Presentation (CTI)

OBJECTIVES: Artificial intelligence (AI) is increasingly applied in health economics and outcomes research (HEOR) to support real-world evidence generation, economic modeling, and payer and health technology assessment (HTA) decision-making. However, ethical AI principles are not often widely discussed and existing guidelines may not fully address ethical risks specific to HEOR, where AI outputs influence pricing, reimbursement, and population-level access to care. The objective of this study is to review any existing ethical frameworks and develop a HEOR-specific ethical framework for AI deployment with methodological and governance considerations relevant to economic evaluation and HTA decision contexts.
METHODS: We conducted a brief desk research and collected expert feedback on ethical guidelines related to use of AI in HEOR. We, then, drafted a conceptual framework synthesizing ethical AI principles, HEOR methodological standards, and HTA evidence requirements. Ethical risks were assessed across HTA lifecycle with particular attention to how data characteristics, model objectives, transparency practices, reproducibility, and governance structures influence downstream economic and access outcomes.
RESULTS: We propose a five-domain ethical framework for AI-enabled HEOR. First, though predictive accuracy is important, bias and equity should also be assessed based on downstream economic and access implications. Second, transparency should be operationalized as decision auditability, enabling traceability of data provenance, assumptions, and modeling choices relevant to HTA review. Third, reproducibility represents ethical concerns in HEOR, given the persistence of pricing and coverage decisions over time. Fourth, robust data governance is required to address privacy and consent challenges associated with large, linked datasets. Fifth, clear accountability and post-deployment monitoring are necessary to identify indirect ethical harms manifested through access and coverage outcomes.
CONCLUSIONS: Integrating economic impact, auditability, objective alignment, and governance into ethical evaluation can support more trustworthy AI-enabled HEOR evidence and more transparent reimbursement and policy decision-making.