Counseling Supporting HIV Self-Testing and Linkage to Care Among Men Who Have Sex With Men: Systematic Review and Meta-Analysis

Background: Counseling supporting HIV self-testing (HIVST) is helpful in facilitating linkage to care and promoting behavior changes among men who have sex with men (MSM). Different levels of counseling support for MSM HIVST users may lead to variance in the linkage to care. Objective: This study aims to synthesize evidence on counseling supporting MSM HIVST users and to conduct a meta-analysis to quantify the proportion of MSM HIVST users who were linked to care. Methods: A systematic search was conducted using predefined eligibility criteria and relevant keywords to retrieve studies from the MEDLINE, Global Health, Web of Science, Embase, APA PsycINFO, and Scopus databases. This search encompassed papers and preprints published between July 3, 2012, and June 30, 2022. Studies were eligible if they reported counseling supporting HIVST or quantitative outcomes for linkage to care among MSM and were published in English. The screening process and data extraction followed the PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) guidelines. The quality of the included studies was assessed by the National Institutes of Health quality assessment tool. Data were extracted using random effects models to combine the proportion of HIVST users who were linked to care. Subgroup analyses and metaregression were conducted to assess whether linkage to care varied according to study characteristics. All analyses were performed with R (version 4.2.1; R Foundation for Statistical Computing) using the metafor package. Results: A total of 55 studies published between 2014 and 2021, including 43 observational studies and 12 randomized controlled trials, were identified. Among these studies, 50 (91%) provided active counseling support and 5 (9%) provided passive counseling support. In studies providing active counseling support, most MSM HIVST users were linked to various forms of care, including reporting test results (97.2%, 95% CI 74.3%-99.8%), laboratory confirmation (92.6%, 95% CI 86.1%-96.2%), antiretroviral therapy initiation (90.8%, 95% CI 86.7%-93.7%), and referral to physicians (96.3%, 95% CI 85%-99.2%). In studies providing passive counseling support, fewer MSM HIVST users were linked to laboratory confirmation (78.7%, 95% CI 17.8%-98.4%), antiretroviral therapy initiation (79.1%, 95% CI 48.8%-93.7%), and referral to physicians (79.1%, 95% CI 0%-100%). Multivariate metaregression indicated that a higher number of essential counseling components, a smaller sample size (<300), and the use of mobile health technology to deliver counseling support were associated with better linkage to care. The quality of the studies varied from fair to good with a low to high risk of bias. Conclusions: Proactively providing counseling support for all users, involving a higher number of essential components in the counseling support, and using mobile health technology could increase the linkage to care among MSM HIVST users.


