The rising rates of depression and anxiety among college students worldwide, further compounded by the effects of the COVID-19 pandemic, emphasize the critical need for effective interventions to address these mental health challenges. In South Korea, the proportion of people in their 20s and 30s at the risk of depression increased 6-folds, to 22.8%, in 2021 compared to 2018 [1]. Similarly, in the United States, anxiety and depression symptoms among college students rose by 75% from 2019 to 2021 [2]. However, this trend is not solely attributable to the pandemic; college students’ mental health has been consistently declining over the past decade [3].

Depression, a potentially fatal illness, is a leading cause of suicide, claiming approximately 800,000 lives worldwide each year [4]. South Korea’s suicide rate surpasses the Organization for Economic Cooperation and Development average, with a notable increase in suicide rates among patients with depression [5,6]. The seriousness of depression is underscored by its high mortality rate; complex etiology involving genetic, environmental, and psychological factors; and the persistent stigma surrounding mental health [4-6].

Social isolation and loneliness are significant risk factors for mortality [7] and are particularly prevalent among college students [8-10]. The college years are a period of heightened emotional vulnerability, with students being particularly susceptible to depression and loneliness [11,12]. Loneliness can lead to social withdrawal and sleep disturbances and exacerbate depressive symptoms [8,13]. Hawkley and Cacioppo [13] demonstrated that loneliness can indeed lead to depression, particularly in young adults. Their research suggests that even when controlling for factors such as stress and social support, loneliness remains a predictor of depression in college students [13]. The correlation between loneliness and depression is partially attributed to changes in neurotransmitters that regulate mood, such as serotonin and dopamine [14]. In addition, loneliness can trigger negative coping behaviors such as alcohol and drug use, further exacerbating feelings of depression [8].

Cacioppo and Hawkley [15] and Matthews et al [9] emphasized the importance of addressing loneliness and depression in young adults for their overall well-being and development. Untreated depression in this age group can lead to academic difficulties, interpersonal relationship issues, and increased risk of self-harm or suicide. Timely intervention is essential to prevent progression to more severe or chronic depression. Various intervention methods suggest that loneliness and depression in young adults can be effectively alleviated. Key to these interventions is social support through meaningful connections formed through family, friends, school, or community activities [10]. The college years are crucial in career development stages and characterized by increased independence, intimacy, and the dynamic exploration of life goals [11,16]. However, college students often face challenges such as stress while adapting to new environments, anxiety about academic and career paths, and difficulties in relationships, all of which can negatively impact their psychological and emotional well-being [11,12]. Despite these challenges, the college years also provide unprecedented opportunities for intellectual growth and development [17,18].

Digital mental health interventions have shown promise in overcoming the limitations of traditional mental health services [19]. Grounded in theories such as self-determination theory and cognitive behavioral therapy (CBT), digital interventions have demonstrated effectiveness in reducing symptoms of depression and anxiety [19,20]. Self-determination theory emphasizes the importance of autonomy, competence, and relatedness in engaging individuals and initiating sustained changes [21], while CBT posits that depression arises from negative thoughts and beliefs, which can be modified through cognitive restructuring and behavioral activation [22]. The anonymity and accessibility of digital interventions make them particularly appealing to digital-native college students [23].

This study explores the feasibility and acceptability of 2 digital therapy apps, Woebot (Woebot Health, Inc) and Happify (Twill Inc), among college students experiencing loneliness. Inspired by the research of Lim et al [24] on +Connect, this study has a larger sample size and an extended intervention period for a comprehensive evaluation. In addition, the research by Boucher et al [25] and Fitzpatrick et al [26] provided a foundation for selecting these specific digital interventions. The primary aim was to determine whether these digital interventions can meaningfully reduce feelings of loneliness and depression among college students in South Korea. The study also investigated the role of participants’ beliefs and behaviors in the effectiveness of the interventions. The hypotheses were as follows:

In addition, qualitative feedback will provide insights into the user experience, perceived effectiveness, and areas for improvement of these digital interventions.

This study aimed to explore the feasibility and acceptability of digital interventions in reducing loneliness and depression among university students. The study was conducted from March to June 2023, over a 4-month period, with participants from Sungkyunkwan University’s Colleges of Natural Sciences and Humanities and Social Sciences in Seoul, South Korea. The study used a mixed methods approach, incorporating both quantitative and qualitative data collection and analyses. The clinical trial was registered with the Clinical Research Information Service, a World Health Organization International Clinical Trials Registry Platform–linked registry, under the registration KCT0009449.

Given the importance of selecting effective digital interventions for addressing loneliness and depression among college students, Woebot and Happify were chosen for their demonstrated efficacy in prior research. Participants were provided with an instruction manual (Multimedia Appendix 1) to guide them through the use of the assigned intervention app. Woebot, a chatbot based on CBT principles, has shown effectiveness in reducing symptoms of depression and anxiety among young adults, as evidenced by studies such as the randomized controlled trial conducted by Fitzpatrick et al [26]. In addition, Happify has demonstrated promise in alleviating loneliness, particularly during challenging periods such as the COVID-19 pandemic, as reported by Boucher et al [25].

