What would you choose: Online or Offline or Mixed services? Feasibility of online HIV counselling and testing among Thai men who have sex with men and transgender women and factors associated with service uptake

Abstract Introduction HIV testing coverage remains low among men who have sex with men (MSM) and transgender women (TGW). We studied characteristics of Thai MSM and TGW who chose online and/or offline platforms for HIV counselling and testing and the feasibility of integrating online technologies and HIV self‐testing to create service options. Methods From December 2015 to June 2017, MSM and TGW enrolled from Bangkok Metropolitan Region and Pattaya could choose between: 1 offline HIV counselling and testing (Offline group), 2 online pre‐test counselling and offline HIV testing (Mixed group), and 3 online counselling and online, supervised, HIV self‐testing (Online group). Sociodemographic data, risk behaviour and social network use characteristics were collected by self‐administered questionnaires. Logistic regression models identified covariates for service preferences. Results Of 472 MSM and 99 TGW enrolled, 202 self‐selected the Offline group, 158 preferred the Mixed group, and 211 chose the Online group. The Online group had the highest proportion of first‐time testers (47.3% vs. 42.4% vs. 18.1%, p < 0.001) and reported highest HIV prevalence (15.9% vs. 13.0% vs. 3.4%, p = 0.001) as compared to Offline and Mixed groups, respectively. Having tested for HIV twice or more (OR 2.57, 95% CI 1.03 to 6.41, p = 0.04) increased the likelihood to choose online pre‐test counselling. Being TGW (OR 6.66, 95% CI 2.91 to 15.25, p < 0.001) and using social media from four to eight hours (OR 2.82, 95% CI 1.48 to 5.37, p = 0.002) or >8 hours (OR 2.33, 95% CI 1.05 to 5.16, p = 0.04) increased selection of online, supervised, HIV self‐testing. Providers primarily used smartphones (79.2%) and laptops (37.5%) to deliver online services. Self‐testing strip image sharpness and colour quality were rated “good” to “excellent” by all providers. Most participants (95.1%) agreed that online supervision and HIV self‐testing guidance offered were satisfactory and well delivered. Conclusions Online HIV services among MSM and TGW are feasible in Thailand and have the potential to engage high proportions of first‐time testers and those with high HIV prevalence. When designing public health interventions, integrating varied levels of online HIV services are vital to engage specific sections of MSM and TGW populations in HIV services. Clinical Trial Number NCT03203265

HIV [4]. In response, Thailand with support from the USAID has implemented the Key Population-Led Health Services (KPLHS) model led by community-based organizations (CBOs) serving MSM and/or TGW. KPLHS are a defined set of HIVrelated health services delivered by trained KP community health workers in partnership with other health sector entities. This KPLHS model has proved extremely successful in engaging MSM and TGW who are at high-risk for HIV infection into early diagnosis, early antiretroviral treatment (ART) linkage, and high pre-exposure prophylaxis (PrEP) uptake [5]. In 2017 alone, KP community health workers contributed to 38% of the 41,386 HIV counseling and testing services and 26% of 4840 new HIV diagnoses among MSM and TGW in Thailand [6].
Given the heterogeneity in psychosocial context and health and digital literacy among MSM and TGW, multiple HIV testing options are needed to ensure that certain subpopulations are not excluded [7]. Motivation and educationbased interventions through peer mobilizers and mass media campaigns could help tackle fear of testing. New technologies including online communication platforms and HIV selftesting have shown potential to overcome structural barriers and increase access to HIV counselling and testing among KPs [8,9].
A systematic review revealed that supervised HIV self-testing (conducted with real-time support from a healthcare provider) and unsupervised HIV self-testing were highly acceptable and preferred among people at risk for HIV infection, although lower sensitivity was found when self-testing was unsupervised [10] Online support to perform HIV self-testing has shown advantage over unsupervised HIV self-testing by the ability to address certain concerns such as lack of preand post-test counseling [11]. Thai MSM and TGW who lead the region in Internet and technology adoption and utilization have shown consistently high preferences for online HIV service delivery [12,13]. Online supervised self-testing with special emphasis on individual-level counselling could help address barriers to self-testing uptake among Thai MSM, such as fear of receiving self-testing kit at home, fear of one's own lack of understanding of self-testing and receiving results alone [14].
We leveraged Online, Offline and Mixed HIV counselling and testing, three distinct service delivery models by integrating online and HIV self-testing technologies, KPLHS and public healthcare services. In this paper, we specifically explored characteristics of Thai MSM and TGW as well as key factors they took into account when choosing service options. In addition, we studied providers' technology skills and utilization levels and their feasibility of delivering online HIV services as well as satisfaction among MSM and TGW clients receiving services.