Background
Globally, the estimated median HIV prevalence among men who have sex with men (MSM) ranges from 5% in Southeast Asia to 12.6% in Eastern and Southern Africa [1].The risk of acquiring HIV is 26 times higher among MSM compared with the general population worldwide [1].In 2022, the proportion of new HIV cases attributed to MSM was 44% in Asia and the Pacific [2], 38.7% in Europe, and 70% in the United States [3][4][5].
HIV testing is one of the key strategies for controlling the spread of HIV [6,7].Both the Joint United Nations Programme on HIV/AIDS (UNAIDS) and the Centers for Disease Control and Prevention recommend MSM to undergo HIV testing every 3 to 6 months [6,7].The UNAIDS established its 95-95-95 targets in 2014 [6].The aim was to diagnose 95% of all individuals testing positive for HIV, provide antiretroviral therapy (ART) to 95% of those diagnosed, and achieve viral suppression for those treated by 2030 [6].High coverage of HIV testing was the first step in achieving the 95-95-95 targets.However, the overall HIV testing coverage among MSM was 86.2% in Africa and 89% in North America [8].There is a need for further improvement.HIV self-testing (HIVST) could serve as an alternative strategy for enhancing HIV testing coverage.Systematic reviews have demonstrated that HIVST can overcome barriers faced by MSM when accessing HIV testing services, such as perceived stigma from providers and inconvenience [9][10][11].Previous meta-analyses consistently showed that HIVST has doubled the frequency of HIV testing compared with facility-based testing [12,13].As a result, the World Health Organization (WHO) recommends offering HIVST as an additional approach in addition to the existing HIV testing services [14].
On the basis of the presence of counseling support, HIVST can be categorized into assisted and unassisted HIVST.Several studies investigated the linkage to care in assisted and unassisted HIVST.Individuals who received positive results through unassisted HIVST faced more difficulties in accessing care than those who were identified by facility-based HIV testing and counseling [15].According to a systematic review, <25% of unassisted HIVST users were able to complete the test without any errors, and many of them had difficulties interpreting the HIVST results [16].A meta-analysis showed that the absence of assistance would lead to a 17% decrease in the linkage to care rate among HIVST users [12].Across countries, studies have consistently shown that implementing assisted HIVST could increase linkage to care among different populations [17][18][19].A very high linkage to care (99%-100%) was observed among users of assisted HIVST in the United States and Zimbabwe [19].Therefore, the WHO recommends that counseling support be provided to HIVST users [20].
As recommended by the WHO, 8 essential components should be included in the pretest and posttest counseling of a standard-of-care client-initiated HIV testing and counseling.The pretest counseling should include (1) assessing the risks and window periods, (2) informing clients of the benefit of taking the test and the implications of both negative and positive results, (3) assuring the clients' right to refuse to take the test, (4) encouraging the clients to anticipate the possibility of beneficial disclosure of serostatus status, and (5) providing preventive information and materials [21].Essential components of posttest counseling include (1) interpreting testing results; (2) offering psychological support to individuals testing positive for HIV, facilitating beneficial disclosure of their positive serostatus, and referring them for further care and support services; and (3) providing HIV-negative individuals with preventive information and materials [21].However, the level of counseling support varied across previous HIVST programs.Some programs proactively provided pretest or posttest counseling support to all HIVST users unless they refused [22].This mode of counseling support was categorized as active counseling [23][24][25].Providing active counseling increased ART adherence among people living with HIV who had an unsuppressed viral load [26].Other programs did not provide active counseling to users.Users could report their results via a web-based platform and obtain optional posttest counseling [27].This mode of counseling support was categorized as passive counseling [24,25].In addition, the number of essential components involved in counseling support varied significantly.Some studies only involved a single essential component (eg, providing additional HIV care for HIVST users who received reactive results), whereas others provided more comprehensive support (eg, assessing the risks and window periods, delivering preventive information and materials, and providing additional HIV care for HIVST users who received reactive results) [26,27].

Objectives
Existing systematic reviews and meta-analyses have investigated the digital support [28], effectiveness [29], and acceptability of HIVST [30].However, few studies have summarized the levels of counseling support for HIVST among MSM.It is also unclear whether different levels of counseling support would result in differences in the linkage to care among MSM HIVST users.To address this knowledge gap, we systematically reviewed global evidence on counseling support for MSM HIVST users.We also summarized the linkage to care under different modes of counseling support, including (1) the proportion of users who reported HIVST results; (2) the proportion of users with reactive results who were linked to laboratory confirmation, ART initiation, and physicians; and (3) the proportion of users with negative results who were given information related to sexual risk behavior reduction, related to pre-exposure prophylaxis (PrEP), and linked to PrEP initiation.

Methods
This systematic review and meta-analysis was registered with PROSPERO (CRD42022346247) and conducted according to the PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) guidelines (Multimedia Appendix 1) [31].

Search Strategy
We searched the MEDLINE, Global Health, Web of Science, Embase, APA PsycINFO, and Scopus databases for studies (including both published papers and preprints) between July 3, 2012 (the date when HIVST was approved by the Federal Drug Administration), and June 30, 2022, in any country or setting [32].Keywords were selected based on the PICOS (participants, intervention, comparison, outcome, and study) criteria to address the research question (where P=MSM, I=HIVST with counseling, C=none, O=linkage to care, and S=randomized controlled trial [RCT] or observational studies).The Boolean operator was used in the search strategy, with "OR" and/or "AND" used to link search terms, whereas the asterisk "*" was used as a wildcard symbol appended at the end of the terms to search for variations of those terms.Full search strategies are available in Multimedia Appendix 2.
Additional studies were identified through the UNAIDS and WHO websites.We also reviewed databases listing ongoing RCTs, such as ClinicalTrials.gov,the WHO International Clinical Trials Registry Platform, and the Pan African Clinical Trials Registry, as well as reference lists of published reviews, meta-analyses, and articles.

