Social Media–Based Secondary Distribution of Human Immunodeficiency Virus/Syphilis Self-testing Among Chinese Men Who Have Sex with Men

Abstract

Background

Social media and secondary distribution (distributing self-testing kits by indexes through their networks) both show strong promise to improve human immunodeficiency virus (HIV) self-testing uptake. We assessed an implementation program in Zhuhai, China, which focused on the secondary distribution of HIV/syphilis self-test kits among men who have sex with men (MSM) via social media.

Methods

Men aged ≥16 years, born biologically male, and ever had sex with another man were recruited as indexes. Banner ads on a social media platform invited the participants to apply for up to 5 self-test kits every 3 months. Index men paid a deposit of US$15/kit refundable upon submitting a photograph of a completed test result via an online submission system. They were informed that they could distribute the kits to others (referred to as “alters”).

Results

A total of 371 unique index men applied for 1150 kits (mean age, 28.7 [standard deviation, 6.9] years), of which 1141 test results were returned (99%). Among them, 1099 were valid test results; 810 (74%) were from 331 unique index men, and 289 tests (26%) were from 281 unique alters. Compared to index men, a higher proportion of alters were naive HIV testers (40% vs 21%; P < .001). The total HIV self-test reactivity rate was 3%, with alters having a significantly higher rate than indexes (5% vs 2%; P = .008). A total of 21 people (3%) had a reactive syphilis test result.

Conclusions

Integrating social media with the secondary distribution of self-test kits may hold promise to increase HIV/syphilis testing coverage and case identification among MSM.

Compared to index men, significantly more alters who received an human immunodeficiency virus (HIV) self-test from an index case were naive HIV testers and HIV reactive, suggesting the potential for a social media–based secondary distribution strategy in improving HIV testing and case identification.

As the entry point into the human immunodeficiency virus (HIV) care continuum [1] and a critical step in the “Treat All” strategy [2], HIV testing remains a cornerstone of HIV prevention and control. However, about 25% of people living with HIV (PLWH) worldwide do not know their serological status [3]. This is even worse among key populations in China, as nearly half of men who have sex with men (MSM) have never been tested for HIV [4–6]. Syphilis testing is also key to HIV prevention, yet 70% of Chinese MSM have never been tested for syphilis [7]. Serostatus unawareness plays a key role in ongoing transmission of HIV and syphilis; therefore, interventions must reach PLWH who are unaware of their status. Though wider availability of facility-based HIV testing services helped increase testing coverage, innovative interventions are needed to enhance coverage of harder-to-reach MSM [8, 9], especially those not receiving testing from a health facility [10].

The development of self-testing services will support HIV and syphilis control. Self-testing is the process whereby a person collects a specimen, performs the test, and interprets the result themselves [11, 12]. Dual HIV/syphilis self-testing is a promising approach for expanding HIV and syphilis testing that enables individuals to perform rapid HIV/syphilis antibody tests simultaneously and privately [13]. The World Health Organization (WHO) recommends that dual HIV/syphilis rapid tests be offered to screen at-risk populations [11, 14]. By July 2019, 77 countries had adopted policies or guidelines for implementation and support of HIV self-testing (HIVST) [15]. In China, the demand for and acceptability of HIVST among MSM has been high, and many Chinese local community-based organizations (CBOs) are working with health bureaus in piloting HIVST among MSM [16, 17]. No policies or guidelines are in place to improve integrative syphilis testing for MSM in China; only 1 cross-sectional survey study with Chinese MSM found that 92.5% of syphilis self-testers did HIVST at the same time, demonstrating the feasibility of integrating syphilis self-testing into HIV self-testing services [18]. Nevertheless, this study did not explore strategies for further expanding testing uptake.

