NHBS conducts surveying and HIV testing in three populations at risk for HIV acquisition: MSM, persons who inject drugs, and heterosexuals at high risk for HIV infection. Survey populations are annually rotated such that each population is surveyed every 3 years. From 2008 through 2014, sampling occurred in up to 21 metropolitan statistical areas (MSAs) with a high prevalence of stage 3 HIV disease (AIDS). Inclusion criteria for participation in NHBS among MSM were: male, self-reported sex with a male (ever), ≥ 18 years age, MSA resident, ability to complete the survey in English or Spanish, and provision of informed consent. MSM participants were recruited using venue-based, time–space sampling as described in prior publications [16, 17]. Briefly, recruitment activities included: (1) formative research to identify venues and optimal times to recruit MSM; (2) development of sampling frames of eligible venues and time periods; (3) random selection of venues and day–time periods; and (4) recruitment, administration of a standardized anonymous questionnaire, and HIV testing during sampled events. HIV testing was based on blood or oral specimens using laboratory testing or rapid testing in the field followed by laboratory confirmation. Based on locally determined rates, participants received compensation separately for the interview and HIV testing components of the study. NHBS was determined to be non-engaged research by the Centers for Disease Control and Prevention (CDC) and was approved by the Institutional Review Boards for each MSA.

Data from MSM surveying conducted in 2008, 2011, and 2014 were aggregated for the present analysis (n = 27,414). Only those participants with a valid positive HIV test result were included (n = 5935). Additionally, to focus on those MSM who pose a potential transmission risk, only those MLWH who reported having a male anal sex partner in the prior 12 months were included (n = 5079). Sexually active MLWH were then classified into 1 of 4 categories: (1) unaware of HIV-positivity; (2) aware, out of care; (3) aware, in care without ART; and (4) aware, in care on ART. Those who self-reported that they were uninfected or unsure of their status but had a positive HIV test as part of NHBS were classified as ‘MLWH, unaware’. Engagement in care was defined as having received HIV care in the 6 months prior to being interviewed. Current ART use was self-reported at the time of the interview.

For the present analysis, sex was defined as anal or vaginal intercourse. Someone with whom the participant had sex and to whom he felt most committed, such as a boyfriend, spouse, significant other, or life partner was defined as a main partner. Someone with whom the participant had sex but did not feel committed to, did not know very well, or had sex with in exchange for something such as money or drugs was defined as a casual partner. Total number of sex partners was the sum of main and casual sex partners. Condomless anal sex with someone of discordant HIV status was defined based on the participant’s self-reported perceived HIV status and knowledge of his last sex partner’s status. Importantly, condomless anal sex between unaware MLWH and perceived negative partners is classified as concordant so that this measure examines behavioral intentions.

To assess differences in sex practices by HIV diagnosis (i.e. awareness of HIV status), categories 2, 3, and 4 were collapsed to collectively represent ‘aware MLWH’ and compared to category 1 (unaware MLWH). Subsequently, to assess differences in sex practices by engagement in care, categories 2, 3, and 4 were compared with each other.

For each of the two research aims, frequencies of sex practices were calculated and compared. Potential clustering effects by recruiting event were examined; however, intraclass correlations (ICC) across outcomes of interest were nominal (i.e. < 0.02) and accounting for ICC had no effect on findings of statistical significance. Therefore, statistical testing for differences were conducted using logistic regression with Bonferroni corrections of outputted p-values to account for multiple comparisons (with Type I error set at < 0.05). Race and survey year were adjusted for to account for potential confounding effects [18]. The median and 5th–95th percentiles for annual number of sex partners are presented (in total and for casual partners only). Due to the highly right skewed distributions for annual number of sex partners, we also quantified the percentage of participants who reported ≥ 10 annual sex partners to examine the tail of each distribution. Results for male partners and female partners are presented separately.

Similar analyses including those who were sexually inactive were also conducted (see Supplement, Tables A to C).

