Reporting guidelines for artificial intelligence in healthcare research.

Area of Science:

• Medical Informatics
• Artificial Intelligence in Healthcare
• Research Methodology

Background:

• Reporting guidelines are crucial for ensuring the quality and transparency of research studies.
• Studies involving artificial intelligence (AI) interventions present unique challenges and potential biases.
• A proliferation of AI-specific reporting guidance necessitates a framework for selection.

Purpose of the Study:

• To provide researchers with principles for selecting appropriate reporting guidelines for AI intervention studies.
• To enhance the quality, design, delivery, and reporting of AI-driven research.
• To address potential sources of bias unique to AI in research.

Main Methods:

• This review synthesizes existing reporting guidelines for research studies.
• It identifies key principles for selecting guidelines applicable to AI interventions.
• The review emphasizes the hierarchy of guideline applicability based on study design.

Main Results:

• The quality of a study is primarily determined by its design, not the intervention itself.
• Reporting guidelines specific to the study design should be prioritized.
• AI-specific reporting guidance should supplement, not replace, design-specific guidelines.

Conclusions:

• Selecting appropriate reporting guidelines is essential for high-quality AI research.
• Prioritizing study design guidelines ensures methodological rigor.
• Integrating AI-specific guidance enhances transparency and mitigates bias in AI studies.