Using Choice-based Conjoint Exercise to Assess Self-reported Serosorting Among Men Who Have Sex with Men (MSM) Cruising the Internet for Sex

Background:
Serosorting among HIV-positive men may be a promising harm reduction approach to prevent the spread of HIV. Profiles on cruising websites often include HIV status, allowing men to serosort. Indeed, several studies have reported high levels of serosorting among MSM. However, various factors may affect the likelihood of serosorting and furthermore, self-reported serosorting may differ from actual practice.

Objective:
To compare self-reported serosorting to serosorting in a conjoint trade-off exercise and describe factors that influence this behavior among MSM cruising the Internet for sex.

Method:
In 2006, 944 MSM were recruited from a popular cruising website and asked to engage in a choice-based conjoint (CBC) trade-off exercise where they were presented with a series of profiles that differed in partner attributes including appearance, serostatus, risk behavior, and condom use. Collecting these decisions in a CBC design produced a model of attributes MSM consider when choosing sex partners and the types of sexual practices they engage in. In addition, respondents were asked a series of self-reported questions around risk behaviors and serosorting.

Result:
16% of respondents were HIV-positive, 72% HIV-negative, and 11% were unsure of their HIV status. In general, the CBC model predicts that men cruising the Internet for sex are influenced most by a potential partner's attractiveness. Among HIV-positive respondents who never used a condom with online partners, 74% reported always serosorting, however, the CBC model predicts that a lesser percentage (55%) always serosort. In addition, the CBC model predicts that the percentage of MSM who serosort drops when potential online partners are highly attractive.

Conclusion:
Self-reported serosorting by MSM may be substantially higher than their actual practice. Serosorting is affected by other factors including the attractiveness of a potential sexual partner.

Implications:
Harm reduction interventions that promote serosorting practices should incorporate strategies to help clients manage other factors that influence their decision-making.