Causally-informed Deep Learning towards Explainable and Generalizable Outcomes Prediction in Critical Care

2025-02-04

AI Summary

Deep learning–driven Early Warning Score (EWS) systems suffer from poor interpretability and limited out-of-distribution generalizability across clinical centers. To address these challenges, we propose the first causally enhanced deep learning EWS framework for intensive care settings, unifying prediction of six acute deterioration events. Methodologically, we innovatively integrate causal discovery algorithms (PC and NOTEARS) into a graph neural network coupled with attention mechanisms, explicitly modeling causal relationships among clinical variables. This design jointly improves predictive interpretability and robustness to distributional shifts. Evaluated on multicenter real-world ICU data, our model achieves statistically significant accuracy gains over state-of-the-art baselines across all six tasks; it improves cross-population and cross-institution generalization by 12.3%–18.7%. Moreover, it generates clinically meaningful causal intervention pathways—enabling actionable, transparent decision support. This work establishes a novel paradigm for trustworthy, causally grounded AI in critical care.

Technology Category

• Machine Learning: Causal Learning
• Reasoning under Uncertainty: Causality
• Knowledge Representation and Reasoning: Action, Change, and Causality

Application Category

• Graph Algorithms and Modeling for the Web: Graph neural networks and deep learning approaches for Web-related graphs
• Responsible Web: Human-perceived consequences of algorithmic deployment on the web
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs

Abstract

Recent advances in deep learning (DL) have prompted the development of high-performing early warning score (EWS) systems, predicting clinical deteriorations such as acute kidney injury, acute myocardial infarction, or circulatory failure. DL models have proven to be powerful tools for various tasks but come with the cost of lacking interpretability and limited generalizability, hindering their clinical applications. To develop a practical EWS system applicable to various outcomes, we propose causally-informed explainable early prediction model, which leverages causal discovery to identify the underlying causal relationships of prediction and thus owns two unique advantages: demonstrating the explicit interpretation of the prediction while exhibiting decent performance when applied to unfamiliar environments. Benefiting from these features, our approach achieves superior accuracy for 6 different critical deteriorations and achieves better generalizability across different patient groups, compared to various baseline algorithms. Besides, we provide explicit causal pathways to serve as references for assistant clinical diagnosis and potential interventions. The proposed approach enhances the practical application of deep learning in various medical scenarios.

Problem

• Develop explainable EWS system
• Enhance DL model generalizability
• Identify causal relationships for predictions

Innovation

• Causally-informed explainable early prediction model
• Leverages causal discovery for prediction
• Enhances generalizability across patient groups