Presentation (CTI)

OBJECTIVES: HTA frameworks built for drugs and devices poorly assess AI-enabled precision medicine—diagnostics, biomarker stratification, and response prediction—because these technologies are data- and context-dependent, often rely on single-arm or surrogate evidence, and evolve post-deployment; we developed Value-to-Access (V2A) to make this value evidence decision-ready for reimbursement.
METHODS: We conducted a structured synthesis of peer-reviewed and grey literature (2018-2026) sourced from ISPOR/Value in Health, npj Digital Medicine, The Lancet Digital Health, the Journal of Medical Internet Research, and EU and US regulatory bodies in five HTA domains: clinical effectiveness, economics, safety, organisational impact, and ethics/equity. We mapped findings to assessment methods—cost-utility analysis, multi-criteria decision analysis (MCDA), outcome-based agreements, and conditional reimbursement—and appraised them against the CHEERS 2022 economic-reporting standard, feasibility, transferability, and applicability.
RESULTS: V2A comprises four modules plus a cross-cutting AI layer. The evidence-readiness module sets clinical-validity and utility thresholds and accommodates single-arm/surrogate evidence and generalizability concerns. The economic module reframes cost-effectiveness around diagnostic-accuracy-driven value, capturing downstream treatment consequences, implementation and infrastructure costs, and lifecycle reassessment for adaptive technologies. The broader-value module applies MCDA to equity, organisational impact, and transparency. The access-linkage module ties evidence maturity to reimbursement pathways—robust evidence to standard appraisal, residual uncertainty to conditional, monitoring-linked agreements—consistent with the EU HTA Regulation. The cross-cutting layer addresses AI-specific traits (data-dependence, retraining, context sensitivity) and supports AI-assisted evidence synthesis for resource-limited agencies. Applied to precision-oncology cases, V2A converts AI-specific uncertainty into decision-ready evidence suitable for conditional access.
CONCLUSIONS: V2A offers agencies a transparent method to assess AI-enabled precision medicine and developers a value-oriented route to market; we recommend prospective, multi-jurisdictional validation as the next step.