Article Text

Download PDFPDF

Original article
Randomised controlled trial of a sexual risk reduction intervention for STI prevention among men who have sex with men in the USA
  1. Lisa A Eaton1,
  2. Seth C Kalichman2,
  3. Moira O Kalichman2,
  4. Daniel D Driffin2,
  5. Robert Baldwin2,
  6. Larissa Zohren2,
  7. Christopher Conway-Washington2
  1. 1 Human Development and Family Studies, University of Connecticut, Storrs, Connecticut, USA
  2. 2 Department of Psychological Sciences, University of Connecticut, Storrs, Connecticut, USA
  1. Correspondence to Dr Lisa A Eaton, Center for Health Intervention and Prevention, University Connecticut, 2006 Hillside Rd, Storrs, CT 06279-1248, USA; lisaanne.eaton{at}gmail.com

Abstract

Objectives Novel interventions to address sexual risk taking and slow rates of STIs are urgently needed, in particular among black men who have sex with men (MSM) in the USA. Serosorting, or limiting condomless sex acts to partners of the same HIV status, is commonly practised among MSM, yet can lead to STI and remains largely unaddressed by public health agencies.

Methods A two-arm, randomised controlled trial was conducted from 2012 to 2015. This trial assessed the effects of a single-session, sexual partner selection and risk decision intervention (experimental arm) versus a single-session, Centers for Disease Control and Prevention-based, sexual risk reduction intervention (control arm) on psychosocial measures, sexual risk taking and STI.

Results At study follow-ups, multiple beneficial changes were observed on sexual risk beliefs measures (ie, changes in serosorting and condom use beliefs, and HIV risk perceptions) and sexual risk taking among the experimental arm relative to the control arm. Overall main effects, however, of the intervention on STI outcomes on year-long follow-ups were non-significant. There was evidence for short-term effects on STI outcomes, and self-report of multiple STIs and STI symptoms demonstrated positive effects over the follow-up period.

Conclusions Brief interventions to address sexual risk taking can result in short-term beneficial outcomes and can be incorporated into currently existing infrastructure at healthcare agencies. Additional intervention will be necessary for demonstrating long-term results.

Trial registration number NCT02128594.

  • BEHAVIOURAL SCIENCE
  • GAY MEN
  • HOMOSEXUALITY
  • INTERVENTION STUDIES
  • PREVENTION

Statistics from Altmetric.com

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.

Men who have sex with men (MSM) continue to comprise the majority of incident HIV/STI diagnoses in the USA; 66% of new HIV infections are among MSM and rates of HIV among MSM are 44 times that of other men.1 Specific subpopulations in the USA, in particular African-American/black MSM populations in the southeastern USA, experience alarming numbers of HIV/STI transmissions without any evidence of the epidemic slowing.2 ,3 Despite relatively high rates of both HIV and STI among MSM, trials of behavioural interventions to reduce sexual risk taking and prevent STI among this population are surprisingly few.4

It is largely acknowledged in the literature that in order to slow the HIV/STI epidemics observed among MSM in the USA, it is necessary to implement combination forms of prevention (eg, behavioural, structural and biomedical interventions).5 ,6 Available infrastructure for delivering these interventions, however, is underfunded and the needs of individuals at-risk for or living with HIV/STI far exceed the resources available for them at local health departments and community-based organisations.7 One underused mode of intervention that meets the demands of resource-limited settings and has demonstrated efficacy in reducing sexual risk taking is that of single-session intervention counselling.8 Intervention counselling delivered within one, brief session that is integrated into the larger infrastructure of existing health agencies (eg, during HIV test counselling) may provide a beneficial option for affecting behaviour change.

One area of sexual risk taking—that can be addressed using a single-session intervention delivery format—is that of engaging in serosorting (ie, limiting condomless sex partners to those of the same HIV status) for HIV prevention. Serosorting is a commonly used harm reduction practise among MSM, but it relies on having accurate information about HIV status which can be difficult to ascertain. Moreover, serosorting may provide individuals with a false sense of security about their likelihood of exposure to HIV/STI. Research on the overall effectiveness of serosorting for preventing HIV transmission is mixed,6 ,9–12 yet there is considerable evidence that reliance on serosorting can lead to HIV/STI transmission due to difficulties (eg, infrequent HIV testing, acute HIV infection, misrepresentation of HIV status) in being certain of one's own and one's partner’s HIV statuses and lack of explicit conversations around HIV status disclosure.13–17 Furthermore, serosorting does not explicitly address STIs, which can also exacerbate HIV transmission risks.

