From clinical decision support to clinical reasoning support systems.

Area of Science:

• Medical Informatics
• Artificial Intelligence in Healthcare
• Epistemology of Medicine

Background:

• The adoption of artificial intelligence (AI) in healthcare is hindered by slow uptake.
• Clinical decision support systems (CDSS) present unique epistemological challenges.
• Understanding the epistemic tasks of medical professionals is crucial for AI integration.

Purpose of the Study:

• To analyze the epistemological issues in developing and implementing AI for clinical practice.
• To reframe CDSS as clinical reasoning support systems (CRSS).
• To define requirements for effective CRSS development and responsible use.

Main Methods:

• Conceptual analysis of medical professionals' epistemic tasks.
• Evaluation of CDSS capabilities in supporting clinical reasoning.
• Exploration of epistemological responsibilities in AI-assisted decision-making.

Main Results:

• CDSS should be conceptualized as CRSS to better support clinical reasoning.
• Effective CRSS require high-quality data and clinician interaction.
• Medical and AI experts must collaborate for successful CRSS development.
• Empirical justification of CRSS-generated data is essential for responsible use.

Conclusions:

• A hybrid intelligence model combining human and AI capabilities is proposed.
• CRSS excel at statistical reasoning and pattern detection.
• Clinicians remain essential for interpreting, integrating, and contextualizing AI-driven insights.
• Responsible CRSS implementation enhances, rather than replaces, clinical expertise.