Abstract
This study measures changes in condomless anal sex (CAS) among HIV-negative men who have sex with men (MSM) who are not taking pre-exposure prophylaxis (PrEP). It considers the 2014–2019 cycles of the American Men’s Internet Survey, a serial, cross-sectional web-based survey of US cisgender MSM aged ≥ 15 years, in which ~ 10% of each year’s sample is drawn from the previous year. Among those surveyed for 2 years who remained HIV-negative and off PrEP, reports of having any CAS and of CAS partner number were compared across years. We disaggregated by partner HIV status, and considered demographic predictors. The overall population saw a significant 2.2 percentage-point (pp) increase in reports of any CAS year-over-year. Sub-populations with the largest year-on-year increases were 15–24-year-olds (5.0-pp) and Hispanic respondents (5.1-pp), with interaction (young Hispanic respondents = 12.8-pp). On the relative scale, these numbers correspond to 3.2%, 7.2%, 7.3% and 18.7%, respectively. Absolute increases were concentrated among partners reported as HIV-negative. Multivariable analyses for CAS initiation found effects concentrated among Hispanic and White youth and residents of fringe counties of large metropolitan areas. CAS partner number increases were similarly predicted by Hispanic identity and young age. Although condom use remains more common than PrEP use, increasing CAS among MSM not on PrEP suggests potential new HIV transmission pathways. Concentration of increases among 18–24-year-old MSM portends future increases in the proportion of newly diagnosed HIV that occur among youth. Concentration among young Hispanic MSM will likely expand existing disparities. Although reducing barriers to PrEP remains vital, condom promotion for MSM remains a key public health practice and appears to be missing key audiences. LGBTQ+-inclusive sex education is one avenue for enhancing these efforts.
Similar content being viewed by others
Avoid common mistakes on your manuscript.
Introduction
HIV transmission can now be prevented through proactive administration of antiviral medications, either by those who do not have HIV (pre-exposure prophylaxis, or PrEP) or by those who do (treatment-as-prevention, or TasP). Together these have widened the array of strategies that men who have sex with men (MSM) may use to balance intimacy and pleasure against risk of HIV acquisition and transmission, beyond the mainstays of partner reduction and condom use. HIV diagnoses have declined significantly in the US since the mid-2010s and the parallel large-scale expansion of PrEP use and of TasP messaging (often referred to as “U = U”, for “undetectable” viral load equals “untransmittable” infection) [1, 2]. Concurrently, evidence has grown for declining rates of condom use among MSM on PrEP, both overall and specifically with partners living with HIV [3,4,5], even though CDC guidance continues to emphasize both PrEP and condoms as components of a comprehensive sexual health plan [6]. Additional evidence shows behavioral shifts among MSM who are virally suppressed, including reduced selection of partners based on concordant HIV status [7]. Such work generally focuses on implications for bacterial STIs, since behavior change among MSM either fully PrEP-adherent or virally suppressed should not significantly impact HIV incidence [8, 9].
Meanwhile, there is a paucity of quantitative research on behavior change among HIV-negative MSM who are not taking PrEP (hereafter “HNM-NP”), a group comprising the majority of MSM [10], and those who as a group likely experience the greatest HIV acquisition risk. In the US, condom use was declining among HIV-negative MSM [11] before PrEP ramp-up. Recent research finds continued declines [12], although whether these differ by respondents’ PrEP use status is unclear. An Australian study found that the proportion of HNM-NP having condomless anal sex (CAS) with casual partners increased from 30% in 2013 to 41% in 2017 [13], further increasing to 45% in 2019 [14]. We know of no study that has quantified changes in CAS among US HNM-NP.
Two hypotheses for why HNM-NP might have increased CAS over the last decade predict different behavioral signatures. One is a greater reliance on indirect protection from partners, i.e. choosing partners who report PrEP use or viral suppression, or opting for CAS with those partners. Indeed, “U = U” messaging rests on the reliability of this specific form of indirect protection. Reliance on partner PrEP usage would be similarly effective if one were certain that all of one’s partners were adherent. However, in practice, not all men on PrEP or treatment are sufficiently adherent for full protection [15,16,17,18,19], and many may not recognize this themselves, or report it accurately to others [20,21,22,23,24].
The second potential explanation is a broader shift in condom use norms among MSM. Condoms were the main HIV prevention tool available for decades, and as such, MSM developed a strong (but certainly not monolithic) normative value around their usage, colloquially called the “condom code” [25]. Advances such as PrEP and U = U mean that the condom code is less relevant as a universal norm, and with beliefs around health decision-making typically spreading through social networks [26], new norms could expand over time to reach MSM other than those who have adopted newer forms of biomedical protection [27, 28]. This explanation, if true, would yield increasing CAS with partners generally, not just those on PrEP or virally suppressed. Such changes could drive new HIV transmissions, regardless of whether they are explicitly caused by PrEP expansion.
We further hypothesize that such shifts could vary along two key dimensions relating to different normative contexts and ongoing HIV disparities: age and race/ethnicity. Specifically, CAS increases are predicted to be largest for the youngest cohort, whose sexual health philosophies and perceptions of MSM cultural norms are newly forming [29]. If so, this could portend increasing early-life HIV risk. For race/ethnicity, the effect is harder to predict; depending on its direction, it could interact with existing structural factors to exacerbate HIV disparities.
Efforts to investigate these hypotheses face a challenge with confounding. In the absence of individual-level behavior change, any amount of disproportionate adoption of PrEP by MSM at higher risk for HIV acquisition (as appears to occur [30,31,32]) will cause the remaining HNM-NP population to increasingly comprise men who would otherwise be at lower risk. Individual behavior change within that HNM-NP population would then occur on top of this substrate of population-level risk decline, making it difficult to detect. Whether the two combined effects would yield overall stable, increasing or decreasing behaviors at the population level for HNM-NP would depend on the relative magnitude of the two effects. Longitudinal data on HNM-NP can help disentangle these, by allowing one to examine behavior change at the individual rather than just the population level.