Inclusion and Exclusion Criteria
Inclusion and exclusion criteria are presented in Table 1.The exposure categories and outcomes of the study are presented in Textbox 1. Textbox 1. Exposure categories and outcomes included in the review of linkage to care following HIV self-testing (HIVST), along with counseling.

Exposure categories
1. Studies that included HIVST along with active counseling support (eg, studies proactively provided pretest or posttest counseling to all HIVST users unless refused)

Data Extraction
Critical information from this study was extracted using a data extraction form, as outlined in Table 2.The study outcome was the proportion of MSM HIVST users who were linked to care.
Two independent reviewers (SC and YF) assessed the eligibility, evaluated the quality, and extracted information from the included publications.Any disagreements during the data extraction and quality assessment process were resolved by a senior reviewer (ZW).

Quality Assessment
The National Institutes of Health quality assessment tool was used to assess the quality of RCTs and observational studies [87].The tool covers 14 domains for RCTs, observational cohorts, and cross-sectional studies, with a total score ranging from 0 to 14. Higher scores indicated better quality, and each study's summary score was categorized as poor (0-4 out of 14 questions), fair (5-10 out of 14 questions), or good (11-14 out of 14 questions).

Data Analysis
Meta-analyses were conducted using random effects models to combine data and calculate pooled proportions and 95% CIs based on the generalized linear mixed effects method [88].Heterogeneity was quantified using I 2 statistic.I 2 values <25%, 25% to 75%, and >75% indicate low, moderate, and high heterogeneity, respectively [89].We used visual inspection to assess the asymmetry of funnel plots and the Egger test to detect potential publication bias [90].Sensitivity analysis was conducted by removing one study at a time.
Subgroup analyses and metaregression were conducted to assess whether the proportion of linkage to care varied and was predicted according to the values of the study characteristics.These included study year (we used 2016 as the cutoff as it was the year when the WHO started recommending HIVST) [91], study sample size (<300 vs ≥300), study countries (high income, upper middle income, lower middle income, and low income based on the new World Bank country classification) [92], HIVST counseling delivery modes (technology and mobile health vs peer and community), presence of pretest and posttest counseling (posttest counseling only vs both pre-and posttest counseling), and quality of counseling.Quality of counseling was measured by the number of essential components involved in the counseling support for MSM HIVST users.Among all studies, we assessed whether the overall linkage to care varied by study characteristics (type of counseling support, study year, study sample size, study countries, HIVST counseling delivery modes, presence of pretest and posttest counseling, and quality of counseling) using the univariate metaregression model.Factors with P<.10 in univariate metaregression analyses were entered into the multivariable metaregression model.Within studies providing active or passive counseling support, univariate and multivariate metaregression were used to examine whether linkage to care varied by study characteristics.All analyses were performed with R (version 4.2.1;R Foundation for Statistical Computing) using the metafor package.

Essential Components Involved in the Pretest Counseling of the Active Counseling Supporting HIVST
In the pretest counseling, 32 studies proactively provided at least one essential component to MSM HIVST users (Multimedia Appendix 4 [27,).Five studies provided only one essential component, such as informing users of the benefits of taking the tests (1/5, 20%), assuring users' rights to refuse HIV testing (2/5, 40%), or providing HIV prevention information (2/5, 40%).Five other studies provided 2 essential components.In addition to informing users of the benefits of taking the tests, these studies provided risk assessment (1/5, 20%), assured users' right to refuse (2/5, 40%), encouraged beneficial disclosure of HIV serostatus (1/5, 20%), or provided HIV prevention information (1/5, 20%).Four other studies provided 3 essential components.In addition to informing the users about the benefits of taking the test, the combination of other components were (1) providing risk assessment and HIV prevention information (1/4, 25%), (2) encouraging beneficial disclosure of serostatus and providing HIV prevention information (1/4, 25%), (3) assuring users' right to refuse HIV testing and providing HIV prevention information (1/4, 25%), and (4) providing risk assessment and assuring users' right to refuse (1/4, 25%).Another study provided 4 essential components: (1) informing users of the benefits of taking the tests, (2) providing risk assessment, (3) assuring the user's right to refuse HIV testing, and (4) encouraging beneficial disclosure of serostatus status.The remaining studies (n=13) provided all 5 essential components.

