To adapt the P3 app to the Thai context, we collected qualitative data through focus group discussions (FGDs) with YMSM and key informant interviews (KIIs) in Bangkok, Thailand, between January and March 2019. Participants in both the FGDs and KIIs (1) described facilitators and barriers to PrEP adherence among Thai YMSM, (2) discussed mobile health (mHealth) app features and functions preferred by YMSM; and (3) provided specific feedback about P3 and offered suggestions for adapting it to create P3T.

Institutional review board approval was granted for this study by the Faculty of Medicine, Chulalongkorn University, with a waiver for parental consent granted. This study was registered with ClinicalTrials.gov (NCT04413708).

All study participants completed a brief computer-assisted self-interviewing survey on patient demographics, technology use, and previous experience using or providing PrEP before their session. All FGDs and KIIs were facilitated by a study staff member using a semistructured interview guide. After exploring PrEP adherence facilitators and barriers and mHealth app preferences among Thai YMSM, the facilitator gave a presentation of key features of P3 and elicited feedback and suggested changes from study participants.

The presentation of P3 features included screenshots and descriptions translated into Thai by the facilitator or interviewer starting with screenshots of onboarding activities (eg, review of community guidelines, app terms and conditions, setting up an account, option to lock screen with a PIN, brief introductory screens describing key P3 features, and profile setup). Next, participants were shown a screenshot of the user dashboard, or home screen, with an explanation of each icon. Five key features of P3 (Textbox 1) were demonstrated using screenshots and explained in detail by the facilitator, including the social wall, daily quests, medication tracking and adherence support, knowledge center, and collections. An additional key feature, adherence counseling, was not presented via screenshots but was explained during the presentation of the user dashboard screenshot. A brief description of these key features is provided in Textbox 1. Detailed information on all P3 features has been described elsewhere [34].

FGD participants were recruited from (1) HIV prevention clinics at the Center for Excellence for Pediatric Infectious Diseases at the Chulalongkorn University, a major teaching hospital in Bangkok; (2) TRCARC, Thailand’s largest HIV voluntary counseling and testing (VCT) center, which serves over 2000 adolescents per year; and (3) Two community-based organization branches of TRCARC, Service Workers IN Group and Rainbow Sky Association Thailand. Potential participants were approached at HIV VCT visits to assess their interest in study participation. In addition, study advertisements were posted in clinics, social media, and via the TRCARC website. The inclusion criteria were as follows: (1) age 16-24 years, (2) male sex at birth, (3) self-identification as MSM, (4) ability to speak and read Thai, (5) familiarity with smartphones, and (6) current or previous PrEP use. Those unable to provide consent due to substance use or psychological conditions were excluded.

FGDs lasting 60-120 minutes were conducted in private conference rooms either at offices in clinics or reserved private conference rooms in cafes in groups of 4-8. All participants were identified using a pseudonym to protect their confidentiality. At the conclusion of the FGDs, participants were provided THB 500 (US $16) to compensate them for their time.

Key informants who had collaborated with the Center for Excellence for Pediatric Infectious Diseases or participated in academic conferences on adolescent health in Bangkok were recruited for KIIs via in-person contact, phone calls, text messages, and email invitations. To participate in a KII, individuals were required to be a current PrEP provider for YMSM in Bangkok. KIIs, which lasted between 40 and 60 minutes were conducted face-to-face at PrEP clinics. The KII participants were not compensated.

Pre-FGD and KII survey data were summarized using Adobe Acrobat. FGDs and KIIs were audio-recorded and transcribed in Thai. A directed content analysis approach, defined as categorization derived directly from text data and transcribed interviews, was used [37]. An initial codebook was developed collaboratively between the US and Thai study teams based on semistructured interview guides that reflected the study research themes and subthemes. An iterative approach was used when new codes arose, as interviews were conducted. New codes emerging were added, and any transcribed code that had already been coded was recoded. Two members of the Thai study team coded the FGD and KII transcripts using NVivo (version 12.5.0) [38]; statistical interrater reliability for coding was not calculated. Instead, divergence among coders was discussed among the analysis groups to reach consensus, resolved by the Thai site principal investigator. Data were then deductively analyzed using thematic analysis by the same Thai study team members to identify key themes and subthemes relevant to the adaptation of P3 for YMSM in Thailand. Quotes illustrative of key themes were identified and translated into English by bilingual Thai English speakers.

This study explored facilitators and barriers related to PrEP adherence and elicited general and specific recommendations from Thai YMSM and key informants to ensure that adaptations of a PrEP adherence app originally developed for YMSM in the United States were relevant, engaging, and useful for the Thai context. To our knowledge, this is one of the few studies conducted outside high-income settings that address the adaptation of an mHealth approach to support PrEP adherence among YMSM [40].

