Reducing substance use and sexual risk behaviour among men who have sex with men in South Africa

FUNDING: US Centers of Disease Control and Prevention (CDC): 5U2GPS001137-05 Men who have sex with men have been identified as a population at risk of acquiring and transmitting HIV. Studies in South Africa have reported a high prevalence of HIV, as well as high levels of alcohol and other drug use, among men who have sex with men, and the use of substances (alcohol and drugs) to facilitate their sexual encounters. Since 2007, interventions focused on prevention have been rolled out to vulnerable men who have sex with men and who also use alcohol or other drugs. The interventions include community-based outreach; provision of information on HIV/AIDS, substance abuse, and safer sex practices; and the development of risk-reduction plans. Among 195 men who participated in our study, there were significant reductions in the proportion who used cannabis and ecstasy, including the use of these drugs during sex. No reduction was observed in the use of any other substances. In general, after the intervention our participants reported less frequent use of alcohol and drugs and greater engagement in safer sexual practices. Despite these encouraging findings, the combination of substance use while engaging in sex had actually increased. The study findings suggest that interventions that target men who have sex with men, and who use alcohol and other drugs, could reduce risk behaviours in this population.


Introduction
South Africa is a country considered to have the worst HIV epidemic worldwide.In 2012, an estimated 12.2% (6.4 million persons) of the population were HIV-positive. 1 Although researchers have not determined the national HIV prevalence among men who have sex with men (MSM) in South Africa, this group has been identified as being at particular risk of acquiring HIV.[4][5] Local research has also documented high levels of drug use among MSM, and the use of drugs to facilitate sexual encounters. 6The link between drug use and HIV risk behaviour among MSM (including multiple sexual partners, unprotected anal intercourse, condom use and transactional sex) has been established in international literature. 7The need to develop and accelerate the rollout of evidence-based interventions to address the related risks in this population has been clearly articulated. 7South African studies have also identified the need for targeted risk reduction interventions to address the link between substance use and HIV risk behaviour among MSM. 8,9Interventions focusing on risk reduction strategies among substance-using MSM have been shown to be successful in reducing risky sexual behavior. 10,11M in South Africa remain under-served in terms of HIV prevention and treatment services, despite studies revealing high risk of HIV infection among this population.In addition to the incidental use of drugs during sexual encounters, drugs are also used specifically to enhance sexual interactions.Aspects of such usage include enhancing the sexual experience, increasing sexual arousal, facilitating sexual encounters, increasing the capacity to engage in particular sexual activities, increasing the length of sexual interactions and facilitating sex work. 12ccording to McIntyre et al. 12 , despite drug use being well documented among MSM in South Africa, there are few programmes that target this aspect of HIV risk.McIntyre et al. argue that a range of possible interventions already exist and could be developed to target drug-using MSM.These include a specific focus on drug use in the context of sex parties, group sex and sex-on-site venues; non-judgemental materials focusing on risk reduction when using drugs; information about the risks of combining different substances (such as 'poppers' -amyl nitrate or butyl nitrate -and erectile dysfunction drugs); and the development of MSM-specific drug use risk reduction counselling training for health workers and people working with drug users. 12uth Africa's second National Drug Master Plan gave prominence to the need to address drug abuse as part of broader HIV prevention efforts. 13However, interventions targeting MSM to address drug use and sexual risk behaviour remain scarce in South Africa.Recently, the feasibility of delivering large-scale interventions to this particular group has been demonstrated locally. 14The aim of our study was to test whether an intervention aimed at MSM who use substances (alcohol and other drugs) could affect risky substance use and sexual behaviour.

Method
In 2007, in collaboration with a local NGO in Pretoria (OUT LGBT), an initiative was begun to implement a number of harm-reduction strategies for MSM who use alcohol and other drugs.This NGO was selected because of its expertise and extensive experience working with the target population.Important components of the intervention included community outreach, distribution of condoms and lubricants, HIV risk assessment and risk reduction counselling, expanded access to HIV counselling and testing, care and treatment of HIV and sexually transmitted http://www.sajs.co.za

Participants and data collection
The intervention was conducted in various areas in and around Johannesburg and Pretoria in Gauteng, and around Nelspruit in Mpumalanga.A number of locations were targeted, including gay and lesbian pride events, gay clubs and areas known to be frequented by MSM, and through site work at the NGO's premises or clinics and wellness centres, where the NGO made use of negotiated spaces within such clinics or centres.
To be included in the study, participants had to be 16 years or older.They also had to be self-reported MSM who were using alcohol and/ or drugs, regardless of whether they self-identified as gay, bisexual or heterosexual, and regardless of HIV status.The quantity of alcohol or drugs used by participants was not set as an exclusion criterion.Participants who were eligible for inclusion and willing to participate signed an informed consent form, and the principles of the Declaration of Helsinki were adhered to.Peer outreach workers completed a face-toface baseline questionnaire with participants, to record risk behaviours.A risk-reduction plan was developed with each participant, which took into account risks related to injection and non-injection drug use, sexrelated risks and HIV testing.
Intervals between follow-up appointments were not specified.In some instances outreach workers made appointments for follow-up at a time that suited the participants, and in other instances follow-up appointments occurred spontaneously when outreach workers met up with participants at various events or social gatherings.At follow-up, the same questionnaire that had been administered at baseline was completed again.Also at follow-up, clients were asked the frequency of risk behaviour or its reduction, and whether they had been tested for HIV or asked their partner to test for HIV or encouraged a friend to be tested.Behaviour change was thus self-reported rather than observed.
In the areas where the intervention was delivered, MSM participants spoke Sepedi, Setswana, Afrikaans or English.Questionnaires were admi nis tered in English, with outreach workers translating difficult concepts where necessary.Our study is part of a bigger study that includes four provinces of South Africa.Ethical approval for conducting the study was granted by the Health Research Committee of the University of Stellenbosch.

