Peer Delivered HIV/Syphilis Self-Testing With Assisted Partner Notification Services for Men Who Have Sex With Men (MSM) in Uganda
Peer Delivered HIV/Syphilis Self-Testing With Assisted Partner Notification Services for Men Who Have Sex With Men (MSM) in Uganda
Background and rationale Men who engage in sexual relations with other men (MSM) are disproportionately affected by HIV and other sexually transmitted infections (STI), such as syphilis. Globally, MSM have a 28 times greater risk of HIV acquisition than adult men (15-49 years) in the general population. In Uganda, HIV prevalence in MSM is 13.2% versus 4.7% in similarly aged heterosexual males, while syphilis prevalence is 8.3% versus 5.8% in males aged 15-64 years. Syphilis and HIV transmission share common sexual risk behaviors, and syphilis increases HIV acquisition risk three-fold. Uganda guidelines recommend annual HIV/syphilis testing for MSM, but uptake of facility-based HIV testing is low (32%) of the estimated MSM population in Kampala. Assisted partner notification (services, i.e., tracing sexual partners of people with HIV and offering them testing services, is the standard of care (SOC) in Uganda. The World Health Organization recommends task shifting to MSM peers to increase access to and availability of HIV/syphilis testing services. Differentiated service delivery models, such as peer-delivered HIV/syphilis self-tests and assisted partner notification with linkage to care, could improve engagement in care by MSM. Our prior work found that peer-delivery of HIV self-tests (HIVST) was feasible and acceptable to 90% of MSM in Uganda. All those newly diagnosed with HIV received confirmatory testing, were linked to care, and started antiretroviral treatment (ART). Other work in Zimbabwe found high acceptability (89.6%) of peer-delivered syphilis self-tests among MSM. Joint delivery of peer-delivered HIV/syphilis self-tests and assisted partner notification is an empowering, innovative approach that could substantially increase testing among MSM in Uganda and help achieve global HIV 95:95:95 targets. However, no studies to our knowledge have evaluated the effectiveness of peer-delivered HIV/syphilis self-tests and assisted partner notification services for MSM in any setting.
Study objectives
Study design Objective 1: Cross-sectional qualitative study design (formative stage). Objective 2: Cluster randomized trial to pilot test the preliminary effectiveness of peer-delivered HIV/syphilis self-tests and partner services versus facility-based testing.
Objective 3: Cost-effectiveness analysis using payers and health sector perspectives.
Primary outcomes for objective 2:
Data analysis Objective 1: Investigators will use content analysis techniques. Two coders will read a random subset of interviews to identify general themes and create a preliminary codebook. A subset of transcripts will be coded together, and findings will be discussed to resolve discrepancies. An additional subset of interviews will be coded using the revised codebook, and Cohen's kappa will be calculated. Themes with kappa values of <0.60 will be redefined.
Objective 2: Sociodemographic characteristics will be summarized using descriptive statistics. The primary outcomes shall be analyzed using intent-to-treat. All participants who receive an HIV/syphilis test will be included in the analysis. Those who receive the self-test kits but do not use them will also be included in intent-to-treat analyses to provide the most generalizable effect measure. Mixed effects/multi-level models using modified Poisson regression with robust standard errors will be used to estimate relative risks of HIV/syphilis testing (yes/no). They permit estimation of relative risks with more stability than log binomial or logistic models. The random effects from mixed effects/multi-level model appropriately adjust for correlation in outcomes within participants and between participants recruited by the same peer (i.e., clustering effect).
Objective 3: Cost-effectiveness analysis of peer-delivered HIV/syphilis self-tests and partner services compared to facility-based testing. Cost-effectiveness will be estimated as cost per additional person reached and cost per additional person linked from the healthcare sector and client perspectives.
