Dear Statausers:
One question regarding the G-estimation that address the "time-dependent
confounding bias.
Currently, I am working on an issue, try to identify the association of HIV
with Overdose mortality among a Injection Drug User Cohort and explore the
potential reason if anwer is yes. One important time-varying confounder
should be considered here is: Drug use behavior. Since HIV seroconversion
status can affect the drug use behavior (people may use less drug after
knowing their infection status), which can also affect on the outcome later
(higher risk associated with heavier drug use). Here I am just wondering: is
there a "time-dependent confounding" bias concern here? If so, is
G-estimation appropriate to solve the problem?
By the way, also wonder whether you know any resource for the analysis with
the alternative of Marginal Structure Model, by using Stata, SAS, or any
other program?
Appreciate your help. Thanks in advance.
Sincerely.
Cunlin Wang, MD, Ph.D Candidate
Johns Hopkins School of Public Health
Tel: 410-614-3632 / Fax: 410-614-9910
*
* For searches and help try:
* http://www.stata.com/support/faqs/res/findit.html
* http://www.stata.com/support/statalist/faq
* http://www.ats.ucla.edu/stat/stata/
