Two-Phase Outcome-Dependent Sampling Design with Interval-Censored Failure Time Data

• Activity Number: 113 - Recent Advances in Design and Analysis of Two-Phase Studies
• Type: Topic Contributed
• Date/Time: Monday, July 30, 2018 : 8:30 AM to 10:20 AM
• Sponsor: Lifetime Data Analysis Interest Group
• Author(s): Qingning Zhou* and Jianwen Cai and Haibo Zhou
• Companies: University of North Carolina at Charlotte and University of North Carolina and University of North Carolina
• Keywords: Bernstein polynomial; Estimated likelihood; Missing data; Response-selective sampling; Semiparametric regression

Abstract

We propose a two-phase outcome-dependent sampling design and an inference procedure for studies that concern interval-censored failure time outcomes. This design enhances the study efficiency by allowing the selection probabilities of the second-phase sample, for which the expensive exposure variable is ascertained, to depend on the first-phase observed interval-censored failure time outcomes. The second-phase sample is enriched by selectively including subjects who are known or observed to experience the failure at an early or late time. We develop a sieve semiparametric maximum likelihood approach that makes use of the available data from both phases. The resulting regression parameter estimator is shown to be consistent and asymptotically normal, and a consistent estimator for its asymptotic variance is derived. Simulation results demonstrate that the proposed design and inference procedure performs well in practical settings and is more efficient than the alternative designs and methods. An application to diabetes data from the Atherosclerosis Risk in Communities (ARIC) study is provided.