All analyses were conducted in SAS software 9.3 (SAS Institute Inc., Cary, NC, USA).

A total of 5079 MLWH were included in the present analysis (Table 1). In comparing the 2008 through 2014 samples, there was a decrease in the percentage of MLWH who were unaware of their HIV-positive status (38.8 to 23.1%, respectively, p-value < 0.0001), as well as an increase in MLWH in care and on ART (41.3 to 66.8%, respectively, p-value < 0.0001). Over half of the sample was younger than 40, identified as Black or Hispanic, attended at least some college, were employed, had health insurance, and were circumcised.

Reports of condomless anal sex with a male partner in the prior 12 months were similar across the two categories of HIV awareness, unaware MLWH and aware MLWH (64.5 and 69.2%, respectively) (Table 2). Of the total sample, 9.1% reported having vaginal or anal sex with females in the prior 12 months. Unaware MLWH were three times as likely to report a female sex partner in the prior 12 months as aware MLWH (17.3 and 5.6%, respectively, p-value < 0.0001). Aware MLWH were significantly more likely to report non-HIV sexually transmitted infections (STIs) during the prior 12 months than unaware MLWH (p-value < 0.0001). The high-risk practice of condomless anal sex with a male partner of discordant or unknown HIV status was more commonly reported among unaware MLWH (25.9%) than aware MLWH (18.0%, p-value < 0.0001).

The median number of annual male anal sex partners was similar across the two categories of HIV awareness for total and casual partnerships overall, as well as total and casual partnerships in which condomless sex occurred. However, when examining the skewed tails for the distributions of annual number of male anal sex partners two findings emerge. First, aware MLWH were significantly more likely to report ≥ 10 male anal sex partners in the prior year when compared to unaware MLWH. Second, this pattern was present when examining total and casual partnerships overall, as well as total and casual partnerships in which condomless sex occurred.

When examining female sex partners, there were no marked differences across the two categories of HIV awareness in terms of the median number of partners (for both total and casual partnerships overall, as well as total and casual partnerships in which condomless sex occurred). However, unaware MLWH were over three times as likely to report ≥ 10 female sex partners as aware MLWH (for both total and casual partnerships overall, as well as total and casual partnerships in which condomless sex occurred). For instance, 12.1% of unaware MLWH reported having ≥ 10 female sex partners in comparison to 4.0% of aware MLWH (p-value = 0.0066). Notably, these observed differences were statistically significant for total and casual partnerships, as well as partnerships in which condomless sex occurred.

When examining sex practices across the continuum of care, we find that reports of condomless anal sex with a male partner in the past year (ranging from 65.0 to 70.0%), and condomless anal sex with a male partner of discordant or unknown HIV status (ranging from 17.7 to 21.3%) were similar across all categories of aware MLWH (Table 3). Aware MLWH who were in care without ART (category 3) were significantly more likely to report having an STI in the past year when compared to aware MLWH who were not in care and aware MLWH who were in care and on ART (categories 2 and 4, respectively). Aware MLWH who were in care on ART (category 4) were significantly less likely to report having a female sex partner in the prior year than MLWH without ART (categories 2 and 3) (ranging from 10.3 to 4.8% across the continuum).

When examining number of male sex partners across the 3 continuum of HIV care categories, three findings emerge. First, the median number of total annual male anal sex partners was similar across all categories of aware MLWH. This pattern was present when examining total and casual partnerships overall, as well as total and casual partnerships in which condomless sex occurred. Second, the percentage reporting ≥ 10 total and casual anal sex partners was consistent across all categories of aware MLWH. Third, aware MLWH who were in care and on ART (category 4) were significantly less likely than MLWH who were out of care to report ≥ 10 total and casual partnerships in which condomless sex occurred.

Across the continuum of care, there were no meaningful differences in the median number of annual female sex partners or the percentage reporting ≥ 10 female sex partners (for both total and casual partnerships overall, as well as total and casual partnerships in which condomless sex occurred).