Men who engage in sexual relations with other men (MSM)(1) are disproportionately affected by HIV and syphilis. Globally, gay men and other men who have sex with men have 28 times greater risk of HIV acquisition than adult men (15-49 years) in the general population.(2). Syphilis and HIV disproportionately affect MSM because of sexual network dynamics that include high-risk sexual practices and multiple sex partners.(3,4). HIV and syphilis co-infection is common because syphilis and HIV have similar modes of transmission(5,6). Having syphilis is associated with a 3-fold increased risk of HIV acquisition.(7,8); syphilis also increases the risk of HIV transmission and disease progression(9). The resurgence of syphilis among MSM globally(10,11) is attributed to increasing rates of risky sexual behaviors, such as anonymous sex, oral and anal condomless sex, sex with multiple partners, and/or sex under the influence of drugs(5,12). In 2013, the United States estimated that 46% of MSM with primary and secondary syphilis had HIV, compared to 8% of heterosexual men and 6% of women.(6). In sub-Saharan Africa, 6% of new HIV infections occurred among MSM in 2021(13). In Uganda, HIV prevalence among MSM was 13.2% in 2012 compared to 4.7% in similarly-aged males. (14,15).. Of the 41,655 syphilis cases reported in the United States in 2020, 43% were among MSM(16). In 2016, syphilis prevalence among MSM in Uganda was estimated at 9.0%(17) compared with 5.8% in males aged 15-64 years(14,15). Ugandan national 2020 guidelines for HIV prevention and care recommend annual STI testing as a strategy for STI prevention among MSM and their sexual networks, but uptake of facility-based STI testing is low (32%) of the estimated MSM population in Kampala. (14,15) ,(18). To achieve the WHO Global Health Sector Strategy goal of reducing syphilis incidence by 90% by 2030(19), and the first UNAIDS 95-95-95 goal of 95% of persons with HIV knowing their status by 2025(20), scaling up HIV/syphilis testing is critical to reaching undiagnosed MSM.
1.2. HIV/syphilis testing is the entry point to treatment and prevention services The World Health Organization (WHO) recommends dual HIV/syphilis testing for MSM.(21). Increasing testing coverage is key to reaching the first UNAIDS 95-95-95 target and reducing disease incidence.(13,20). Early detection, timely linkage to care and other prevention services, and identification of sexual contacts reduced HIV transmission by 89% among MSM in Bangkok.(22). However, current testing strategies in sub-Saharan Africa only reach men who are not at high risk of HIV and/or already tested for HIV and are thus low yield: 2.4% HIV positivity for provider-initiated testing, 2.9% positivity for voluntary counseling and testing, and 3.1% positivity for mobile testing approaches.(7). Thus, identifying high-yield strategies for dual HIV/syphilis testing that effectively reach MSM in the riskiest sexual networks is key.(23). In Uganda, HIV/syphilis testing uptake is still low among MSM despite efforts to scale up prevention services for this population.(23). A recent bio-behavioral survey in 2020 estimated that there were 202,343 undiagnosed HIV infections in Uganda, of which 44% were in men, mostly MSM and their sexual contacts.(24). Data on syphilis test uptake and coverage among MSM in Uganda are lacking.(15). In contrast, syphilis test coverage among pregnant women attending antenatal clinics is >80%(25). This high coverage is part of a comprehensive dual HIV/syphilis elimination strategy using the SD BIOLINE HIV/syphilis Duo assay.(26,27). This assay is available at health facilities that are not friendly to MSM because of stigma and discrimination. Decentralized testing and community-based approaches could improve dual HIV/syphilis test uptake in this population.
1.3. Self-testing is feasible and acceptable to MSM Self-administered HIV self-tests (HIVST) and syphilis self-tests (permit the user to interpret test results by themselves in private (28). A scoping review of 11 studies in sub-Saharan Africa found that HIVST was feasible and acceptable to MSM (29). Studies among MSM in China(30,31) have shown that syphilis self-testing may increase testing frequency by empowering MSM to test and reducing the impact of structural barriers(32), but data from sub-Saharan Africa are limited. Combining HIVST and syphilis self-testing with peer delivery offers a unique opportunity to reach MSM with user-friendly technology in their communities. This strategy could overcome some of the barriers associated with facility-based testing, and promote early diagnosis and linkage to prevention services. As demonstrated with at-home COVID-19 testing, HIV and syphilis self-testing could sustain testing services when health facilities are inaccessible or inconvenient(33).
1.4. Peer-delivered self-tests can efficiently reach MSM and their social networks Peer approaches are recommended by the WHO for HIV testing.(30,34) and can be leveraged for dual HIV/syphilis testing. Peer delivery is a person-centered approach that could maximize the coverage, effectiveness, efficiency, and impact of HIV/syphilis services for MSM in Uganda. Peers are non-clinically trained persons with similar background characteristics to the beneficiary population. (34). Peer-led outreach services reach more MSM than traditional approaches by generating demand in the community and tapping into MSM social networks.(35,36). In Zimbabwe, peer-delivered syphilis self-testing was highly acceptable (89.6%), increased privacy, convenience, and autonomy, and helped circumvent social and healthcare provider stigma.(37). Our prior work found that peer-delivery of HIV self-tests was acceptable to 90% of MSM in Uganda.(38,39). All newly diagnosed participants with HIV infection were linked successfully to receive confirmatory testing at a friendly health facility and linked to treatment using a peer-assisted linkage model.(38). Thus, HIVST and syphilis self-testing linked with partner services may empower MSM and expand testing uptake and coverage.(1).
