Similarity assessment of quality attributes of biological medicines: the calculation of operating characteristics to compare different statistical approaches

Abstract

The comparison of quality attributes is a key element in the evaluation of both biosimilars and manufacturing process changes for biological medicines. Different statistical approaches are proposed to facilitate such evaluations. However, there is no regulatory consensus on a quantitative and scientifically justified definition and an underlying hypothesis of a statistically equivalent quality. The latter is essential to calculate operating characteristics of different approaches. This article proposes a hypothesis for establishing statistically equivalent quality which is concordant with current regulations. It also describes a tool which allows comparisons of different statistical approaches or tests by calculating the operating characteristics for false acceptance and false rejection rates of a claim for statistically equivalent quality. These error rates should be as low as possible to allow a meaningful application of a statistical approach in regulatory decision making. The described tool can be used to compare different statistical approaches for their suitability and may also facilitate the discussion and development of statistical approaches for comparing quality attributes in similarity assessments in general.