Optimal auxiliary‐covariate‐based two‐phase sampling design for semiparametric efficient estimation of a mean or mean difference, with application to clinical trials. (9th October 2013)

• Record Type: Journal Article
• Authors: Gilbert, Peter B.; Yu, Xuesong; Rotnitzky, Andrea
• Is Part Of: Statistics in medicine. Volume 33:Number 6(2014)
• Page Start: 901
• Page End: 917
• Publication Date: 2013-10-09
• Subjects: Medical statistics -- Periodicals; Statistique médicale -- Périodiques; Statistiques médicales -- Périodiques; 610.727
• DOI: 10.1002/sim.6006
• Languages: English
• ISSNs: 0277-6715

Abstract

<abstract abstract-type="main" id="sim6006-abs-0001"> <title> <x xml:space="preserve">Abstract</x> </title> <p id="sim6006-para-0001">To address the objective in a clinical trial to estimate the mean or mean difference of an expensive endpoint <italic>Y</italic>, one approach employs a two‐phase sampling design, wherein inexpensive auxiliary variables <italic>W</italic> predictive of <italic>Y</italic> are measured in everyone, <italic>Y</italic> is measured in a random sample, and the semiparametric efficient estimator is applied. This approach is made efficient by specifying the phase two selection probabilities as optimal functions of the auxiliary variables and measurement costs. While this approach is familiar to survey samplers, it apparently has seldom been used in clinical trials, and several novel results practicable for clinical trials are developed. We perform simulations to identify settings where the optimal approach significantly improves efficiency compared to approaches in current practice. We provide proofs and R code. The optimality results are developed to design an HIV vaccine trial, with objective to compare the mean 'importance‐weighted' breadth (<italic>Y</italic>) of the T‐cell response between randomized vaccine groups. The trial collects an auxiliary response (<italic>W</italic>) highly predictive of <italic>Y</italic> and measures <italic>Y</italic> in the optimal subset. We show that the optimal design‐estimation approach can confer anywhere between absent and large efficiency gain (up to 24 <italic>%</italic> in the examples) compared to the approach with the same efficient estimator but simple random sampling, where greater variability in the cost‐standardized conditional variance of <italic>Y</italic> given <italic>W</italic> yields greater efficiency gains. Accurate estimation of <italic>E</italic>[<italic>Y</italic> | <italic>W</italic>] is important for realizing the efficiency gain, which is aided by an ample phase two sample and by using a robust fitting method. Copyright © 2013 John Wiley & Sons, Ltd.</p> </abstract>