Introduction

Men who have sex with men (MSM) remain disproportionately affected by HIV infection in the United States. Despite representing between 2 and 3% of the United States population, MSM accounted for 67% of new HIV diagnosis in 2015 [1]. High HIV infection rates among MSM have necessitated the development of more effective HIV prevention tools. Preexposure prophylaxis (PrEP) is a recent biomedical HIV prevention strategy that involves HIV-negative individuals taking antiretroviral medication prior to HIV exposure to prevent HIV infection. It has demonstrated safety and efficacy in reducing HIV acquisition among MSM and other high-risk groups [2,3,4].

Research assessing actual PrEP use is evolving; studies suggest low PrEP use among MSM with rates ranging from 1.5 to 12% [5,6,7,8,9,10,11,12], with one study reporting a prevalence rate as high as 23% [13]. Prior studies have identified barriers to PrEP use including high cost, concerns about side effects, accessibility, and lack of knowledge [12, 14]. There are also concerns that PrEP use may be associated with increased HIV sexual risk behaviors [14, 15]. However, data on socio-structural factors including income, unstable housing, and internalized homonegativity and their association with PrEP use have not being thoroughly described. The aim of this analysis was to understand the socio-structural and behavioral correlates of PrEP use among a sample of high-risk HIV-negative MSM in Los Angeles.

Methods

Participants

The Men who have sex with Men & Substance Use Cohort at UCLA Linking Infections, Noting Effects (mSTUDY) is an ongoing 5-year prospective cohort study measuring factors linked to substance use and HIV transmission dynamics for HIV-positive and HIV-negative male-identified MSM in Los Angeles. The mSTUDY is focused on enrolling Black/African American and Latino/Hispanic MSM between 18 and 45 years of age who were born male. In addition, HIV-negative MSM were eligible if they reported unprotected anal intercourse with a man in the past 6 months. Participants in the mSTUDY return every 6 months for physical examinations, laboratory testing, and completion of a survey collecting sociodemographic, psychosocial, and behavioral data. All behavioral questions were assessed using computer-assisted self-interview (CASI). The current analysis includes data from 185 HIV-negative MSM enrolled in the mSTUDY between February 2015 and January 2017. The mSTUDY protocols confirmed HIV-negative status at enrollment via rapid HIV antibody test. The UCLA Institutional Review Boards approved the mSTUDY protocols, and all participants provided informed consent.

Measures

The survey included questions about participant’s age, racial/ethnicity status, sexual orientation, current employment status, current insurance status, and annual income. ‘Outness’ was measured with the question “Who have you told that you have sex with men?.” Five dichotomous (not mutually exclusive) variables were created from their response to indicate being out to doctors, family members, their priest, straight friends, and work colleagues. To measure internalized homonegativity, we used an adapted measure originally developed by Herek et al. [16]. Participants indicated their level of agreement with nine statements (such as, “I wish I were not gay/bisexual/attracted to men”) using a five-point Likert scale. Response options for each item ranged from 1 = strongly disagree to 5 = strongly agree. Scores were summed and ranged from 9 to 45, with higher scores indicating greater internalized homonegativity. The survey also asked about sexual behaviors in the past 6 months, history of sexually transmitted infections (STI), and substance use in the past 6 months. PrEP use was assessed by self-reported recent (past 6 months) use of an anti-HIV medication. Participants who reported to have taken PrEP or both PrEP and postexposure prophylaxis were classified as recent PrEP users.

Data Analyses

We used Chi-square or Fisher’s exact tests to compare the distributions of sociodemographic, psychosocial, and behavioral characteristics by recent PrEP use. We performed multivariable log-binomial regression models on data from all available participant study visits using generalized estimating equations. The final multivariable model included variables previously associated with PrEP use [11, 12] and variables significant (p < 0.10) in the bivariable analyses. We also adjusted for participant enrollment, because some mSTUDY participants were enrolled from a prior study facilitating PrEP use.

Results

Sample Characteristics

The current study includes data from 185 HIV-negative MSM who contributed 429 person-visits. The mean age of the sample was 29 years [standard deviation (SD) = 6.5] (Table 1). Most participants were African American (40%) or Hispanic (41%), reported current health insurance coverage (80%), and earned $9, 999 or less in annual income (57%).

Table 1 Baseline characteristics of mSTUDY participants by recent PrEP use (N = 185)
Full size table

Bivariable Associations

In bivariable analysis, participants who reported recent PrEP use compared to non-PrEP users were significantly more likely to have been enrolled in a prior study facilitating PrEP use (27 vs. 6.9%; p = <0.01), to have inhaled poppers in the past 6 months (56.8 vs. 24.8%; p = <0.01), to have reported anal sex with six or more partners in the past 6 months (48.6 vs. 24.1%; p = <0.01), and to be ‘out’ to both their doctor (97.2 vs. 74.8%; p = <0.01) and colleagues (85.7 vs. 62.4%; p = <0.01). Recent PrEP users were also less likely to have slept in a place not designed for sleep (10.8 vs. 29.7%; p = <0.01) than non-PrEP users. There was no significant difference in levels of internalized homonegativity and recent PrEP use.