Essential Components Involved in the Posttest Counseling of the Active Counseling Supporting HIVST
In posttest counseling, 50 studies proactively provided at least one essential component to MSM HIVST users (Multimedia Appendix 4).Six studies only provided one essential component, offering additional HIV care to MSM who received reactive HIVST results.Moreover, 24 studies provided 2 essential components.These combinations included (1) additional HIV care for users with reactive HIVST results and interpretation of HIVST results (23/24, 96%) and ( 2) additional HIV care for users with reactive HIVST results and HIV prevention information for users with negative HIVST results (1/24, 4%).The rest of the studies (n=18) provided all 3 essential components (eg, interpretation of testing results, HIV prevention information for users with negative HIVST results, and additional HIV care for users with reactive HIVST results).

Essential Components Involved in the Pretest Counseling of Passive Counseling Supporting HIVST
In the pretest counseling, 3 studies offered at least one essential component upon request.One study only informed the benefits of taking the tests.The other 2 studies included 2 essential components: informing the participants of the benefits of taking the tests and providing HIV prevention information.In addition to these essential components, one study provided local data, news, and policies regarding HIV and sexually transmitted infections among MSM.

Essential Components Involved in the Posttest Counseling of the Passive Counseling Supporting HIVST
In posttest counseling, all 5 studies provided at least one essential component upon request.Three studies provided one essential component, such as interpretation of the HIVST results (1/5, 20%), or provision of psychological support and referral to HIV care for users with positive results (2/5, 40%).The other 2 studies provided 2 essential components.In addition to the interpretation of HIVST results, these studies provided referral to HIV care for users with positive results (1/5, 20%) or HIV prevention information for users with negative results (Multimedia Appendix 4).

Meta-Analysis of Linkage to Care Among MSM HIVST Users Along With Active and Passive Counseling
The main findings of the meta-analysis of the linkage to care among MSM HIVST users are summarized in Figure 2.

Linkage to Information Related to Sexual Risk Behaviors Reduction and PrEP and PrEP Initiation
Overall, 7 studies reported linkage to information related to sexual risk behavior reduction and PrEP among users with negative results.In studies with active counseling support, the pooled proportion of linkage to sexual risk behaviors reduction was 100% (n=7; 95% CI 0%-100%; I 2 =0%; Figure 7 [34,49,60,63,71,81,85]).The pooled proportion of PrEP initiation was 27% (n=6; 95% CI 10.2%-54.6%;I 2 =97%; Figure 7) [48,60,72,81,84,85] in studies providing active counseling support.No studies with passive counseling support reported a linkage to information related to sexual risk behaviors and PrEP or PrEP initiation among users with negative results.

Publication Bias
Upon examination of the funnel plots (Figure 8; Table 3), there was a publication bias in studies reporting the proportion of linkage to laboratory confirmation (P=.02), referral to physicians (P<.001), and prevention strategies (P<.001).Furthermore, outliers were identified in studies reporting the proportion of reporting test results [51] and linkage to laboratory confirmation [73,82], ART initiation [63], physicians [35], and PrEP initiation [85] (Figure 8).

Sensitivity Analysis
For studies providing active counseling support, the pooled proportion of reporting test results, and linkage to laboratory confirmation, ART initiation, physicians, information related to sexual risk behaviors reduction and PrEP, and PrEP initiation changed slightly after removing one study at each time.
With regard to the studies providing passive counseling support, the pooled proportion of linkage to ART initiation did not change after removing one study at each time.However, after removing the study conducted by Jin et al [50], the pooled proportion of linkage to laboratory confirmation changed from 78.7% (95% CI 17.8%-98.4%)to 55.5% (95% CI 27%-80.8%).As there were only 2 studies with passive counseling support that reported linkage to physicians, a sensitivity analysis was not conducted.Details of the sensitivity analysis results are presented in Multimedia Appendix 5 [27,.

