Department seminar by Liqun Diao

Please Note: This seminar will be given online.

Adaptive Response-Dependent Two-Phase Designs: Some Results on Robustness and Efficiency

Large cohort studies routinely create biobanks in which biospecimens are stored for use in future biomarker studies. In such settings, response-dependent two-phase sampling designs involve sub-sampling individuals in the cohort, assaying their biospecimen to measure an expensive biomarker, and using the data to estimate key parameters of interest under budgetary constraints. In this talk, I will introduce a new scheme of optimal sub-sampling within the framework of adaptive two-phase designs, in which a preliminary phase of sub-sampling based on a standard design facilitates approximation of an optimal selection model for a second sub-sampling phase. I will describe how the new sub-sampling scheme can be used when analyses are based on inverse probability weighted estimating functions, likelihood or conditional likelihood, and consider the setting of a continuous biomarker where the nuisance covariate distribution is estimated nonparametrically at the design and analysis stages. Efficiency and robustness under different design strategies and analysis approaches will be discussed.