In the current study, MSM reporting sexual risk taking in the past year and testing HIV negative were randomised to a two-arm trial for counselling that included: (a) a single-session, Conflict Theory of Decision Making (CTDM)-based18 intervention focused on sexual risk decision making including the limitations of serosorting (experimental arm) or (b) a single-session, Centers for Disease Control and Prevention (CDC)-based, sexual risk reduction intervention (control arm). It was hypothesised that participants receiving the CTDM-based, experimental intervention would report decreased agreement with serosorting strategies and sexual risk behaviours, and demonstrate reduced STI relative to participants receiving the control intervention at follow-ups. CTDM guides decision making under conditions of uncertainty by highlighting the importance of gathering all relevant information, weighing the costs and benefits of each possible decision and choosing the decision that best represents an individual's needs. CTDM is based on the perspective that individuals will benefit from greater risk awareness and understanding of behavioural risks posed to them, which will guide them to more effective behavioural decisions.

Methods

Participants and setting

Participants were 600 men residing in and around Atlanta, Georgia, USA who were recruited at lesbian, gay, bisexual and transgender identified venues (eg, bars/clubs/parties) and online sites (eg, dating sites and apps), and over the phone through fliers and word-of-mouth. Enrolment occurred between December 2012 and November 2014. All study protocols received Institutional Review Board approval and the trial was registered in the clinical trials registry, clinicaltrials.gov (NCT02128594). Study entry criteria included: 18 years of age or older, assigned male gender at birth and identify as male or transgender female, and report HIV-negative or unknown serostatus, two or more male sex partners in the past year and condomless anal sex in the past year.

Study procedures

Prior to study enrolment, all participants were tested for HIV using OraQuick ADVANCE Rapid HIV 1/2 Antibody Test. Participants testing HIV negative and meeting eligibility criteria were offered enrolment into the trial, participants testing HIV positive were linked to care and offered enrolment in other available studies. All study procedures occurred at a community-based research site. Using randomisation software, all participants were assigned to one of two trial counselling arms. Counsellors were trained in client-centred counselling and motivational interviewing, and intervention sessions were audio-recorded and evaluated (using a checklist rubric) for fidelity. Participants completed baseline, and 3-month, 6-month and 12-month follow-up appointments (inclusive of a computerised assessment and STI testing) and were compensated US$45 for each assessment. Using nucleic acid amplification testing (NAAT), participants were tested for Chlamydia trachomatis and Neisseria gonorrhoeae through self-collected samples of urine and rectal swabs at each assessment (baseline, and 3-month, 6-month and 12-month follow-ups).19 Participants testing positive for Chlamydia or Neisseria were linked to treatment and followed-up for confirmation of treatment by the project manager. Study materials are accessible in the online supplementary appendix.

Experimental condition

The single-session, experimental condition lasted on average 45 min.20 The main foci of the intervention were to highlight misbeliefs about selecting sex partners, to shape accurate beliefs and perceptions of risk about the use of serosorting for HIV/STI prevention and to determine a practical, risk reduction strategy tailored for each participant. A graphic novel was created to convey messages about serosorting. The novel depicted a fictitious (but evidence-based) story of a man who tests HIV negative, uses serosorting as an HIV prevention strategy and then tests HIV positive at the end of the story. This activity led to a discussion about how the main character could have become HIV positive (eg, acute HIV infection, non-explicit disclosure of HIV status, misrepresentation of HIV status, infrequent HIV testing) or transmitted HIV to his partners. Guided by CTDM, the counsellor and participant worked together to identify and discuss these varying scenarios with a focus on what the main character could have done to reduce his risk for HIV. CTDM guides decision making by highlighting the importance of gathering all relevant information, weighing the costs and benefits of each decision and choosing the decision that best addresses an individual's needs. Next, participants were asked to create a sexual network diagram, wherein they provided information on their recent sex partners and sex acts. Participant diagrams were compared with the character's diagram, thereby allowing the participants to observe how their behaviours related to those of an evidence-based character who tests HIV positive. Through this activity, participants reflected on instances in which they potentially exposed themselves to HIV/STI. Participants used their sexual network diagram as a guide to forming a plan they could carry out to reduce their risks. In keeping with informed decision making, participants were guided towards a risk reduction plan that they considered agreeable. Plans included increases in condom use, reductions in sex partners and acts, alternatives to condomless anal intercourse and greater inquiry into a sexual partner's HIV status and testing history. Thus, the participant along with the counsellor generated a menu of harm reduction options by weighing the relative costs and benefits of each and deciding on the optimal choice for the participant.

