Pharmacoeconomics

Assessing Cancer Drugs for Reimbursement: Methodology, Relationship between Effect Size and Medical Need

¹ Laboratoire Roche, Neuilly sur Seine, France
² Centre Hospitalier Universitaire de Besançon, Besançon, France
³ Centre Hospitalier Universitaire de Bordeaux, Bordeaux, France

Correspondence and offprints : Rima de Sahb-Berkovitch, Affaires ´Economiques, Laboratoire Roche, 52 boulevard du Parc, 92521 Neuilly sur Seine Cedex, France. E-mail: Cette adresse e-mail est protégée contre les robots spammeurs. Vous devez activer le JavaScript pour la visualiser.

Received: 10 February 2010
Accepted: 26 April 2010

Abstract

Reimbursement is assessed by the Transparency Commission from the Health Authority (HAS) using a medical benefit (SMR) score that gives access to reimbursement, an “improvement of medical service rendered” (ASMR) that determines the added therapeutic value, and the target population. Assessing cancer drugs for reimbursement raises the same issues as other therapeutic classes, with some key differences. Overall survival (OS) is considered by the Transparency Commission as the endpoint for assessing clinical benefit, and yet it is not an applicable primary endpoint in all types of cancer. Later lines of treatment, particularly during the development process, may make it difficult to interpret OS as the primary endpoint. Therefore, progression-free survival (PFS) for metastatic situations and disease-free survival (DFS) in adjuvant situations are wholly relevant endpoints for decisions on the reimbursement of a new cancer drug. Effect size is assessed using actuarial survival curves of the product versus the comparator, and it is difficult to summarise them into one single parameter. Results are generally interpreted based on median survival, which is fragmented because it only measures one point of the curve. The hazard ratio measures the effect of treatment throughout the duration of survival and is therefore more comprehensive in quantifying clinical benefit. Determining an effect size threshold for granting reimbursement is difficult given the diversity of cancer settings and the level of medical need, which influences assessment of the clinical relevance of the observed difference. Rapid progress in comparators (700 molecules in development) and the identification of predictive factors of efficacy (biomarkers, histology, etc.) during development may lead to different ASMR scores per population, or to the restriction of the target population to a subgroup of the marketing authorisation (MA) population in which the expected effect size is greater. To address these issues, the roundtable recommends the possibility of early scientific opinions by the office of the Transparency Commission in order to discuss comparators and the relevance of responder subgroups. It also recommends the possibility of granting a temporary ASMR, on condition of subsequent confirmation by production of data, when reimbursement appears justified in a subpopulation of the MA for which only subgroup analysis is available.