This study was approved by the Institutional Review Board of the first author’s affiliation (4-2022-1192), and the protocol was registered with the Clinical Research Information Service (KCT0008014). All interventions and assessments were conducted through non–face-to-face contact, maintaining the privacy and confidentiality of the participants. Self-reported assessments were administered only after participants read and clicked on the informed consent via a web-based platform. Participants self-downloaded the mobile app from Google Play or the App Store and registered it on their mobile device. The app was available for both iOS and Android platforms, ensuring accessibility and minimizing potential selection bias in participation. Participants took part anonymously and could withdraw from the study at any time by simply choosing not to respond. To protect confidentiality, the app used a secure user authentication system that did not require personally identifiable information, using an anonymized ID and password. All self-recorded data, including diary entries and symptom tracking, were stored in a deidentified format. Users also had the option to delete their records at any time, giving them full control over their data. After completing the interventions, participants were compensated with gift coupons worth US$ 70 (KRW ₩100,000) for the app group and US$ 35 (KRW ₩50,000) for the link group.

This study was a 2-arm randomized clinical trial. We compared the link group, which received text messages of URL links for accessing information, with the app group, which used a mobile app for exploring information and self-recording health outcomes (Figure 1). The 2 interventional arms were implemented using participants’ smartphones. The contents of the information provided were the same in both groups; however, the delivery methods were different. The link group clicked on the URL address of the weekly received text messages, and the app group logged into the app to view the content posted at all times.

A priori sample size was estimated using G*power 3.1.9.2 software. Based on a previous study evaluating a similar mHealth intervention [16], a sample size of 60 participants was required to detect an effect size of 0.44, with a type 1 error rate of 5% and a power of 0.80. Considering an expected dropout rate of approximately 30%, the target sample size was set at 86 participants. Block randomization (block size of 4) was used to ensure 1:1 balanced enrollment. Research staff who generated the random numbers and block sequences (ie, AABB, ABAB, BAAB, BABA, BBAA, and ABBA) using Microsoft Excel and those who registered participants were different. However, the research staff could not be blinded after the arm assignment to be involved in data control. Participants were unable to blind the assignment in the app-based trial. The research staff and participants did not know each other before and had never met. All contacts from the research team were made via text messages or email, and a phone number was provided for participants to call 24/7 if needed. The period of the mHealth intervention was 4 weeks, with assessments at baseline (T0), 4 weeks postintervention (T1), and a follow-up at 8 weeks, after an additional 4-week voluntary use period (T2).

The participants were recruited between July 19 and October 18, 2023. The eligibility criteria were as follows: (1) people diagnosed with HIV, aged 19 years and older, (2) individuals who owned and used a smartphone, and (3) those who had no difficulty reading and responding to the questionnaire via smartphone. Participants were excluded if they had a medical diagnosis that required immediate treatment and counseling, other than HIV infection. Participants were recruited using an online access method. A prior study to identify the mHealth needs of people living with HIV was conducted at 5 hospitals, and about half of the participants expressed their willingness to participate in further mHealth interventions. A recruitment poster was emailed to 138 individuals who had left their email addresses and expressed interest in participating in the intervention study based on a previous service-needs assessment conducted in 2020‐2021. In addition, 340 small leaflets that fit into a pocket were distributed to visiting outpatients “people living with HIV” by medical staff at the 2 hospitals. The poster and leaflets specified the study purpose and participation method, and included a QR code that could access the web-based baseline survey. Participants who completed the baseline survey were randomly assigned to the app or link groups. The research staff sent the app group a text message explaining how to download the app from the App Store or Google Play. The app allowed participants to log in only after receiving approval from the research team. The link group received a text message mentioning that they would receive a weekly text message including a link to access the information.

These participants received text messages, including the link to access information, once a week for the 4-week intervention period. When participants click the URL address in the text message, the screen is linked to the educational material, a PDF file. This information is the same as in the poster-templated content of the ESSC app. The 6 domains were categorized into 4 sessions: HIV management, psychological and sexual health care, ART care, and health promotion and family and social life. When participants had questions about information or personal concerns, they could send text messages or emails to the research team.