While dual HIV/syphilis self-testing is a promising strategy for expanding testing coverage, it is critical to develop and evaluate outreach and distribution strategies for self-tests, such as secondary distribution [19]. Secondary distribution is a social network–based approach that involves giving 1 individual (index) multiple self-testing kits for distribution to people within their social networks (alters) [20]. Two studies conducted in Kenya have shown that secondary distribution has strong potential to promote HIV testing among the male sexual partners of women seeking antenatal and postpartum care [21, 22]. Three other cohort studies showed feasibility and effectiveness of secondary distribution of HIVST among male fishermen in Uganda, female sex workers, and women seeking perinatal care, as well as MSM in South Africa [20, 23, 24]. However, most previous studies in low- and middle-income countries (LMICs) focused on the sexual network of the index participants. One study in the United States (US) revealed that 59% of alters who received a self-test were friends of the index MSM [25]. This indicates that we can further use this opportunity to leverage nonsexual social contacts of MSM and reach more people who have high risk but have not been reached by conventional services in LMIC settings. In addition, the testing results of the alters in most of the previous studies were self-reported by the index participants, which raises concerns about privacy and reliability of the test results. None have explored secondary distribution of syphilis self-testing as an integrated service into existing HIV self-testing programs. A social media–based secondary distribution of dual HIV/syphilis self-tests may hold potential in overcoming these barriers and add value to existing literature.

In this study, we evaluated a social media–based secondary distribution of HIV/syphilis self-testing among Chinese MSM. We defined “social media–based secondary distribution” as the implementation of the secondary distribution (including participant recruitment, screening, survey, kit application, and returning of results) is based on social media platforms. By leveraging the social contacts of participants, we explored whether this enhanced model can expand the coverage of HIV testing and help identify new HIV/syphilis cases. We assessed the feasibility of having the testers upload photographs of their completed self-tests for verification by study staff, and evaluated alters’ sexual intercourse with an index at the point of self-testing.

METHODS

Study Setting and Infrastructure Establishment

This implementation study was conducted in Zhuhai, southern China, with an estimated 17 000 MSM living in the city. It has an HIV prevalence of 7% among MSM and was one of the first sites to pilot HIVST among MSM in China [26]. The Zhuhai Center for Diseases Control and a gay community-led organization (Zhuhai Xutong Voluntary Services Center, hereafter “Xutong”) initiated a social-media based online system in 2016 for MSM to apply for free dual HIV/syphilis self-test kits. MSM were able to apply for HIVST kits using Xutong’s public WeChat account. WeChat is a multifunctional social app that can be used for messaging, public surveys, and monetary transactions.

Secondary Distribution

Beginning on 17 June 2018, index men could request up to 5 HIV/syphilis self-test kits (HIV/Syphilis Duo, Standard Diagnostics) per application (with a limit of 1 application per 3-month period). Figure 1 shows the secondary distribution model of the program. Test kits were mailed to index men after they paid the deposit (US$15 per kit) and provided shipping information. All test kit packages mailed to a participant were packed in 1 parcel and assigned an identical numerical identification number. Each dual test kit package contained instructions and a WeChat QR code for the tester to upload a photograph of the test result anonymously. Index MSM were informed that they could distribute the kits to other social contacts, including partners or friends (referred to as “alters”). Following the instructive notes, alters were encouraged to upload an anonymous photograph of the test result themselves. Upon receiving a photographed test result, the deposit was immediately refunded to index men through WeChat.

Figure 1.

The social media–based secondary distribution model in Zhuhai, China, 2018–2019. Abbreviation: HIV, human immunodeficiency virus.

Consented MSM participants provided their mobile numbers for follow-ups. Volunteers at Xutong, mostly gay men, were trained as lay health workers to verify the photographed test results. Once participants’ test results were verified as reactive for HIV and/or syphilis, our CBO volunteers contacted and suggested that participants with a reactive result seek timely confirmatory testing and treatment services at a local health facility via telephone calls. These participants were provided with information about locations and open hours of local health facilities for MSM-friendly sexual healthcare.

