Best Linear Weighted Estimation for Two-phase Designs, with Application to Secondary Trait Association in Case-control Genetic Studies

Abstract

The inverse selection weighted estimator has been well applied to two-phase designs. Inverse-probability weighted estimators typically are less efficient than likelihood-based estimators, but are robust against model misspecification. In this paper, we propose a best linear weighted estimator for two-phase designs. Under some situations, the proposed best linear weighted estimator is equivalent to an augmented weighted estimator. One advantage of the proposed estimator is that there is no need to calculate the augmented term of the augmented weighted estimator. The estimator can be applied to general missing data problems. We extend the method to case-control genetic association studies, which are often nested in large-scale cohort studies and with collection of data on secondary traits such as biomarkers and quantitative traits. Key words: Case-control sampling; Inverse selection weighting; Missing at random.