Assessing the behavioral impact of a diagnostic decision support system

Area of Science:

• Medical Informatics
• Clinical Decision Support
• Human-Computer Interaction

Background:

• Physician behavior is influenced by diagnostic decision support systems (DDSS).
• Quantifying the impact of DDSS on clinical practice is crucial for effective implementation.
• Existing research lacks standardized methods for evaluating DDSS behavioral impact.

Purpose of the Study:

• To describe a research prototype for evaluating the impact of DDSS on physician behavior.
• To propose indices for quantifying the magnitude of DDSS impact.
• To assess the influence of different inference models within a DDSS on physician actions.

Main Methods:

• Development of a research prototype for DDSS impact evaluation.
• Utilization of the Iliad knowledge base, a comprehensive internal medicine diagnostic knowledge base.
• Evaluation of behavioral impact across two distinct case domains and physician specialties.
• Assessment of different inference models against the knowledge base.

Main Results:

• The prototype allows for the quantification of DDSS impact on physician behavior.
• Variations in behavioral impact were observed based on different inference models.
• The study identified specific case domains and physician specialties where DDSS impact differs.
• Proposed indices provide a scalable method for impact assessment.

Conclusions:

• The developed prototype offers a viable framework for assessing DDSS influence on clinical decision-making.
• Quantifiable metrics are essential for understanding and optimizing DDSS in medical practice.
• Further research is warranted to refine inference models and expand evaluation across diverse medical fields.