Design and Analysis of a Biosimilar Bridging Study with a Prediction Interval-Based Consistency Test

The development of biosimilars has substantially increased in recent years. A biosimilar is a biological product which is highly similar to a licensed biological product (reference product), with no clinically meaningful differences between the proposed biosimilar and the reference product. A bridging study is a viable strategy to bring an approved biosimilar product from the original region(s) to new regions or countries. While the bridging concept and the principles of conducting bridging studies for innovative products outlined in ICH E5 guidance are informative for biosimilar development, existing statistical strategies for designing and analyzing a bridging study may not be directly applicable when the biosimilar global study compares the biosimilar product to a reference product as control. In this paper, we present a novel and practical statistical methodology for designing and analyzing a biosimilar bridging study based on a prediction interval-based consistency test for assessing consistency between the bridging study and the biosimilar global study. We prove analytically that the method has the desired statistical properties (i.e., high power when the two studies are perfectly consistent or nearly consistent in efficacy and type I error control when the two studies are clearly inconsistent in efficacy). We present a numerical example to illustrate that the implementation of the method is straightforward in practice. We also discuss the application of the method to various endpoints and metrics.