PNS263 THE ROLE OF SURROGATES FOR PATIENT-RELEVANT ENDPOINTS IN THE EARLY BENEFIT ASSESSMENT IN GERMANY

Abstract

Clinical trials conducted for regulatory approval mainly evaluate the efficacy of drugs by objective biomarkers rather than by patient-relevant endpoints. In the German early benefit assessment however, only patient-relevant endpoints are accepted. If surrogate endpoints are used instead, they need to be validated. We investigated the role of surrogate endpoints for early benefit assessment by the Institute for Quality and Efficiency in Health Care (IQWiG) and the Federal Joint Committee (G-BA). All early benefit assessments since the Act on the Reform of the Market for Medicinal Products (AMNOG) in 2011 until September 2018 were examined for surrogate endpoints in the manufacturer’s dossiers or defined as such during dossier assessment by IQWiG/G-BA. For all identified surrogate endpoints, the type of justification by the manufacturer and their final acceptance were evaluated. All 331 early benefit assessments were analyzed. In total 363 surrogate endpoints (multiple mentions possible) in 84 (25%) assessments were identified. Of these surrogate endpoints 9% were taken into account for the benefit assessment and 14% were accepted as surrogate or patient-relevant endpoint for the final decision by the G-BA. The following surrogate endpoints were accepted: virologic response and CD4 cell count in human immunodeficiency virus infection, reduction of oral corticosteroids in asthma and systemic lupus erythematosus, and HbA1c level in type 1 diabetes. However, reasons for acceptance sometimes differed from the manufacturer’s argumentation. Acceptance of surrogate endpoints for the added benefit decision is the exception. Although surrogate endpoints are commonly included in the manufacturer’s dossiers to claim benefit, only few manufacturers attempted to validate them. Instead, surrogates are typically either justified qualitatively or claimed to be patient-relevant endpoints. To increase the future acceptance of surrogate endpoints, more rigorous validation processes are necessary.