An application of propensity score weighting to quantify the causal effect of rectal sexually transmitted infections on incident HIV among men who have sex with men

• DOI: 10.1186/s12874-015-0017-y
• Publication Date: 2015-03-20T11:34:36Z
• Authors: Adam S Vaughan; Colleen F Kelley; Nicole Luisi; Carlos del Rio; Patrick S Sullivan; Eli S Rosenberg
• Keywords: Inverse probability weighting; Marginal structural model; Censoring (clinical trials)

Exploring causal associations in HIV research requires careful consideration of numerous epidemiologic limitations. First, a primary cause HIV, unprotected anal intercourse (UAI), is time-varying and, if it also associated with an exposure interest, may be on confounding path. Second, rare outcome, even high-risk populations. Finally, for most causal, non-preventive exposures, randomized trial impossible. In order to address these limitations and provide practical illustration efficient statistical control via propensity-score weighting, we examine the association between rectal STI acquisition InvolveMENt study, cohort Atlanta-area men who have sex (MSM). We hypothesized that, after controlling potentially behavioral demographic factors, significant STI-HIV would attenuate, but yield estimate effect.The interest was incident gonorrhea or chlamydia infection; outcome infection. To adjust confounding, while accounting limited infections, used inverse probability treatment weighted (IPTW) Cox proportional hazards (PH) model HIV. Weights were derived from propensity score modeling as function potential confounders, including UAI interval acquisition/censoring.Of 556 HIV-negative MSM at baseline, 552 (99%) included this analysis. 79 diagnosed 26 6 HIV-infected previously STI. unadjusted analysis, significantly subsequent (HR (95%CI): 3.6 (1.4-9.2)). final adjusted model, attenuated more precise (95% CI): 2.7 (1.2-6.4)).We found risk behaviors time-invariant diagnosis prior CT GC. Our analysis lends support effect provides framework similar analyses incidence.