Basic information variables at baseline are presented by the groups using descriptive analysis (counts with percentages or mean values with SDs). Changes in primary and secondary outcome variables between baseline and postintervention were analyzed using paired 2-sample t tests, which examined the differences in the mean values of outcomes by group during the intervention. The primary outcome variables were compared between the app and link groups using a generalized estimating equation (GEE). It is an appropriate statistical method that analyzes repeated measures data reflecting the interaction between group and time and performs intention-to-treat analysis. To compare the effectiveness of group differences across time, GEE analysis was performed, including the results of the 3 assessment points (T0, T1, and T2). Here, the GEE analysis was adjusted for 3 covariates: number of years after HIV diagnosis, psychological domain of the WHOQoL-HIV BREF, and SF-HLS. Previous studies have shown that the HIV self-management ability was affected by the duration of HIV diagnosis, psychological status, and literacy [33-36]. In addition, health literacy scores were significantly different between the 2 groups in the homogeneity test of basic information using the chi-square test or independent t test. Therefore, the research team acknowledged that these 3 covariates should be included in the evaluation of HIV-SE effectiveness. For all analyses, the threshold for statistical significance was a P value of .05. According to the statistical method, the GEE analysis included all data of the withdrawn participants, but the paired t test could not include withdrawn participants. Additionally, open-ended qualitative feedback collected during the usability test was analyzed using a descriptive content analysis approach. Two researchers independently reviewed the responses, grouped them into common themes, and categorized the feedback as either positive or negative from the user perspective. Key patterns were summarized narratively to complement the quantitative usability findings.

Figure 2 illustrates a flow diagram for trial recruitment. Overall, 87 individuals accessed online sites to participate. After responding to the baseline survey, 4 were excluded as they met the exclusion criteria: had a psychiatric diagnosis (ie, bipolar disorder or depression) with symptoms, and were under treatment. The remaining 83 people living with HIV were assigned to the 2 groups, of which 73 completed the 8-week follow-up. When required to self-report assessments after completing the intervention at 4 weeks, 12% (10/83) of participants did not respond to the research team contact.

The baseline characteristics are shown in Table 1. The mean age of the participants was 39.42 years (range: 21‐69 years), 96.4% (80/83) men, and the majority, 85.5% (71/83), were men who had sex with men, including homosexual and bisexual gender orientation. A total of 9 in 10 participants had never been married, and half lived alone. There was an average of 9.51 years after HIV diagnosis, and 6 in 7 participants had viral suppression and no HIV-related symptoms. Only 7 perceived themselves as having an AIDS status. The results of the homogeneity test showed that the SF-HLS scores (t(81)=2.667, P=.009) were significantly different between the 2 groups, while the other variables showed no significant differences.

The difference in mean values of each group of primary and secondary outcomes between baseline (T0) and postintervention (T1) is shown in Table 2. During the intervention period, the overall outcomes improved in the app group, except for medication adherence. The results of the paired t test showed that the total score and the managing fatigue domain of the HIV-SE significantly improved in the app group. The link group also showed significant improvements in 3 domains: managing depression or mood, managing fatigue, and getting support or help. However, there was no significant improvement in secondary outcomes in either group. At 8 weeks (T2), the voluntary use for 4 weeks after completing the intervention, the overall primary outcomes decreased from the measured score at T1 (Table 3), but managing depression or mood in the link group increased over 8 weeks.

GEE was used for group comparisons of differential changes from T0 at T1 and T2. The effects of the app group compared with the changes in the link group over time points, as a reference at T0 in self-efficacy for HIV self-management, are shown in Table 3. Regarding symptom management, the app group showed a significant improvement at T1, but this effect was not observed at T2. On the other hand, a significant reduction was observed at T2 in scores for managing depression or mood compared with the link group. There were no statistically significant findings for the other outcome variables in the GEE analysis.

Regarding the quantitative usability of the ESSC app during the 4-week intervention period, 42.9% (18/42) of app group participants completed the in-app logs fully, with medication adherence at 54.8% (23/42), mental diary at 54.8% (23/42), sexual diary at 64.3% (27/42), and treatment retention at 97.6% (41/42) (Table 4). Only 9.5% (4/42) maintained daily medication records from the completion of the intervention to the voluntary use period of 4‐8 weeks. The periods of app use per participant ID are presented in Figure 3. Individuals who recorded daily medication adherence tended to record weekly mental and sexual diaries.

