Problems and Barriers Related to the Use of AI-Based Clinical Decision Support Systems: Interview Study

Area of Science:

• Health Informatics
• Artificial Intelligence in Healthcare
• Clinical Decision Support Systems

Background:

• Digitalization is transforming healthcare globally.
• Artificial intelligence (AI) offers significant potential to support healthcare providers.
• AI integration in clinical decision support systems (CDSSs) is particularly promising but faces challenges.

Purpose of the Study:

• To identify challenges and barriers of AI-based CDSSs.
• To gather perspectives from diverse expert disciplines.
• To inform the optimization of AI development and implementation in healthcare.

Main Methods:

• Conducted semistructured expert interviews with 15 stakeholders (patients, physicians, developers, researchers, legal experts, ethicists, etc.).
• Utilized qualitative content analysis with MAXQDA software.
• Systematized identified problems through a project consortium workshop.

Main Results:

• Identified 309 expert statements on problems and barriers related to AI-based CDSSs.
• Categorized problems into technology (14.9%), data (19.1%), user (33%), studies (5.5%), ethics (6.5%), law (10.7%), and general (10.4%).
• User-related issues were the most frequently cited barriers.

Conclusions:

• Numerous challenges exist in the development and clinical use of AI-based CDSSs.
• Problems can be categorized by occurrence (general, development, clinical use) or by domain (technology, data, user, studies, ethics, law).
• Further investigation of these barriers is crucial for improving AI-based CDSS development, acceptance, and utilization.