The individual, societal, and systemic level facilitators and barriers to PrEP adherence among Thai YMSM identified in this study have also been reported in other studies of MSM and YMSM in the United States, Thailand, and other countries. Consistent with other studies [7,41-45], adherence facilitators included awareness of HIV risk and desire to protect themselves, concerns related to sexual partners, support from friends, sexual partners, family members, and adolescent-friendly providers, and PrEP affordability (PrEP is provided free of charge through universal health coverage for key populations in Thailand). Key barriers, also identified in other studies [41-44,46-49], included sexuality, HIV and PrEP stigma, attending regular PrEP medical appointments, forgetting to take PrEP daily or while intoxicated, not having PrEP with them at certain times, running out of PrEP, side effects, and low HIV risk perception. Because the PrEP facilitators and barriers identified in this study had been identified in prior research and were already addressed by the original P3 app, its content was already designed to foster these facilitators and address the barriers and consequently no content was changed; only app feature changes were made in adaptations in this trial.

In general, the main features, such as the social wall, daily quests, and collections of the P3 app were found to be acceptable among the study participants. It is likely that this is in part because of the ways in which the key app features address the factors identified by participants in this study as the most important PrEP adherence facilitators and barriers among YMSM in Thailand. For example, most participants expressed sexuality, PrEP, and HIV stigma as barriers to PrEP adherence and social wall of P3T provides an opportunity for positive social interaction and support related to these issues to take place, which has been shown in previous studies to be associated with reduced stigma [50,51].

However, the study participants provided critical recommendations for P3T app adaptations in the way the core features were presented that address cultural and contextual factors specifically relevant to Thai YMSM, which are expected to substantially enhance app acceptability, relevance, user engagement, and utility. For example, Thai YMSM voiced a strong preference for more visual content, such as infographics and video content, as opposed to text, similar to findings from formative work to develop P3 [34]. This may be because of the interest in consuming information via different mediums [35,36,52-54]. Therefore, these factors highlight the importance of understanding the reading ability and reading norms of the target population when developing app interventions [55]. The desire to keep the app fun and light through the use of stickers, a lighter color scheme, and ongoing seasonal promotions and games also featured prominently in both FGDs and KIIs. These suggestions reflect Thai culture and thus emphasize the importance of understanding cultural preferences and popular marketing practices when adapting mHealth interventions for use in different countries [56-58].

Both YMSM and key informants noted the strong influence peers have on the health beliefs, decision-making processes, and behaviors of Thai YMSM. This phenomenon has been identified in research in other cultures [59,60], and a study of Thai YMSM found that health beliefs and self-efficacy were associated with HIV prevention behaviors [52]—factors known to be influenced by peer groups [61]—were associated with HIV prevention behaviors [52]. App interventions focused on HIV prevention for Thai YMSM should assess whether functions that acknowledge the importance of peer relationships increase the likelihood of achieving the desired behavior. In P3T, peer interactions are possible via the social wall, and positive peer norms are presented in the videos added to the knowledge center that feature peer influencers. Future development of P3T will include the ability to visually represent the popularity of content among users via functions such as number of likes on comments and most popular topics, which could further capitalize on the influence of peers to promote behavior change.

There have been a small number of mobile apps that have tested YMSM at risk for HIV to support PrEP adherence and HIV prevention, mostly based in the United States. The availability of having reliable information on HIV and sexual health, reminder features for users to access relevant health care such as drug adherence and medical appointments, esthetically pleasing features for target users, and ease of use are common findings with what was found in this study from multiple studies previously conducted, including the LYNX and MyChoices apps [62-64]. Similar needs for medication reminders and education on their condition have been found in adolescents with other chronic health conditions, such as sickle cell disease [65]. No striking contrasts in YMSM needs using PrEP adherence apps in other studies compared with findings from those found in our study are apparent so far; however, many studies are underway that may produce differing outcomes to what has been seen so far [62,63,66]. Given the future potential scalability of the use of mobile phone apps in supporting PrEP adherence, understanding differences in feature preferences in different contexts, such as with age, gender identity, urban versus rural dwellers, and specific user profiles are valuable in adjusting mobile health technologies to maximize their benefits to specific populations [67]. Given the time limitations with data collection in adapting mobile technologies to user needs, future studies may benefit from sequential adaptive research trial designs to address this challenge.

We identified the benefits and challenges of adapting P3 for YMSM in Thailand. First, the remarkable similarities in PrEP adherence facilitators and barriers among YMSM in our study, and in other studies, along with the general acceptability of the key features of P3 among our study participants support the growing trend of adapting evidenced-based apps, such as P3, for different cultures and contexts rather than developing new apps from scratch [31]. However, extensive formative research is critical to ensure that an app is appropriately adapted to meet the unique needs and population preferences in different settings [67,68]. Adapting an existing app with a common behavior change goal can increase the speed of app development, reduce costs, and ultimately expedite scalability [31-33].