Intervention
The NGO received training on an intervention that was based on a local adaptation of the World Health Organization's Training Guide for HIV Prevention Outreach to Injecting Drug Users. 16The adaptation lessened the focus on injection drug use related behaviours, and placed greater focus on substance-related sexual HIV risk behaviour.The adapted manual also emphasised drugs commonly used in South Africa.The NGO recruited peer outreach workers on a volunteer basis and they were paid a modest stipend.Project coordinators were appointed in the organisation to ensure the smooth running of the project, and to liaise between the NGO and the Medical Research Council project manager.
The intervention included community-based outreach and provision of information on HIV/AIDS, substance use, and safer sex practices.Each MSM who participated developed a personalised HIV risk reduction plan with the help of the outreach workers.The plans were followed up and reassessed with the participants.In addition to the provision of risk reduction counselling, the intervention included the provision of HIV counselling and testing services and referrals to treatment and care for substance abuse, HIV or sexually transmitted infections, and other social services.Intervention services were monitored on a monthly basis and evaluated biannually.

Data analysis
Demographic information was collected at the first contact (baseline), and descriptive statistics were calculated for these variables.Data on substance use and sexual risk behaviour were also collected at baseline and at each follow-up appointment.We assessed the distribution of behavioural data using the Shapiro-Wilks test of normality, and the results indicated that the data were non-normally distributed (p<0.001).Therefore, we conducted bivariate analysis for continuous variables (number of sex partners, number of times engaged in sex, number of times had unprotected sex, number of times traded sex, and number of times had sex under the influence of substances).The Wilcoxon signed rank test was used to determine whether selfreported differences between baseline and follow-up were statistically significant.For categorical variables (types of substances used, either in general or during sex), we assessed the differences in proportions with the chi-square test of association.All statistics were analysed at 95% confidence intervals, and data analysis was performed with SPSS version 21.

Sample characteristics
The early years of implementation of the intervention were 2007 to 2009.This paper reports on implementation activities between 2010 and 2012.Baseline information was collected from 195 MSM who were recruited into the study, whose substance abuse and HIV risk profile was assessed.The participants then received an intervention and were subsequently followed up.Of these, 27.7% had one repeat contact with outreach workers, 30.3% had two repeat contacts, 29.7% had three repeat contacts and 10.8% had four repeat contacts.The data presented are for first contact (baseline) and the final follow-up contact for a particular participant.The median age of the MSM participants was 27 years, and participants had a median of 12 years of formal education.The majority of participants in our study were employed (56.4%) and single (66.7%).Table 1 shows the demographic data.

Substance use
We calculated the proportion of participants who used each substance at baseline (time 1 ) and at last follow-up (time 2 ), and the difference between these two time points (Table 2).Substances used by less than 5% of the sample were excluded.Only four of the 195 participants reported injection drug use (at baseline).There were significant reductions in the proportion of MSM who used cannabis (p=0.009) and ecstasy (p=0.043)from time 1 to time 2 .No significant differences were observed for the use of alcohol, cocaine, inhalant, amyl or butyl nitrate, and overthe-counter or prescription drugs from time 1 to time 2 .There was also a highly significant change in the frequency of substance use (p≤0.001), with participants using substances less frequently after the intervention.At time 1 , 30.3% reported daily use of substances, whereas only 13.8% reported daily use at time 2 , with a greater number of participants using substances only once a week or less (Table 3).The use of more than one substance (polydrug use) was recorded at time 1 and time 2 .After the intervention, 35.9% of participants reported no change in the number of different substances used, and 28.2% had actually increased the number of different substances they used.However, 35.9% had decreased the number of different substances used between time 1 and time 2 (Table 4).

Sexual risk and substance use
As shown in Table 6, there were significant reductions in the proportion of participants who used cannabis (p=0.005) and ecstasy (p=0.04)during sex, from time 1 to time 2 .No significant differences were observed between time 1 and time 2 for the use of alcohol, cocaine, inhalants, amyl or butyl nitrate, and over-the-counter or prescription drugs during sex.However, participants engaged in sex while using drugs and alcohol more frequently after the intervention (median sexual events: time 1 =10; time 2 =18; z=-3.465;p=0.001).In total, 34.4% of participants reported no change in the number of different substances used during sex, whereas 28.7% had increased the number of different substances used during sex, and 36.9% had decreased the number of different substances used during sex (Table 7).