CDC-based, sexual risk reduction, control condition

Participants in the control condition received a contact-matched, standard-of-care, HIV/STI risk-reduction counselling session consistent with CDC post-HIV test guidelines.21 The session was tailored to address sexual risk taking, safer sex practises and inhibitors and triggers to sexual risk taking. Substance use in the context of sexual decision making was a primary focus. Counselling regarding serosorting-related practises was not included in order to avoid contamination with the experimental arm.

Measures

Sociodemographic characteristics

Participants reported on age, race/ethnicity, gender identity, sexual orientation, relationship status, education, income, substance use and depression (10-item Center for Epidemiological Studies Depression Scale Revised,22 ≥10 was used as cut-off, Cronbach's α=0.82).

Sexual risk beliefs measures

To assess serosorting beliefs, participants were asked five questions including: “If my partner tells me his HIV status is HIV negative then I worry less about HIV”, Cronbach's α=0.72 (1=strongly disagree, 6=strongly agree).23 HIV risk perceptions was assessed using five items and included asking participants to rate sexual risk taking scenarios for perceived risk of HIV,24 Cronbach's α=0.80 (0=no or low risk to 10=very high risk). Condom use beliefs related to serosorting 23 was assessed using a nine-item measure including: “I did not use a condom because my partner has the same HIV status as me so it doesn't matter”, Cronbach's α=0.85 (1=strongly disagree, 6=strongly agree). Finally, HIV status disclosure was assessed using three items including: “I am certain I can have a discussion about HIV status before having sex with a new partner”,25 Cronbach's α=0.76 (1=strongly disagree, 6=strongly agree).

Sex behaviour and self-reported STI diagnoses and symptoms

Sexual risk taking was measured using three main variables: (1) proportion of anal sex acts with condoms used, (2) number of condomless anal sex acts (on the whole and separated by insertive and receptive position) and (3) number of male sex partners.26 All variables were based on a 3-month recall period. For self-report of STI diagnosis, participants were asked to report whether they had been diagnosed with chlamydia, gonorrhoea, syphilis, herpes or genital warts in the past 3 months, and whether they were experiencing burning during urination, discharge from their penis, sores on their penis or sores in and around their rectum.

Biologically assessed STIs

Participants self-collected urine samples and rectal swabs for gonorrhoea and chlamydia testing. Results were reported both as a composite of STI and separately by site of infection. NAAT was conducted for all STI tests.

Data analysis

Descriptive data were provided including means and SDs, or numbers and percentages for all our study variables. For all analyses (ie, sexual risk belief outcomes, sexual risk behaviour outcomes and STI outcomes), generalised estimating equations (GEE) were conducted using a negative binomial distribution for count data, a binary logistic distribution for dichotomous data and a linear distribution for normal data. An intent-to-treat data-analytic approach was employed. Baseline behavioural data were entered as covariates, and condition, time and condition× time were modelled. IBM SPSS Statistics V.21.0 (SPSS, Chicago, Illinois, USA) was used for all of the analyses.

Results

Figure 1 presents the complete study screening/eligibility data and retention rates (87% at 3-month, 86% at 6-month and 83% at 12-month assessments). There were no significant differences between trial arms on any demographic variables. A majority of the sample identified as black/African-American/African diaspora (N=530, 88.8%), male (N=554, 92.8%), same gender loving/gay (N=260, 44%), college educated (N=360, 60.3%), having ≤US$20 000 in income (N=438, 74.0%) and not being in a relationship (N=371, 62.1%). Half of the sample reported depressive symptoms indicative of needing further evaluation for mental healthcare (N=298, 50.1%) (table 1).