Subgroup Analysis and Metaregression
Subgroup analysis by study sample size (<300 vs ≥300) revealed different levels of linkage to reporting test results (98.5% vs 82.7%), laboratory confirmation (93.8% vs 85.6%), physicians (92% vs 86%), and PrEP initiation (35.6% vs 23.9%).The subsequent subgroup analysis (posttest counseling only vs both pre-and posttest counseling) also found different proportions of users who reported test results (87.1% vs 93.2%) and initiated PrEP (20.2% vs 53%).Subgroup analysis by other study characteristics did not reveal a large difference in the linkage to care.Among all studies, univariate metaregression analysis demonstrated that the type of counseling (active vs passive), a smaller sample size (<300 vs ≥300), and a higher number of essential components involved in the counseling support were significantly associated with better linkage to care.Furthermore, the multivariate metaregression analysis confirmed that a larger sample size was linked to a lower linkage to care (P=.03).In contrast, mobile health technology counseling (P=.05) and a higher number of essential components involved in the counseling support were associated with increased linkage to care (P=.04).
In studies providing active counseling support, univariate metaregression analysis indicated a smaller sample size (<300 vs ≥300), provision of both pretest and posttest counseling (vs XSL • FO RenderX posttest counseling only), and a higher number of essential components were significantly associated with better linkage to care.The findings of multivariate metaregression analysis revealed that a smaller sample size (P=.03) and using mobile health technology for counseling (P=.05) were associated with a higher linkage to care.With regard to different outcomes related to linkage to care, a larger sample size was correlated with a lower linkage to laboratory confirmation (P=.03) and prevention strategies (P<.001) with active counseling support (Multimedia Appendix 6).

Principal Findings
This systematic review and meta-analysis aimed to summarize the global evidence on counseling support and synthesize the proportion of linkage to care among MSM HIVST users.We categorized counseling support in assisted HIVST into active or passive.More than 90% of the MSM HIVST users with reactive results were linked to laboratory confirmation and ART initiation in studies implementing active counseling support.Such a proportion was higher than that of the studies with passive counseling support (78.7%-79.1%).Therefore, the provision of active counseling support may be helpful in improving the linkage to HIV care and treatment for MSM HIVST users with reactive results.
Relatively few studies (7/55, 13%) provided information related to sexual risk behaviors reduction and PrEP for MSM HIVST users with negative results.One possible explanation was that most resources were used to provide support for users with reactive results, which was considered a priority for some HIVST programs [93].Hence, there are constraints in resources to provide support for the large number of users with negative results [93].In addition to identifying individuals testing positive for HIV, facilitating behavior changes is an important purpose of HIV testing and counseling.Future studies should consider providing more comprehensive support for MSM HIVST users with negative results.
The metaregression results identified some significant determinants of the linkage to care among MSM HIVST users.First, more essential counseling components were associated with better linkage to care, which aligns with findings from a previous study [94].Incorporating a higher number of essential components would enhance counseling quality.Previous studies suggest that delivering high-quality counseling improves linkage to care, reduces risky behaviors, and prevents new infections [95,96].However, our study found that only 26% to 36% of the studies provided all essential active pretest and posttest counseling support.As a result, future programs should consider offering comprehensive counseling to MSM HIVST users.
Second, a larger sample size was associated with a lower linkage to care among MSM HIVST users.Providing active counseling support for HIVST users was resource demanding.For example, it took 1 hour to prepare and implement one session of real-time pretest and posttest counseling support for each MSM HIVST user [49,71].Therefore, it is challenging to provide counseling support to a larger number of HIVST users.Furthermore, mobile health technology counseling was associated with a better linkage to care compared with peer and community counseling.Our study found that mobile health technology counseling is the predominant method to support MSM HIVST users, which aligns with a previous study [28].Owing to high smartphone ownership among MSM (>94%) [97], mobile health technology presents a viable strategy for counseling among MSM HIVST users.In addition, previous studies have demonstrated that using mobile health technology for counseling support reduces the workload of HIV testing administrators [28,98].
Stigma and discrimination against MSM impede access to HIV testing and counseling services [99].Systematic reviews have shown that perceived stigma remained a significant obstacle to engaging in assisted HIVST [100,101].Future programs should consider increasing the empathy of health workers who provide counseling to support HIVST.A previous study suggested that the negative effects of perceived stigma or discrimination on HIV testing use could be offset by increasing the empathy of service providers [102].A recent study applied computerized programs (instead of people) to provide active counseling supporting HIVST [98].Such an approach could also reduce the concerns of stigma or discrimination when using assisted HIVST among MSM.

