Estimating the magnitude of STD cofactor effects on HIV transmission - Howwell can it be done?

Background: If sexually transmitted disease (STD) cofactor effects are stro ng and STDs are highly prevalent, STD control can be a strategy for HIV pre vention. Objective: To review possibilities for estimating cofactor effects of STDs on HIV transmission based on observational studies. Study Design: This study consisted of an analysis of factors influencing as sociations between HIV and STDs, which can bias STD cofactor studies, from a sexual network perspective. Study designs that reduce distortions and met hods to improve estimates in the presence of confounding are discussed. Results: Standard statistical adjustments of cofactor estimates are insuffi cient because they ignore clustering between HIV and STDs in partners of st udy subjects, resulting from population heterogeneity in risk factors and a ssortative mixing. Reverse causation due to HIV-related immunosuppression m ay further inflate cofactor estimates. Misclassification of STD and cluster ing between STD can bias estimates in either direction. This study demonstr ates quantitatively that ignorance of sexual network effects may result in considerable overestimation of cofactor magnitudes. Conclusion: The limitations of observational studies complicate quantitativ e inferences on the role of STD in HIV transmission.