This paper examines behavior change among HNM-NP over the period 2014–2019, an interval of major PrEP expansion in the US [10, 33]. Data are drawn from the American Men’s Internet Survey (AMIS), which contains a sub-sample of respondents repeated across consecutive years. Analyses examine trends overall and by key demographic covariates, and disaggregate by respondents’ knowledge or perception of partner HIV status, allowing for insight into whether changes are occurring selectively or generally, and thus to their potential causes.
Methods
Data are drawn from the 2014 to 2019 cycles of AMIS, a serial, cross-sectional web-based survey of cisgender MSM aged ≥ 15 years, administered by Emory University’s PRISM Health; additional methodological details have been published elsewhere [34, 35]. The study is conducted annually with ~ 10,000 respondents/year; each year ~ 10% of the previous year’s sample self-selects into re-enrollment. Online recruitment protocols remain consistent across years, facilitating analysis of trends. Rounds from 2020 onward were excluded given confounding effects from the COVID-19 pandemic. We re-coded several AMIS variables a priori based on power considerations, including creating a four-level race/ethnicity variable: Hispanic (any race) and Black, White, and Another race (all non-Hispanic). Additional data-processing details are in the online supplement. Unfortunately, AMIS does not include information on the respondent’s perception of partner biomedical prevention usage (PrEP or TasP) over this time period; as far as we can determine, no large-scale study of US MSM with a longitudinal component from this pivotal time period does.
Our analyses used a series of restricted subsets of the data, described and named with abbreviations in Table 1; Fig. S1 visualizes relationships among analyses. All work occurred in R v.4.1.0, using an alpha-level of 0.05. Comparisons with a direction of effect hypothesized a priori used one-tailed tests; all others used two-tailed tests.
The analysis began by estimating the proportion of HNM-NP (defined in Table 1) respondents who reported having CAS each year, and conducted a one-tailed Cochran–Armitage trend test to identify significant increases over time, within the full sample and stratified by age and race/ethnicity. While measuring the trend here is an important first step, we reiterate that it should logically represent the combined effect of differential selection into PrEP use and potential behavior change. The next step thus tested whether there is indeed evidence for selectivity of the HNM-NP population into PrEP use in our data; this entailed the subset of HNM-NP respondents with a second year of data, comparing CAS rates in the first year for those subsequently initiating PrEP vs. not.
The remaining analyses focused on the second potential contributor, individual behavior change while remaining off PrEP. This employed the HNM-NP-2Y (Table 1) sample, with one-tailed McNemar tests to compare the proportion reporting CAS. Data were combined into year-on-year analyses regardless of calendar year to obtain power while also maintaining coverage across the period of PrEP scale-up in the US. Comparisons were conducted on the sample overall and by predictors, including a priori interactions between age and race/ethnicity, given our central concern for these dimensions of identity. Analyses were then disaggregated into CAS specifically with partners of positive, negative, and unknown HIV status. Results are reported as absolute (percentage point) differences in proportions reporting CAS between years, since many of the sub-analyses involve small values, and ratios on these are unstable. However, a few key sub-populations are also presented as proportional changes for ease of interpretation.
Individual-level comparisons of those who reported CAS in the second survey year to those who did not were conducted using bivariate logistic regression on the HNM-NP-2Y-NC (Table 1) sub-sample. We call the former “CAS initiation,” using the term relative to the previous year and not necessarily the lifetime. Correlates with < 5% missing data were included in a multivariable logistic regression model, as was the a priori interaction on age and race/ethnicity. Differential adoption of CAS while remaining off PrEP could be counteracted by higher cessation in the same groups, reflecting more temporal variability in risk rather than increasing risk, something one may imagine to be especially true for young MSM. Thus, a parallel analysis on the HNM-NP-2Y-YC (Table 1) sub-sample was used to estimate odds ratios for reporting no CAS in the second survey year (“CAS cessation”). Finally, the HNM-NP-2Y-YC2 (Table 1) sub-sample yielded measures of the 75th, 90th, and 95th quantiles for number of CAS partners in the past 12 months, values selected since partner counts are generally strongly right-tailed. This analysis included only those with ≥ 1 partner to distinguish changes in partner counts from the behavior already considered (no vs. any CAS partners). Wilcoxon signed-rank tests were conducted on each group with > 30 observations.
Results
Table 2 shows the sample size and characteristics for HNM-NP respondents. The 2014 sample is roughly half subsequent years’ because only a subset of respondents was randomized to receive PrEP questions. The sample is diverse across multiple socio-demographic dimensions. In all years > 80% of respondents who provided sufficient data reported being HNM-NP (Fig. 1). Notably, in the most recent year (2019), condom use remained more common than PrEP, with the proportion of HIV-negative respondents reporting current PrEP use (12%) or any PrEP use in the last 12 months (15%, including the current 12%) lower than that reporting completely consistent condom use for the year (17%) or at least some condom use (at least 45%—the questionnaire design details mean only a lower bound can be determined here). An additional 14% did not have anal sex.
The subsample of 2-year data by respondents represents the basis for most of our analyses; Table S1 examines this subsample’s composition to identify potential selectivity. Repeat respondents were disproportionately older, wealthier, more educated, more urban, and more likely to report White race and use of PrEP than the single-year eligible sample from which they were drawn. Nonetheless, the large sample sizes combined across years ensures that important subgroup analyses retain power.
