Factors Associated With HIV Testing in Teenage Men Who Have Sex With Men

Data were collected in 2018 and 2019 as part of SMART, an ongoing pragmatic trial³⁴  of a suite of HIV prevention interventions for AMSM. SMART uses a sequential multiple-assignment randomized trial design³⁵  to assess the effects of a package of increasingly intensive HIV prevention programs^24,36–38  on sexual risk behaviors among racially diverse AMSM across the United States.

Eligibility criteria included the following: (1) age 13 to 18 years; (2) assigned male at birth; (3) identify as gay, bisexual, or queer or attracted to cisgender men; (4) some sexual experience (ie, genital contact); (5) able to read English or Spanish; and (6) Internet access. Participants were largely recruited via social media campaigns. Prospective, eligible participants were routed to an online consent form and capacity-to-consent assessment. A video chat with staff confirmed eligibility and consent and completed enrollment. Survey data were collected in Research Electronic Data Capture,³⁹  and participants were paid $25. All procedures were approved by the institutional review board with waivers of parental permission.⁴⁰  Data for this report came from the baseline assessment of SMART and were collected before any intervention participation. At the time of analysis, 699 participants completed the baseline measures.

Participants self-reported their age, race, and ethnicity. Socioeconomic status was assessed by asking, “Do you or anyone in your family receive any form of public assistance, such as reduced-price meals at school, SNAP [Supplemental Nutrition Assistance Program], or welfare?”⁴¹  Rural residence was assessed by categorizing zip codes into rural-urban commuting area codes⁴² ; zip codes with 30% or more of their workers going to a US Census-defined urbanized area were considered urban, and all others were considered rural. Outness about sexual orientation was measured with a single item, an approach that has been found to perform well compared with multi-item scales.⁴³ 

HIV transmission risk behaviors and testing behaviors were assessed with the HIV Risk Assessment of Sexual Partnerships⁴⁴  and previous research on HIV testing among AMSM.²⁰  Items assessed lifetime engagement in vaginal or anal sex and lifetime engagement in condomless anal sex (CAS) with a male partner. Items assessed lifetime and previous-3-month HIV testing and if their parents were present when they were tested.

On the basis of the CDC’s Youth Risk Behavior Surveillance System,⁴⁵  participants were asked about receipt of HIV education in school, if they spoke to parent or other adults in their family about HIV, and if their parents ever talked to them regarding expectations about what to do or not do when it comes to sex. Response options were “yes,” “no,” and “not sure,” and for the purposes of these analyses, “no” and “not sure” were combined.

Three items assessed participants’ communication with health care providers about topics relevant to their sexual health.³²  Items began with, “In the last 12 months, I have spoken to a doctor, nurse, or other health care provider about…” and asked about sexual orientation, sex with male partners, and HIV testing. Response options were “yes” or “no.” An additional item asked, “Do you have 1 person you think of as your regular doctor, like a pediatrician or family doctor?”

Scales and items assessed aspects of the IMB model that, according to the theory, would be related to HIV testing.²⁸  HIV knowledge was assessed with 11 items that asked for true or false responses regarding facts about HIV transmission and prevention⁴⁶  (eg, “A person will not get HIV if (she or) he is taking antibiotics”). Two summed items assessed perceived vulnerability to HIV infection⁴⁷  (“How likely do you think it is you will get HIV if you have sex without a condom or PrEP [preexposure prophylaxis]?” and “How likely do you think it is that you will become HIV-positive at some point in your lifetime?”), with responses ranging from 1 (very unlikely) to 5 (very likely). To capture attitudes toward HIV testing, participants were prompted with, “Getting an HIV test would be…” and asked to respond on a scale ranging from 1 (very unimportant) to 5 (very important). We used 2 items to capture self-efficacy for relevant HIV-related behaviors. For HIV testing, we asked, “How hard would it be for you to get an HIV test?” with responses options ranging from 1 (very hard) to 5 (very easy). To assess self-efficacy in managing HIV treatment if diagnosed, participants were asked, “If you were HIV-positive, how confident are you that you could get treatment for HIV where you live?” with responses ranging from 1 (not at all confident) to 4 (very confident).