Table 1

Baseline demographics and health characteristics of clinical trial participants

Figure 1

Flow chart of participants in randomised controlled trial to reduce sexual risk taking and prevent STI. Note: Recruitment efforts for multiple study opportunities occurred simultaneously. The co-occurrence of multiple study enrolment periods affected (increased) the number of participants reporting HIV-positive status at screening.

Sexual risk beliefs measures

Participants in the experimental arm were significantly less likely to endorse serosorting as a form of HIV prevention at study follow-ups than control arm participants (Wald χ2=8.29, p<0.01) (table 2). HIV risk perceptions were significantly higher among participants in the experimental arm compared with the control arm at study follow-ups (Wald χ2=10.91, p<0.01) and HIV risk perceptions decreased among both groups over the follow-up period (Wald χ2=7.52, p<0.05). Participants in the experimental arm were significantly less likely to endorse beliefs that condoms are unnecessary if engaging in serosorting behaviours than participants in the control arm (Wald χ2=8.29, p<0.05). There were no differences in HIV status disclosure at study follow-ups. All analyses controlled for the baseline assessment of each examined variable (ie, serosorting beliefs, risk perceptions, condom beliefs and HIV status disclosure).

Table 2

Sexual risk beliefs, sex behaviour and self-reported STI diagnosis outcomes at 3-month, 6-month and 12-month follow-up

Sex behaviour, and self-reported STI diagnoses and symptoms outcomes

The proportion of condom-protected sex acts was significantly higher among participants in the experimental arm than in the control arm (Wald χ2=4.09, p<0.05) (see table 2). Likewise, the number of condomless anal sex acts reported at follow-up was significantly lower among the experimental group compared with the control group (Wald χ2=4.00, p<0.05). Condomless insertive sex acts was significantly lower among the experimental group (Wald χ2=10.53, p<0.01), while condomless receptive sex acts was non-significant between groups (Wald χ2=.24, p=not significant). Number of male sex partners decreased over time for both groups (Wald χ2=14.90, p<0.01). Participants in the experimental arm reported reduced self-reported STI diagnoses (Wald χ2=2.87, p=0.09) and STI symptoms (Wald χ2=3.59, p=0.06) results that were trending towards significance compared with participants in the control arm. All analyses controlled for baseline sex behaviour/STI diagnoses or symptoms, relationship status and sexual orientation.

STI outcomes

At baseline, across study arms, 12.3% (N=73) of sample was diagnosed with an STI (3.4% urine, N=11 chlamydia, N=9 gonorrhoea and 9.5% rectal, N=36 chlamydia, N=30 gonorrhoea (N=9 with dual diagnosis)). There were no significant STI test result differences at baseline. At follow-ups, results of the analyses indicated no overall effect of intervention condition, time or interaction on urine STI diagnoses (table 2). Post hoc pairwise analyses did demonstrate significant difference in urine diagnoses between conditions at the 3-month follow-up (1.5% intervention and 4.9% control, p<0.05). For rectal STI, there were no overall effects of intervention condition, time or interaction on STI diagnoses (table 2).

Discussion

Findings from the current study, among a sample of primarily black/African-American MSM, demonstrate varying results regarding intervention effects on psychosocial variables, sexual risk behaviour and STIs. Multiple psychosocial and behavioural outcomes demonstrated positive changes in beliefs regarding reliance on serosorting, HIV risk perceptions, condom use and sex behaviours. Overall main effects of the intervention on STI outcomes, however, were non-significant. There was evidence for short-term effects on urine STI diagnosis, and self-report of STIs and STI symptoms demonstrated modest effects over the follow-up period. On the whole, additional interventions (eg, more intensive counselling and multilevel interventions) are needed to strengthen the effects of the current intervention. Findings should be considered in light of the fact that the experimental intervention was a brief, single counselling session. Unlike many behavioural interventions for affecting behavioural change,27 the experimental intervention was low burden (ie, a single, brief intervention) and can be implemented within currently existing infrastructure including during post-HIV test counselling. The low burden of the intervention bolsters its likelihood of community uptake and the possibility of sustainability of outcomes (through repeated intervention exposure via repeat HIV testing).

