B0226
Title: Penalized doubly robust regression-based estimation of adaptive treatment strategies
Authors: Erica Moodie - McGill University (Canada) [presenting]
Zeyu Bian - McGill University (Canada)
Sahir Bhatnagar - McGill University (Canada)
Susan Shortreed - (United States)
Sylvie Lambert - (Canada)
Abstract: Adaptive treatment strategies (ATSs) are often estimated from data sources with many covariates measured, only a subset of which are useful for tailoring treatment or control of confounding. Including all available covariates in the analytic model could yield a needlessly complicated treatment decision, with poor statistical efficiency. Hence, we aimed to incorporate variable selection techniques into ATSs. Variable selection with the objective of optimizing treatment decisions has been the subject of very little literature. We will present a regression-based estimation method that can naturally incorporate variable selection through a penalization approach that incorporates sparsity while ensuring strong heredity, and show how we can additionally incorporate confounder selection into the approach. We illustrate the methods by analyzing a pilot sequential multiple assignment randomized trial of a web-based, stress management intervention using a stepped-care method for cardiovascular diseases patients to determine useful tailoring variables while adjusting for chance imbalances in important covariates due to the smaller sample size in the pilot.
