Validity of sexual histories in a prospective study of male sexual contacts of men with AIDS or an AIDS-related condition.

As part of a prospective study in Toronto, Canada of male sexual contacts of men with either acquired immunodeficiency syndrome (AIDS) or an AIDS-related condition (primary cases), separate interviews were conducted in 1984-1985 with primary cases and their corresponding sexual contacts. Seventy-five primary cases and their corresponding sexual contacts were asked details concerning the sexual activities involved in their sexual encounters. Spearman correlation coefficients were calculated for the responses regarding frequency of various sexual activities between the primary case and his corresponding contact. Comparison of responses revealed good to excellent correlations for the majority of the sexual activities. Regression analysis of responses (dependent variable = contact's response; independent variable = primary's response) revealed the tendency for primary cases to report greater numbers of various activities than sexual contacts, as illustrated by regression coefficients that were consistently less than unity. Agreement between primary responses and sexual contact responses appeared to be affected by the time lapsed from the last sexual encounter with the contact and the date of the primary case's interview. No effect on agreement was observed when primary case diagnosis and the total number of other sexual partners of primary cases were considered. The authors conclude that such data can be captured with sufficient reliability and validity for use in epidemiologic investigations to assess the risk of acquisition of human immunodeficiency virus for the more common forms of sexual activity. However, some caution may be necessary in interpreting the results for some activities (anal digit insertion, anilingus) since agreement is less satisfactory for these activities.