Logistic regression models were used to assess whether each independent variable was associated with ever having received an HIV test. Odds ratios (ORs) with 95% confidence intervals were calculated for each independent variable. Next, a multiple logistic regression model was fitted to predict ever having received an HIV test. This full model included all aforementioned independent variables. Adjusted ORs with 95% confidence intervals were calculated for this multivariable model. All analyses were performed by using R version 3.5.1.⁴⁸ 

The sample had strong representation of racial and/or ethnic minority participants and youth from low–socioeconomic-status backgrounds (Table 1). The mean age of the sample was 16.6 (SD = 1.3) years. One-quarter of the participants (n = 162; 23.2%) had ever had an HIV test in their lifetime, and few reported >1 test (n = 81; 11.6%). More than half of those who were ever tested received an HIV test in the 3 months before the survey (n = 95 [58.6%]; 13.6% of total sample). Testing with a parent was rare among those who had been tested in their lives, with 74.7% (n = 121) reporting that their parents were never present when they received their HIV test(s), and another 4.9% of participants reported that their parents were present for some, but not all, of their HIV tests.

Table 2 reports descriptive statistics on independent variables; 64.2% of participants had vaginal or anal sex in their lifetime, and 45.5% had ever engaged in CAS. Most participants learned about HIV in school (66.4%), but fewer spoke with parents about HIV (32.5%) or parental expectations about sexual behavior (38.3%). Most participants had a regular doctor or pediatrician (67.5%), but few had conversations with a health care provider about their sexual behavior with males (21.3%) or HIV testing (19.2%) or were asked about their sexual orientation (29.2%).

In terms of IMB factors, the mean percentage correct for the HIV knowledge test was 66.5%. For motivational factors, participants fell at the midpoint of the scale for perceived vulnerability to HIV (ie, infection was neither unlikely nor likely). Attitudes toward the importance of HIV testing were positive within the sample, with the mean falling closest to the value for “very important.” Participants were divided on their self-efficacy for getting an HIV test (41.1% were not at all or somewhat confident) or treatment if they received an HIV-positive diagnosis (42.5% were not at all or somewhat confident).

Table 3 reports bivariable associations between these factors and ever having received an HIV test. In terms of demographic factors, rates of HIV testing significantly increased with age, and there was a trend of African American AMSM being more likely to report having received an HIV test than white AMSM (P = .07). AMSM who identified with a sexual orientation other than gay or bisexual were significantly less likely than those who identified as gay to have received an HIV test. Youth who were more out about their sexual orientation were significantly more likely to have been tested for HIV.

In terms of HIV risk, youth who had sexual intercourse (vaginal or anal) or CAS were significantly more likely to have had an HIV test. School-based discussions of HIV were not significantly associated with HIV testing, but discussions with parents about HIV and expectations about sexual behavior were both significantly associated. Each of the variables assessing discussions with physicians had large effects, including discussing sex with men, sexual orientation, and HIV testing. In terms of IMB skills factors, HIV knowledge was significant, with each percent increase in HIV knowledge test score producing a 3% increase in the odds of HIV testing (OR = 1.03; P < .001). Individuals who believed HIV testing was important and that it was easy and who were confident they could obtain an HIV test were significantly more likely to have been tested. The largest effect in these analyses was speaking to a doctor about HIV testing (OR = 25.29; P < .001), with 75.4% who had such conversations having been tested, compared to only 10.8% of those who had not had such conversations.

Table 3 reports adjusted ORs from a multivariable model that included all factors. Because speaking to a health care provider about HIV testing could be caused by testing rather than leading to HIV testing, the variable assessing ever having discussed an HIV test with a health care provider was not included in the multivariable model. Adjusting for other factors, differences in HIV testing by age were substantial (OR = 5.48; P < .001; comparing 13–14-year-olds to 17–18-year-olds). Physician discussions about sexual orientation or sex with a male partner had the largest effect on HIV testing. Variables assessing information, motivation, and self-efficacy for testing were next in terms of their importance; however, in this multivariable model, self-efficacy for receiving HIV treatment was no longer significant. In terms of engagement in HIV risk behaviors, lifetime anal or vaginal sex was significant (OR = 3.24; P < .01), but lifetime engagement in CAS was no longer significant. Factors related to families were no longer significant, including having spoken to parents about HIV or expectations about sexual behavior.

