Think as a Doctor: An Interpretable AI Approach for ICU Mortality Prediction
Qingwen Li, Xiaohang Zhao, Xiao Han, Hailiang Huang, Lanjuan Liu
TL;DR
ProtoDoctor addresses the need for intrinsically interpretable ICU mortality prediction by embedding three ICU decision-making elements—clinical course identification, demographic heterogeneity, and prognostication awareness—into a prototype-based framework. It introduces two innovations: the Prognostic Clinical Course Identification module and the Demographic Heterogeneity Recognition module, which jointly learn clinical-course and demographic prototypes and enable cross-modal interactions through a demographic–clinical interaction matrix. The model achieves state-of-the-art predictive performance, notably in AUPRC on the MIMIC-III dataset (AUROC ≈ $0.867$ and AUPRC ≈ $0.542$ at $T=48$), and demonstrates clinically meaningful interpretability via prototype visualizations and human expert evaluations. Ablation and robustness analyses corroborate the importance of both the prognostication-aware regularization and the D–C interaction for identifying mortal cases under class imbalance. The work has practical impact for ICU decision support by delivering transparent reasoning aligned with clinical practice, potentially improving trust, regulatory compliance, and adoption of AI in critical care.
Abstract
Intensive Care Unit (ICU) mortality prediction, which estimates a patient's mortality status at discharge using EHRs collected early in an ICU admission, is vital in critical care. For this task, predictive accuracy alone is insufficient; interpretability is equally essential for building clinical trust and meeting regulatory standards, a topic that has attracted significant attention in information system research. Accordingly, an ideal solution should enable intrinsic interpretability and align its reasoning with three key elements of the ICU decision-making practices: clinical course identification, demographic heterogeneity, and prognostication awareness. However, conventional approaches largely focus on demographic heterogeneity, overlooking clinical course identification and prognostication awareness. Recent prototype learning methods address clinical course identification, yet the integration of the other elements into such frameworks remains underexplored. To address these gaps, we propose ProtoDoctor, a novel ICU mortality prediction framework that delivers intrinsic interpretability while integrating all three elements of the ICU decision-making practices into its reasoning process. Methodologically, ProtoDoctor features two key innovations: the Prognostic Clinical Course Identification module and the Demographic Heterogeneity Recognition module. The former enables the identification of clinical courses via prototype learning and achieves prognostication awareness using a novel regularization mechanism. The latter models demographic heterogeneity through cohort-specific prototypes and risk adjustments. Extensive empirical evaluations demonstrate that ProtoDoctor outperforms state-of-the-art baselines in predictive accuracy. Human evaluations further confirm that its interpretations are more clinically meaningful, trustworthy, and applicable in ICU practice.