| METHODS
From December 2015 to June 2017, Thai MSM and TGW were consecutively recruited and enrolled into a 12-month cohort study, with six-monthly visits to assess the preferences and feasibility of online and offline HIV counselling and testing strategies (NCT03203265). The Institutional Review Board of the Faculty of Medicine, Chulalongkorn University, and the Bangkok Metropolitan Administration Ethics Committee approved this study. Inclusion criteria included Thai national, aged >18 years, being men or TGW, engaged in unprotected anal sex with men at least once in the past 6 months, living in Bangkok Metropolitan Region or Pattaya, and not known to be HIV positive. The study was supported by amfAR GMT Initiative grant and conducted by the Thai Red Cross AIDS Research Centre (TRCARC), Service Workers IN Group (SWING) Foundation, Rainbow Sky Association of Thailand (RSAT), and Sisters Foundation. SWING, RSAT and Sisters were CBOs serving MSM and TGW. SWING had two community health centres in Bangkok and Pattaya, RSAT had one community health centre in Bangkok, and Sisters had one in Pattaya. We reported baseline data from this study.

| Recruitment strategies and informed consent process
Online recruitment strategies included dissemination of study recruitment posters, text-based messages, and an online HIV self-testing video promoted through organizations' websites, banners on popular websites, and platforms commonly used by MSM and TGW such as Facebook, LINE chat groups, Camfrog video chat rooms, Hornet, and Jack'D. Offline recruitment was conducted by CBOs at hot spots using study posters and flyers. Participants interested in joining the study were scheduled for either online or offline informed consent process, based on individual preference. The online informed consent process was conducted using a real-time video chatting platform enabling the participant information sheet to be reviewed via shared screen.

| HIV counselling and testing via online and/or offline strategies
To understand the types of service delivery models that appeal to various sub-groups of MSM and TGW participants, the study allowed participants to self-select from three distinct strategies ( Figure 1) including: (1) conventional offline HIV counselling and testing (Offline group), (2) online pre-test counselling and offline HIV testing (Mixed group), and (3) completely online counselling and supervised HIV self-testing (Online group). Services in the Offline group were delivered by staff at TRCARC and the four community health centres. In the Mixed group, pre-test counselling was conducted online by study staff and participants were scheduled to receive HIV testing and post-test counselling at a by-appointment-only clinic in Bangkok operated by TRCARC. Online group participants received totally online pre-test counselling, HIV testing, and post-test counselling. Study staff provided linkage to care and ART initiation support to all HIV-positive participants, regardless of CD4 count as per the Thailand National Guidelines [15].