In the HNM-NP sample, the proportion of respondents reporting any CAS was consistently high (62–66% per year), with no significant temporal increase (Fig. 2; Table S2) either overall or stratified by age or race/ethnicity. We reiterate, however, that this test’s null hypothesis is for no increase, whereas, in the absence of behavior change among HNM-NP, one should anticipate a decline in CAS among HNM-NP concordant with disproportionate PrEP adoption among those having CAS. This hypothesized differential selection into PrEP use is indeed confirmed in our data: among 2-year respondents not using PrEP in Y1, prevalence of CAS in Y1 was 79% among those who adopted PrEP in Y2 (N = 220), and 68% among those who did not (N = 2421). Among those with any CAS partners in Y1 (a cutoff chosen to distinguish changes in partner counts from the none/any distinction already considered), the 75th and 90th quantiles for partner number were 10 and 30 for subsequent PrEP adopters (N = 41) and 5 and 10 for non-adopters (N = 244).
Comparing the HNM-NP-2Y sub-sample (N = 2365) between their respective 2 years of data (Table 3) reveals a small (69.9–67.7% = 2.2 percentage-point, or pp) but significant (McNemar odds ratio [MOR] = 1.27, p = 0.008) absolute increase in reports of any CAS year-over-year. This corresponds to a 3.2% (= 69.9%/67.7% - 1) increase on the relative scale. Sub-populations with the largest year-over-year increases include 15–24-year-olds (5.0-pp, 7.2% relative, MOR = 1.60, p = 0.007) and Hispanic respondents (5.1-pp, 7.3% relative, MOR = 1.70, p = 0.038). Smaller but significant effects occurred among those with low-middle incomes ($20–40k, 3.9-pp, 5.8%, MOR = 1.50, p = 0.046), Northeast residents (4.5-pp, 7.3%, MOR = 1.66, p = 0.020), West residents (3.4-pp, 5.1%, MOR = 1.49, p = 0.038), and those with “some” college education (3.8-pp, 5.6%, MOR = 1.45, p = 0.016). The increase among White MSM was also significant (MOR = 1.24, p = 0.027), although with a smaller effect size than in the total population (1.9-pp, 2.8%). Disaggregating race/ethnicity by age reveals a notable 12.8-pp absolute increase (18.7% relative) in any CAS among young Hispanic MSM (MOR = 3.50, p = 0.003). Increases were concentrated in the middle time periods 2015–2016 and 2017–2018 (each a 4.6-pp increase, MOR2015-2016 = 1.60 and MOR2017–2018 = 1.64, each p = 0.009); other year-pairs all saw small, non-statistically significant changes.
Table 4 disaggregates these results by reported partner HIV status. Absolute increases are concentrated among partners reported as HIV-negative, with a significant (MOR = 1.37, p = 0.005) 3.6-pp increase, compared to 1.8-pp for partners living with HIV and 0.4-pp for those of unknown status. On a relative scale, these numbers correspond to 6.5%, 32.1%, and 1.9% increases, respectively, indicating that by far the greatest relative increase is with partners known to be living with HIV. Significant absolute increases among sub-groups were also concentrated with HIV-negative partners, and largely replicate significant predictors from overall increases. Significant predictors of increasing CAS with partners living with HIV were a distinct set: having a bachelor’s degree (2.9-pp, MOR = 2.33, p = 0.003), and residence in the Northeast (2.1-pp, MOR = 3.33, p = 0.046) or South (3.5-pp, MOR = 1.85, p = 0.049; relative increases not reported for these due to small denominators).
The HNM-NP-2Y-NC sub-sample (N = 764) provides evidence for individual predictors for CAS initiation over time while remaining off PrEP (Table 5). In bivariate analyses, significant predictors include younger age, low household income, lower education, and residence in a fringe county of a large metropolitan area as defined by the National Center for Health Statistics (hereafter “fringe metro” residence). Notably, Hispanic ethnicity was not significant here, in contrast to above. In the multivariable analysis, the largest point estimate for increased odds of initiating CAS was being a young Hispanic respondent (relative to older White respondent), at 2.91 (95% CI 1.42, 6.03; p = 0.004). Other large, significant effects were among younger White respondents (1.81, 9% CI 1.14, 2.87; p = 0.011), those having completed some college (1.45, 95% CI 1.01, 2.07; 0.041), relative to a 4-year college degree), and fringe metro residence (1.70, 95% CI 1.10, 2.62; p = 0.016).
The proportion of men ceasing CAS (12%) was lower than that initiating (32%, HNM-NP-2Y-YC sub-sample, Table S3). The only significant predictor of cessation was high school education or less, with an odds ratio (relative to college degree) of 1.81 (95% CI 1.08, 2.92; p = 0.019). This paucity of significant predictors occurred despite the sample size here being more than twice that for CAS initiation (1601 vs. 764, Table 1).
The final analysis considers the number of CAS partners in the last 12 months among those reporting at least one CAS partner (HNM-NP-2Y-YC2 sub-sample, N = 836, Table S4). The overall sample saw the highest two quantiles increase (90th = 4–5, 95th = 7–8), but no significant shift upward in the distribution (p = 0.099). In stratified analyses, significant upward shifts occurred among younger respondents (p = 0.001), Hispanic respondents (p = 0.027), those making $20–40k (p = 0.032), those in the Northeast (p = 0.023), and those with fringe-metro residence (p = 0.001). The inclusion criteria make this sample non-overlapping with that in Table 5, making the similarity in predictors especially notable.