Limitations
There are several limitations in this study.First, there was high heterogeneity among the studies that reported outcomes on linkage to laboratory confirmation, physicians with the provision of passive counseling support, linkage to PrEP initiation, and reporting test results.Heterogeneity pertains to the diversity observed in the design of studies, the effects of interventions, or the outcomes obtained across different studies.Persistent heterogeneity could not be resolved using sensitivity analysis.Second, publication bias was found in relation to the pooled proportion of linkage to laboratory confirmation, physicians, and prevention strategies, which could impact the validity and generalization of conclusions.Third, because half of the included studies were cross-sectional, a causal relationship could not be established.Furthermore, the use of nonprobabilistic sampling for MSM in all the included studies limited the generalizability of the findings.Finally, the small number of studies in the comparison group (those providing passive counseling) would result in bias when comparing the study outcomes between active and passive counseling support.

Conclusions
This study synthesized evidence on active and passive counseling support for MSM HIVST users and quantified the proportion of linkage to care.As compared with passive counseling support, active counseling support had a better linkage to care.Having a higher number of essential counseling components, a smaller sample size, and using mobile health technology to deliver counseling support were also associated with a better linkage to care.As our results showed, proactively providing counseling support for all users, involving more essential components in the counseling support, and using mobile health technology should be considered to increase the linkage to care among MSM HIVST users.
c CDC: Centers for Disease Control and Prevention.d HIVST: HIV self-testing.e PrEP: pre-exposure prophylaxis.f MSM: men who have sex with men.g ART: antiretroviral therapy.

Figure 1 .
Figure 1.Flowchart outlining the systematic review process.
antitretroviral therapy.bPrEP: pre-exposure prophylaxis.c Pooled proportion was not performed because of fewer than 2 studies.

Figure 2 .
Figure 2. Main findings in the meta-analysis of the linkage to care among users.ART: antiretroviral therapy; PrEP: pre-exposure prophylaxis.

Figure 3 .
Figure 3. Forest plot of the pooled proportion of reporting HIV self-testing results.

Figure 4 .
Figure 4. Forest plot of the pooled proportion of linkage to laboratory confirmation among users with reactive results: (A) studies were provided active counseling along with HIV self-testing and (B) studies were provided passive counseling along with HIV self-testing.

Figure 5 .
Figure 5. Forest plot of the pooled proportion of linkage to antiretroviral therapy among users who were confirmed HIV positive: (A) Studies were provided active counseling along with HIV self-testing and (B) studies were provided passive counseling along with HIV self-testing.

Figure 6 .
Figure 6.Forest plot of the pooled proportion of linkage to physicians among users with reactive results: (A) studies were provided active counseling along with HIV self-testing and (B) studies were provided passive counseling along with HIV self-testing.

Figure 7 .
Figure 7. Forest plot of the (A) pooled proportion of linkage to information related to sexual risk behaviors reduction and pre-exposure prophylaxis (PrEP) among users with negative results and (B) pooled proportion of PrEP initiation among users with negative results.

Figure 8 .
Figure 8. Funnel plots for assessing the publication bias among the included studies: (A) linkage to reporting test results; (B) linkage to laboratory confirmation with active counseling support; (C) linkage to laboratory confirmation with passive counseling support; (D) linkage to antiretroviral therapy (ART) initiation with active counseling support; (E) linkage to ART initiation with passive counseling support; (F) linkage to physicians with active counseling support; (G) linkage to physicians with passive counseling support; (H) linkage to information related to sexual risk behaviors reduction, and pre-exposure prophylaxis (PrEP); and (I) linkage to PrEP initiation.

Table 1 .
Summary of the inclusion and exclusion criteria.

Table 2 .
Characteristics, active counseling, and passive counseling support of included studies.
b RCT: randomized controlled trial.

Table 3 .
Results for studies that assessed linkage to care.