This study was a single-blinded randomized controlled trial (RCT) with a pre-post assessment design (trial registration: KCT0008887; Clinical Research Information Service [CRIS]). Participants were recruited via online and poster advertisements, and from the outpatient department of Catholic Kwandong University International St. Mary’s Hospital, from October 2023 to December 2023. Eligible participants provided informed consent and underwent a screening process, leading to the exclusion of those who did not meet the inclusion criteria. To maintain single-blind conditions, participants were blinded to their assigned study condition (intervention group [IG] or CG).

Participants were randomly assigned to either the IG or CG prior to completing the initial online assessment. Randomization was performed using a list of random numbers generated via 4-block randomization in R (version 4.0.2). This list, securely stored as an encrypted file to ensure allocation concealment, was managed by a researcher independent of participants’ recruitment and assessment, who assigned participants to either the IG or CG based on their registration order. After the assignment, participants completed an online survey collecting sociodemographic, medical, pregnancy-related, and psychiatric information, as well as baseline measures for primary and secondary outcomes.

Upon completion of the baseline assessment, both the IG and CG commenced their respective 4-week interventions. At the conclusion of this period, participants were notified of study completion and requested to complete an online survey for postintervention assessment. The postintervention assessment was designed to mirror the baseline assessment.

The sample size was determined using G*Power software, version 3.1.9.4 (Franz Faul, University of Kiel, Germany). Based on power analysis aiming to detect differences between 2 groups over time, the calculation assumed a large effect size and an α level of .05, with a statistical power of 95%. This yielded a required sample size of 70 participants (35 per group). To account for potential attrition, which has been reported to reach up to 25% in a previous study [28], the recruitment target was set at 88 participants at baseline.

Potential participants were screened according to the following inclusion criteria: (1) pregnant women aged 19‐65 years; (2) experiencing mild to moderate depression, as indicated by a score of 10‐20 on the depression subscale of the Depression, Anxiety, and Stress Scale-21 (DASS-21) [32,33] or a score of 9‐11 on the Edinburgh Postnatal Depression Scale (EPDS) [34,35]; (3) possession of a smartphone compatible with the required mobile app; (4) meeting the minimum operating system requirements for the mobile app; and (5) ability to actively use the mobile app and provide feedback regarding its functionality.

The exclusion criteria were as follows: (1) gestational age exceeding 34 weeks at the time of enrollment or expected delivery within 6 weeks of enrollment; (2) diagnosis of severe psychiatric disorders, including bipolar disorder, major depression disorder with psychotic features, posttraumatic stress disorder, and obsessive-compulsive disorder; (3) presence of other severe medical or psychiatric conditions that would impede study participation or safety; and (4) inability to download a mobile app.

To investigate whether a mobile MBI app is more effective than self-monitoring alone, the IG used both the Avecmom app and the Big 4+ app for 4 weeks, whereas the CG used only the Big 4+ app for the same duration.

The Avecmom app (Digital Medic Co, Ltd), developed by Cho et al [36], is a freely accessible self-help mobile MBI app used exclusively in the IG. This intervention is designed to support the self-management of depressive symptoms among pregnant women. The Avecmom app aims to enhance mindful awareness and compassion to help women manage perinatal stress, depression, and anxiety. It comprises 4 modules involving mindfulness training, structured for completion within 4 weeks. The 4 core mindfulness trainings—breathing meditation, body scan, emotion-awareness meditation, and loving-kindness meditation—unlocked successively. Each subsequent module became accessible after participants completed the primary training within the preceding module two or more times. Participants were instructed to practice mindfulness at least twice a week using the Avecmom app. To support adherence, they were encouraged to set personalized in-app meditation reminders, and daily automated notifications were sent on days when nonengagement was detected to promote continued engagement.

The mindfulness training content offered by the Avecmom app was specifically designed to alleviate discomfort among perinatal women. It introduces basic mindfulness training focusing on breathing and bodily sensations in modules 1‐2 and then progresses to mindfulness training integrated with techniques of emotional and cognitive behavioral therapy in modules 3‐4. To address pregnancy-specific discomforts, such as concerns about fear of miscarriage or increased physical strain [37], the training also included content aimed at fostering maternal-fetal attachment and managing pregnancy-related bodily sensations. This content was preliminarily reviewed by perinatal women to ensure it was appropriately tailored to their needs. The detailed content and usage process of the Avecmom app are described in the report by Park et al [28], and the app is freely available to the public on both iOS and Android platforms.

