Mortality Prediction
Overview
Mortality prediction estimates who is at risk of death and, in survival analysis, when an event may occur despite incomplete follow-up or censoring. Researchers use statistical survival models such as Cox regression alongside recurrent and transformer networks, graph models, and multimodal methods that combine electronic health records with images, notes, or bedside signals. Studies span ICU outcomes commonly evaluated on MIMIC-III and MIMIC-IV, as well as population risks such as heat-related and overdose mortality. Models can improve discrimination, but reliable calibration, fairness across patient groups, missing data, and performance across hospitals remain important challenges.
Latest papers
30 papers3 papers in the last four weeks, level with the four weeks before. 0.0% of all new papers.