However, even minor app changes or the addition of a few new features can require significant time and budgetary resources. Planning for a budget that is sufficient for formative research and app adaptations may ensure that the adapted app is culturally and contextually relevant and may shorten the time between adaptation, pilot study, and, if feasible and acceptable, a larger trial. Owing to the limited time and funding of a pilot grant and working with a commercial app developer based in North America, we were unable to implement all design changes that would better align the app with cultural preferences identified in the FGDs and KIIs. These changes focused on stylistic preferences and the inclusion of additional functions within the existing features. The main changes suggested included making the app color scheme lighter, offering customizable skins, adding stickers to enhance communication in the social wall, adding an app search function, adding engagement statistics to visualize content popularity among peers (eg, number of times a knowledge center article or social wall post has been read), adding health behavior tracking features that would allow for determination of HIV risk level based on data entered into the app, and regular content updates to make the app more dynamic. An additional need identified in using an app such as P3T in Thailand is the ability of the more features of the app to function offline. According to the study participants, most Thai YMSM are unable to afford data plans that would allow for consistent and reliable internet access. To maximize access of users to P3T, more extensive offline functionality is critically important in the Thai context. If justified by the findings from the pilot study, the remaining adaptations, along with additional findings anticipated from the pilot study, are planned for a subsequent development phase that could also inform the adaptation of other apps.

Including local designers and developers on the team may help stretch limited resources and ensure that local expertise leads to the cultural adaptation process [68]. For example, through the support of a Thai collaborator skilled in graphic design, we were able to make important design changes to the app that reflected a rich understanding of the esthetic preferences of Thai YMSM. The local designer rapidly developed avatars that were highly responsive to study participant preferences. We also faced unforeseen formatting challenges translating the content of the app from English to Thai, which substantially increased the amount of time required for this stage. Having a fluent Thai speaker on the development team could help address these and similar barriers in the adaptation process. Local app designers and developers also provide a capacity-building opportunity for local collaborators, which can then reduce cost and time during both adaptation and scale-up phases. Given that, particularly in the context of limited resources such as where this study was performed, it must be acknowledged that there are limited data to support the cost-effectiveness of digital health interventions [69]. Such considerations are important when considering future scalability and sustainability to inform policies on investments in this area of health care interventions [70].

In the COVID-19 era, it must also be noted that with its considerable challenges, there are opportunities to optimize digital approaches as well as psychosocial intervention health care delivery in adolescents. Multiple areas regarding this with relevance to future studies related to the P3T app could be explored, including the impact of the app on preventive behaviors, medication adherence, self-management, and cost-effectiveness [71,72].

A major strength of this study was that it addressed a knowledge gap in the feasibility and acceptability of mHealth in low-middle income settings [29] and also in adolescent populations [28]. In addition, it provided practical guidance for adaptation of an existing mHealth platform in low-middle income settings rather than starting from scratch, a practical solution in the context of limited financial and staff resources as well as difficult-to-reach populations. It also highlighted the importance of using local expertise and capacity building in producing well-suited and sustainable interventions where mHealth interventions are adapted to very different settings for which they were originally intended. Finally, this study involved target user input from the start of its adaptation to the Thai context, an approach known to improve target user short- and long-term engagement [65,73-75].

This study was conducted only in Bangkok; therefore, data collected from study participants may not reflect the perspectives of YMSM and PrEP providers nationwide. In addition, the majority of FGD participants were enrolled in full-time education, which in the context of Thailand signifies a higher socioeconomic status. Therefore, these findings may not be generalizable to the Thai YMSM of all socioeconomic groups. As parental consent was required for FGD participants under the age of 18 years, we were unable to enroll participants in this age group because of the unwillingness of potential participants to reveal their sexual orientation or risk behaviors to their parents. This may have led to findings biased toward the views, needs, and developmental requirements of older adolescents. In most countries, formidable challenges remain in conducting research with those under the age of 18 years, despite the need to gather information to inform interventions that address the current global HIV epidemic, which disproportionately affects young people [5,76]. In Thailand, parental consent is exempted for individuals under 18 years of age to receive sexually transmitted infection care, HIV testing, or PrEP initiation. However, the decision to allow those under 18 years to participate in HIV prevention research without parental consent is at the discretion of the ethical review boards. On the basis of our experience, some review boards in Thailand are willing to allow exemption of parental consent for adolescents in HIV prevention trials where study involvement is a direct benefit to participants from this key population.

Despite these limitations, the adapted P3T app represents a significant step toward expansion and scale up of an app-based intervention to support PrEP adherence in Thailand, where PrEP has recently become available free of charge for YMSM through a national health coverage [5,77]. Additional studies in different geographical areas of Thailand with a high YMSM population density and studies focusing on young transgender women who have sex with men and other key populations would help further the adaptation processes of mHealth apps for those disproportionately affected by HIV in Thailand.

As observed in this study, most facilitators and barriers for PrEP adherence in the Thai context are similar to those observed in other global settings. The core features of the P3T app address the main facilitators and barriers related to PrEP adherence in Thailand for YMSM. The majority of the improvements suggested in this study focus on customizing the app to suit the stylistic presentation preferences of Thai YMSM and tailoring the app content to suit the cultural context of Thailand and the cultural nuances of the YMSM demography. Given that many of the facilitators and barriers are similar globally, adapting existing successful mHealth solutions will be the most effective and cost-efficient way forward as opposed to developing new apps. However, to obtain optimal results, it is vital that local app developers are included as partners in the adaptation process.