Article Text

Download PDFPDF

Short report
High HIV incidence among young black men who have sex with men: constructing a retrospective cohort from a community health programme
  1. M Reuel Friedman1,2,
  2. Nayck B Feliz3,
  3. Jess Netto4,
  4. Brian J Adams1,2,
  5. Derrick D Matthews1,2,
  6. Ron D Stall2,5,
  7. Ken S Ho6,
  8. Sarah E Krier1,2,
  9. Anthony J Silvestre1,2
  1. 1 Department of Infectious Diseases and Microbiology, Graduate School of Public Health, University of Pittsburgh, Pittsburgh, Pennsylvania, USA
  2. 2 Center for LGBT Health Research, Graduate School of Public Health, University of Pittsburgh, Pittsburgh, Pennsylvania, USA
  3. 3 Division of Adolescent Medicine, Children's Hospital of Pittsburgh, University of Pittsburgh Medical Center, Pittsburgh, Pennsylvania, USA
  4. 4 Community Human Services, Inc., Pittsburgh, Pennsylvania, USA
  5. 5 Department of Behavioral and Community Health Sciences, Graduate School of Public Health, University of Pittsburgh, Pittsburgh, Pennsylvania, USA
  6. 6 Pittsburgh AIDS Center for Treatment, Division of Infectious Diseases, University of Pittsburgh Medical Center, Pittsburgh, Pennsylvania, USA
  1. Correspondence to Dr M Reuel Friedman, P.O. Box 7319, Pittsburgh, PA 15213, USA; mrf9{at}


Objectives We sought to calculate HIV incidence in a retrospective cohort of young (13–29 years old) black men who have sex with men (YBMSM) accessing repeated HIV-antibody testing in a mid-size city in the USA.

Methods We aggregated site-specific HIV-antibody testing results from the project's inception among YBMSM who received an initial negative result and accessed at least one additional HIV-antibody test. From these data, we assessed number of seroconversions and person-years and calculated HIV incidence using a mid-P exact test to estimate 95% CIs.

Results Five seroconversions were documented over 42.3 person-years (the mean age at first onsite test: 19.7 years), resulting in an HIV incidence rate of 11.8% (95% CI 4.3% to 26.2%). The mean age at seroconversion was 20.4 (±3.0) years.

Conclusions Even in mid-size cities with low HIV prevalence rates in the general population, HIV incidence among YBMSM may be high. Community-based HIV-antibody testing organisations serving YBMSM should be encouraged and trained to track repeated HIV testing and calculate HIV incidence rates. Increased resources should be deployed to develop and encourage regular HIV testing in community health sites serving YBMSM.

  • HIV

Statistics from


In the USA, young black men who have sex with men (YBMSM) aged 13–29 years experience the most sharply increasing rate of new HIV diagnoses.1 A recent meta-analysis estimated an annual HIV incidence rate of 4.16% among black men who have sex with men (BMSM), which if sustained until age 40 would yield a prevalence rate of 60%.2 Because YBMSM do not report higher levels of HIV behavioural risks (eg, condomless anal sex, substance use) than other men who have sex with men (MSM) populations, other factors have been proposed to explain this population's disproportionate HIV incidence.3 This disparity has been attributed largely to higher background HIV viral load in BMSM communities and high racial homophily in their sexual networks.2–5 Lowering HIV incidence rates among YBMSM will require significantly improved linkage, re-engagement and retention in care in order to effectuate viral suppression, along with other combination prevention efforts tailored for YBMSM communities.3 Until treatment as prevention within these communities reaches greater saturation, other dedicated primary prevention approaches are urgently needed.

HIV incidence studies are resource-intensive and, by their nature, lengthy; it may be for these reasons that the literature so far contains only a small number that have calculated HIV incidence rates among YBMSM in the USA, though several currently funded studies may soon provide further data.2 ,6–8 However, HIV testing and counselling programmes exist across the USA, and several new funding initiatives from the Centers for Disease Control and Prevention (CDC) have encouraged a focus on HIV testing provided specifically to YBMSM populations. By using data already being collected in sites such as these, public health programmes can augment existing HIV epidemiology. Published studies have focused on HIV incidence among YBMSM in large cities, and none have sampled solely from existing community-based public health practices dedicated to serving YBMSM. We calculated HIV incidence among YBMSM using HIV-antibody testing services at a recreation-based community health site between 2013 and 2015 in Pittsburgh (population 305 704) located in Allegheny County, Pennsylvania, where overall HIV prevalence is a relatively low 0.23%.9