Discussion
This paper examines behavior changes among MSM who are HIV-negative and not using PrEP (HNM-NP)—a population that accounts for the majority of US MSM—in the age of PrEP expansion and U = U messaging. Most men in this population have CAS each year, but most also use condoms at least some of the time. Most importantly, the findings suggest that the apparent flat rates of CAS for this population over time combine two countervailing phenomena, as hypothesized: selective adoption of PrEP among those with the most CAS partners, and increases over time in CAS among a subset of those remaining off PrEP. These increases appear both in the proportion of MSM having any CAS, and CAS partner number among those who do.
The meaning of the observed CAS increases—and the extent to which they are concerning from an HIV prevention standpoint specifically—depends on the partners with whom they occur. Little of the increase was with partners of unknown HIV status, suggesting that increases may not reflect an across-the-board decrease in condom-use culture. A small portion of the CAS increase occurred with partners living with HIV. Although partners’ viral suppression status cannot be determined here, one distinct predictor—high education—suggests that this effect may be concentrated among those who are most well-positioned to trust U = U messaging, a public health-endorsed form of indirect protection. The disproportionately large increase in the South in this measure might simply reflect the region’s higher HIV prevalence. U = U has a clear role to play in reducing stigma and discrimination towards persons living with HIV, with myriad psychosocial and public health benefits [37, 38]; the findings here warrant further investigation into variation in appreciation of U = U among US MSM, which is rapidly evolving [39].
On an absolute scale, most of the observed CAS increase occurred with partners who the respondent reported as HIV-negative; on the relative scale, the largest increase was with partners reported as living with HIV. Without knowing partners’ biomedical status, one cannot fully distinguish here between indirect protection or more general condom reductions; however, the evidence across partner HIV status seen above tilts towards the former, as does a recent report on hypothetical condom use likelihood by partner biomedical status [40]. If the increase in CAS with partners believed to be HIV-negative were entirely with partners perfectly adherent to PrEP, there would be little concern from an HIV transmission perspective. However, we consider this unlikely for two reasons. First, PrEP use remains the minority status among MSM generally, with no more than 15% of HIV-negative respondents here reporting PrEP use in any year. Second, the predictors for increasing CAS with HIV-negative partners (younger age, Hispanic ethnicity, fringe metro residence, and lower education and income) likely represent a pattern of MSM who are less well-positioned than others to navigate reliance on indirect protection through partner PrEP, reflecting well-known structural barriers to accessing care within such communities [41]. The significance of newly reporting CAS with HIV-negative partners depends on whether such respondents were not previously having CAS, or were having CAS with partners living with HIV or of unknown status. A post-hoc analysis determined that 92% of respondents who reported CAS initiation with HIV-negative partners reported no CAS in the previous year.
The strongest, most consistent predictor of initiating CAS or increasing CAS partner number while remaining off PrEP was young (18–24) age. This supports our initial hypothesis that condom use practices are changing most rapidly among the youngest cohort newly forming their sexual health strategies. Consequently, special attention should be paid to determine if young MSM not on PrEP represent a rising proportion of all MSM diagnosed with HIV in the US in coming years.
Hispanic respondents accounted for the highest increases in any CAS among racial/ethnic groups, and the only significant increase in CAS partner number. This is particularly concerning, since Hispanic MSM were already disproportionately burdened by HIV before PrEP expansion, and PrEP use lags in this community relative to non-Hispanic White MSM [42]. Our multivariable model confirmed that this was not simply driven by age confounding given the younger age profile of Hispanic persons in the US; nevertheless, these effects were concentrated among young Hispanic respondents. Disproportionate behavior change in this population provides additional explanation for the slower decline in new HIV diagnoses among Hispanic MSM relative to other racial/ethnic groups, even after accounting for changes in their share of the US population [1]. It is also consistent with work showing variation in how MSM interpret testing negative for HIV, with one study finding that Hispanic/Latino MSM exhibited a relatively low sense of reinforced safety and high sense of luck or invulnerability compared to non-Hispanic respondents—a pattern associated with subsequent increases in CAS [43]. Still, more research is needed to determine if this pattern has continued into the present day and, to the extent that these young Hispanic MSM (or, indeed, any MSM) are basing their decisions on indirect protection, what is the basis and accuracy of their understanding of each partners’ biomedical prevention status.
Our full set of significant predictors for CAS increases while off PrEP include not just age and ethnicity but also fringe metro residence and lower education and income. If these were predictors of PrEP uptake, we might attribute them to well-known barriers to PrEP access [44,45,46]. Their role in predicting increases of CAS conditional on being off PrEP is perhaps less expected, but could still reflect lower access to MSM-oriented HIV prevention messaging and services. That said, we note that Black MSM, who are also disproportionately impacted by HIV and experience reduced service access [47,48,49], did not report significant CAS increases while not on PrEP. Although this is itself encouraging news, it is worth noting that the persistent segregation of sexual networks and disparities in viral suppression and PrEP uptake together still imply that the same behavior is more likely to generate higher HIV acquisition risk for Black MSM than for other groups [50,51,52].
Collectively our findings support the familiar and persistent need to identify feasible, acceptable, scalable, and culturally-sensitive ways to deliver HIV prevention services to sub-populations of MSM who experience higher risk for HIV and lower PrEP uptake. In particular, this work demonstrates that this need still applies to the traditional mainstay of both HIV and STI prevention—condoms—alongside newer biomedical prevention forms. Some men may choose not to use condoms consistently or ever, despite CDC recommendations for a comprehensive sexual health strategy encompassing both HIV and STI prevention. These men remain prime candidates for PrEP, including regular STI testing. However, even with increases in CAS, our data show that a sizeable fraction of HIV-negative MSM report at least some condom usage (at least 45% in the latest survey year, 2019, representing 52% of those having any anal sex), suggesting that condoms retain some level of feasibility and acceptability for many MSM. Nevertheless, the concentrated increase in CAS among MSM who are young, Hispanic, and/or outside urban cores points to particular sub-populations where re-invigorated public health promotion and access for condoms may serve the greatest HIV prevention need, alongside PrEP promotion and access. All such efforts must contend with the challenge that, unlike PrEP or viral suppression, condom decisions are explicitly dyadic, making them especially prone to cascading social network feedback, which creates both specific challenges and opportunities [26].