The Big 4+ app (Digital Medic Co, Ltd) is a self-monitoring app that includes single-item assessments and a diary function. It enables users to monitor various psychological states, such as mood, sleep duration, appetite, and other relevant psychological parameters, providing users with easy and personalized access to their self-reported data. Users of the Big 4+ app were encouraged to record their status once daily using the single-item measures provided within the app. Upon completing the measurement, users received self-feedback through data visualizations presented in graphs. The Big 4+ app is accessible on both iOS and Android platforms, and all participants in both the IG and CG were asked to use the Big 4+ app.

For intervention delivery, the IG participants were provided with URLs for downloading both the Avecmom app and Big 4+ app, along with manuals on how to install and use them. Participants’ understanding of app usage, initially guided by the manuals, was verified by a responsible researcher via phone or in-app messenger. Throughout the intervention period for the IG, a separate researcher continuously monitored participants’ engagement rates via the Avecmom app’s administrator page. In-app notifications were sent to encourage participants who did not meet the predefined engagement standard (eg, practicing at least twice a week). Similarly, the CG received a URL to download the Big 4+ app with its respective manuals, and their understanding of the app’s use was verified by a researcher via phone or in-app messenger.

All statistical analyses were conducted using IBM SPSS Statistics for Windows, version 25 (IBM Corp), with the significance level set at an α level of .05 (2-tailed). Initially, independent t tests and chi-square tests were performed to compare baseline sociodemographic and clinical variables between the IG and CG. Analysis of covariance (ANCOVA) was then employed to analyze the pre-post data for primary and secondary outcomes. In these analyses, postintervention scores served as the dependent variable, with corresponding baseline scores included as covariates. This approach was chosen because it provides unbiased intervention effect estimates and enhances estimation precision [51]. Specifically, ANCOVA was conducted to examine group differences in scores of the EPDS, CESD-R, DASS-21, CAMS-R, PANAS, MFAS, and MHC-SF.

To further explore relationships and underlying mechanisms of change, change scores for all clinical variables were calculated by subtracting baseline scores from postintervention scores. We then examined correlations among these change scores across all participants. Additionally, correlations were investigated between the clinical change scores and the recorded duration of audio-guided mindfulness training within the Avecmom app, specifically for participants in the IG. Subsequently, serial mediation analysis was performed using PROCESS MACRO by Hayes, Model 6 (version 4.2) [52]. This analysis examined the serial mediating effects of changes in CAMS-R and CESD-R scores on the relationship between Avecmom app use and changes in MHC-SF scores. The mediation model was tested using 95% CIs generated from 5000 bootstrap samples.

This study was conducted in accordance with the principles of the Declaration of Helsinki. Prior to commencing the study, all participants provided informed consent. All study protocols were approved by the Institutional Review Board of Catholic Kwandong University International St. Mary’s Hospital (IRB number: IS23EIMI0054). To ensure participant privacy and data security, all collected data were deidentified prior to analysis and stored in password-encrypted files, accessible only to authorized research staff. Participants received compensation of KRW 100,000 (US $71.35) to cover transportation expenses.

As depicted in Figure 1, a total of 158 participants were screened for eligibility. Of these, 68 were excluded based on the established criteria, and 90 were subsequently randomized to either the IG or CG. During the postintervention assessment period, 4 participants were excluded from the study (3 from the IG and 1 from the CG) due to reasons such as loss of contact, insufficient adherence to the intervention protocol (defined as a Big 4+ app response rate below 70%), and data collection errors. Finally, 86 participants completed the study and were included in the final analysis. The CONSORT-EHEALTH checklist is provided in Checklist 1.

Participants in the IG (n=42) had a mean age of 34.10 (SD 3.04) years, while those in the CG (n=44) had a mean age of 33.80 (SD 3.42) years. There was no significant difference in age between the 2 groups (t₈₄=0.43; P=.67). Similarly, at baseline, the mean gestational period was 25.43 (SD 6.77) weeks for the IG and 23.40 (SD 6.88) weeks for the CG, with no significant difference observed (t₈₄=1.37; P=.17). Table 1 summarizes the group differences in other sociodemographic variables. Consistent with the age and gestational period findings, no significant differences were observed across any of these variables, including education level, pregnancy planning, smoking habits, alcohol use, history of depression, and other relevant medical problems.

The homogeneity test revealed no significant differences in the clinical variables between the IG and CG at baseline. Descriptive statistics and t values are presented in Table 2.