Survey Instruments and Data Collection

Inclusion criteria for index participants included (1) men who were aged ≥16 years; (2) born biologically male; (3) ever had sex with another man; (4) were willing to provide a contact number; and (5) agreed to participate in a follow-up survey in 3 months. Eligible index participants completed a baseline online survey that collected information about the indexes’ sociodemographic characteristics (including age, education, income, marital status, residence, and occupation), sexual biography (including gender identity, sexual orientation, or sexual orientation disclosure), and HIV testing history.

Index men were followed up 3 months after the recruitment to collect information about their process of distributing dual test kits to individuals within their social networks. A short online survey was also administered to alters upon uploading their photographic evidence of a completed test. A small incentive of US$3 was provided to all participants who completed a questionnaire.

Outcomes

Primary outcomes included proportion of new testers (defined as naive HIV testers) among unique indexes and alters as well as HIV and syphilis reactivity rates among unique indexes and alters. Those who had a reactive result (for either HIV or syphilis or both) were followed up via telephone calls by the CBO volunteers and recommended to seek confirmatory testing and medical treatment at local facilities (ie, linkage-to-care services). We also calculated the rate of returning a photograph of the completed test within 30 days after receiving the test kit(s). Additionally, among those who reported a sexual relationship with the index, we measured their sexual intercourse behaviors at the point of testing, aiming to understand how participants used HIVST to inform their sexual behaviors.

Statistical Analyses

Duplicate applications were identified by duplicate mobile numbers used for the application and assumed to be from the same index participant, and only the baseline characteristics from the first application by calendar date were included for analysis for each index. We conducted descriptive analyses of the outcomes for both the unique index and the alter participants. The χ ² or Fisher exact test was used to assess differences in proportion of naive HIV testers and HIV/syphilis reactivity rates between indexes and alters, and to compare alters’ reports of sexual decisions based on whether alters shared their HIV test result with their index (the alter shared a reactive result, a nonreactive result, or the alter did not share the result). Statistical analyses were conducted using SPSS version 25 software (IBM SPSS, Armonk, New York).

Ethical Statement

The study protocol was approved by the Zhuhai Center for Disease Control and Prevention. Informed consent was collected from each participant online.

RESULTS

Test Kit Application and Distribution

This study was conducted between 17 June 2018 and 12 November 2019. Figure 2 shows the flowchart of the number of applications and dual test kits distributed. Overall, there were 851 application attempts, of which 649 applications met eligibility criteria, and 1150 test kits were distributed to 371 unique index men, with 48% of indexes requesting 2 or more kits.

Figure 2.

Flowchart of the number of application attempts, human immunodeficiency virus/syphilis self-test kits distributed, test results returned, and completed follow-up surveys in China, 2018–2019. Abbreviation: MSM, men who have sex with men.

Return of Photographed Test Results

In total, 1141 test results were returned with photographic evidence, resulting in an overall return rate of 99% (1141/1150) (Figure 2). A total of 1099 test results were valid, among which 810 (74%) were from 331 unique index men, and 289 (26%) were from 281 unique alter testers. On average, each index successfully reached 0.85 (281/331) alters.

Participant Characteristics

Table 1 shows the characteristics of study participants. The mean age for indexes and alters was 28.7 (standard deviation [SD], 6.9) years and 29.4 (SD, 7.0) years, respectively. Proportions of testers who obtained university or higher education were similar between the 2 groups. More index men than alter men (70% vs 60%, respectively) had disclosed their same-sex behavior to individuals other than their sexual partners. There were no significant differences in sociodemographic characteristics between index men who returned a result and those who did not (Table 2).

Table 1.

Characteristics of the Index Men and Alters in China, 2018–2019 (N = 652)