Project Silk, an academic/community partnership between the University of Pittsburgh and Community Human Services, was established as a public health practice demonstration project in 2012. It operates under an exempt approved non-research (quality assurance/evaluation) determination from the University of Pittsburgh's institutional review board (IRB) (PRO12100115). All participants who receive confidential HIV testing and counselling provide consent. Located in a mixed-use building in an accessible location and open evenings and weekends, Project Silk functions to provide a recreation-based community health space for young MSM and transwomen of colour and their social and sexual partners. This population is diverse, including students, House and Ball Community members, artists and designers, homeless and runaway youth and their friends and family (biological and chosen). Every day that a participant enters the space, he or she logs into an electronic, password-protected intake system to request services offered onsite, including HIV and STI testing, linkage to medical care and broad-spectrum social services. When receiving HIV-antibody tests, participants are encouraged to undergo regular testing (e.g., testing every three months). Participants consenting to receive HIV-antibody tests are given a choice of HIV-antibody test and counsellor. Unique identification codes differentiate participants receiving multiple HIV-antibody tests over time.

HIV-antibody testing procedures

Between the initiation of HIV-antibody testing at Project Silk in May 2013 and the end of the demonstration period in December 2015, HIV-antibody testing choices included INSTi HIV-1/HIV-2 Rapid Antibody Test (bioLytical Laboratories, Richmond, British Columbia, Canada), OraSure HIV-1 Oral Collection Device and OraQuick ADVANCE Rapid HIV-1/2 Antibody Test (OraSure Technologies, Bethlehem, Pennsylvania, USA). For indeterminate or positive ELISA results, confirmatory (western blot) testing on oral mucosal or serum samples was conducted by Pennsylvania Department of Health Bureau of Laboratories (2013), the Allegheny County Health Department (2014) and the Center for Disease Detection (2015).

Statistical analysis

We extracted HIV-antibody test records from May 2013 to December 2015. Participants were included in this analysis if they were 13–29 years old at time of first onsite test; identified as man (non-transgender) and Black/African-American; reported sexual activity with men; whose first HIV-antibody test result onsite was negative and who received at least two HIV tests onsite. All others were excluded from analysis. For this secondary data analysis, we considered this group of YBMSM to have constituted a naturally occurring cohort, for example,a cohort that developed from a group of individuals who used the safe community space provided by Project Silk and who elected to receive regular HIV testing there. From this subsample, we extracted data for test results: age at each test and days between first test and last test. We aggregated data to assess seroconversions and person-time. Incident HIV cases were confirmed with the Pennsylvania Department of Health, Division of HIV. We used publicly available software to calculate HIV incidence, using the mid-P exact test to estimate 95% CIs.10 Post hoc analyses assessed rates of condomless anal intercourse (CAI), condomless vaginal intercourse (CVI) and sex with HIV-positive partners, using self-reported data collected during participants' most recent HIV testing and counselling sessions.


Thirty-nine people fit our inclusion criteria (table 1), comprising 42.3 person-years of observation. A total of 149 HIV-antibody tests were performed, with each participant receiving an average of 3.8 tests over a mean observation period of 1.1 years. The mean age at first test was 19.7 years (SE: ±3.7; range 15–29); at most recent test, 20.8 years (SE: ±3.6; range 16–29) and at seroconversion, 20.4 years (SE: ±3.0). There were five seroconversion events. HIV incidence was calculated to be 11.8 per 100 person-years (95% CI 4.3 to 26.2).

Table 1

HIV incidence among young black men who have sex with men receiving repeated HIV-antibody tests: data from a recreation-based community health site, 2013–2015 (n=39)

At their most recent test, 23 (59.0%) reported CAI with a male partner, and three (7.7%) reported CVI. Three individuals (7.7%) reported at most recent test having had sex with an HIV-positive male partner, while 11 (28.2%) reported being unsure whether any male sexual partners were HIV positive. Twenty-four separate ZIP codes were represented by participants; 96% were from the Pittsburgh region (data not shown).


