Augmented case-only designs for randomized clinical trials with failure time endpoints

摘要: Under suitable assumptions and by exploiting the independence between inherited genetic susceptibility treatment assignment, case-only design yields efficient estimates for subgroup effects gene-treatment interaction in a Cox model. However it cannot provide of main effect baseline hazards, that are necessary to compute absolute disease risk. For two-arm, placebo-controlled trials with rare failure time endpoints, we consider augmenting random samples controls from both arms, as classical case-cohort sampling scheme, or sample active arm only. The latter is motivated vaccine cost-effective use resources specimens so host genetics vaccine-induced immune responses can be studied simultaneously bigger set participants. We show these designs identify all parameters model estimator incorporated two-step plug-in procedure. Results simulations data example suggest incorporating estimators improves precision estimated parameters; only attains similar level efficiency.