1.5. Integrating assisted partner notification with peer-led self-testing approaches could accelerate prevention uptake among MSM Assisted partner notification is an evidence-based strategy in which a trained provider encourages persons with an HIV/STI diagnosis to disclose their status to their sexual partners. Assisted partner notification is recommended by WHO, the U.S. President's Emergency Plan for AIDS Relief (PEPFAR), and the Uganda Ministry of Health and offers testing services to sexual partners of index cases with their consent(40,41). Assisted partner notification is acceptable and feasible(42-45) and increases testing uptake among sexual partners(42,45,46), enables early diagnosis, and linkage to care and other prevention services, motivates behavior change in index cases and partners, and may reduce STI and HIV burden(42-45). A pilot partner services program in rural Uganda that enrolled 464 people with HIV (PWH) and tested their recent sexual partners found that 32% (61/193) of sexual partners who were traced and tested had HIV(46). A mathematical modeling study in South Africa found that assisted partner notification could enable the identification of new HIV diagnoses among MSM(47). Similarly, a combination of assisted partner notification, HIVST, and community worker outreach was acceptable and feasible in Kenya(48) and increased uptake of HIV testing and linkage to care and PrEP among MSM and their sexual partners(48,49). A systematic review of 26 trials found that assisted partner notification identified more undiagnosed syphilis cases than passive notification(50). Thus, partner services could efficiently identify MSM with undiagnosed HIV and syphilis and help reach the first UNAIDS 95 target and global syphilis elimination goals.
1.6. Cost-effectiveness analysis is critical in designing a scalable prevention strategy HIVST and syphilis self-testing was highly acceptable to MSM in 10 studies(37,51-60). Peer-delivered combination prevention campaigns have identified undiagnosed HIV infections, linked people with HIV to care and those without HIV to other prevention strategies such as pre-exposure prophylaxis (PrEP)(61,62). However, in many regions like Uganda, where HIV prevalence has experienced substantial declines and is concentrated in key populations (KP)(63)Standard HIV testing approaches are inefficient.(7), incurring higher costs and yielding fewer HIV and syphilis infections detected per number of KP tested(7,64). However, targeted testing for MSM and assisted partner notification is cost-saving (median of $20 per life-year saved) compared with the secondary distribution of self-tests to partners of pregnant women (median of $130 per life-year saved)(47). Peer approaches can improve efficiency and contain costs by testing MSM at high risk of HIV and syphilis. Therefore, cost-effectiveness analyses evaluating peer-delivered combination approaches linked with partner services are needed to inform guidelines and interventions that can be scaled up.
1.1. Background and rationale Joint delivery of peer-delivered HIV/syphilis self-tests with assisted partner notification is an empowering, innovative approach that could substantially increase testing among MSM in Uganda and help achieve global HIV 95:95:95 targets (95% of people who know their status on treatment and 95% of people on treatment with suppressed viral loads). However, no studies to our knowledge have evaluated the effectiveness of peer-delivered HIV/syphilis self-tests and assisted partner notification services for MSM in any setting. The proposed work is uniquely positioned to improve prevention uptake for MSM, a high-risk, marginalized, and underserved population. This pilot study will be among the first to measure the benefits and cost-effectiveness of using a peer-delivered combination prevention strategy among MSM to scale up HIV, syphilis self-testing, and assisted partner notification strategies and will inform program costs.
1.7. Preliminary data This approach builds from our pilot study of MSM peer-driven distribution of HIVST to sexual networks.(54,65,66) (Table 1). Our work found that peer-delivery of HIVST was acceptable to 90% of MSM in our population tested in Uganda(54). All undiagnosed MSM with HIV received confirmatory testing, were linked to care, and started on treatment. Overall, peer delivery of HIVST resulted in higher HIV positivity yield (4.9 vs. 1.4%) and could avert more HIV infections per quarter compared with facility-based testing(66).