Discussion
MSM participants in our study engaged in multiple alcohol or other drug use and sexual risk behaviours.However, as reported in an earlier phase of this project 14 , participants were receptive to NGOs delivering HIV prevention and substance risk reduction services, and were willing to develop individual risk reduction strategies with outreach workers.Other intervention studies with substance-using MSM have also demonstrated that it is possible to enrol and retain MSM in intervention studies. 7r study showed that the only significant reduction after the intervention was the proportion of MSM who used cannabis and ecstasy, including the use of these drugs during sex.However, participants did report less frequent alcohol or other drug use at Time 2 .Additionally, more participants had reduced the diversity of substances used relative to the number of participants who increased their polysubstance use (including the use of substances during sex).Most importantly, participants reported safer sex practices following the intervention, including having fewer sexual partners and more frequent condom use during sex.However, after the intervention, participants engaged in sex while using drugs and alcohol more frequently than they had before the intervention.Other studies have similarly found that risk reduction, rather than 'abstinence only' strategies, are an effective approach for this population. 17idence suggests that behavioural interventions to reduce risk for sexual transmission of HIV among MSM can be effective.A metaanalysis that reviewed HIV behavioural interventions for reducing sexual risk behaviour of MSM found, as did our study, that interventions were associated with a significant decrease in unprotected anal intercourse and number of sexual partners, and with a significant increase in condom use during anal intercourse. 11A Cochrane review found that these interventions can lead to significant risk reduction among MSM, specifically a reduction in unprotected anal intercourse. 18However, the authors distinguish between individual, group and community-level interventions; not all interventions are equally effective. 18randomised controlled trial was conducted in the U.S. to test a cognitivebehavioural group intervention outside treatment centres, focusing on reducing sexual risk behaviour of substance-using MSM. 19The study found the intervention was no more efficacious than the control condition in reducing sexual risk behaviours or substance use.Instead, substantial reductions occurred in both the intervention and control groups. 19In another trial examining the effects of motivational interviews on club drug use and risky sex, with MSM participants who did not seek treatment, motivational interviews were not found to be effective in reducing risky sexual behaviour. 20However, the motivational interviews did result in significant reductions in club drug use in the intervention group compared with the control group -but only among participants with lower severity of drug dependence. 20Based on these findings, it appears that although risk-reduction interventions have demonstrated their ability to reduce risky behaviour, further research is needed to determine which interventions are most efficacious. 7e findings of our study are subject to the following limitations.Firstly, risk behaviours were self-reported by the MSM participants and might be subject to social desirability or other biases.Although peer outreach workers established relationships with the participants and this may have facilitated honest self-reporting, we did not confirm the self-reports by biological testing for substance use.Secondly, questionnaires were administered in English, with outreach workers translating difficult concepts where necessary.This approach is not as robust as written and tested translation, and interpretation errors can occur when English-medium tools are used with non-native English speakers.
Thirdly, there was no comparison group and all participants received the same intervention.Thus, it was not possible to conclude that changes in risky behaviour were the result of the intervention.Finally, we recruited participants only from certain areas in and around Gauteng and Mpumalanga, and the results should not be generalized to other provinces and areas.As reported in other studies, our participants were recruited by outreach workers from local MSM NGOs, which further limits the generalizability of the findings. 21spite these limitations, the findings of our study (and others) demonstrate that interventions targeting substance-using MSM can reduce risk behaviour among this population.In South Africa, programmes that target MSM for HIV prevention, treatment and care are not well developed.Where they do occur, they are provided mostly by NGOs or community-based organisations in specific geographical areas and have not been up-scaled across the country. 12As a result there remains a need to offer widespread integrated services related to alcohol or drug use and HIV risk, especially where community-based outreach is available and is tailored for MSM who use alcohol or other drugs.What the current study has shown is that while interventions with MSM can have a positive effect in reducing HIV risk behaviours (such as having sex without a condom), the effects on substance-use behaviour are modest.More work is needed to determine how substance abuse behaviour among MSM can be reduced effectively.

substance use and sexual risk behaviour among MSM Page 2 of 5 diseases
, and referrals to substance abuse treatment and other social services.The intervention also aimed to build referral networks, decrease stigma among service providers, and influence NGO practices and government policies on drugs and HIV.These activities were in line with the six core elements defined in the Technical Guidance on Combination HIV Prevention for effective MSM programmes as released by the Office of the U.S. Global AIDS Coordinator in May 2011, as part of the U.S. President's Emergency Plan for AIDS Relief's overall prevention strategy.15

Table 2 :
Prevalence of substance use at time 1 and time 2

Table 3 :
Frequency of substance use at time 1 and time 2

Table 4 :
Change in drug and alcohol use Key: AOD, alcohol and other drugs

Table 5 :
Sexual risk behaviour reported at time 1 and time 2

Table 7 :
Change in alcohol and drug use during sex