Our findings, indicating that 11.8% of HIV-negative YBMSM in this cohort seroconvert each year, provide more evidence that, in the USA, these communities are at great risk for HIV infection, even in mid-size cities not included in the CDC's 12 targeted urban areas. Few populations worldwide experience such high HIV incidence rates. Nonetheless, our results align with HIV incidence rate estimates for YBMSM in other American cities, which have ranged from 5.1% to 14.7%.2 ,6–8 Even the lower-tail 95% CI in our analysis (4.3%) is consistent with a projected HIV prevalence of 60% by the time this cohort reaches age 40.2 These data demonstrate that even cities with low overall HIV prevalence rates can be settings in which very high HIV incidence rates can exist among vulnerable populations such as YBMSM, and that community-based organisations implementing regular HIV testing can contribute valuable incidence information to locally supplement any existing regional or state data. While the proportion of YBMSM in this sample reporting CAI (59.0%) appear high, the rate reported here is lower than in other recent studies demonstrating that large majorities of MSM, including those who are black, engage in CAI.8 This suggests that the YBMSM in this sample are not engaging in risky sexual behaviours at higher rates than their counterparts in other US locales.

This analysis contains limitations. Although we were able to capture a reasonably long time frame, the small number of participants contributed to a wide CI, and the results should be interpreted with caution. First, we were unable through existing data-sharing protocols to identify and include any individuals who tested HIV negative at our site but who later seroconverted at other HIV testing sites. Second, sexual behaviour data collected during rapid HIV testing events are streamlined in this region; secondary data analysis offers few opportunities to analyse risk or protective correlates, such as social determinants, partner number or PrEP (pre-exposure prophylaxis)/nPEP (nonoccupational postexposure prophylaxis) use. In addition, participants disclose sexual behaviour data to test counsellors, which may heighten social desirability bias in responses. Participants receiving HIV-antibody testing onsite are offered risk assessments, STI testing, sexual health counselling and access to biobehavioural interventions, as well as social support, social service and social capital initiatives intended to facilitate HIV risk reduction; the results may not be generalisable to YBMSM in the region who do not have these opportunities, nor to YBMSM outside the region. However, by restricting analyses to health-seeking participants with verifiable biomedical data, the above limitations engender more conservative HIV incidence rate calculations. Because this site operates as a public health practice, we did not set out to create a research cohort; rather, we evaluated the natural cohort that developed as participants requested repeated HIV-antibody testing over the demonstration project period. While likely extensible to other evaluations of HIV incidence in community-based settings, we recommend that these limitations not deter other organisations from similar calculations, given the scarcity of published HIV incidence data specific to YBMSM in the USA and the sustainability of community health programmes relative to research studies.

This analysis provides additional evidence that YBMSM face extraordinarily high risk of HIV acquisition, highlighting that even cities with low prevalence of HIV infection can be settings in which ‘hidden’ epidemics occur among YBMSM.2–4 ,6 ,8 Prioritising need for HIV prevention services solely on HIV prevalence rates of the general populations in which YBMSM reside will likely result in continuing high rates of infection in new generations of YBMSM as they come of age. The analysis demonstrated here provides a model by which data evaluating front-line HIV prevention services can be used to calculate real-time, real-world HIV incidence rates that can then guide investments in HIV prevention services as the National HIV/AIDS Strategy intends. Increased resources to develop and encourage regular HIV testing and linkage-to-care initiatives in community health spaces serving YBMSM should be deployed in all settings with demonstrably high HIV incidence.


This manuscript is dedicated to the memories of Andre (Pok'ie) Gray (1980–2014) and Omar Islam (1992–2013). The authors would like to thank Jill Garland and Ken McGarvey of the Pennsylvania Department of Health; John Beltrami at the CDC for demonstration project coordination and technical support and Roderick Joiner of the CDC. For essential project contributions, the authors are grateful to Daphne Beers, Michael Brookins, Aria Copeland, Dalen Hooks, Clarisse Jordan, Kenneth McDowell, Terrance McGeorge, Michael Morgan, Adrienne Walnoha, Mona Whitley and the Youth Advisory Board.



  • Handling editor Jackie A Cassell

  • Contributors MRF conducted the analyses and wrote the majority of the manuscript. NBF, JN, BJA and SEK assisted with design, data collection and manuscript refinement. DDM, RDS, KSH and AJS contributed to the introduction and discussion sections of the manuscript.

  • Funding National Center for HIV/AIDS, Viral Hepatitis, STD, and TB Prevention (PS12-1201c). DDM and RDS were funded by the National Institutes of Health (R01 NR013865). Additional support was provided by Pennsylvania Department of Health and Allegheny County Health Department.

  • Competing interests None declared.

  • Ethics approval University of Pittsburgh IRB PRO12100115.

  • Provenance and peer review Not commissioned; externally peer reviewed.

  • Data sharing statement Data sharing is subject to brokered de-identification of data collected and a Data Use Agreement brokered between respective institutions.

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.