ISSUE:

Advancements in therapies with novel mechanisms of action drive our need to consider emerging, time-varying network meta-analysis (NMA) approaches during HTA submissions. These methods, relaxing the proportional hazards (PH) assumption, can appropriately and maximally capture benefits over legacy treatments or placebo. However, detailed guidance including precedent use of time-varying NMA methods supporting HTAs, and their subsequent impact on cost-effectiveness analysis (CEA), is limited.

OVERVIEW:

Our panel addresses a recent NICE technology appraisal (TA858) in advanced renal cell carcinoma. Due to PH violations in the network, time-varying NMA were performed. However, their results were not applied for lack of interpretability and an inability to judge clinical plausibility for the trial and long-term survival extrapolation. In response, we present an enriched scenario of the MTA, demonstrating the advantages, limitations and differential impact on a CEA applying different NMA methods: PH, fractional polynomial, restricted cubic spline and parametric.

Increasing needs to account for time-varying relative efficacy are reflected in both NICE HTA requirements and the European Union (EU) Joint Clinical Assessment (JCA) guidelines. While crucial to accurately capture changes in comparative effectiveness, the complexity of associated methods compounds familiar challenges such as study and population heterogeneity in the network. This compounded project scope has implications both for the JCA and CEA-based HTA systems such as NICE.

Within the EU HTA setting, a crucial unknown is the individual member states’ receptivity of novel comparative efficacy methods. As the current role of HTA is very different per country, there may be differences between member states’ levels of preparedness to assess their application. Thus, exploration and reporting on how these methods can reduce uncertainties, what uncertainties remain, and how they impact CEA is key, so that some level of uniformity across member states may be achieved.