One context with clear opportunity for these efforts is the continued expansion of comprehensive LGBTQ+-inclusive sexual education. Roughly half of US high schools lack this resource [53], a number which is almost certainly higher outside major urban cores, where the largest increases in CAS among young MSM were observed. This may be complemented by further promotion of online, interactive, and engaging sexual health education materials catering specifically to YMSM, which have been shown to reduce CAS in this population [54, 55], as has sexual health education that explicitly incorporates discussions of pleasure [56]. Another area for opportunity is the further development and promotion of comprehensive, coordinated, patient-specific approaches to sexual health that address all prevention modalities, such as New York City’s PlaySure network [57].
Respondents who supplied multiple years of data and who are thus central to our analysis disproportionately reported White race, older age, and urban residence than the single-year samples from which they were drawn. However, the largest CAS increases were among younger and Hispanic MSM and those in the metro fringe. Thus, one would expect the sampling bias in AMIS to underestimate the population-level effect in the overall sample, portending a larger public health concern. Similarly, Black MSM are under-represented in AMIS, reducing our ability to discern trends in this highly impacted population, although the absolute number of surveys from Black respondents was still large. Racial and ethnic identity in the US is complex, and it is difficult to know a priori which elements will be most salient for a given phenomenon. The 4-category version of race/ethnicity used here is commonly employed in demographic research for power considerations; considering the profound, multi-level racial and ethnic disparities in the HIV incidence and care continuum in the United States, it is conceivable that using more granular categories in future work will provide more nuanced insight into potential future trends in the epidemic. AMIS did not ask about country of birth for much of this time period, limiting our ability to disaggregate effects between native-born and immigrant MSM overall, or within the young Hispanic population who saw the largest behavioral changes in particular. We did not employ Type I error corrections, instead focusing our interpretation on predictors that appeared across multiple analyses, including those relying on non-overlapping sub-samples. The available data cannot determine the extent to which PrEP expansion is the cause of observed behavior change or simply coincident with it; however, the epidemiological impact of that change does not actually depend on this.
A small portion of our sample (1.9%) provided 3 years of data, which were treated as two observations of 2-year comparisons despite non-independence, given the small sample size and since our nested analyses are already complex. To test the impact of this decision, we repeated our analysis using only the first 2-year pair for each of these respondents. Our central findings remain qualitatively similar. The most notable difference is that the main effect for Hispanic respondents in Table 3 retains the same effect size (5.1-pp) but falls out of significance given the smaller sample (p = 0.072). However, the effect for young Hispanic respondents remains highly significant (p = 0.007) with a similar effect size (12.9-pp), and this combination of identities also retains similar explanatory power in the logistic regression results (Table 5; multivariable OR estimate 2.91, CI 1.42–6.03, p = 0.004). The significant increase for Hispanic respondents in CAI with partners perceived to be HIV-negative also remains (Table 4; 8.5-pp increase; p = 0.010). We note that this approach is also imperfect as it discards useful information (the third year of reports from some respondents) and thus loses power. We anticipate that more refined analyses retaining all data and accounting for dependence would produce results in line with those already presented.
Our study’s largest limitation is the lack of data on partner PrEP or viral suppression status. We chose to analyze AMIS data regardless, since the study covers the roll-out period for both PrEP and U = U messaging, which were hypothesized to be a critical period to observe behavior change among MSM not on PrEP, and contains longitudinal observations of a sufficiently large number of respondents, a key element for the analysis. Moreover, no study identified fulfilled all desiderata, which may help explain the relative dearth of research on this topic; however, we considered the question crucial enough to warrant whatever forms of investigation were possible with existing data. Efforts to ameliorate this limitation included disaggregating by partner status, and then identifying the distinct predictors, including education level, of CAS initiation by partner status. Although this approach is unable to demonstrate precisely what proportion of the increase is with partners not known to be on PrEP—or believed to be on PrEP but not fully adherent—it still highlights a strong need for increased attention to the question of behavior change among MSM who themselves are not on PrEP. While HIV diagnoses have declined considerably in the US in recent years, there are still more than 20,000 new cases among MSM each year [1]; and MSM who are not using PrEP presumably make up a large proportion of those newly infected. Understanding the contexts behind these transmission events is crucial for extending progress. Both quantitative and qualitative studies will be useful to understand the nuanced meanings and motivations of partner selection and condom use decision-making for MSM who remain off PrEP, especially young MSM who have only ever navigated sexual health in the PrEP and U = U era.
To our knowledge, this study is the first since the onset of PrEP and U = U to document CAS increases specifically among MSM who are HIV-negative and off PrEP in the United States. Considerable attention has rightly been paid to the many barriers to PrEP uptake, adherence, and retention. In the US, these include internalized stigma, distrust or discomfort with medical systems, avoiding PrEP-use disclosure to health plan providers, and lack of insurance or other resources to access PrEP, all of which fall disproportionately to MSM who are younger or of color [44, 58, 59]. While long-acting injectable PrEP may relieve pill burden [60], its current form brings additional challenges, including increased medication cost and more frequent provider visits [61]. For these and other reasons, a decade after PrEP’s approval most MSM are not on PrEP at any moment, and young MSM in particular are likely to continue experiencing sizeable HIV exposure risk before they initiate PrEP, if they ever do. While it is crucial that we continue working to remove PrEP barriers for all, it remains just as crucial to find ways to support MSM, and especially young MSM, who are not on PrEP in adopting or maintaining strategies that they find feasible and sustainable for balancing HIV prevention with intimacy and pleasure.
