G-Estimation of Structural Nested Model Parameters for Optimal Dynamic Treatment Regimes: Looking for Problems

• Activity Number: 451
• Type: Topic Contributed
• Date/Time: Wednesday, August 4, 2010 : 8:30 AM to 10:20 AM
• Sponsor: Biometrics Section
• Author(s): Erica E.M. Moodie*+
• Companies: McGill University
• Keywords: Structural nested models ; Optimal dynamic treatment regimes ; G-estimation ; Non-regularity ; Residuals

Abstract

To estimate the sequence of actions that optimizes response in a longitudinal setting, it is important to study the actions as a set rather than individually. Many statistical challenges arise in the estimation of optimal dynamic regimes including model choice and non-regularity. While semiparametric methods have found favour in recent years, little attention has been paid to model adequacy. In this talk, residual diagnostic plots for G-estimation of structural nested mean models which can aid in variable selection are described. Non-regularity and methods to suggest its existence are also discussed. Motivation is drawn from the estimation when to wean in the Promotion of Breastfeeding Intervention Trial (PROBIT). There are many covariates to consider in this study and previous work indicates non-regularity.