Beyond evidence hierarchies: governing clinical uncertainty in ATMP assessment in Europe. A focus group study

Introduction Advanced Therapy Medicinal Products (ATMPs) pose significant challenges for health technology assessment (HTA) due to high levels of clinical uncertainty at the time of assessment. These uncertainties are often driven by reliance on single-arm trials, surrogate endpoints, and limited long-term data. While methodological guidance exists, less is known about how such uncertainty is interpreted and managed in real-world decision-making across stakeholders. This study aims to explore the perspectives of European stakeholders on key evidentiary challenges and approaches to clinical uncertainty management in ATMP assessment. Materials and methods Two online focus group discussions were conducted with stakeholders involved in ATMP development and assessment, including representatives from academia, industry, HTA bodies, and payer organisations. The first focus group (n = 7) addressed challenges related to comparator selection, while the second (n = 5) focused on long-term outcomes and surrogate endpoints. Discussions were guided by a semi-structured topic guide informed by a literature review and were audio- and video-recorded, transcribed verbatim, and analysed using thematic framework analysis. Results Across both focus groups, uncertainty emerged as a central and persistent feature of ATMP assessment. According to participants, evidence is evaluated within a broader clinical and contextual framework, where factors such as disease severity, unmet need, and expected treatment effect shape the acceptability of uncertainty. Single-arm trials and surrogate endpoints were considered conditionally acceptable, particularly when randomised evidence is infeasible, but were associated with important limitations. Indirect treatment comparisons were viewed as methodologically constrained yet often necessary in the absence of comparative data. Evidence generation was described as a question-driven and iterative process, balancing methodological rigour with feasibility and ethical considerations. Biological plausibility and anticipated clinical benefit influenced evidence interpretation, while real-world evidence was seen as important but limited in its ability to resolve uncertainty post-launch. Participants noted that uncertainty is rarely eliminated and is instead managed throughout the product lifecycle. Conclusion Clinical uncertainty in ATMP assessment appears to be an inherent feature of decision-making rather than a temporary consequence of incomplete evidence. Addressing this requires more flexible, context-sensitive, and lifecycle-oriented approaches to evidence generation and assessment.