eICU
A public multicenter electronic health record dataset covering intensive care units at multiple hospitals in the United States; because of its cross-institutional heterogeneity, it is often used to evaluate the external transferability of clinical prediction models.
Timeline (last 90 days)
The model was fitted on 7,762 MetaVision patients and externally evaluated on 21,682 MIMIC-IV patients and 43,878 eICU patients covering 204 hospitals.
Physiological trajectory features (R2 relative to R1) brought consistent improvement in external prediction on eICU, with a log-loss difference of -0.00922 (95% CI -0.01148 to -0.00704).
All Facts (2)
The model was fitted on 7,762 MetaVision patients and externally evaluated on 21,682 MIMIC-IV patients and 43,878 eICU patients covering 204 hospitals.
50%UnverifiedPhysiological trajectory features (R2 relative to R1) brought consistent improvement in external prediction on eICU, with a log-loss difference of -0.00922 (95% CI -0.01148 to -0.00704).
50%