Characteristic	Unique Index Men (n = 371)	Unique Altersa (n = 281)	P Valueb
Age, y, mean (SD)	28.7 (6.9)	29.4 (7)
Sex
Male	371 (100)	264 (98)
Female	…	4 (2)
Marital status
Never married	312 (84)	191 (72)
Engaged or married	51 (14)	60 (22)
Separated, divorced, or widowed	8 (2)	16 (6)
Residence
Rural	164 (44)	91 (34)
Urban	207 (56)	177 (66)
Ethnicity
Han	367 (99)	260 (97)
Other minorities	4 (1)	8 (3)
Highest education
High school or less	74 (20)	64 (24)
Some college	99 (27)	62 (23)
University	198 (53)	142 (53)
Employment status
Employed	283 (79)	227(81)
Unemployed	74 (21)	52 (19)
Annual income levels, US dollars
<2518	48 (13)	33 (12)
2518–5038	42 (11)	27 (10)
5038–8393	99 (27)	77 (29)
8393–13 430	103 (28)	81 (30)
>13 430	79 (21)	50 (19)
Gender identityc
Male	358 (96)	254 (96)
Female	3 (1)	7 (3)
Transgender	6 (2)	1 (0.4)
Unsure/other	4 (1)	2 (1)
Sexual orientationc
Gay	259 (70)	178 (67)
Heterosexual	2 (0.5)	13 (5)
Bisexual	91 (24)	56 (21)
Unsure	19 (5)	17 (6)
Disclosed same-sex behavior with another man to othersc	260 (70)	158 (60)	.007
Anal sex with a male partner in the past 6 moc	314 (85)	181 (69)	<.001
Consistent condom use	184 (59)	103 (57)	.71
Used a condom during last sexual intercourse with a male partner	262 (83)	146 (81)	.43
Sex with a female partner in the past 6 moc	31 (8)	35 (13)	.046
Consistent condom use	13 (42)	13 (37)	.69
Used a condom during last sexual intercourse with a female partner	18 (58)	15 (43)	.22
Ever tested for HIV beforec	294 (79)	158 (60)	<.001

Data are presented as no. (%) unless otherwise indicated.

Abbreviations: HIV, human immunodeficiency virus; SD, standard deviation; US, United States.

^aTotal unique alter testers, including male and female participants. Responses of variables may not add up to the total due to missing values, and valid percentages were calculated.

^bχ ² test.

^cMale participants only.

Table 2.

Characteristics of Index Men Who Returned a Result and Who Did Not Return a Result in China, 2018–2019

Characteristic	Did Not Return a Result (n = 40)	Returned a Result (n = 331)	P Value
Age, y, mean (SD)	28.5 (7.8)	28.7 (6.8)	.87
Marital status
Never married	33 (82)	279 (84)	.49
Engaged or married	7 (18)	44 (13)
Separated, divorced, or widowed	0 (0)	8 (2)
Residence
Rural	15 (38)	149 (45)	.40
Urban	25 (62)	182 (55)
Highest education
High school/below	7 (18)	67 (20)	.92
Some college	11 (28)	88 (27)
University	22 (55)	176 (53)
Employment status
Student	8 (20)	54 (16)	.51
Nonstudent	32 (80)	277 (84)
Annual income levels (in US dollars)
<2518	7 (18)	41 (12)	.50
2518–5038	5 (12)	37 (11)
5038–8393	7 (18)	92 (28)
8393–13 430	14 (35)	89 (27)
>13 430	7 (18)	72 (22)

Data are presented as no. (%) unless otherwise indicated.

Abbreviations: SD, standard deviation; US, United States.

HIV/Syphilis Test Results

Overall, 20 of 612 unique testers (3%) had a reactive HIV self-test result. Fifteen of them were alters, and 5 were indexes. The HIV-reactive rate among alters was significantly higher than that among indexes (15/281 [5%] vs 5/331 [2%], respectively; P = .008). All men who had reactive results were first-time testers. Four of 5 index cases (80%) and 12 of 15 (80%) alter cases who had a reactive result were linked to facility-based confirmatory testing and enrolled in HIV treatment.

Twenty-one of 612 (3%) testers had a reactive syphilis result, with 13 among 331 (4%) indexes and 8 among 281 (3%) alters, with no significant difference. One index participant (0.3%) was found to be reactive for both HIV and syphilis. Among all syphilis-reactive participants, 8 were previously diagnosed with and formally treated for syphilis, and 8 were newly syphilis reactive and referred to confirmatory testing at a local facility. The remaining 5 were lost to follow-up.