1.8. Study setting This study will be conducted in Kampala and Wakiso districts as study recruitment districts. In these districts, through a PEPFAR grant. IDI manages a network of 200 key population peers across Kampala and Wakiso districts to improve HIV outcomes (identification, linkage, retention, and viral suppression) along the care cascade. Peers are assigned to high-volume key population facilities, drop-in centers and community venues. Peer capacity building occurs through tailored HIV-related training, mentorships, and site support supervision. Peers also support the demand for the creation and distribution of HIVST kits to their social network members. Peer distributors receive three-day training focused on self-test kit use, the importance of HIV-positive result confirmation, inventory management of test kits including storage and documentation, HIV counseling, and linkage to care. Telephone numbers for both peers and key populations that received HIVST kits are recorded with consent for follow-up purposes. All HIVST kits are recorded in the national HIV self-test kit distribution log. Recipients who don't report results within two days are followed up through a phone call by a trained health worker. Recipients who reported HIV-positive results are offered confirmatory testing at the nearest health facility of their choice. Those who cannot go to the facility for confirmatory testing within seven days are followed up and offered confirmatory testing through a community visit. Between 2018 and 2020, IDI distributed 2,192 HIVST kits to MSM at 35 public health facilities.(67). These are the networks investigators shall leverage to facilitate study recruitment and retention. The study will identify study peers from networks already offering HIV care, prevention, and support services. The identified peers will offer self-testing. For the facility arm, participants will be tested at the IDI Kasangati research clinic or IDI-identified key population health facilities in Kampala and Wakiso, which investigators are mapping with key population focal persons.Study Design: Randomized cluster trial (Fig 2). Twenty MSM peers from the interviewees and those recommended as peers in Aim 1 will be trained in intervention delivery (test performance, result interpretation, linkage to care, and protection of human subjects). Trained peers will be randomized 1:1 to the intervention (10 peers) or control arms (10 peers).
In this study, a peer will be a non-clinically trained MSM identified by social network members with similar background characteristics to the MSM population. Each peer will distribute self-testing kits or offer invitation letters.
A research participant will be a person to whom the peer will distribute the self-testing kits and invitation letter after they provide informed consent.
A research assistant is a study staff member who supervises peers. Each randomized peer and recruited participant will be regarded as a cluster. One cluster equals to (one peer + ten MSM). The diagram below shows a relationship between a research assistant, a peer, and a participant.
Intervention peers will: a) deliver HIVST/syphilis self-testing kits to MSM within their networks using SNS and target never or infrequent testers; b) actively link MSM who test positive to confirmatory testing and care; c) offer sexual partners voluntary assisted partner notification and HIV/syphilis testing while maintaining confidentiality.
Control arm peers will: a) distribute invitation letters to social network members for fast-track HIV/syphilis testing at a health facility. All MSM will receive: i) free confirmatory HIV and syphilis testing, condoms, and treatment and ii) psychosocial peer support at the health facility.
Peer selection and recruitment: If feasible in Aim 1, investigators will use SNS, an established method for sampling MSM(69) to recruit the sample for Aim 2. Using the results of the formative research in Aim 1, investigators will initiate recruitment by purposively selecting 20 peers from respondents identified in Aim 1 to ensure the representation of key variables (age, geographic location). investigators will identify MSM peers with diverse characteristics and social relationships, comprising multiple types of relationships, subgroups, and activities to help ensure the recruitment of a social mix of subjects that represents the target MSM population.
Peer training: Peers will be trained on how to maintain confidentiality, HIVST and syphilis self-testing kit testing and delivery, assisted partner notification, and peer-assisted linkage to care of MSM who test positive. They will also be trained in the use of HIVST, syphilis self-testing, and assisted partner notification according to Ministry of Health guidelines(18,70). Peers will demonstrate how to perform HIVST and syphilis self-testing. Peers will build rapport and obtain the trust of potential subjects, perform recruitment and retention, and intervention delivery including promotion of self-testing and assisted partner notification. Peers will be instructed not to provide medical advice, but instead refer participants (and sexual partners) to study staff as needed. Peers will be invited to participate in monthly study team meetings to address challenges with study implementation.
Inclusion criteria: In both arms, peers will recruit network members who are
1) Participants who already know that they have HIV and those who are on treatment for syphilis 2. Those enrolled in other HIV prevention trial 3. Investigators shall exclude participants who don't speak English and Luganda.
Intervention procedures: Participants will be given a demonstration and instructions on how to use self-test kits (written and pictorial instructions in Luganda) and provided with ten serialized kits (five HIVST and five syphilis self-testing) to distribute to eligible social and sexual network members. Peers effective at reaching infrequent and non-testers will be asked to distribute an additional three to five test kits. Each peer will receive an IRB-approved $5 for each kit distributed and accounted for as in prior studies. (62). Respondents will be compensated for each successful recruitment, and SNS recruiting will continue until the sample size of 200 is attained. Each peer will receive a REC-approved transport facilitation of $30 per month since most MSM who have never tested are hard to trace. Control group peers will only distribute invitation letters for fast-track HIV/syphilis at the study clinic to minimize contamination (i.e., use of the intervention by control group participants).
