|
Abstract:
|
Randomized clinical trials are often designed to assess whether a test treatment prolongs survival relative to a control treatment. Increased patient heterogeneity, while desirable for generalizability of results, can weaken the ability of common statistical approaches to detect treatment differences, potentially hampering the regulatory approval of safe and efficacious therapies. A novel solution to this problem is proposed. First, blinded to patient-level treatment assignment, subjects are segmented into prognostically homogenous subgroups (risk strata). After unblinding, time ratios from accelerated failure time model fits with model averaging are computed within each stratum and stratum-level results are combined for overall statistical inference. Our proposed 5-step stratified testing and amalgamation routine (5-STAR) satisfies both estimand-alignment and assumption-reduced analysis per ICH E9/R1 guidance and has increased power relative to that of the logrank test and other common approaches that do not leverage structured patient heterogeneity, as demonstrated by simulations and real data applications.
|