Results demonstrate the efficacy of using a single-session intervention to affect sexual risk reduction beliefs and behaviours, but not in overall reduction of STI. The urine STI reduction among experimental group participants observed at the 3-month follow-up is consistent with the reduction in insertive condomless anal sex that was also observed. However, reductions in receptive condomless anal sex and rectal STI were not found. These findings suggest that the intervention demonstrated efficacy in changing participant's own condom use, but not that of their partner. This distinction in condom usage is noteworthy and underscores the need for intervention content to further address the challenges in influencing partner condom use. Likewise, the individual-level focus of the intervention also limits the extent to which behaviour change will yield consistent and sustained reductions in STI.28 Combining this individual-level focused intervention with higher-level focused interventions such as a sexual networks-level (eg, sexual-network risk reduction, partner notification of STI), or community-level (eg, mobile HIV/STI testing and treatment, addressing logistical barriers to care linkage and retention) intervention would likely provide additional strength in preventing STI beyond behavioural change.

Findings from the study advance the literature in important ways. Currently, there are 96 risk reduction interventions available in the CDC's compendium of evidence-based risk reduction interventions, yet only 2 of these interventions were developed primarily for HIV-negative black MSM.29 The limited focus on black MSM is problematic when considering the alarmingly high rates of HIV/STI transmission among this population. Furthermore, of the available interventions, both rely on multiple sessions which limits their likelihood of uptake for community agencies with limited resources.

There are multiple important points of consideration in interpreting the current findings. First, half of the sample reported depressive symptoms indicating the need for greater mental health assessment. This rate is alarmingly high and the need for comprehensive mental healthcare among this sample is urgent. Second, alcohol, marijuana and other drug use were frequently reported among the sample. The elevated rates of substance use are consistent with the high rates of depressive symptoms and suggest that substance use counselling, for many participants, is also needed. Regarding demographic data, it is worth noting that a majority of the sample reported receiving at least some college as their educational level, yet a majority of the sample reported making ≤US$20 000. The impact of concentrated disadvantage (eg, limited community resources and social networks) may explain the discrepancy between education level and limited income. With these findings, it is apparent that mental health and substance use interventions, along with community targeted interventions to improve opportunity for advancement, remain sorely unaddressed among communities hardest hit by HIV/STI epidemics.

The outcomes of the trial should be considered in light of their limitations. Findings are limited to a sample of MSM reporting sexual risk taking and residing in and around the Atlanta, Georgia, USA area and may not be generalisable to all MSM. The effect of the experimental intervention may have been muted by using a control arm focused on sexual risk reduction. Lab-based STI testing focused on chlamydia and gonorrhoea, therefore, intervention effects on other biologically assessed STIs were not included.

Novel approaches to sexual risk reduction counselling, such as the serosorting focused intervention described here, offer an important component of combination prevention needed for MSM, in particular black MSM. Brief, single-session interventions for MSM, however, will need additional support to strengthen effects and longevity. Efforts to slow STI/HIV epidemics remain an urgent need and developing comprehensive approaches to prevention must be prioritised.

Key messages

  • In the USA, rates of STI among men who have sex with men (MSM), in particular, black MSM, are alarmingly high and demand urgent attention.

  • Sexual risk reduction interventions are a critical component of combination STI prevention among MSM, however, interventions for MSM that can be implemented with minimal resources are limited.

  • Sexual risk reduction interventions that address the risks associated with serosorting and other partner selection strategies can affect psychosocial factors associated with sexual risk taking.

Acknowledgments

The study authors acknowledge the contributions of Megan McNerney, Tamar Grebler, Christopher Linton, Cynthia Merly, Ginger Hoyt, Brandi Welles, Chauncey Cherry, Kevin English and Harlan Smith.

References

Footnotes

  • Handling editor Jackie A Cassell

  • Contributors LAE conceptualised study, conducted data analyses and wrote manuscript; SCK conceptualised study; MOK contributed to study implementation; DDD conducted study management and RB and LZ implemented study activities.

  • Funding This study was funded by the National Institutes of Health grants R01MH094230, R01NR013865 and P30AI050409 (Emory University Center for AIDS Research).

  • Competing interests None declared.

  • Ethics approval University of Connecticut (IRB-H10-299).

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