An ANCOVA was conducted to compare postintervention scores between the groups, adjusting for baseline scores. The detailed ANCOVA results are presented in Table 3, and the significant findings for each variable in the IG and CG are visually represented in Figures 2 and 3, respectively.

For primary outcomes, the IG demonstrated significantly lower depression severity (CESD-R; F_1,84=5.69; P=.02; partial ƞ²=0.06) and anxiety (DASS-21 anxiety subscale; F_1,84=5.28; P=.02; partial ƞ²=0.06) compared to the CG. Conversely, no significant differences were observed between the groups for depressive symptoms (EPDS and DASS-21 depression subscale scores), stress (DASS-21 stress subscale scores), or total DASS-21 scores.

Regarding secondary outcomes, the IG exhibited significantly higher levels of mindfulness (CAMS-R; F_1,84=11.18; P=.001; partial ƞ²=0.12), mental well-being (MHC-SF; F_1,84=8.79; P=.004; partial ƞ²=0.10), and maternal-fetal attachment (MFAS; F_1,84=5.54; P=.02; partial ƞ²=0.06) compared to the CG. Positive affect scores were significantly higher in the IG than in the CG (PANAS-PA; F_1,84=7.21; P=.009; partial ƞ²=0.08), while negative affect (PANAS-NA) scores did not show significant differences between the groups.

Participants in the IG were instructed to engage with the Avecmom app at least twice weekly to progress to subsequent sessions. The app was used, on average, 12.60 times (SD 2.19), with individual usage ranging from 10 to 23 times. The mean total duration of app use was approximately 110.57 (SD 18.08) minutes. Module-specific mean usage durations (in minutes) were as follows: module 1, 23.66 (SD 5.66); module 2, 33.65 (SD 5.49); module 3, 27.93 (SD 4.48); module 4, 25.32 (SD 6.65). The temporal distribution of app usage indicated that most sessions occurred in the morning (6:00 AM to 12:00 PM; 231/552 sessions, 41.9%), followed by the afternoon (12:00 PM to 6:00 PM; 157/552 sessions, 28.4%), evening (6:00 PM to 12:00 AM; 130/552 sessions, 23.6%), and late night (12:00 AM to 6:00 AM; 34/552 sessions, 6.2%).

Correlation analyses were conducted between the duration of engagement with each Avecmom app module and the clinical variables, with details presented in Table 4. No significant correlations were found between clinical variables and the durations of modules 1, 2, and 3. Significant correlations were observed between the duration of module 4 (which primarily features loving-kindness meditation) and negative affect (r=−0.39; P=.01) and mindfulness (r=0.33; P=.04). Additionally, a positive correlation trend was noted between the duration of module 4 and both mental well-being (r=0.30; P=.05) and maternal-fetal attachment (r=0.30; P=.05), though these correlations did not reach statistical significance.

Serial mediation analysis was conducted using PROCESS Macro by Hayes, Model 6 (version 4.2), to investigate whether changes in mindfulness and depression severity serially mediated the relationship between the intervention (the combined use of the Avecmom app and Big 4+ app) and changes in mental well-being. As depicted in Figure 4, the mediation model revealed several significant paths. The intervention led to significantly greater changes in mindfulness compared to the control condition (b=3.48; P=.002). Increased mindfulness was significantly associated with a reduction in depression severity (b=−0.82; P<.001). This enhanced mindfulness also directly influenced changes in mental well-being, showing a positive effect (b=0.78; P=.003). Changes in depression severity were negatively associated with changes in mental well-being (b=−0.28; P=.03).

Table 5 presents the specific indirect effects. The results revealed a significant serial indirect effect of the intervention on the changes in mental well-being through sequential changes in mindfulness and depression severity (CESD-R; standardized β=.07; Boot SE=0.05, 95% CI 0.003-0.19). Additionally, a significant indirect effect was found for the path from the intervention through changes in mindfulness alone to changes in mental well-being (standardized β=.23; Boot SE=0.11, 95% CI 0.06-0.49). In summary, Avecmom app use was found to improve mental well-being in perinatal women. This improvement appears to be primarily driven by the enhancement of mindfulness, which, in turn, contributes to a reduction in depression and directly fosters mental well-being.

