Skip to main content
Fig. 2 | Critical Care

Fig. 2

From: Real-time machine learning model to predict short-term mortality in critically ill patients: development and international validation

Fig. 2

The mean alarm count per day (MACPD) for the study cohorts. MACPD is determined to indicate the average alarm count per bed every day. The MACPD values at NEWS 4, 5, and 6 cutoffs are compared with our model’s cutoff while maintaining the same sensitivity. A internal validation on SNUH testing dataset B external validation on MIMIC C external validation on eICU-CRD D external validation on AmsterdamUMCdb. SNUH Seoul national university hospital, MIMIC Medical information mart for intensive care, eICU-CRD eICU collaborative research database, AmsterdamUMCdb Amsterdam university medical center database, SPTTS single-parameter weighted “track and trigger” systems, NEWS national early warning score, MEWS modified early warning score

Back to article page

Critical Care

ISSN: 1364-8535