5.You are interested in assessing the quality of a number of studies designed to examine the role of male circumcision in preventing acquisition of HIV by men through heterosexual intercourse. For each of the following study descriptions, state whetherinformation bias,selection bias,orconfoundingis MOST likely to affect the measure of association and explain why.

a)In a cross-sectional study in the general population of the relationship between circumcision and HIV acquisition, 1014 men were eligible for the study and invited to participate, and 626 agreed to participate. Of those, 124 men were excluded from the analysis because they reported not being sexually active.

b)In a case-control study among a group of truck drivers at high risk for HIV, a rapid HIV test was used without any confirmatory tests. The rapid test is known to be less accurate than more laboratory-intensive testing methods. People carrying out the HIV testing were blinded to circumcision status.

c)In a cohort study of a group of men in Mozambique at high risk for HIV, all men who reported having had sex with a sex worker in the past month and were HIV-uninfected were invited to participate. Of those, 370 men were enrolled. There was unequal follow-up between the circumcised and uncircumcised groups (follow-up 62% and 27%, respectively).

d)You look at 16 cross-sectional studies to compare the clinical confirmation versus self-report of circumcision status. All 6 of the studies that use clinical confirmation indicate a benefit of circumcision in terms of reduced HIV infection (OR range = 0.28-0.95). In the 10 studies based on self-report, the effect estimates varied, with 4 reporting a protective effect of circumcision (OR range = 0.40-0.97) and 6 reporting an increased risk (OR range = 1.2-1.9). What is likely to account for the discrepancies between the studies? (Note: This is hard because it is discussing many studies and we have only been working with one at a time.Try to use the information that you understand about biases in single studies to answer this question)

e)A prospective cohort study followed HIV-uninfected men in Botswana for 8 years to identify risk factors for HIV infection. The investigators were surprised to find that circumcision was not protective against infection. They also found that men who were circumcised had a significantly higher number of sex partners than men who were uncircumcised.