This study evaluated the effectiveness of the Avecmom app, a mobile MBI app specifically designed for perinatal women, in reducing depression severity among pregnant women with mild to moderate depression. The findings indicated that the use of the Avecmom app significantly reduced depression severity and anxiety levels compared to self-monitoring alone. Furthermore, the use of the mobile MBI app enhanced mindfulness, overall mental well-being, maternal-fetal attachment, and positive affect compared to self-monitoring alone. Notably, module 4, which focused on loving-kindness meditation, showed potential for greater improvements in mindfulness, negative affect, maternal-fetal attachment, and overall mental health compared to other modules, such as breathing meditation, body scan, and emotion-awareness meditation. Serial mediation analysis showed that Avecmom app use enhanced mindfulness, which, in turn, reduced depression severity and ultimately improved overall mental well-being.

The current findings demonstrate the effectiveness of the Avecmom app in reducing depression severity and anxiety among perinatal women with mild to moderate depression. Despite its relatively brief protocol, with weekly engagement averaging less than 30 minutes, the mobile MBI app produced meaningful changes in both depression severity and anxiety, comparable to other MBIs with more extended protocols [53,54]. These results are particularly noteworthy given the rigorous RCT design and the inherent benefits of self-monitoring. Self-monitoring, a common active component in various digital mental health interventions, is known to enhance self-awareness, thereby potentially contributing to improved outcomes. Our study distinctively showed that the combined approach of self-monitoring and mindfulness training yielded significantly greater benefits than self-monitoring alone.

According to the Monitor and Acceptance Theory proposed by Lindsay and Creswell [55], mindfulness inherently involves both monitoring and acceptance components. In the context of our study, the CG effectively practiced monitoring alone through the Big 4+ app, while the IG engaged in both monitoring and acceptance through the combined use of the Avecmom and Big 4+ apps. This distinction likely contributed to the observed differences between the groups. Supporting this, a previous study found that engaging in both monitoring and acceptance improved positive affect more substantially than practicing either component in isolation [56]. Consistent with this, our current findings also revealed a more positive affect in the IG compared to the CG, further suggesting the effectiveness of mindfulness as an intervention strategy for enhancing positive affect.

An improvement in maternal-fetal attachment was also observed in this study. Maternal-fetal attachment, defined as the emotional bond between a mother and her fetus [57], has been consistently reported to be negatively associated with perinatal depression symptoms and maternal anxiety [58,59]. Furthermore, strong maternal-fetal attachment is known to positively influence the social and emotional development of infants after birth [60]. These findings suggest that the Avecmom app may exert a long-term beneficial effect by influencing not only users’ depression severity but also other critically related developmental factors. Therefore, future studies should be designed to evaluate the long-term effectiveness of the Avecmom app and to determine whether additional booster sessions are necessary to sustain its beneficial effects.

Among the various sessions of mindfulness training offered within the Avecmom app, correlation analysis revealed that engagement with module 4 (loving-kindness meditation) had a positive correlational trend with maternal-fetal attachment. Loving-kindness meditation aims to cultivate unconditional love and kindness toward oneself and others [61], which may have contributed to enhancing maternal-fetal attachment in this population. While previous studies have reported an association between loving-kindness meditation and enhanced mental health outcomes, such as positive emotions, its specific effectiveness in perinatal depression has not been extensively explored [62,63]. Furthermore, loving-kindness meditation is not yet frequently featured in mobile MBI apps designed for maternity care. The current findings thus suggest that integrating loving-kindness meditation into perinatal care may be particularly beneficial for mothers experiencing depression and, by extension, for their children.

Another notable finding of this study was that the overall total time spent on mindfulness training was not significantly associated with changes in clinical variables. This finding aligns with a meta-review suggesting that increased hours of mindfulness training might not necessarily yield proportionally better outcomes [64]. The recommended weekly training duration within the Avecmom app is relatively short, at approximately 30 minutes. Nonetheless, most participants exhibited improvements in psychiatric outcomes, even when adhering to the minimum standard of training twice a week. However, it is premature to definitely conclude that there is no relationship between the extent of mindfulness training and the improvement in psychological outcomes, as some studies have reported that a longer mindfulness training duration may be associated with greater benefits [65,66]. In particular, long-term meditators have been shown to have different brain connectivity and amygdala activation levels in response to emotional stimuli compared to those engaging in mindfulness training for the first time [67,68]. Therefore, further research is needed to examine the effects of increasing the frequency and duration of practice.

