Partner concordance in reports of joint risk behaviors

Objective: To evaluate the accuracy of self reports on sexual and drug use behaviors. Methods: Data from a network study of HIV transmission among a sample of dr ug users and nonusers are used to compare reports of sexual and drug use be haviors by partners who engaged in those behaviors. Partner concordance (se lf-report agreement between two people) was used as an estimate of validity . Results: Results showed that persons are able to recall and report about 85 % of their recent partners (15%-20% less for recent drug use partners). For relationships that were reported by both partners, a high degree of concor dance existed about recent behaviors (83%-96%) and variable agreement about frequency (0.48 less than or equal to r less than or equal to 0.86). When the time between interviews was longer, then recall, behavioral concordance , and agreement about frequency were lower. Conclusions: Underreporting of sex and drug partners may make HIV preventio n and intervention efforts more difficult if risk partners cannot be identi fied. The ability to reach out to all affected partners is critical in the effort to contain any epidemic. Underreporting may also skew epidemiologic projections on which many policy decisions are made.