FROM POLICY TO PRACTICE: ASSESSING THE AI GOVERNANCE IMPLEMENTATION GAP ACROSS HEALTH TECHNOLOGY ASSESSMENT AGENCIES: A CROSS-NATIONAL READINESS REVIEW
Author(s)
Louise Perrault, PhD Candidate1, Nihad Khiat, PhD2, Nicole Tunstall, BSc, MSc3.
1International MarketAccess Consulting, Montreal, QC, Canada, 2Internationale MarketAccess Consulting, Montreal, QC, Canada, 3International Market Access Consulting, Montreal, QC, Canada.
1International MarketAccess Consulting, Montreal, QC, Canada, 2Internationale MarketAccess Consulting, Montreal, QC, Canada, 3International Market Access Consulting, Montreal, QC, Canada.
OBJECTIVES: Major health technology assessment (HTA) agencies have issued statements and preliminary guidance on artificial intelligence (AI) integration into their review processes. However, the gap between published policy intentions and actual operational readiness remains unmeasured. This study aimed to systematically map existing AI governance commitments across leading HTA agencies and identify domains where operational frameworks remain absent.
METHODS: A rapid structured literature review was conducted across four major HTA agencies (NICE, HAS, CDA, CMS) using publicly available guidance documents, strategic plans, policy statements, and peer-reviewed literature published between 2023 and 2026. Databases searched included PubMed, ISPOR.org, HTAi, and Cambridge Core. Key sources included the ISPOR Working Group Report on Generative AI (2025), the ELEVATE-GenAI reporting guidelines (2025), and the HTAi Global Policy Forum report (2026).
RESULTS: The review identified five major publications addressing AI in HTA published between February 2025 and January 2026, suggesting a rapidly emerging but nascent literature. All four agencies reviewed had issued AI-related strategic commitments; however, none had published an operational AI governance framework specific to HTA review processes. Transparency and explainability were the most frequently addressed domains across agency statements. Enforcement mechanisms, staff capacity requirements, and validation standards for AI-assisted outputs were consistently absent across all agencies reviewed.
CONCLUSIONS: A substantial gap exists between AI governance policy announcements and operational practice across leading HTA agencies. No agency has yet translated strategic commitments into enforceable, standardized governance requirements. These findings highlight an urgent need for the HTA community to move beyond principle-level statements and develop practical, auditable frameworks, a prerequisite for trustworthy AI-assisted decision-making in health technology assessment
METHODS: A rapid structured literature review was conducted across four major HTA agencies (NICE, HAS, CDA, CMS) using publicly available guidance documents, strategic plans, policy statements, and peer-reviewed literature published between 2023 and 2026. Databases searched included PubMed, ISPOR.org, HTAi, and Cambridge Core. Key sources included the ISPOR Working Group Report on Generative AI (2025), the ELEVATE-GenAI reporting guidelines (2025), and the HTAi Global Policy Forum report (2026).
RESULTS: The review identified five major publications addressing AI in HTA published between February 2025 and January 2026, suggesting a rapidly emerging but nascent literature. All four agencies reviewed had issued AI-related strategic commitments; however, none had published an operational AI governance framework specific to HTA review processes. Transparency and explainability were the most frequently addressed domains across agency statements. Enforcement mechanisms, staff capacity requirements, and validation standards for AI-assisted outputs were consistently absent across all agencies reviewed.
CONCLUSIONS: A substantial gap exists between AI governance policy announcements and operational practice across leading HTA agencies. No agency has yet translated strategic commitments into enforceable, standardized governance requirements. These findings highlight an urgent need for the HTA community to move beyond principle-level statements and develop practical, auditable frameworks, a prerequisite for trustworthy AI-assisted decision-making in health technology assessment
Conference/Value in Health Info
2026-11, ISPOR Europe 2026, Vienna, Austria
Value in Health, Volume 29, Issue 12S
Code
EE150
Topic
Economic Evaluation, Health Technology Assessment, Study Approaches
Disease
No Additional Disease & Conditions/Specialized Treatment Areas