Efficient semiparametric inference under two-phase sampling with applications to genetic association studies

摘要

In clinical studies where it is expensive to measure a large number of subjects, two-phase design is a cost-effective solution in which information on outcome and inexpensive covariates gathered in the first phase is used to select subjects for measurements of expensive covariates during the second phase. General two-phase designs are considered where the outcome can be continuous or discrete and inexpensive covariates can be continuous and at the same time they are correlated with expensive covariates. A semiparametric inference procedure is proposed to regression analysis by approximating the conditional density functions of expensive covariates given inexpensive covariates with B-spline sieves. A computationally efficient and numerically stable expectation-maximum algorithm is developed for maximizing the sieve likelihood. Extensive simulation studies are conduced to compare the performance of the proposed inference procedure with the existing ones. The asymptotic properties such as consistency, normality and efficiency of the estimators are also studied. Real time examples are presented to illustrate the application of the proposed semiparametric inference under two-phase sampling designs.