (P < 0.0001). Kaplan-Meier survival curves for each graph with a relevant SMEWS or RS showed a significant reduction in mean survival time compared to non-occurrence cases. Furthermore, the mean time until death decreased with an increase in the score.

Conclusions: Calculating the SMEWS based on the "24-6 rule" from routinely measured vital sign data in general ward predicted survival probability in the subsequent 24 hours. Our findings suggest that it is a useful method for monitoring patients.