Integrating Causal Machine Learning into Clinical Decision Support Systems: Insights from Literature and Practice

2026-03-25

AI Summary

This study addresses a critical gap in current clinical decision support systems (CDSS), which predominantly rely on correlational analyses and lack both causal reasoning capabilities and user-centered interfaces tailored for clinicians. Integrating design science research methodology, the authors conduct a structured literature review, expert interviews, and human factors analysis to develop the first design framework for CDSS interfaces powered by causal machine learning. The resulting framework articulates eight design requirements, seven guiding principles, and nine actionable design features, emphasizing human-AI collaboration, trust calibration, and seamless integration into clinical workflows. This work provides both theoretical grounding and practical pathways for developing trustworthy, interpretable, and clinically viable causal AI systems in healthcare settings.

Technology Category

Machine Learning: Causal LearningReasoning under Uncertainty: CausalityHumans and AI: Planning and Decision Support for Human-Machine Teams

Application Category

Responsible Web: Machine-in-the-loop, human agency and autonomySemantics and Knowledge: Data modeling to support human-machine intelligence, including LLMs agents, intelligent system behavior, explanations, and user-friendly interactionsUser Modeling, Personalization and Recommendation: Fairness-aware retrieval and ranking

Abstract

Current clinical decision support systems (CDSSs) typically base their predictions on correlation, not causation. In recent years, causal machine learning (ML) has emerged as a promising way to improve decision-making with CDSSs by offering interpretable, treatment-specific reasoning. However, existing research often emphasizes model development rather than designing clinician-facing interfaces. To address this gap, we investigated how CDSSs based on causal ML should be designed to effectively support collaborative clinical decision-making. Using a design science research methodology, we conducted a structured literature review and interviewed experienced physicians. From these, we derived eight empirically grounded design requirements, developed seven design principles, and proposed nine practical design features. Our results establish guidance for designing CDSSs that deliver causal insights, integrate seamlessly into clinical workflows, and support trust, usability, and human-AI collaboration. We also reveal tensions around automation, responsibility, and regulation, highlighting the need for an adaptive certification process for ML-based medical products.

Problem

• clinical decision support systems
• causal machine learning
• human-AI collaboration
• design requirements
• causality

Innovation

• Causal Machine Learning
• Clinical Decision Support Systems
• Human-AI Collaboration
• Design Science Research
• Interpretable AI