Alters’ Testing Experiences

Among 281 alters who returned their test results, 268 (95%) completed an online survey, among whom 264 (98%) were male and 4 (2%) were female. Of the 264 male alters, 114 (43%) were a sexual partner of the index. The remaining 150 alters self-reported to be nonsexual contacts: 130 (49%) identified as a gay friend of the index, 13 (5%) as a family member, and 4 (2%) as a straight friend. Of the 264 male alters, 106 (40%) were new testers, compared to only 77 of 371 (21%) index men (P < .001) (Supplementary Figure 1). Among the 114 alters who reported a sexual relationship with the index, 33 (29%) were new testers, while the 150 nonsexual contacts had a higher proportion of new testers at 49% (P = .001).

Sex With the Index at the Point of Self-testing

Of the 114 male alters who were sexual partners of the index, 108 (95%) reported that they shared their test results with the index, including 5 reactive and 103 nonreactive results (Table 3). Sixty-one of the 114 alters (54%) reported having sexual intercourse with the index on the day of performing the self-test, of which 41 of 61 alters (67%) reported using a condom with intercourse.

Table 3.

Sexual Decision Making Among Male Alters Who Reported a Sexual Relationship With the Index, Based on Alters’ Sharing Status of a Human Immunodeficiency Virus Self-test Result in China, 2018–2019

			Alters’ Test Result Sharing Status With the Index (n = 114)
Question	Variable	Total	Shared a Reactive Result (n = 5)	Shared a Nonreactive Result (n = 103)	Not Shared (n = 6)
Did you have sex with the index on the day of testing?	Yes	61 (54)	3 (60)	58 (56)	0 (0)
	No	53 (46)	2 (40)	45 (44)	6 (100)
If you had sex with the index on the day of testing, when was it?	Before the test	27 (44)	2 (67)	25 (43)	…
	After the test	34 (56)	1 (33)	33 (57)	…
Did you or your partner use a condom during the sexual encounter?	Yes	41 (67)	3 (100)	38 (66)	…
	No	20 (33)	0 (0.0)	20 (34)	…

Data are presented as no. (%).

DISCUSSION

Effective strategies to promote HIV/syphilis dual self-test uptake are underexplored. We introduced a social media–based secondary distribution model to expand dual HIV/syphilis testing among Chinese MSM. Our study breaks new ground by assessing the feasibility and effectiveness of an integrated social media–based secondary distribution model to expand HIV/syphilis testing coverage, finding new HIV/syphilis cases, and testing the feasibility of test verification by returning of photographed self-test results.

Our model demonstrated 2 important benefits over previously studied offline models of HIVST secondary distribution. First, a high results return rate demonstrated the feasibility of the model. Verifying self-test results is often challenging [27], but important for ensuring the accuracy of HIV case reporting and linking reactive testers to care further. We addressed this issue through 3 components of our model: a gay community-led organization; anonymous upload of test results followed by free telephone counseling services for reactive cases; and a refundable cost-level deposit mechanism. These components created a trusting relationship between testers and lay health providers, protected testers’ privacy against unwanted disclosure of HIV status, and monetarily incentivized positive behavior. Second, this modified model was one of the few secondary distribution studies that included a direct assessment of alters’ test results and experiences. This was mainly possible due to the social media component of our model, which allowed us to reach alters directly and provide them an electronic channel to participate in the study via individually packed test kits.

Overall, <40% of the index men distributed self-tests to their social contacts. This percentage is lower than reported in previous interventional studies [20, 22]. This is likely because we assessed voluntary distributing behaviors of the index MSM. However, the study demonstrated that our model reached those who have never been tested for HIV, identified reactive HIV cases, and facilitated linkage to facility-based care. The rate of first-time testers among alters reached in our study (40%) was significantly higher than that of an earlier secondary distribution study among fishermen (26%) in Uganda [24]. The study results suggest that by tapping into the existing influence of MSM within their own social networks, a social media–based platform holds promise to lower the barrier for hard-to-reach MSM to access the HIV testing and care continuum.