Change history
15 April 2024
A Correction to this paper has been published: https://doi.org/10.1007/s10461-024-04327-w
References
Centers for Disease Control and Prevention. HIV surveillance report, 2021, vol. 34. 2023. http://www.cdc.gov/hiv/library/reports/hiv-surveillance.html. Accessed 7 July 2023.
Smith DK, Sullivan PS, Cadwell B, et al. Evidence of an association of increases in pre-exposure prophylaxis coverage with decreases in human immunodeficiency virus diagnosis rates in the United States, 2012–2016. Clin Infect Dis. 2020;71(12):3144–51.
Montano MA, Dombrowski JC, Dasgupta S, et al. Changes in sexual behavior and STI diagnoses among MSM initiating PrEP in a clinic setting. AIDS Behav. 2019;23(2):548–55.
Oldenburg CE, Nunn AS, Montgomery M, et al. Behavioral changes following uptake of HIV pre-exposure prophylaxis among men who have sex with men in a clinical setting. AIDS Behav. 2018;22(4):1075–9.
Traeger MW, Schroeder SE, Wright EJ, et al. Effects of pre-exposure prophylaxis for the prevention of human immunodeficiency virus infection on sexual risk behavior in men who have sex with men: a systematic review and meta-analysis. Clin Infect Dis. 2018;67(5):676–86.
Centers for Disease Control and Prevention. Preexposure prophylaxis for the prevention of HIV infection in the United States—2021 Update: a clinical practice guideline. 2021. https://www.cdc.gov/hiv/pdf/risk/prep/cdc-hiv-prep-guidelines-2021.pdf.
Chow EPF, Phillips TR, Bradshaw CS, Chen MY, Fairley CK. Brief report: sexual mixing patterns by HIV status and PrEP use among men who have sex with men partnerships in Melbourne, 2011–2018. J Acquir Immune Defic Syndr. 2020;83(2):99–102.
Chou R, Evans C, Hoverman A, et al. Preexposure prophylaxis for the prevention of HIV infection: evidence report and systematic review for the US Preventive Services Task Force. J Am Med Assoc. 2019;321(22):2214–30.
Eisinger RW, Dieffenbach CW, Fauci AS. HIV viral load and transmissibility of HIV infection: undetectable equals untransmittable. J Am Med Assoc. 2019;321(5):451–2.
Centers for Disease Control and Prevention. Monitoring selected national HIV prevention and care objectives by using surveillance data—United States and 6 dependent areas, 2019. In: Division of HIV/AIDS Prevention (DHAP), editors. HIV Surveillance report supplemental report, vol. 26. http://www.cdc.gov/hiv/library/reports/hiv-surveillance.html2021.
Paz-Bailey G, Mendoza MCB, Finlayson T, et al. Trends in condom use among MSM in the United States: the role of antiretroviral therapy and seroadaptive strategies. AIDS. 2016;30(12):1985–90.
Khosropour CM, Dombrowski JC, Barbee LA, Kerani RP, Berzkalns A, Golden MR. Changing patterns of sexual behavior and HIV/STI among men who have sex with men in Seattle, 2002 to 2018. J Acquir Immune Defic Syndr. 2021;87(4):1032–9.
Holt M, Lea T, Mao L, et al. Community-level changes in condom use and uptake of HIV pre-exposure prophylaxis by gay and bisexual men in Melbourne and Sydney, Australia: results of repeated behavioural surveillance in 2013–17. Lancet HIV. 2018;5(8):e448–56.
Holt M, Broady TR, Mao L, et al. Increasing preexposure prophylaxis use and ‘net prevention coverage’ in behavioural surveillance of Australian gay and bisexual men. AIDS. 2021;35(5):835–40.
Centers for Disease Control and Prevention. Monitoring selected national HIV prevention and care objectives by using HIV surveillance data—United States and 6 dependent areas, 2017. HIV surveillance supplemental report, vol. 24. http://www.cdc.gov/hiv/library/reports/hiv-surveillance.html2019.
Li Z, Purcell DW, Sansom SL, Hayes D, Hall HI. Vital signs: HIV transmission along the Continuum of Care—United States, 2016. MMWR Morb Mortal Wkly Rep. 2019;68(11):267–72.
Quinn KG, Voisin DR. ART adherence among men who have sex with men living with HIV: key challenges and opportunities. Curr HIV/AIDS Rep. 2020;17(4):290–300.
Singh S, Mitsch A, Wu B. HIV care outcomes among men who have sex with men with diagnosed HIV infection—United States, 2015. MMWR Morb Mortal Wkly Rep. 2017;66(37):969–74.
Wood S, Gross R, Shea JA, et al. Barriers and facilitators of PrEP adherence for young men and transgender women of color. AIDS Behav. 2019;23(10):2719–29.
Amico KR, Mehrotra M, Avelino-Silva VI, et al. Self-reported recent PrEP dosing and drug detection in an open label PrEP study. AIDS Behav. 2016;20(7):1535–40.
Baker Z, Javanbakht M, Mierzwa S, et al. Predictors of over-reporting HIV pre-exposure prophylaxis (PrEP) adherence among young men who have sex with men (YMSM) in self-reported versus biomarker data. AIDS Behav. 2018;22(4):1174–83.
Mustanski B, Ryan DT, Remble TA, D’Aquila RT, Newcomb ME, Morgan E. Discordance of self-report and laboratory measures of HIV viral load among young men who have sex with men and transgender women in Chicago: implications for epidemiology, care, and prevention. AIDS Behav. 2018;22(7):2360–7.
