Share
Title
A social network-based intervention increases HIV self-testing and linkage to health facilities among fishermen in Kenya
Presenter
Carol S. Camlin
Authors
C.S. Camlin1, L. Sheira2, Z.A. Kwena3, E.D. Charlebois4, K. Agot5, J. Moody6, B. Ayieko5, S.A. Gutin7, A. Ochung5, P. Olugo5, J. Lewis-Kulzer1, H. Nishimura4, M. Gandhi4, E.A. Bukusi3, H. Thirumurthy8
Institutions
1University of California, San Francisco, Obstetrics, Gynecology & Reproductive Sciences, San Francisco, United States, 2University of California, San Francisco, Epidemiology & Biostatistics, San Francisco, United States, 3Kenya Medical Research Institute, Nairobi, Kenya, 4University of California, San Francisco, Medicine, San Francisco, United States, 5Impact Research & Development Organization, Kisumu, Kenya, 6Duke University, Durham, United States, 7University of California, San Francisco, Community Health Systems, San Francisco, United States, 8University of Pennsylvania, Philadelphia, United States

BACKGROUND: Engaging men in HIV prevention and treatment is crucial to ending the AIDS epidemic in sub-Saharan Africa, but many men are unaware of their serostatus and highly-mobile men like Lake Victoria fishermen have particularly low uptake of prevention and treatment. We conducted a cluster randomized trial to determine if an HIV status-neutral, social network-based approach could improve testing and linkage outcomes among fishermen.
METHODS: The Owete study (NCT04772469) mapped the male social networks of fishermen in three beach communities in Siaya, Kenya, and identified distinct social networks (“clusters”) with a highly-connected, network-central man (“promoter”) in each network. Clusters were randomized to an intervention group in which promoters were trained and offered (a)multiple HIV self-tests (HIVST) to distribute to cluster members, and (b)transport vouchers (US$4) to encourage cluster members to link to HIV treatment or pre-exposure prophylaxis (PrEP). In control clusters, promoters received HIV information and referral vouchers for free self-tests in nearby clinics, and encouraged to offer them to cluster members. We compared self-reported HIV testing in past three months among participants in intervention and control clusters at three-month follow-up visit, using a cluster-adjusted two-sample test of proportions. Participants with missing data were coded as failure unless known to be living with HIV (per health facility records). Secondary analyses examined testing via any modality (counselor or HIVST) and linkage to facility.
RESULTS: A total of 934 men in 156 social network clusters were mapped. Of these, 733 completed baseline and 666 follow-up surveys, with 14 deaths due to study-unrelated causes. Participants’ average age was 37 years; 78% were married, 22% in polygynous relationships. Self-reported HIV testing via HIVST at three months was significantly higher in intervention clusters (60% vs. 10%, p<0.001, intent-to-treat). HIV testing via any modality was also significantly higher in intervention clusters (47% vs. 27%, p<0.001). Following testing, linkage to facility for HIV treatment or PrEP was significantly higher in intervention clusters (70% vs. 17%, p<0.001).
CONCLUSIONS: A social network-based, status-neutral intervention in Kenya improved men’s HIV testing and linkage outcomes and is a promising way to engage hard-to-reach populations of men in prevention or treatment.