Follow-up: A toll-free phone number and WhatsApp chatbot will be established solely for study online consultation and follow-up. MSM who have been given self-test kits will be contacted by the research assistant via mobile phone to report their test results, participate in a phone-based short messaging system used in one of our studies, and upload photos of used self-test kits using WhatsApp. Following participant HIV testing, health workers will discuss the next steps and facilitate assisted partner notification, and peers will facilitate linkage to care for those testing positive and negative to PrEP-accredited health facilities. Peers will address participant concerns with study staff as needed.
Data collection: Investigators will employ Ministry of Health data collection tools currently used among MSM in Uganda (48,71). Study staff will collect basic sociodemographic data and conduct a brief behavioral HIV/syphilis risk assessment. Study data will be entered into tablet phones.
Randomization: Peers will be randomized 1:1 to either the intervention or control arm (using block randomization) through REDCap, a secure web-based application(72).Each peer randomized will recruit 10 participants to make one cluster
Contamination: To minimize contamination, the peers will be trained on the importance of maintaining randomization and providing services intended for each arm. Investigators will inquire about the use of intervention components in the control group at the end of the study while keeping in mind that self-report of HIV/syphilis tests may be an underestimate. If several control group participants indicate intervention component use, Investigators will consider increasing the sample size to address the loss of statistical efficiency due to contamination. The use of intervention components by control group participants could reduce the point estimate of intervention preliminary effectiveness and lead to a type II error - rejection of intervention as ineffective if it exceeds 30% (Torgerson BMJ 2001). To minimize contamination, Investigators will carefully review study procedures with peers and stress that HIV/syphilis kits are for the use of intervention participants only and should be "valued" by those groups randomized to receive them. Moreover, HIV/syphilis self-tests are not routinely available to MSM through the public health system, which will further reduce the risk of contamination.
Primary outcomes: proportion testing HIV/syphilis positive: Intervention acceptability (assessed using 5-point Likert scales), uptake of HIVST, syphilis self-testing and assisted partner notification, and linkage to care (Table 3).
Sample size and power calculations: Investigators used an inequality test to calculate the sample size (PASS 15.0, NCSS, LLC). Preliminary pilot RCT data showed that the mean proportion of social network members reached by an index was 0.05 through a standard of care delivery and a mean proportion of 0.60 through peer delivery (59). Therefore, Investigators assumed that the variances of the two groups were equal, with the same standard deviation of 0.8. Assuming the number of clusters will be allocated 1:1 to each arm, with cluster size of 10 per cluster, alpha of 0.05, power of 0.8, inter-class correlation of 0.01, and loss to follow-up rate of 0.20, Investigators will need a minimum of 10 clusters per arm. To integrate the sample size calculations and to consider practicality, the final sample size would be ten clusters per arm with a total cluster size of 20 peers (200 participants).
Sample size calculations
Intervention ICC Mean difference Standard deviation COV of cluster sizes Number of people per cluster Alpha Power Total sample size per arm Number of clusters per arm 0.01 0.55 0.8 0.65 10 0.05 0.8 60 6 Control arm ICC Mean difference Standard deviation COV of cluster sizes Number of people per cluster Alpha Power Total sample size per arm Number of clusters per arm
0.01 0.4 0.8 0.65 10 0.05 0.8 80 8 ICC (intra-class correlation), COV (coefficient of variation)
To integrate the sample size calculations and to consider practicality, the final sample size would be ten clusters per arm with a total cluster size of 20 peers (200 participants). I am allowing for a transition in testing and linkage to care up to 20% so that the study remains sufficiently powered to answer the study objectives.
Data analysis: Sociodemographic characteristics will be summarized using descriptive statistics. The primary outcomes shall be analyzed using an intent-to-treat analysis. All participants who receive HIV/syphilis tests will be included in the analysis. Also, those who receive self-test kits but do not use them will be included in intent-to-treat analyses to provide the most generalizable effect measure. Mixed effects/multi-level models using modified Poisson regression analysis with robust standard errors will be used to estimate relative risks of HIV/syphilis testing (yes/no) because they permit the estimation of relative risks with more stability than log-binomial models. Mixed effects/multi-level models appropriately adjust for correlation in outcomes within participants and between participants recruited by the same peer (i.e. clusters).
Inclusion Criteria:In both arms, peers will recruit network members who are
Exclusion Criteria:We will exclude participants