| HIV testing and other services provided in
Online, Mixed and Offline groups Blood samples were collected by venipuncture in the Offline group for HIV testing. Other sexually transmitted infection testing were conducted based on provider's judgment. HIV testing was conducted according to Thailand National Guidelines [15], starting with either machine-based 4th generation (at TRCARC) or rapid third generation assay (at the community health centers) with confirmation of reactive result made by another two assays. Confirmed HIV status was available within one to two hours. To identify acute HIV infection cases, all non-reactive samples were also sent for nucleic acid testing (NAT) by Aptima HIV-1 RNA qualitative assay (Gen-Probe Inc., San Diego, CA, USA).
Participants in the Mixed group were scheduled to receive fingerprick blood collection at the clinic. Rapid third generation assay was performed with confirmation of reactive result by another two assays which confirmed HIV status within one to two hours.
In the Online group, participants had an HIV self-testing package and an online HIV self-testing video URL link [16] sent to them on a pre-scheduled date. Study staff contacted the participants to ensure package delivery and scheduled date and time for the online, supervised, HIV self-testing process.
Step-by-step, real-time guidance was provided through a video chatting platform preferred by the participants. For selftesting, we used a rapid third generation assay (Alere Determine TM HIV 1/2, Alere Medical Co., Ltd., Matsuhidal, Matsudoshi, Chiba, Japan) and fingerprick blood sample which allowed for 15 minutes diagnosis result. Post-test counselling for participants with reactive results emphasized immediate linkage to confirmation and ART initiation at preferred hospitals or clinics.
PrEP and post-exposure prophylaxis (PEP) were offered to all participants and assistance to access PrEP and PEP services offline was given by study staff. Participants in the Mixed and the Online groups were also provided access to the Adam's Love Electronic health record (EHR) system which enabled secure access to laboratory results, post-test summaries and appointment scheduling for HIV test, as previously described [17].

| Data collection and questionnaires
A self-administered questionnaire was used at baseline visit to collect sociodemographic, risk behaviour, social network use characteristics, perceived barriers and facilitators for HIV testing, and experiences around stigma and discrimination. Online group participants also completed a questionnaire assessing their decision-making and reasons for choosing online, supervised, HIV self-testing, over clinic-based testing, and feelings post-service utilization. Study staff delivering online services completed a questionnaire that assessed their technology skills and utilization, and the feasibility and acceptability of delivering online HIV services.

| Statistical analysis
Data from the questionnaires were reported overall and by self-selected groups and by gender as frequency and proportion for categorical parameters and mean (standard deviation, SD) and median (interquartile range, IQR) for continuous parameters. Comparison of continuous variables between groups was made by Two-sample t test, ANOVA techniques or non-parametric tests. Chi-square or Fisher's exact test was used for comparison of proportion of characteristics between groups.
The associations of covariates with a preference for online vs. offline pre-test counselling service and online vs. offline HIV testing and post-test counselling services were modelled using binary logistic regression, adjusting for confounders as appropriate. Factors showing significant level of 0.10 in the univariate model were adjusted in the multivariate model.

| RESULTS
The study recruited 437 MSM and 66 TGW through online and 37 MSM and 33 TGW via offline channels ( Figure 1). Informed consent process was conducted online with 320 MSM and 54 TGW and offline with 152 MSM and 45 TGW. Of 571 participants enrolled, 202 selected the Offline group, 158 chose the Mixed group, and 211 preferred Online group. Seven individuals were further excluded (Figure 1) and data analyses were subsequently performed on 564 participants (Table 1). Mean (SD) age was 27.9 (7.2) years. In the past six months, 68.3% had multiple sex partners, 77.4% inconsistently used condoms, 6.2% used amphetamine-type stimulants, and 18.2% engaged in group sex. Basic characteristics of MSM and TGW participants are summarized in Table 1.

| First-time tester rate and HIV prevalence by self-selected groups
All participants in the Offline group completed HIV testing process although proportions were lower in the Mixed group (94.3%) and Online group (92.4%) ( Figure 1). Nine participants in the Mixed group and 11 in the Online group did not show up for the offline or the online scheduled visit, and four could not complete online, supervised, selftesting process.
First-time testing rate was 42.4% in the Offline group, 18.1% in the Mixed group, and 47.3% in the Online group, p < 0.001. HIV prevalence was lowest in the Mixed group followed by the Offline and Online groups (3.4% vs. 13% vs. 15.9%, p = 0.001). One HIV-positive case was diagnosed by NAT in the Offline group. The Mixed group also had lowest proportion of TGW when compared to the Offline and Online groups (1.3% vs. 22.5 vs. 25.0%, p < 0.001).