This study further explored how the Avecmom app could enhance the overall mental well-being of pregnant women with depression, focusing on the sequential and direct roles of changes in mindfulness and depression severity. The serial mediation analysis demonstrated that Avecmom app use contributed to improved overall mental well-being. Crucially, our findings from this model illuminated how enhanced mindfulness, a primary outcome of the intervention, appears to impact mental well-being not through a single, isolated path, but rather through a sophisticated interplay of effects. Specifically, increased mindfulness was observed to directly foster improvements in mental well-being. Simultaneously, it also sequentially contributed to a reduction in depression severity, which, in turn, further mediated the positive change in mental well-being. This finding aligns with broader evidence suggesting that mobile MBI apps may positively influence not only overall mental well-being but also a range of clinical variables that may impact it in diverse populations [69-71]. Consistent with these findings, previous discussions on the mechanisms of traditional MBIs have suggested that mindfulness training enhances mental well-being by increasing mindfulness and reducing reactivity to emotional stimuli [13]. By integrating these insights with our results, it appears that mobile MBI apps operate through mechanisms similar to those of traditional face-to-face MBIs.

Lastly, the dropout rate in this study was 4%, which is markedly lower than the 30% dropout rate reported in a previous study that used the same Avecmom app and a similar design [28]. One possible explanation for this discrepancy is the difference in participant characteristics; while the earlier study targeted the general perinatal population, this study specifically recruited participants experiencing mild to moderate depression. Compared to other mobile mindfulness interventions, the low dropout rate in this study may also be attributed to the relatively short intervention duration and the app’s high perceived user experience. The Avecmom app features 4 mindfulness modules designed to be completed within a 4-week period, with each module unlocking after participants complete its primary training session two or more times. Participants were instructed to complete at least two training sessions per week. Compared to other mobile MBIs, this structured yet flexible design likely reduced participant burden and improved engagement, which consequently led to a positive influence on user satisfaction ratings. In a prior development study of the Avecmom app, participants rated its usability using the USE Questionnaire [28]. On a 7-point Likert scale, the average scores were as follows: usefulness, 5.4 (SD 0.79); satisfaction, 5.6 (SD 0.79); ease of learning, 6.4 (SD 0.67); and ease of use, 5.9 (SD 0.68), consistently indicating above-average evaluations across all domains.

In summary, this study addressed a significant research gap by evaluating the effectiveness of a 4-week mobile MBI app tailored for pregnant women experiencing mild to moderate depression, a population often overlooked in digital health research. Unlike many previous studies that predominantly relied on commercial apps, such as Headspace and Calm, or focused on general and nonclinical populations [23,24,72-75], the Avecmom app was uniquely designed for pregnant women with depression and demonstrated significant effectiveness, including a reduction in depression severity and anxiety symptoms despite the short duration. Notably, a previous review indicated limited effects of digital interventions on anxiety [76]. The Avecmom app showed promise as an effective and accessible option for perinatal women, particularly given their limited medication options. Its digital nature also enables automated data sharing with health care professionals and integration with electronic health records, which can support coordinated care and continuous monitoring, thereby suggesting its potential utility within stepped care models for perinatal depression. To achieve broader clinical implementation, however, future research would importantly need to evaluate its effectiveness when used concurrently with pharmacotherapy and further explore its potential to serve as either a first-line self-help intervention or an adjunct to established treatments, ultimately facilitating its systematic integration into health care infrastructure.

A primary limitation of this study is the absence of an untreated CG, which precludes definitively distinguishing between changes attributable to the intervention and those occurring naturally over time, especially given the dynamic psychological landscape during pregnancy [77]. Thus, there is a need for future research incorporating designs that enable such crucial distinctions. Another limitation lies in the integrated nature of the active components, namely self-monitoring and loving-kindness meditation. However, both are recognized for their independent benefits [78-82]. Furthermore, as an initial investigation of a newly developed intervention, the applicability of the findings of this study in broader clinical settings is limited by the lack of evidence regarding the long-term effects and the effectiveness in individuals with severe depression. Therefore, future research is crucial to address these limitations for the Avecmom app to achieve broader clinical implementation.

The findings of this study demonstrate the effectiveness of a mobile MBI app in enhancing the mental health of pregnant women with mild to moderate depression, compared to self-monitoring alone. Specifically, the intervention led to enhanced mindfulness that was found to influence mental well-being through multiple pathways, including directly fostering mental well-being and sequentially contributing to a reduction in depression severity, which then improved mental well-being. Furthermore, its relatively short protocol and high accessibility contributed to higher adherence. Mobile MBI apps offer a promising avenue to overcome barriers to accessing traditional psychotherapeutic treatments, presenting an effective and accessible option for perinatal women experiencing depression.