Sewell J, Daskalopoulou M, Nakagawa F, et al. Accuracy of self-report of HIV viral load among people with HIV on antiretroviral treatment. HIV Med. 2017;18(7):463–73.
Stephenson R, Bratcher A, Mimiaga MJ, et al. Brief report: accuracy in self-report of viral suppression among HIV-positive men with HIV-negative male partners. J Acquir Immune Defic Syndr. 2020;83(3):210–4.
Chambers DL. Gay men, AIDS, and the code of the condom. Harv CR-CLL Rev. 1994;29:353.
Valente TW. Social networks and health: models, methods, and applications. Oxford/New York: Oxford University Press; 2010.
Carson B. The informal norms of HIV prevention: the emergence and erosion of the condom code. J Law Med Ethics. 2017;45(4):518–30.
Alcantar Heredia JL, Goldklank S. The relevance of pre-exposure prophylaxis in gay men’s lives and their motivations to use it: a qualitative study. BMC Public Health. 2021;21(1):1829.
Fontenot HB, Krakower D, White BP, et al. Condom use philosophy and behaviors among young men who have sex with men: variations among HIV pre-exposure prophylaxis users and nonusers. J Am Assoc Nurse Pract. 2020;33(7):521–8.
Rao DW, Carr J, Naismith K, et al. Monitoring HIV preexposure prophylaxis use among men who have sex with men in Washington State: findings from an internet-based survey. Sex Transm Dis. 2019;46(4):221–8.
Okafor CN, Gorbach PM, Ragsdale A, Quinn B, Shoptaw S. Correlates of preexposure prophylaxis (PrEP) use among men who have sex with men (MSM) in Los Angeles, California. J Urban Health. 2017;94(5):710–5.
Kuhns LM, Hotton AL, Schneider J, Garofalo R, Fujimoto K. Use of pre-exposure prophylaxis (PrEP) in young men who have sex with men is associated with race, sexual risk behavior and peer network size. AIDS Behav. 2017;21(5):1376–82.
National Center for HIV VH, STD, and TB Prevention. AtlasPlus. 2021. https://www.cdc.gov/nchhstp/atlas/. Accessed 1 Aug 2022.
Wiatrek S, Zlotorzynska M, Rai R, Sullivan P, Sanchez T. The annual American men’s internet survey of behaviors of men who have sex with men in the United States: key indicators report 2018. JMIR Public Health Surveill. 2021;7(3): e21812.
Sanchez TH, Zlotorzynska M, Sineath RC, Kahle E, Tregear S, Sullivan PS. National trrends in sexual behavior, substance use and HIV testing among United States men who have sex with men recruited online, 2013 through 2017. AIDS Behav. 2018;22(8):2413–25.
Sullivan PS, Sanchez TH, Zlotorzynska M, et al. National trends in HIV pre-exposure prophylaxis awareness, willingness and use among United States men who have sex with men recruited online, 2013 through 2017. J Int AIDS Soc. 2020;23(3): e25461.
Rendina HJ, Talan AJ, Cienfuegos-Szalay J, Carter JA, Shalhav O. Treatment is more than prevention: perceived personal and social benefits of undetectable = untransmittable messaging among sexual minority men living with HIV. AIDS Patient Care STDS. 2020;34(10):444–51.
Ford OG, Rufurwadzo TG, Richman B, Green I, Alesi J. Adopting U = U to end stigma and discrimination. J Int AIDS Soc. 2022;25(3): e25891.
Rendina HJ, Cienfuegos-Szalay J, Talan A, Jones SS, Jimenez RH. Growing acceptability of undetectable = untransmittable but widespread misunderstanding of transmission risk: findings from a very large sample of sexual minority men in the United States. J Acquir Immune Defic Syndr. 2020;83(3):215–22.
Algarin AB, Shrader CH, Hackworth BT, Ibanez GE. Condom use likelihood within the context of PrEP and TasP among men who have sex with men in Florida: a short report. AIDS Care. 2022;34(3):294–300.
Garney W, Wilson K, Ajayi KV, et al. Social-ecological barriers to access to healthcare for adolescents: a scoping review. Int J Environ Res Public Health. 2021;18(8):4138.
Sullivan PS, Whitby S, Hipp P, et al. Trends in PrEP inequity by race and census region, United States, 2012–2021. Paper presented at: AIDS 2022; Montreal. https://programme.aids2022.org/Abstract/Abstract/?abstractid=12943.
Feinstein BA, Johnson BA, Parsons JT, Mustanski B. Reactions to testing HIV negative: measurement and associations with sexual risk behaviour among young MSM who recently tested HIV negative. AIDS Behav. 2017;21(5):1467–77.
Bonett S, Dowshen N, Bauermeister J, et al. Characterizing the PrEP continuum for black and Latinx sexual and gender minority youth. AIDS Behav. 2022;26(4):1211–21.
Allan-Blitz LT, Mena LA, Mayer KH. The ongoing HIV epidemic in American youth: challenges and opportunities. mHealth. 2021;7:33.
Sullivan PS, Siegler AJ. Getting pre-exposure prophylaxis (PrEP) to the people: opportunities, challenges and emerging models of PrEP implementation. Sex Health. 2018;15(6):522–7.
Russ S, Zhang C, Liu Y. Pre-exposure prophylaxis care continuum, barriers, and facilitators among black men who have sex with men in the United States: a systematic review and meta-analysis. AIDS Behav. 2021;25(7):2278–88.
Johnson J, Radix A, Copeland R, Chacon G. Building racial and gender equity into a national PrEP Access Program. J Law Med Ethics. 2022;50(S1):55–9.
