Fig. 2.

The flowchart depicts the entire confirmatory testing strategy. This implementation will allow the investigator to make inferences on the overall population and/or on a prespecified subgroup (cold storage). The methodology developed by Alosh and Huque [31], which requires meeting some consistency constraints as a prerequisite for testing the next hypothesis, will be used. This proposed methodology requires first testing global efficacy at a reduced alpha level α0. Based on this result and on a consistency constraint within the complementary subgroup (machine perfusion), the efficacy for the cold storage subgroup may also be tested. The variability (standard deviation) is assumed to be similar in the two subgroups. Therefore, the significance level for the subgroup of interest αS is derived from the proportions of the subgroups, which are expected to be balanced within the global sample, i.e., 50% cold storage and 50% machine perfusion. This yields an αS equal to 0.02633 (two-sided) to control for the familywise error rate (FWER) strongly