Introduction
For many ATMPs, clinical benefit is expected to persist over long periods following one-time or short-course administration, while pivotal evidence is often based on short follow-up, limited sample sizes, and sometimes non-randomised or single-arm designs.
Durability of effect and long-term safety are therefore frequent sources of uncertainty in access decision-making. The access professional therefore has to address topics including: the framing of long-horizon outcomes within clinical development programmes; the use of modelling and evidence synthesis to extrapolate beyond observed data; and post-authorisation evidence generation, including registries and real-world evidence, to support reassessment over time. These activities have to be carried out in the context of how payers and HTA bodies make decisions linked to evidence maturity, and with an understanding of how outcomes-linked contracting approaches depend on endpoint definition, data capture, and administrative feasibility.
Why durability is a central evidential issue for ATMPs
For one-time therapies, uncertainty about long-term efficacy and long-term safety is often highlighted as a core barrier to reimbursement decisions. Short pivotal follow-up can be difficult to reconcile with claims of lifelong benefit, and HTA may need to judge whether observed effects are likely to persist, wane, or require subsequent interventions.
Uncertainty is reinforced when pivotal datasets rely on small samples, single-arm designs, surrogate endpoints, or indirect comparators. These features are not unique to ATMPs, but they occur frequently because many ATMPs target rare conditions and have development programmes constrained by feasibility.
Evidence strategy choices that affect perceived durability
Durability arguments usually depend on the internal logic of the clinical programme as much as the observed follow-up. Key design choices include the clinical relevance of endpoints, the plausibility of the biological mechanism for sustained benefit, and the extent to which early response is predictive of longer-term outcomes.
The strategy also involves planning the transition from pre-authorisation evidence to post-authorisation evidence, including clarity on which uncertainties are expected to be resolved later, and how. In practice, this means specifying what will be measured, for how long, and in which patient groups, rather than treating “long-term follow-up” as a generic commitment.
Regulators may also shape evidence strategy directly through long-term follow-up expectations for gene therapy products, reflecting concerns about delayed adverse events and durability.
Evidence synthesis and modelling for long horizons
Economic models for ATMPs can be highly sensitive to assumptions about duration of effect, time-to-event extrapolation beyond observed data, and downstream healthcare utilisation. In this context, modelling is not simply a calculation step, but an explicit representation of uncertainty and its decision consequences.
Uncertainty in long-term relative effectiveness is a recurring challenge for medicines with limited follow-up, and approaches to manage this uncertainty include scenario analyses, structured sensitivity analysis, and iterative reassessment as new data emerge.
Real-world evidence (RWE) is often positioned as a means to reduce uncertainty after launch, but its usefulness depends on whether real-world endpoints, timing, and data quality align with the specific decision uncertainties. NICE’s RWE framework, for example, is explicitly oriented towards deciding when RWE can help and what “good practice” looks like for planning and reporting.
Post-authorisation evidence generation and registries
Post-authorisation evidence plans commonly rely on registries, routine data, and structured follow-up protocols. This is partly a scientific need and partly a system requirement when payment or reassessment depends on measured outcomes.
A widely cited example is the use of monitoring registries linked to managed entry agreements in Italy, where outcomes tracking and staged payment approaches have been used for some high-cost therapies. The practical limits are also repeatedly noted: outcome specification may be contested, data capture may be incomplete, and administrative workload can be material for providers and payers.
Long-term follow-up is also treated as a safety and effectiveness requirement in gene therapy contexts. The United States Food and Drug Administration (FDA) guidance sets expectations for long-term monitoring after administration of human gene therapy products.
Implications for access decisions and reassessment
Across countries, mechanisms for reassessment and conditional approaches vary, but a common pattern is that initial decisions may be coupled to explicit evidence collection, with later review points. This matters because durability is often not proved at launch. Instead, payers determine whether the remaining uncertainty is acceptable given disease severity, alternatives, and the credibility of the evidence-generation plan.
There is also a pragmatic interaction between evidence strategy and contracting. Outcomes-based agreements and staged payment approaches are often described as ways to accommodate uncertainty, but they only function when outcomes are measurable, attributable, and recorded consistently over time.
Summary
ATMPs frequently require access decisions to be made with limited long-term follow-up, making durability a central driver of evidential uncertainty. Evidence strategy therefore depends on both pre-authorisation design choices and a credible plan for post-authorisation evidence generation, often using registries and routine data. Modelling is an important intermediary because it converts limited observation into long-horizon estimates, while making assumptions about duration of effect explicit and testable. Regulatory expectations for long-term follow-up in gene therapy further reinforce the need for structured monitoring beyond pivotal trials. Finally, where contracting or reassessment depends on outcomes, evidence plans become operational requirements, and limitations in data capture and administration can become binding constraints.
Links
Italian Medicines Agency (Agenzia Italiana del Farmaco, AIFA): Monitoring Registers
UK National Institute for Health and Care Excellence (NICE): NICE real-world evidence framework
Scientific publications
Avşar F, Roubin L, Sermet C, et al. Linking reimbursement to patient benefits for advanced therapy medicinal products: A review of outcomes-based agreements in Europe. Value Health. 2024. doi: 10.1016/j.jval.2024.10.023.
Greco A, Frederix GWJ, Hooft L, Ten Ham RMT. A systematic review of challenges and opportunities in the implementation of managed entry agreements for advanced therapy medicinal products. Clin Ther. 2025;47(2):e16-e26. doi: 10.1016/j.clinthera.2024.11.019.
Rohde M, Huh S, D'Souza V, Arkin S, Roberts E, McIntosh A. Practical and Statistical Considerations for the Long Term Follow-Up of Gene Therapy Trial Participants. Clin Pharmacol Ther. 2024;115(1):139-146. doi:10.1002/cpt.3087
Versteeg JW, Vreman R, Mantel-Teeuwisse A, Goettsch W. Uncertainty in Long-Term Relative Effectiveness of Medicines in Health Technology Assessment. Value Health. 2024;27(10):1358-1366. doi:10.1016/j.jval.2024.05.023.
Wagner, T.D., Riposo, J.W., Gould, K.M. et al. Innovative Contracting for Gene Therapies: Current Landscape and Perspectives on the Future of Gene Therapy Financing in the USA.PharmacoEconomics (2025). doi:10.1007/s40273-025-01563-3.