| Factors associated with selecting "online
pre-test counselling" (Table 2) In a logistic regression model, having a bachelor degree or higher (OR 2.02, 95% CI 1.02 to 4, p = 0.045), having monthly income between 429 and 857 USD (OR 2.31, 95% CI 1.02 to 5.21, p = 0.04), and having had HIV tested twice or more (OR 2.57, 95% CI 1.03 to 6.41, p = 0.04) increased online pre-test counselling selection chances. Identifying inconvenient service hours as a barrier to HIV testing (OR 2.92, 95% CI 1.32 to 6.45, p = 0.008) influenced MSM to choose online pre-test counselling. Among TGW, positive attitude towards HIV testing (OR 8.75, 95% CI 1.56 to 49.01, p = 0.01) was associated with a higher chance to select online pre-test counselling, although prioritizing NAT-based HIV testing (OR 0.13, 95% CI 0.02 to 0.76, p = 0.02) decreased such chances.

| Other factors associated with the decision to join the Online group and feelings post serviceutilization
Of the 208 participants who joined the Online group, 160 responded to the survey. For most (87.5%), it was an easy decision to choose online, supervised, HIV selftesting. Preference for joining was guided by logistic/time convenience (46.9%), privacy and confidentiality (19.4%), altruism (16.9%), and scheduling flexibility (15.6%). Positive perceptions ("it's good and convenient", "it's amazing", "I'm glad to join HIV self-testing") increased from 67% before the process to 82% after. Negative perceptions ("fear of fingerprick", "I feel anxious", "I am nervous with the procedure") decreased from 23.2% to 8.8%. Most agreed that the HIV self-test kit mailed to them was complete (95%), the video and study brochure helped them comprehend the procedure (92.9%), and online

| Technology skills and utilization of counsellors and feasibility of providing online services
There were four MSM, one TGW and three cisgender women staff who responded to the survey. Median (IQR) age was 28.5 (24.5 to 31.5) years. Half had counselling experience of two years or less. Majority (75.0%) spent >4 hours using Internet in a day. All reported having used video calls for communication and felt comfortable using technology in their daily lives. Smartphones (79.1%) and laptops (37.5%) were the primary devices used for delivering services. Primary apps for conducting video calls included LINE (87.5%), Facebook messenger (20.8%), ClickDesk (8.3%), Zoom (8.3%) and Facetime (8.3%). Majority (75%) reported experiencing a maximum of two Internet glitches per video call session. On a 5-point Likertscale, majority (83.4%) rated the image quality of HIV selftesting strip image as "very good" or "excellent" and all agreed that HIV test result image captured and displayed was sufficient and conclusive for providing post-test counselling.

