OBJECTIVES

: A systematic review of all oncology products’ approvals by the European Medicines Agency between 2009 and 2013 showed that most drugs entered the market without proven evidence of benefit on overall survival (OS). This research aims to explore the methods currently used to address this issue when submitting a health economic evaluation to support a reimbursement request, and their acceptability by 4 European Health Technology Assessment authorities.

METHODS

: A comparison of the methods of publicly available cost-effectiveness evaluations of two CAR-T cell therapies and two PARP inhibitors performed by HAS for France, NoMA for Norway, TLV for Sweden and NICE for England was performed. For each therapy, the available efficacy data and the methods used by the industrial to implement them into a cost-effectiveness model were reviewed. The analysis focused on four topics: the model structure, the assessment of relative effects, the use of surrogate and the extrapolation of data beyond trial follow-up. Then, a transversal qualitative analysis of the methodological expertise to assess the acceptability of each method by HTA authorities was conducted.

RESULTS

: Magnitude of effect benefit was highly uncertain mainly due to the lack of direct comparative data and to immature data. No model, no statistical method was optimal and the Incremental Cost-Effectiveness Ratios (ICER) were driven by several strong assumptions. To deal with this uncertainty, HTA authorities generally applied changes through more conservative approaches, which can lead to great higher ICER.

CONCLUSIONS : Since it is not possible to resolve the uncertainty about the OS benefit at an early assessment stage, the implementation of immature and/or indirect efficacy data in cost-effectiveness model should be based on rigorous justification and validation process and on comprehensive sensibility analysis. Furthermore, HTA authorities largely apply the principle of conservative choices in economic evaluations.