Re: st: Propensity Score Matching on Individuals with Multiple Records (Time-Varying Covariates)

["Ariel Linden, DrPH" <ariel.linden@gmail.com>]

You can perform propensity score-based weighting on longitudinal data which
accounts for time varying covariates. Jamie Robins at Harvard has made a
cottage industry of this approach (which he refers to as marginal structural
modeling). See the following references (to name a few):

Robins JM, Hernán MA, Brumback B. Marginal structural models and causal
inference in epidemiology. Epidemiol 2000;11:550?60.

Robins JM. Marginal structural models. In: 1997 Proceedings of the Section
on Bayesian Statistical Science. Alexandria, VA: American Statistical
Association, 1998:1?10.

Hernán MA, Brumback B, Robins JM. Marginal structural models to estimate the
causal effect of zidovudine on the survival of HIV-positive men. Epidemiol
2000;11:561?70.

Date: Thu, 10 May 2012 14:42:44 -0500
From: "Weichle, Thomas" <Thomas.Weichle@va.gov>
Subject: Re: st: Propensity Score Matching on Individuals with Multiple
Records (Time-Varying Covariates)

I suppose the matching could be performed without reference to the
time-dependent variables.  However, after the match, I would like to
perform the survival regression including the time-dependent variables.

In contrast, the only work that I was able to come across dealing with
this situation is by Bo Lu (2005), summarizing that it is desirable to
balance the distribution of the covariates at every time point in
observational studies with a time-dependent treatment and time-dependent
covariates.

Lu, B. (2005), Propensity Score Matching with Time-Dependent Covariates.
Biometrics, 61: 721-728.

Tom Weichle
Math Statistician
VA Information Resource Center (VIReC)
Edward Hines Jr. VA Hospital, Bldg 18, 202E
708-202-8387 ext. 24261
Thomas.Weichle@va.gov 




*
*   For searches and help try:
*   http://www.stata.com/help.cgi?search
*   http://www.stata.com/support/statalist/faq
*   http://www.ats.ucla.edu/stat/stata/