Physiological trajectories transfer better than observation patterns in ICU mortality
Original title: Physiological Trajectories and Observation Patterns in ICU Mortality Prediction: Transportability Under Changing Measurement Conditions
Physiological trajectories improve external eICU prediction while observation patterns remain sensitive to measurement timing.
The study separates transferable value of physiological trajectories from observation patterns for 48-hour ICU mortality prediction. Models trained on MetaVision data were evaluated on MIMIC-IV and eICU cohorts. Trajectories transferred consistently; observation features benefited only local models and were harmed under measurement constraints.
- Trajectories yield -0.00922 log-loss gain on eICU transfer
- Observation patterns show no benefit after cross-hospital transfer
- Timing of measurements affects R3 more than R2 at fixed counts
From the original article
1 min read at arxiv.orgObjective: To distinguish the transferable predictive value of physiological trajectories from that of observation patterns in ICU mortality prediction, and to assess whether measurement timing affects performance independently of observation quantity. Materials and Methods: In-hospital mortality was predicted at 48 hours using MIMIC-III, MIMIC-IV, and eICU.
By Benjamin T. Pike · arxiv.org
A short excerpt; the article and its images belong to the publisher. Summary, key facts and assessment above are AI-written from the article and can be wrong.