| DISCUSSION
Our implementation research study results illustrate that conducting online pre-test counselling and online, supervised, HIV self-testing and post-test counselling among MSM and TGW are feasible in Thailand, when conducted by healthcare professionals and trained KP community health workers. Prior HIV testing experiences, privacy and confidentiality needs, HIV testing attitudes, and social network use patterns are significant factors driving the choice to select online, mixed or offline services. Online supervised HIV self-testing significantly engaged first-time testers and those with highest HIV prevalence, further emphasizing the need for large-scale implementation of such model. Our study also demonstrated that an implementation research offering online, mixed and offline HIV testing options with self-selection by participants is feasible and allows for near real-life situations and lessons learned.
Having prior HIV testing experience may help facilitate one's decision to seek unconventional HIV testing services [18]. MSM previously tested for HIV were more likely to choose online pre-test counselling in our study. Inconvenient location and service hours are commonly cited barriers to scaling-up HIV testing among MSM and other KP [8,19]. Those who cited such barriers were more likely to choose online pre-test counselling, offering opportunities to foster participant-counsellor relationship and enable appointment scheduling.
MSM participants with interest in home-based testing and having concerns around confidentiality showed higher preference for online, supervised, HIV self-testing, which supports findings among Chinese MSM who prioritized privacy and confidentiality when selecting online HIV services [20]. In contrast to a previous study [21], seeking sex partners online did not influence participant's decision to seek online HIV testing. Although, consistent with an earlier finding, high social media usage increased participants' selection of online HIV services [20].
Being TGW was another strong predictor for choosing online, supervised, HIV self-testing in our study. High Internet and social media usage patterns among TGW and its potential to reach and provide non-judgmental support to TGW was reported in earlier studies [22][23][24]. TGW who considered taking an HIV test as a way to living healthy life were also more likely to choose online services. This is vital to designing public health interventions and research projects targeting TGW.
Our results harmonize well with previous study by Flowers et al. highlighting the need for diverse approaches to HIV testing interventions for maximum public health benefit [7]. It is plausible that first-time testers or testers with high risk would access a friendly offline HIV testing clinic where immediate confirmatory testing and linkages to PrEP/PEP or ART are available. Testers who perform regular check up every three to six months may prefer HIV self-testing with an initial supervision by providers.
HIV self-testing is not yet registered in Thailand as the Thai Food and Drug Administration lacks clarity around the level of support optimal for Thai users. Findings from this study and growing literature on the need for online counselling support for HIV self-testers [10] should provide necessary guidance for shaping Thailand's national policy around online HIV self-testing.
Concerns around quality of HIV services and intention to have immediate confirmatory testing and linkage to ART were main factors driving MSM and TGW away from selecting online HIV services. This may be particularly challenging for In the past 12 months, ever experienced stigma and discrimination in healthcare setting (n = 169)   [25,26]. In addition, to implement online HIV services, efforts should be made to ensure that the KP communities clearly perceive high quality of online service delivery and trust that adequate support can be provided for linkage to clinical services. Almost 40% of MSM and TGW in our study felt that the availability of NAT affected their decision to take HIV testing, pointing to the need for KPs' access to an HIV testing assay which could allow for early detection of HIV infection. The availability of rapid 4th generation antigen-antibody assay highlights the near-future feasibility to have a self-testing assay which could detect HIV infection earlier than the existing assays [27,28].
Level of technology skills among healthcare providers has appeared to be the most important predictor of technology use in the workplace and is related to higher nursing competency [29,30]. Our findings showed that the level of technology ownership, skills and use was innately high among Thai providers who used them based on experience and self-learning, and thus, did not require intensive technology use training to deliver online services. High satisfaction for online counselling guidance reported by participants, and satisfactory video and self-testing result image quality expressed by providers have significant implications in terms of broader scale-up of online, supervised, HIV self-testing services in Thailand. However, among sub-populations or in settings with inadequate digital literacy, efforts should focus on simplifying digital health technologies, which help to bridge, not exacerbate, health and social inequities [31].
Our study has a few limitations. Although participants were allowed to self-select the groups, we could not fully avoid the possibility of bias by study staff at each study site who might have unintentionally influenced participants decision. All three options offered involve some level of social exposure and may not be inclusive enough for those who could not overcome even limited level of social exposure in the Online group. Secondly, due to unavailability of KPLHS clinics in other regions we did not enroll MSM and TGW beyond Bangkok Metropolitan Region and Pattaya to avoid possible bias to select the Online group. The result from this study, thus may not be generalized to MSM and TGW in other provinces. A recent online survey found that 30% of MSM in Southeast Asia had never been tested for HIV and were likely to be young, high-risk, non-gay-identified MSM [32]. Given these groups could be reached online for web-based surveys, there is a high potential of such outreach platforms for delivering online HIV services. To reduce barriers in healthcare access, public health experts and programme  implementers are encouraged to adapt online HIV service models explored in our study.

| CONCLUSIONS
In summary, we demonstrated the feasibility of conducting online HIV counselling and testing services among Thai MSM and TGW. The online, supervised, HIV self-testing service was particularly preferred by TGW, MSM who had privacy and confidentiality concerns, and those who spent more time using social media per day. Results from this study are vital in designing public health interventions targeting segments of MSM and TGW populations with preference towards online HIV services delivery.

R O L E O F T H E F U N D I N G S O U R C E
The funders of the study had no role in study design, data collection, data analysis, data interpretation, or writing of the report. The corresponding author had full access to all the data in the study and had final responsibility for the decision to submit for publication.