Theory of radiologist interaction with instant messaging decision support tools: A sequential-explanatory study

Area of Science:

• Medical Informatics
• Human-Computer Interaction
• Radiology

Background:

• Clinical decision support systems (CDSS) and AI show limited integration within radiology workflows.
• Current CDSS development focuses on exam-centric time points: post-acquisition, intra-report, post-report, and workflow-adjacent interventions.

Purpose of the Study:

• To understand the unique interaction patterns of radiologists with software interventions.
• To develop a theoretical framework for radiologist-decision tool interaction.
• To propose novel solutions for enhancing CDSS integration and user satisfaction in radiology.

Main Methods:

• A sequential explanatory study design combining quantitative and qualitative phases.
• Phase 1: A quantitative survey utilizing the User Acceptance of Information Technology (UTAUT) framework to compare average users and radiologists.
• Phase 2: Semi-structured interviews to explore the reasons behind observed differences in user acceptance.

Main Results:

• Identified significant differences in software intervention acceptance between average users and radiologists.
• Gathered qualitative data providing context and explanations for these user-specific differences.
• Proposed Radibot, a conversational agent integrated with instant messaging, as a novel CDSS delivery method.

Conclusions:

• Understanding radiologist-specific user behavior is crucial for effective CDSS design.
• The proposed Radibot system offers a potential pathway to improve CDSS satisfaction and integration in radiology.
• This research provides actionable insights for developers to create more user-centered radiology CDSS tools.