(210) Navigating artificial intelligence in evidence synthesis: A scoping review of current and developing guidelines

Background: : The integration of artificial intelligence (AI) is transforming health economics and outcomes research and market access by offering efficiencies in evidence synthesis, including systematic and pragmatic literature reviews. As AI tools become more sophisticated, it is essential that transparent, ethical guidelines are in place to promote the responsible application of those tools and maintain the integrity of managed care decision-making processes.

Objective: : This scoping review aimed to identify and characterize existing global guidelines and recommendations for the use of AI in evidence synthesis, with a particular focus on applicability and gaps relevant to systematic and pragmatic literature reviews.

Methods: : A scoping review was conducted in January 2025 and updated in April 2025, drawing from diverse sources including PubMed, Cochrane, the Centre for Reviews and Dissemination (CRD), Responsible AI in Evidence Synthesis (RAISE), and ISPOR - The Professional Society for Health Economics and Outcomes Research. The scoping review sought to delineate recommended AI workflows (e.g., human-in-the-loop), specific phases of evidence synthesis where AI is advised, and any identified limitations or ethical considerations.

Results: : Current guidelines broadly advocate for AI as an augmentative tool rather than a replacement for human expertise in evidence synthesis. Recommendations predominantly address systematic literature reviews, particularly focusing on tasks such as title/abstract screening and data extraction. However, substantial gaps were identified in existing guidelines regarding the application of AI to less structured, pragmatic literature reviews. Limitations on the use of AI for literature reviews included concerns about scientific validity, potential bias, and ethical implications.

Conclusions: : The landscape of AI in evidence synthesis is rapidly evolving, necessitating continuous development and refinement of guidance. This scoping review highlights a critical need for standardized, comprehensive, and adaptable guidelines that specifically address the unique challenges and opportunities of AI in pragmatic literature reviews and ensure robust, transparent, and equitable evidence generation.

Sponsorship:: OPEN Health HEOR & Market Access