Moreover, both Woebot and Happify are readily downloadable in South Korea, providing convenient and accessible platforms for individuals seeking to improve their mental well-being. These chatbots were also selected based on the performance results of other mental health chatbots that have been verified for effectiveness and satisfaction in clinical experimental studies. Building on the promising findings of a recent study by Li et al [27], regarding their top-tier performance in terms of engagement (Woebot: mean 4.10, SD 0.14; Happify: mean 4.60, SD 0.00), functionality (Woebot: mean 4.65, SD 0.21; Happify: mean 4.25, SD 0.00), and overall quality (Woebot: mean 4.31, SD 0.09; Happify: mean 4.15, SD 0.02) on a 5-point scale, Woebot and Happify were selected as the interventions for this study. The features and instructions of use of Woebot and Happify are shown in Table 1.

Given the lack of Korean-language versions of Woebot and Happify, all interactions with the apps were conducted in English. This decision was made to ensure uniformity across all intervention groups, despite the potential challenge it posed to participants with lower English proficiency. No special cultural adaptations for Korean users were implemented, although compensation was provided in the form of a choice between equivalent-value coffee coupons and cash, with all participants opting for the cash incentive.

The target sample size was based on previous studies on digital mental health interventions. Initially, 63 undergraduate and graduate students aged 18 to 27 years were recruited through the university portal (see Multimedia Appendix 2 for the recruitment notice and survey questionnaires). Participants with very low loneliness scores (<34) were excluded. While the recruitment notice mentioned that participants should be confident in using English, as the Woebot and Happify apps do not have Korean versions, no applicants were excluded based on their English proficiency. To ensure fairness, all participants, including those in the control group using Bondee, were required to interact with the apps in English. Due to high attrition rates, the final sample consisted of 27 participants (Woebot: n=15, 55%; Happify: n=9, 33%; Bondee: n=3, 11%). To enhance retention, participants received a US $30 incentive for completing the study and an additional US $30 for completing the postintervention assessment [28].

Participants were assigned to 1 of the 3 groups (Woebot, Happify, or Bondee) using a stratified randomization method. The participants’ names were blinded, and they grouped such that there were similar distributions of age, sex, and loneliness scores across the groups. The decision to have fewer participants in the control group was made to ensure sufficient power in the intervention groups, given the limited overall sample size and anticipated high attrition rates in the control group. This decision was made based on the intention of the research to ensure the study focused on the effectiveness of the interventions.

The following questions were asked during the focus group interviews.

Textbox 1 provides the list of statistics considered and tests conducted during qualitative and quantitative analyses.

This study was approved by the institutional review board (IRB) of Sungkyunkwan University, Seoul, South Korea (IRB file: SKKU 2023-02-043; approval date: February 27, 2023; see Multimedia Appendix 6 for the IRB approval letter). All participants provided informed consent before their involvement in the study (see Multimedia Appendix 7 for the informed consent form). Data were deidentified to ensure confidentiality. Participants were compensated for their time and participation in the study. Compensation varied based on participation, with a maximum amount of up to KRW 200,000 (US $154). Participants received KRW 10,000 (US $7.70) per survey, KRW 20,000 (US $15.40) per interview, up to KRW 30,000 (US $23.10) for experimental participation, and an additional reward at the end of the study (amount to be determined). Compensation was calculated based on participation frequency and compliance, and payments were made at the end of each month after the interviews were completed. The details of this compensation were clearly outlined in the informed consent form provided to all participants.

This study followed the CONSORT (Consolidated Standards of Reporting Trials) reporting guidelines for randomized controlled trials. The CONSORT flow diagram, illustrating the flow of participants through the study, is presented in Figure 1. The CONSORT-EHEALTH (Consolidated Standards of Reporting Trials of Electronic and Mobile Health Applications and Online Telehealth) checklist (version 1.6) [33] was also completed and is available as Multimedia Appendix 8.

Table 3 presents the demographic and clinical variables of participants at baseline. The table includes measures of central tendency and dispersion, such as mean, SD, median, and IQR for both the UCLA Loneliness Scale and PHQ-9.

Among the initial 63 applicants, the average PHQ-9 score was 9.23 (SD 6.15), with 25 (40%) individuals classified as being at risk for depression (PHQ-9≥10). At the baseline survey, 23 (43%) out of 53 participants were categorized as being at risk for depression, which is considered a very high proportion according to the Ministry of Health and Welfare’s criteria [1]. These findings highlight the urgent need for interventions to address the mental health crisis among college students.

In addition, participants with a UCLA Loneliness Scale score of ≥65, indicating severe loneliness, and those with a PHQ-9 score of ≥20, indicating severe depression, were advised to seek further counseling.

Figure 1 displays the CONSORT flow diagram, illustrating the flow of participants through the study. Of the initial 63 respondents, 10 (16%) were excluded due to loneliness scores <34. The remaining 53 (84%) participants were randomly assigned to the intervention groups (Woebot: n=22, 42%; Happify: n=21, 40%) and the control group (Bondee: n=10, 19%). Due to attrition, the final sample consisted of 27 participants (Woebot: n=15, 55%; Happify: n=9, 33%; Bondee: n=3, 11%).

