|
Activity Number:
|
263
|
|
Type:
|
Topic Contributed
|
|
Date/Time:
|
Tuesday, August 4, 2009 : 8:30 AM to 10:20 AM
|
|
Sponsor:
|
Section on Health Policy Statistics
|
| Abstract - #303925 |
|
Title:
|
Maximum Likelihood Estimation of the Structural Nested Mean Model Using SAS PROC NLP
|
|
Author(s):
|
Daniel Almirall*+ and Cynthia Coffman and William S. Yancy, Jr. and Susan Murphy
|
|
Companies:
|
Duke University and Duke University Medical Center and Duke University and University of Michigan
|
|
Address:
|
101 Thomas Lane, Unit E2, Carrboro, NC, 27705,
|
|
Keywords:
|
causal inference ; structural nested mean model ; maximum likelihood estimator ; time-varying effect moderation
|
|
Abstract:
|
We review Robins' Structural Nested Mean Model (SNMM) for assessing the effect of predictors that vary over time. The SNMM is used to study the effects of time-varying predictors (or treatments) in the presence of time-varying covariates that are moderators of these effects. We describe a maximum likelihood estimator (MLE) of the parameters of a SNMM (implemented in SAS PROC NLP). The MLE requires correct model specification of the distribution of the primary outcome given the history of time-varying moderators and predictors, including proper specification of both the causal and non-causal portions of the SNMM. Consistency and sequential ignorability are required for causal inference. We illustrate the method by assessing the impact of early versus later weight loss on health-related quality of life as a function of time-varying covariates thought to be moderators of these effects.
|
