Measuring Sex Partner Concurrency: It’s What’s Missing That Counts

Publisher Website
Google Scholar
Abstract
Sex partner concurrency is an important determinant of STI transmission dynamics, yet its measurement is not standardized. : We assessed the agreement, compared correlates, and investigated data quality and completeness between 2 common concurrency measures. Young adults (ages 18-26) attending public STD clinics between 2001 and 2004 in Seattle, St. Louis, and New Orleans, provided data on 2 or more sex partners in a computer-administered survey interview (N = 680). Concurrency with last partner was measured in 2 ways: (a) a direct question about other sexual contacts during the most recent sexual relationship and (b) overlapping start and end dates of the 2 most recent relationships. Although 56% reported concurrency by direct questioning and 54% by overlapping dates, the kappa statistic for agreement between measures was only fair (0.395). Indeed, 29% of those reporting concurrent partners by the direct question did not do so by overlapping dates and 26% of participants concurrent by overlapping dates were not concurrent by the direct question. Each of the measures had dissimilar correlates, and concurrency data were missing or uninterpretable more often for the overlapping dates measure (21.3%) than the direct question (1.8%). Concurrency was common by both measures but the measures were not interchangeable. Although the overlapping dates measure provided information about partnership duration, it is subject to missing or uninterpretable data. The direct question substantially minimized the amount of missing data and may be more appropriate for use with computer-administered survey interview.