In this study, we examined HIV testing and associated factors among AMSM, an understudied group with high HIV incidence and some of the lowest levels of HIV status awareness.^3,4  Fewer than one-quarter of the AMSM participants reported having received an HIV test in their lifetime. Rates were lowest at the youngest ages. AMSM who engage in condomless sex are recommended to receive an HIV test at least annually, if not more frequently.^13,14  These recommendations notwithstanding, we found that testing prevalence was low even among AMSM who reported CAS (34.8%). This finding is of significant public health concern because low levels of HIV testing in populations with high HIV incidence will prevent the achievement of goals to get to 0 new HIV infections in the United States.

Among a set of theoretically derived barriers and facilitators of HIV testing, physician conversations about HIV, same-sex behavior, and sexual orientation were the most influential. Although most participants reported having a regular doctor, only 20% to 30% reported having a conversation with their doctor about these topics. This is consistent with previous studies of AMSM and young adult MSM^10,26,32  and physicians⁴⁹  that found a similarly low frequency of reporting these conversations. Our findings suggest that pediatricians can encourage AMSM HIV testing by discussing sexual orientation with patients. Moreover, such conversations are recommended by the American Academy of Pediatrics.⁵⁰  Although adolescent-patient–physician communication about sex may be difficult for some, simple structural changes within the practice can facilitate these discussions. For example, pediatricians can update their intake forms to include a section on sexual orientation, and they can signal safety to adolescents through visual cues such as Safe Zone or lesbian, gay, bisexual, transgender, queer ally posters in examination rooms.⁵¹  Overall, such small office-environment adaptations can precipitate meaningful dialogues while simultaneously improving public perceptions of inclusivity.⁵² 

In addition to these nonverbal ways to elicit sexual orientation discussions, behavioral and communication adjustments to the pediatrician-patient interaction may be effective. Recent research suggests that AMSM are unlikely to broach sexual health topics with their pediatricians because of fears about lack of acceptance and privacy from parents.^32,53  In response, doctors should actively articulate to patients that their office is a safe space to discuss sexuality. They should reinforce doctor-patient confidentiality, which can be accomplished by asking patients’ parents to exit the room during part of the patient history.^51,52  Finally, for pediatricians who do not want to engage in conversations about sexual orientation, defaulting to HIV testing with an informed “opt out” option would be effective as well. A combination of some or all of these recommendations could improve testing among AMSM.

IMB skills, which are core components of behavioral HIV prevention theories, showed significant associations in our bivariable and multivariable models. These constructs represent actionable targets for pediatrician and public health interventions that can increase HIV testing among AMSM. In fact, in a sample of AMSM, a randomized controlled trial of an intervention based on this theoretical perspective showed that providing information about HIV testing, motivational messages about testing, a proximity-based HIV testing locator, and a video that demystified and instilled self-efficacy in HIV testing produced threefold greater odds of HIV testing compared with a control arm.⁵⁴  Research is needed on how to scale up and expand such interventions to promote HIV testing among AMSM. This is particularly important for adolescent men who have fewer routine interactions with the health care system⁵⁵  and, as such, may be more likely to seek testing as a standalone service from a community organization.

This study is not without limitations. First, the data used are cross-sectional, so directionality and causal relationships cannot be determined. Second, although we used a large, diverse sample of AMSM, it is not a population-based sample and is therefore not representative of the entire population of AMSM in the United States. Third, AMSM in this sample were enrolled in a research trial, so they may differ from AMSM who are not eligible for, or would not be willing to enroll in, the trial. However, all AMSM in this sample were sexually experienced, so they are a group for which HIV testing is highly relevant.

This study reports unique data from the understudied, but high-HIV-risk, group of AMSM. We found that despite high rates of engaging in anal sex and CAS, few participants reported ever having received an HIV test. Our models indicate an important role for pediatricians as well as foci for interventions to increase testing rates. Data collected in 2014 from an identically recruited sample of AMSM showed only slightly lower rates of ever receiving HIV testing,⁵⁴  suggesting we have made little headway in achieving Healthy People 2020 goals of increasing HIV testing rates.⁵⁶  Findings also suggest a major need for implementation research to increase HIV testing among AMSM if we are to achieve the goals of the new federal initiative to end the HIV epidemic in the United States in the next decade.

 
• AMSM

  adolescent men who have sex with men

 
• CAS

  condomless anal sex

 
• CDC

  Centers for Disease Control and Prevention

 
• IMB

  information-motivation-behavioral

 
• MSM

  men who have sex with men

 
• OR

  odds ratio