The mean scale of Health-ITUES after the intervention period was 3.98 (SD 0.75) out of 5 points. The 4 subscale scores were as follows: impact (3.93, SD 0.95), perceived usefulness (3.86, SD 0.99), perceived ease of use (4.32, SD 0.71), and user control (3.95, SD 0.80). While it was easy for the participants to become skillful users of the ESSC app, the impact on a healthy life and the usefulness of supporting HIV self-management did not exceed 4 points. Qualitative feedback regarding the ESSC app or the link is shown in Textbox 1. Participants believed that the information was easy to understand and important; they wanted to regularly receive updated information, such as up-to-date treatment (eg, long-acting injectable HIV ART).

This study is the first trial to conduct fully non–face-to-face digital interventions for community-dwelling people living with HIV in Korea. These digital interventions have been designed to expand access to health information resources that are limited to hospital settings. Two-arm interventions were performed using a mobile app or link in text messages, which are popular implementation methods for mHealth. Comprehensive information that helps people living with HIV self-manage their HIV disease and a healthy life was provided by logging into the app or accessing links in the text messages. While the mobile link provided only information, the mobile app had functions to self-record and view their recorded health data. After 4 weeks, the total self-efficacy score for HIV management increased from the baseline in both groups, and a statistically significant improvement was found in the app group. Although the GEE analysis revealed no significant between-group difference in total self-efficacy score over 8 weeks, domain-specific comparisons indicated that the app group experienced a significant positive effect in managing symptoms at 4 weeks and a negative effect in managing depression or mood at 8 weeks. Our findings suggest that the information-driven mobile app helps improve self-efficacy for HIV disease management and that the function of self-recording health life has benefits for symptom management.

Informational support is the most frequently requested and provided form of assistance within online support groups [37]. In this study, 9 in 10 participants were single men, and half lived alone; they spent more than 4 hours a day using the app or accessing the internet via smartphone. Despite being well-positioned to engage in online support groups, only a quarter of participants had experienced participating in self-help groups. Previous studies on online-based relationships among people living with HIV have identified challenges such as difficulty maintaining relationships, concerns about HIV disclosure, and the spread of inaccurate information [38]. In line with a prior mHealth needs survey, it was found that people living with HIV desired evidence-based and trusted information first, and easy-to-understand content [11]. They reported that the need to communicate with other people living with HIV was low, but the need for HCP was high [11]. In this study, mHealth interventions through a mobile app or link provided validated information from HCP across 6 domains: HIV management, ART care, psychological health care, sexual health care, health promotion, and family and social life, which covered a holistic approach to healthy life. While both groups demonstrated improvements in total self-efficacy scores, no improvements were observed in the domains of managing medication or communication with HCP. Similarly, medication adherence did not improve, which aligns with previous studies showing that digital interventions often have limited impact on adherence among people living with HIV who are already taking ART consistently [39,40]. Most participants had maintained HIV treatment and viral suppression for years, with an average of 10 years since HIV diagnosis, reflecting long-term retention in care. While therapeutic communication with HCPs has been emphasized for individuals newly diagnosed with HIV [41], information provision alone may not enhance communication self-efficacy among those already familiar with their HCPs. Furthermore, although the informational content addressed diverse domains, such as sexual health care and social life, it did not improve self-efficacy in condom usage and health-related quality of life.

The app group showed a significant improvement in self-efficacy for managing symptoms at 4 weeks compared with the link group. According to self-recorded symptoms on the ESSC app, a quarter of participants experienced symptoms that lasted from a minimum of one day up to 19 days. As symptom experiences are adverse signs of HIV progression and strongly related to the quality of life of people living with HIV, previous studies have developed mobile apps aimed at symptom management [42,43]. Similar to our study, a 4-week intervention using an app with educational content, self-assessment, and health tracking was shown to enhance symptom management knowledge [43]. Unlike that study, which included weekly group sessions on symptoms, our intervention did not provide customized interactive sessions. Instead, participants engaged in self-reflection through the “My Page” feature, where they could review their own health data. Interestingly, in another intervention study, web-based interactive sessions for patients did not significantly improve symptom management [44]. The authors explained that skill-learning processes may not effectively enhance self-efficacy in patients with high treatment adherence. Despite the absence of structured interactions, self-efficacy for managing symptoms improved, suggesting that self-recording alone may contribute to symptom self-management. Meanwhile, recording the mental diary of the ESSC app had no significant effect on managing depression or mood compared with the mobile link group. Rather, improvement was seen at 8 weeks in the link group that provided only psychological health care information. However, it is difficult to determine whether the mental diary had an adverse effect on mood management because the number of records was low between 4 and 8 weeks. Other apps specializing in symptom management [42,43], including psychological coping strategies for depression and anxiety severity, have been shown to help reduce psychological distress. User-customized psychiatric content beyond a mental diary is suggested for managing depressed moods.

