Effectiveness of sexual health influencers identified by an ensemble machine learning model in promoting secondary distribution of HIV self-testing among men who have sex with men in China: study protocol for a quasi-experimental trial

Abstract

Background: HIV self-testing (HIVST), especially the secondary distribution of HIVST (SD-HIVST) initiated by sexual health influencers (SHIs), has been recognized as an effective strategy in promoting HIV testing, especially among men who have sex with men (MSM). This quasi-experimental study aimed to evaluate whether SHIs identified through the ensemble machine learning approach can distribute more HIVST than those who identified by the empiricalscale.

Methods: We will recruit eligible adults (≥18 years old) who were assigned male gender at birth, and willing to participate in potential SD-HIVST online. Participants will be assigned randomly to two groups (scale group or machine learning group), followed by a separate process of SHI identification based on the group assignment. After identification, all index participants (defined as identified SHIs who are verbally consented to participate in SD-HIVST or who directly order HIVST kits) will follow the same procedure for SD-HIVST acquisition and distribution. Index participants can order HIVST online and distribute them to members within their social networks (defined as alters) in-person or virtually through a personalized peer referral link. Once a unique alter uploads a photographed test result to the platform, both the alter and the corresponding index participant will receive a fixed incentive of 3 USD. The index MSM can order up to five HIVST in the first three months and ten HIVST in the following three months. Each index participant will need to complete a baseline survey at the first-time ordering and one to two follow-upbased on the times of ordering,, three months after ordering. This trial will be comparing 1) the mean number of alters motivated by each index participant in each group and 2) the mean number of newly-tested alters motivated by each index participant in each group.

Discussion: In promoting the efficacy of identifying SHIs for SD-HIVST, our study has the potential to enhance testing coverage, particularly among marginalized individuals and those who are reluctant to for HIV and other sexually transmitted infections.

Trial registration: We registered the study on the Chinese Clinical Trial Registry website on 4th November 2021, with registration number ChiCTR2000039632 .

Keywords: China; Ensemble machine learning; HIV self-testing; Men who have sex with men; Secondary distribution; Sexual health influencers.