Predicting Short-Term Mortality in Elderly ICU Patients with Diabetes and Heart Failure: A Distributional Inference Framework

2025-06-18

AI Summary

Existing short-term mortality prediction models for elderly ICU patients aged 65–90 years with comorbid diabetes and heart failure lack personalization and uncertainty quantification. Method: We propose the Distributed Risk Estimation and Assessment Modeling (DREAM) framework—a novel distributional inference approach tailored to this high-risk subgroup, yielding posterior risk distributions rather than point estimates. DREAM integrates two-stage feature selection, CatBoost-based modeling, Accumulated Local Effects (ALE) interpretability analysis, and ablation studies to ensure clinical transparency and robustness. Results: Evaluated on MIMIC-IV (n = 1,478), DREAM achieves an AUROC of 0.863 and identifies 19 clinically relevant predictors, with APS III score, oxygen flow rate, GCS eye response, and Braden activity subscore emerging as the strongest mortality indicators. This work enables individualized, probabilistic, and interpretable short-term mortality risk assessment for critically ill older adults.

Technology Category

• Machine Learning: Calibration & Uncertainty Quantification
• Knowledge Representation and Reasoning: Diagnosis and Abductive Reasoning
• Reasoning under Uncertainty: Relational Probabilistic Models

Application Category

• User Modeling, Personalization and Recommendation: Explainable and interpretable methods for personalization
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Semantics and Knowledge: Data modeling to support human-machine intelligence, including LLMs agents, intelligent system behavior, explanations, and user-friendly interactions

Abstract

Elderly ICU patients with coexisting diabetes mellitus and heart failure experience markedly elevated short-term mortality, yet few predictive models are tailored to this high-risk group. Diabetes mellitus affects nearly 30% of U.S. adults over 65, and significantly increases the risk of heart failure. When combined, these conditions worsen frailty, renal dysfunction, and hospitalization risk, leading to one-year mortality rates of up to 40%. Despite their clinical burden and complexity, no established models address individualized mortality prediction in elderly ICU patients with both diabetes mellitus and heart failure. We developed and validated a probabilistic mortality prediction framework using the MIMIC-IV database, targeting 65-90-year-old patients with both diabetes mellitus and heart failure. Using a two-stage feature selection pipeline and a cohort of 1,478 patients, we identified 19 clinically significant variables that reflect physiology, comorbidities, and intensity of treatment. Among six ML models benchmarked, CatBoost achieved the highest test AUROC (0.863), balancing performance and interpretability. To enhance clinical relevance, we employed the DREAM algorithm to generate posterior mortality risk distributions rather than point estimates, enabling assessment of both risk magnitude and uncertainty. This distribution-aware approach facilitates individualized triage in complex ICU settings. Interpretability was further supported via ablation and ALE analysis, highlighting key predictors such as APS III, oxygen flow, GCS eye, and Braden Mobility. Our model enables transparent, personalized, and uncertainty-informed decision support for high-risk ICU populations.

Problem

• Predict short-term mortality in elderly ICU patients with diabetes and heart failure
• Develop a probabilistic mortality prediction framework using clinical variables
• Enhance clinical relevance with distributional risk assessment and interpretability

Innovation

• Probabilistic mortality prediction using MIMIC-IV database
• Two-stage feature selection for 19 clinical variables
• DREAM algorithm for posterior risk distributions