Only one in 10 participants maintained their daily visits after completion of the intervention, during the voluntary use period between 4 and 8 weeks. In other app interventions requiring a query response for medication adherence, the response rate was 47.7% over 2 years [45]. The app provides weekly summary reports, and its most popular feature is the community message board, which allows users to share and interact with other users on social networks. The ESSC app has features for viewing summary reports at any time on “My Page,” but there is no community board for user interaction. This reflected the low need for peer communication among people living with HIV in the mHealth needs survey [11]. However, participants in this study gave feedback that they lacked motivation to continue using the app and expressed a desire for a community feature. By sharing experiences and common issues, people living with HIV can tap opportunities to perceive the usefulness of support groups and reduce stigma [46]. The community board should be operated in a safe space so that users do not encounter interpersonal conflicts [37], which will be refined in the next version of the ESSC app improvement. In addition, to maintain engagement with the app, participants responded that information should be updated regularly and interesting content should be provided to people living with HIV, such as a case of cure and the latest treatment development. HIV treatments have continued to develop in the evolution of ARTs, including injectable medication, and this progress has extended life expectancy and improved the quality of life [47]. Accordingly, it is natural that people living with HIV desire to stay up-to-date on the treatment, and the ESSC app should incorporate a dedicated news section with the latest information.

The link group, which accessed digital information via individual text messages without signing up for the app, showed significantly improved self-efficacy in managing depression or mood, fatigue, and getting support or help. While the mHealth app is an emerging technology for empowering patients’ self-management of chronic illnesses, it must also address potential risks to confidentiality and ensure patient autonomy [48]. Some users in this study perceived the ESSC app as an unnecessary device for good patients. The study by Jacomet et al [10] of people living with HIV and their physicians’ use patterns, benefits, and perceived barriers to different aspects of eHealth reported that 26% of people living with HIV and 43% of physicians were enthusiastic about eHealth, whereas 31% and 37%, respectively, expressed skepticism. To adopt mHealth for patients, multiple factors should be considered, including user characteristics, literacy, system utility, technology infrastructure, collaboration, and regulatory settings [49]. Diversifying the range of mHealth services available to patients can reduce barriers to accessing digital health resources and contribute to health equity. Link-based access can be proposed as an alternative mHealth strategy to bridge the gap between people living with HIV who are hesitant to use apps and those who use app in the context of digital information technology.

There are several limitations in interpreting the results. First, the inclusion criteria for smartphone users may not be representative of the generalized findings for people living with HIV. To ensure representativeness and health equity, further studies are needed to enhance community-based enrollment without limiting smartphone literacy. Second, although the final sample size exceeded the minimum required (n=60) based on a priori power analysis, it did not reach the target of 86 participants. While the final sample size was sufficient to detect moderate effect sizes, the reduced sample size may have limited the ability to identify smaller differences between groups. This limitation should be considered when interpreting the findings. Third, the 4-week intervention period may have been too short to fully assess the effectiveness and usability of the app. To observe sustained engagement patterns, an additional 4-week voluntary-use period was provided. However, voluntary use during the follow-up period was notably low. We designed a relatively short intervention period, as the primary aim of this study was to evaluate the impact of 2 types of fully non-face-to-face, information-driven mHealth delivery formats. To enhance usability and engagement for longer-term interventions, future versions of the app should incorporate additional features, such as user-interactive community boards, that go beyond information provision and self-recorded healthy living. To determine the practical usefulness of the ESSC app, it is necessary to evaluate its long-term effectiveness following future app improvements.

The provision of integrated information using mobile devices improves HIV-SE among people living with HIV. Both mobile apps and link accessing in text messages are useful methods for providing non-face-to-face information. People living with HIV who have already been retained in care may not prefer to become online members of HIV-related apps, and mobile-based interventions should be applied according to individual preferences. While an informational mobile app that allows users to self-record and reflect on their healthy life improved self-efficacy for managing symptoms compared with link accessing in text messages, the app has challenges in maintaining long-term use. Usability findings suggest that up-to-date news sections and user-to-user community features are required for long-term engagement with informational mobile apps.