Correlates of Recent PrEP Use in Multivariable Analyses

In the multivariable model (Table 2), those with current health insurance had significantly greater prevalence of recent PrEP use compared to those with no health coverage (adjusted prevalence ratio [aPR] = 2.02, 95% confidence interval [CI] 1.01 to 4.01; p = 0.04). Participants with annual income > $30,000 compared to those who made <$9, 999 were significantly more likely to report recent PrEP use (aPR = 2.56, 95% CI 1.15 to 5.69; p = 0.02). The men who reported sleeping in a place not designed for sleep were less likely to report recent PrEP use (aPR = 0.44, 95% CI 0.22 to 0.90; p = 0.02). Having sex with six or more anal sex partners in the past 6 months compared to less than five anal sex partners was significantly associated with greater prevalence of recent PrEP use (aPR = 2.20, 95% CI 1.26 to 3.82; p = <0.01). Sex with a HIV-positive partner compared to sex with a HIV-negative/unknown status partners was also associated with greater prevalence of recent PrEP use (aPR = 3.63, 95% CI 1.45 to 9.10; p = 0.01). Finally, the men who reported popper use in the past 6 months had a higher prevalence of recent PrEP use compared to those who had not used poppers (aPR = 2.76, 95% CI 1.58 to 4.84; p = <0.01). No other substance use variables were significantly associated with PrEP use.

Table 2 Multivariable log-binomial analyses of predictors of recent PrEP usea among MSM in the mSTUDY (N = 185)
Full size table

Discussion

In this sample of predominantly high-risk HIV-negative MSM in the Los Angeles area, indicators of higher socio-structural status including having health insurance, greater annual income, and having a stable place to sleep were positively associated with recent PrEP use. Also, practice of behaviors such as many anal sex partners, sex with a HIV-positive partner, and popper use was associated with greater prevalence of PrEP use.

Prior studies have documented associations between higher income [14] and health insurance coverage [17] as facilitators of PrEP use. Our study also found that sleeping in a place not designed for sleep—an indicator of unstable housing—was associated with lower rates of recent PrEP use. Collectively, these findings emphasize that stable structural factors of health are key to PrEP access in this group. Yet, in this group nearly 80% had health insurance, which suggests those without coverage, as those whose survival needs trump those over their sexual health, limiting access for low-income and uninsured MSM to PrEP services. Manufacturer medication assistance programs [18]—which provide support for medication costs and co-pays for PrEP—may help to minimize cost barriers to PrEP access. But these findings signal that, even with assistance, many low-income and uninsured MSM are simply not able or not interested in accessing PrEP.

In our study, we did not find a significant association between internalized homonegativity and recent PrEP use. One prior study found that higher levels of internalized homonegativity were significantly associated with increased PrEP use [5]. However, because it has been suggested that higher levels of internalized homonegativity can manifest as reduced self-worth and self-care and thus diminished motivations to use PrEP [19], links between internalized homonegativity and PrEP use warrant further investigation.

We also found that elevated individual risk behaviors, such as many male anal sex partners and sex with an HIV-positive partner, were significantly associated with greater PrEP use. This is consistent with prior research [12] and indicates that those MSM on PrEP are precisely the group that would most likely benefit from prophylaxis. Interestingly, poppers emerged as the only substance use significantly associated with PrEP use. While we have no data on this, it is possible that poppers are being used to enhance pleasure during receptive anal sex. Popper use has also been associated with increased practice of behaviors that enhance risk of HIV exposure, such as condomless anal sex [20]. Thus, the men who reported using poppers and high individual risk behaviors for HIV exposure perceived themselves at elevated risk for HIV and engaged in PrEP. Therefore, these correlates of recent PrEP use identified in our study provide direction for intervention development to increase PrEP uptake among MSM with low income and unstable housing.

Our study has some limitations. The data are correlational and cannot assess causality. These analyses do not include specific sexual risk behaviors linked to HIV transmission (e.g., condomless anal sex with serodiscordant or unknown serostatus partners) that are criteria for PrEP eligibility [12]. It is possible that there is some misclassification of PrEP use because we relied on self-reports. Finally, the recruitment of some men from a prior study facilitating PrEP use may have inflated our rates of PrEP use in this population.

Conclusion

In summary, in this sample of predominantly racial/ethnic minority MSM, we found that socio-structural factors such as health insurance coverage, greater income, and stable housing were important factors that were associated with PrEP use. Interventions that address these factors may help enhance PrEP use.