Keywords: Causal estimand, principal stratification, ratio of means, PP, SACE
In clinical endpoint bioequivalence (BE) studies, the observed per-protocol (PP) population (compliers and completers in general) is usually used as the primary analysis for equivalence assessment. However, intercurrent events, i.e., missingness and noncompliance, are not properly handled and the resulting estimand is not causal. Previously, we proposed the first causal framework to assess equivalence with missing data and non-compliance by proposing a causal Survivor Average Causal Effect (SACE) estimand for difference of means (DOM). In equivalence assessment, DOM is not as widely-used as ratio of means (ROM). However, ROM is computationally more complicated and no existing formula links the observed PP estimand to the SACE estimand as DOM does. Herein we propose a similar causal frame for ROM, still using a principal stratification approach, one of the strategies recommended by the International Conference on Harmonisation (ICH) E9 (R1) addendum. We quantify the bias between the observed PP estimand and SACE estimand, which provides a basis to identify three conditions under which the two estimands are equal and for a sensitivity analysis method to evaluate the robustness of the current PP estimator to estimate the SACE estimand. Simulation demonstrates that the PP estimator is biased and may inflates Type 1 error and/or change power when the three identified conditions are violated. Our work can be applied to comparative clinical biosimilar and non-inferiority studies.