Brooks RA, Nieto O, Landrian A, Fehrenbacher A, Cabral A. Experiences of pre-exposure prophylaxis (PrEP)-related stigma among black MSM PrEP users in Los Angeles. J Urban Health. 2020;97(5):679–91.
Millett GA, Peterson JL, Wolitski RJ, Stall R. Greater risk for HIV infection of black men who have sex with men: a critical literature review. Am J Public Health. 2006;96(6):1007–19.
Goodreau SM, Rosenberg ES, Jenness SM, et al. Sources of racial disparities in HIV prevalence in men who have sex with men in Atlanta, GA, USA: a modelling study. Lancet HIV. 2017;4(7):e311–20.
Rosenberg ES, Millett GA, Sullivan PS, Del Rio C, Curran JW. Understanding the HIV disparities between black and white men who have sex with men in the USA using the HIV care continuum: a modeling study. Lancet HIV. 2014;1(3):e112–8.
Centers for Disease Control and Prevention. School Health Profiles 2020: characteristics of health programs among secondary schools. Atlanta. 2022. https://www.cdc.gov/healthyyouth/data/profiles/pdf/2020/CDC-Profiles-2020.pdf.
Mustanski B, Parsons JT, Sullivan PS, Madkins K, Rosenberg E, Swann G. Biomedical and behavioral outcomes of keep it up!: an eHealth HIV prevention program RCT. Am J Prev Med. 2018;55(2):151–8.
Mustanski B, Garofalo R, Monahan C, Gratzer B, Andrews R. Feasibility, acceptability, and preliminary efficacy of an online HIV prevention program for diverse young men who have sex with men: the keep it up! intervention. AIDS Behav. 2013;17(9):2999–3012.
Zaneva M, Philpott A, Singh A, Larsson G, Gonsalves L. What is the added value of incorporating pleasure in sexual health interventions? A systematic review and meta-analysis. PLoS ONE. 2022;17(2): e0261034.
NYC Health. PlaySure network for HIV prevention. 2023. https://www.nyc.gov/site/doh/providers/resources/playsure-network.page. Accessed 20 Oct 2023.
Ezennia O, Geter A, Smith DK. The PrEP care continuum and Black men who have sex with men: a scoping review of published data on awareness, uptake, adherence, and retention in PrEP care. AIDS Behav. 2019;23(10):2654–73.
Blashill AJ, Brady JP, Rooney BM, et al. Syndemics and the PrEP cascade: results from a sample of young Latino men who have sex with men. Arch Sex Behav. 2020;49(1):125–35.
Biello KB, Mimiaga MJ, Santostefano CM, Novak DS, Mayer KH. MSM at highest risk for HIV acquisition express greatest interest and preference for injectable antiretroviral PrEP compared to daily, oral medication. AIDS Behav. 2018;22(4):1158–64.
Morton T, Chege W, Swann E, et al. Advancing long-acting and extended delivery HIV prevention and treatment regimens through behavioural science: NIH workshop directions. AIDS. 2021;35(8):1313–7.
Acknowledgements
The authors thank the members of the Coalition for Applied Modeling for Prevention (www.emorycamp.org), especially Monica Trigg and Abby Tighe. We also thank the members of our Public Health Advisory Group, especially those who provided feedback on this work: Drs. Thomas Bertrand, Daniel Driffin, Gregory “Grey” Felzien, Jane Kelly, and Ron Valdiserri. Our gratitude to all of our CDC partners, including Taiwo Abimbola, Terrika Barham, and Ranell Myles. Thank you to the entire AMIS team and to Dr. Jane Lee.
Funding
This study was supported by National Center for HIV/AIDS, Viral Hepatitis, STD, and TB Prevention (Grant No. U38-PS004646), National Institute of Allergy and Infectious Diseases (Grant Nos. R01-AI138783, P30-AI027757), National Institute of Child Health and Human Development (Grant No. P2C-HD042828).
Author information
Authors and Affiliations
Corresponding author
Ethics declarations
Competing Interests
The authors have no competing interests to declare that are relevant to the content of this article. The findings and conclusions in this report are those of the authors and do not necessarily represent the official position of the Centers for Disease Control and Prevention.
Additional information
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
The original online version of this article was revised: In the sentence beginning “Meanwhile, there is a paucity of quantitative research...” in this article, the text “An Australian study found no significant changes in the proportion of HNM-NP having condomless anal sex (CAS) with casual partners from 2013 (30%) to 2017 (29)% [13].” should have read “An Australian study found that the proportion of HNM-NP having condomless anal sex (CAS) with casual partners increased from 30% in 2013 to 41% in 2017 [13], further increasing to 45% in 2019 and in the sentence beginning “To our knowledge” the text “this study is the first since the onset of PrEP and U = U to document CAS increases specifically among MSM who are HIV-negative and off PrEP” should have read “this study is the first since the onset of PrEP and U = U to document CAS increases specifically among MSM who are HIV-negative and off PrEP in the United States.”
Supplementary Information
Below is the link to the electronic supplementary material.
Rights and permissions
Springer Nature or its licensor (e.g. a society or other partner) holds exclusive rights to this article under a publishing agreement with the author(s) or other rightsholder(s); author self-archiving of the accepted manuscript version of this article is solely governed by the terms of such publishing agreement and applicable law.
About this article
Cite this article
Goodreau, S.M., Barry, M.P., Hamilton, D.T. et al. Behavior Change Among HIV-Negative Men Who Have Sex with Men Not Using PrEP in the United States. AIDS Behav 28, 1766–1780 (2024). https://doi.org/10.1007/s10461-024-04281-7
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s10461-024-04281-7