The qualitative findings from the open-ended survey questions and focus group interviews provided valuable insights into participants’ experiences with the digital interventions and their coping mechanisms for loneliness.

This study adds to the research on digital mental health interventions by evaluating the effectiveness of Woebot and Happify in reducing loneliness and depression among college students. Compared to the study by Fitzpatrick et al [26], our study had a longer intervention period and a postintervention follow-up for a more comprehensive evaluation. While the study by Lim et al [24] had a similar intervention period, our study had a larger sample size and more extensive analyses. Unlike Boucher et al [25], who focused on qualitative insights during a shorter period, our mixed methods approach that spanned a longer duration provided a better understanding of the interventions’ efficacy and user experiences. Our study uniquely compared Woebot and Happify, offering insights into their effectiveness and user preferences.

The quantitative data showed a moderate positive correlation between loneliness and depression (r=0.37; P<.001), highlighting their complexity. This suggests that while related, loneliness and depression are distinct, and other factors may also influence college students’ mental health outcomes (see the PHQ-9 and UCLA Loneliness Trend Analysis code in Multimedia Appendix 5).

The moderate correlation between loneliness and depression highlights their intricate, bidirectional relationship. Loneliness can lead to depressive symptoms due to a lack of social connection, while depression can increase loneliness through social withdrawal. This suggests that interventions targeting either issue may positively affect the other. However, loneliness and depression are distinct constructs that require different approaches: social skills and connections for loneliness and CBT and emotion regulation for depression.

In addition, the moderate correlation suggests that other factors, such as stress, coping mechanisms, and resilience, also play a role in mental health outcomes. This underscores the need for comprehensive, multifaceted interventions that address the diverse challenges faced by college students. Even within the small control group using Bondee, the slight decrease in loneliness and depression scores suggests that any form of structured interaction can have beneficial effects, emphasizing the importance of social engagement and connection.

Several factors might explain why the interventions did not yield significant effects. First, the lack of a Korean version of the apps may have reduced engagement, especially for those less proficient in English.

In addition, the demanding academic schedule and competitive environment, coupled with daily stressors, may have exacerbated depression and loneliness, reducing intervention effectiveness. This underscores the need for holistic digital mental health interventions considering these factors.

In South Korea, research on digital interventions for reducing loneliness and depression among college students is extremely rare. There are very few studies, and none involve structured intervention experiments. This highlights the importance of developing and validating Korean-language digital mental health interventions. Given the severe mental health crisis among Korean college students, immediate intervention is crucial.

South Korean college students highly valued conversational elements, positive reinforcement, educational features, and reward mechanisms in digital interventions. This study serves as a critical pilot for the future development of Korean-language support features as well as reward-based CBT chatbots integrating positive psychology and empathetic support. The outcomes of this research provide a foundational understanding that will guide the creation of more effective and culturally relevant digital mental health tools for South Korean college students.

The findings of this study have several important implications for future research and practice in the field of digital mental health interventions for college students (Textbox 3).

This study has several limitations that should be acknowledged. First, the small and unbalanced sample sizes across the study groups may have limited the statistical power and generalizability of the results. The notably smaller control group compared to the intervention groups could have influenced the nonsignificant findings. In addition, due to the limited sample size and the need to ensure sufficient power in the intervention groups, an uneven distribution of depression scores was observed at baseline (P=.78). Although the mean depression scores were within a similar range across the groups, this limitation should be considered when interpreting the results, as the groups may not be perfectly comparable in terms of initial depression severity. Future studies with larger and more balanced samples are needed to confirm the efficacy of digital interventions for college students.

Second, the reliance on self-report measures may have introduced response bias. Incorporating objective measures and diagnostic assessments could enhance the validity of future studies. Third, the use of English-language interventions may have influenced engagement and satisfaction among participants with lower English proficiency. Future research should explore the development and evaluation of interventions in participants’ native languages. Finally, the study’s duration may have been insufficient to capture long-term effects. Longer follow-up periods are needed to assess the sustainability of treatment effects and identify potential barriers to long-term engagement with digital mental health tools.

This mixed methods study provides valuable insights into the efficacy and user experiences of digital mental health interventions for college students. While the quantitative findings did not show significant effects, the qualitative data revealed the potential of well-designed digital interventions to positively impact various aspects of college students’ mental health and well-being. Participants’ experiences with gratitude journaling, meditation, and other app features suggest that consistent use of these tools could lead to meaningful improvements in emotional, cognitive, and physical health.

However, the study also underscores the importance of considering contextual factors and individual preferences when designing and implementing digital mental health interventions. Future research should focus on developing personalized, user-centered interventions that address the unique needs of college students and explore ways to integrate these tools with professional support systems. Moreover, a comprehensive approach to promoting mental health among college students is needed, which includes mental health education and the provision of holistic support systems. Further large-scale, longitudinal studies are necessary to confirm the efficacy and sustained impact of digital interventions on college students’ mental health outcomes.