We found that a high proportion of the alters (95%) who had a sexual relationship with the index shared their test results with the index man. More than half of alters had had sexual intercourse at the point of self-testing. However, concerns have been raised about sex at point of self-testing because of the inability of HIV self-testing to detect the HIV infection status within a window period, thus promoting unprotected sex between discordant partners at the time of highest infectivity [28]. More research and efforts are needed to understand the role of self-tests in sexual decision making, especially in the early stage of introducing self-tests into Chinese settings [16].

Our study also demonstrated that secondary distribution is a feasible approach for promoting syphilis self-testing among MSM with a combined HIV/syphilis testing kit. Integrating syphilis testing into existing HIV testing services has been recommended by the WHO [14, 29]; however, few studies have examined effective strategies to accelerate the uptake of syphilis self-testing integrated with rapid HIV self-tests. While recent evidence showed that integrating syphilis self-testing into HIV self-testing services may be a feasible strategy to improve syphilis test uptake among MSM [18, 30], secondary distribution of a dual HIV/syphilis test kit is a promising strategy for further expanding syphilis test coverage among this at-risk population. The model may also have implications for integrating multiple sexually transmitted infection (STI) self-testing services into HIVST secondary distribution programs.

Our study findings have implications. We demonstrated that a social media–based secondary distribution model was feasible and acceptable, with significant potential to expand HIV/syphilis testing among this hard-to-reach subgroup. Furthermore, our study captured the potential of MSM for promoting test behaviors among their social networks. Collecting verifiable HIV/STI self-test results has proven to be challenging, and our model provides a successful example of obtaining test result photographs for better result verification and follow-up care. Since index men were refunded their deposit only when the alters returned a test result, a refundable deposit mechanism may incentivize index men to encourage alters to upload their test result photos. This may be an important strategy for encouraging alters, who were significantly more likely to be new testers, to report their test results. Being led and maintained by a gay community organization as well as supported by health authorities, our collaborative platform empowered the community, gained trust, and secured financial sustainability, showing promise of replicability and scalability in other similar settings.

The study has several limitations. First, our study was an evaluation of an implementation project and did not include a comparison group. Second, this study recruited participants who are relatively well-educated with higher-than-average income levels. This may limit the generalizability of the study, and further exploration of feasibility and acceptability of secondary distribution among MSM with lower economic payment capacity should be considered for future research. Finally, there was a possibility of alters further distributing the kits within their social networks, and these data were not captured in this study.

CONCLUSIONS

This study demonstrated that a social media–based secondary distribution of HIV/syphilis testing model is feasible and acceptable among MSM. It can be used to identify first-time testers, promote HIV/syphilis case identification, and encourage linkage to care. Further implementation studies are needed that expand this service model among MSM and other key populations.

Supplementary Data

Supplementary materials are available at Clinical Infectious Diseases online. Consisting of data provided by the authors to benefit the reader, the posted materials are not copyedited and are the sole responsibility of the authors, so questions or comments should be addressed to the corresponding author.

Notes

Acknowledgments. The authors are grateful to all of the participants who participated in this study.

Financial support. This work was supported by the National Key Research and Development Program of China (grant number 2017YFE0103800); the Academy of Medical Sciences and the Newton Fund (grant number NIF/R1/181020); the National Institutes of Health (grant numbers National Institute of Allergy and Infectious Diseases [NIAID] 1R01AI114310-01, NIAID K24AI143471, and R25 AI140495); the University of North Carolina Center for AIDS Research (grant number NIAID 5P30AI050410); the National Science and Technology Major Project of China (grant number 2018ZX10101-001-001-003); the National Natural Science Foundation of China (grant numbers 81903371 and 81772240); and the Zhuhai Medical and Health Science and Technology Plan Project (grant number 20181117A010064).

Potential conflicts of interest. The authors: No reported conflicts of interest. All authors have submitted the ICMJE Form for Disclosure of Potential Conflicts of Interest.