Introduction
Perinatal mental health issues represent urgent global concerns (Howard, Piot, & Stein, Reference Howard, Piot and Stein2014) from both maternal and fetal perspectives. Prenatal depression increases the risks of preterm labor, low birth weight (Howard & Khalifeh, Reference Howard and Khalifeh2020; Jarde et al., Reference Jarde, Morais, Kingston, Giallo, MacQueen, Giglia, Beyene, Wang and McDonald2016), and suicidal behaviors (Gavin et al., Reference Gavin, Tabb, Melville, Guo and Katon2011; Yu et al., Reference Yu, Shen, Bränn, Yang, Oberg, Valdimarsdóttir and Lu2024), and it is also a predictor of postpartum depression (Robertson, Grace, Wallington, & Stewart, Reference Robertson, Grace, Wallington and Stewart2004), which significantly affects maternal mental health (Slomian et al., Reference Slomian, Honvo, Emonts, Reginster and Bruyère2019). Therefore, preventive psychological interventions are recommended to prevent perinatal depression according to international guidelines (Motrico et al., Reference Motrico, Moreno-Peral, Uriko, Hancheva, Brekalo, Ajaz, Apter, Bramante, Conejo-Cerón, Christoforou, Dikmen-Yildiz, Evagorou, Fonseca, Lupattelli, Radoš, Al Maach, Rodriguez-Muñoz, Žutić and Lambregtse-Van Den Berg2022; O’Connor et al., Reference O’Connor, Senger, Henninger, Coppola and Gaynes2019; US Preventive Services Task Force et al., Reference Curry, Krist, Owens, Barry, Caughey, Davidson, Doubeni, Epling, Grossman, Kemper, Kubik, Landefeld, Mangione, Silverstein, Simon, Tseng and Wong2019). Systematic reviews and meta-analyses have demonstrated that interpersonal psychotherapy (IPT) (Sockol, Reference Sockol2018) and cognitive–behavioral therapy (CBT) (Sockol, Reference Sockol2015) are effective for treating and preventing perinatal depression; however, access to these in-person interventions remains limited because of time constraints, provider shortages and costs (Goodman, Reference Goodman2009; Kim et al., Reference Kim, La Porte, Corcoran, Magasi, Batza and Silver2010). Interactive interventions, delivered via a smartphone, can offer accessible and scalable support, potentially helping maintain mental well-being during stressful conditions, even among individuals without clinically manifesting depressive symptoms (Furukawa et al., Reference Furukawa, Tajika, Toyomoto, Sakata, Luo, Horikoshi and Noma2025; Toyomoto et al., Reference Toyomoto, Tajika, Luo, Sakata, Akechi, Nakayama, Kondo, Fukuma, Horikoshi, Wason, Noma and Furukawa2025). Such low-intensity measures can thus provide scalable public health approaches to support mental well-being among women with nonclinical perinatal symptoms. However, evidence regarding the effectiveness of smartphone-delivered and digitally based interventions remains mixed (Miura et al., Reference Miura, Ogawa, Shibata, Kamijo, Joko and Aoki2023; Tsai et al., Reference Tsai, Kiss, Nadeem, Sidhom, Owais, Faltyn and Lieshout2022), and whether their potential benefits extend from pregnancy into the postpartum period remains unclear. IPT targets psychosocial stressors relevant to the perinatal period, such as the transition to parenthood and interpersonal conflicts within close relationships (e.g. partners), making it suited to address common interpersonal challenges during the perinatal period (Spinelli, Reference Spinelli2017; Weissman, Markowitz, & Klerman, Reference Weissman, Markowitz and Klerman2017). Previous studies suggested the feasibility and acceptability of guided Internet-based IPT modules and a smartphone application incorporating CBT and IPT components in pregnant and postpartum women with nonclinical symptoms from communities (Bright et al., Reference Bright, Stuart, Mcneil, Murray and Kingston2022; Suharwardy et al., Reference Suharwardy, Ramachandran, Leonard, Gunaseelan, Lyell, Darcy, Robinson and Judy2023). However, evidence regarding smartphone-based IPT-informed interventions specifically designed for pregnant women remains limited. To address this gap, we developed a smartphone intervention informed by IPT targeting the period from pregnancy through the first postpartum year for people with nonclinical manifestations, initially focusing on primiparous women with a partner. This low-intensity intervention was intended to maintain and promote mental well-being rather than treat clinical depression. We hypothesized that the intervention would achieve high acceptability and efficacy. Efficacy was assessed by changes in depressive symptoms. We therefore conducted a randomized controlled trial (RCT) to assess the ability of the intervention to support mental health in pregnant and postpartum women.
Methods
Study design and participants
This study was performed as a collaborative research project between Nagoya City University and Shionogi & Co., Ltd. The study design and statistical analysis plan were jointly developed by both institutions. Shionogi served as the study sponsor, overseeing study procedures and the collection of adverse event (AE) data. Participant recruitment, data management, and random allocation were conducted independently by the contracted third-party organizations Integrity Healthcare Inc. and DCT Japan Inc., while statistical analyses were conducted independently by Emeraid Inc. Study procedures were conducted in accordance with the prespecified study protocol and research operations manual, and statistical analyses were performed according to the prespecified statistical analysis plan.
We performed a two-arm, nonblinded, parallel-group RCT to evaluate the acceptability and efficacy of the smartphone intervention informed by IPT, which was designed to support mental well-being during pregnancy, with depressive symptoms at 8 weeks after enrolment serving as the primary outcome. Participants were randomly allocated (1:1) to the intervention group, which received 8 weeks of application use alongside routine health check-ups, or the control group, which only received routine health check-ups. Depressive symptoms were assessed prenatally and during follow-up. Prespecified secondary outcomes included changes in the Patient Health Questionnaire-9 (PHQ-9) scores from baseline to weeks 2, 4, and 6 and at 1 month postpartum (28 days after the expected date of delivery), and changes in the Edinburgh Postnatal Depression Scale (EPDS) scores from baseline to week 8 and 1 month postpartum.
Eligible participants were primiparous women aged 18–39 years in the second trimester of pregnancy (from 16 weeks, 0 days to 28 weeks, 6 days) who had a partner/spouse and sufficient Japanese language literacy to use the application and complete questionnaires. The application was originally developed for a relatively homogeneous target population, focusing on psychosocial adaptation during the transition to first-time parenthood. Therefore, the study enrolled primiparous women aged 18–39 years. This population was selected because primiparous women experience the transition to parenthood for the first time and are at higher risk of perinatal depression and anxiety than multiparous women (Kim et al., Reference Kim, Kim, Lee, Lee, Stubbs, Koyanagi, Dragioti, Jacob, Carvalho, Radua, Thompson, Smith, Oh, Yang, Fornaro, Stickley, de Pablo, Tizaoui, Yon and Fusar-Poli2022). Women younger than 18 years and those aged 40 years or older were excluded because adolescent pregnancy and advanced maternal age are associated with additional obstetric and psychosocial challenges that extend beyond the primary focus of the intervention (Moutos, Ly, Moutos, & Clark, Reference Moutos, Ly, Moutos and Clark2022; Muraca & Joseph, Reference Muraca and Joseph2014; Siegel & Brandon, Reference Siegel and Brandon2014). The exclusion criteria were as follows: current receipt of treatment at a psychiatric or psychosomatic clinic (regardless of medical anamnesis); suicidal ideation (defined as score ≥ 2 on item 9 of PHQ-9) (Kroenke, Spitzer, & Williams, Reference Kroenke, Spitzer and Williams2001; Muramatsu et al., Reference Muramatsu, Miyaoka, Kamijima, Muramatsu, Tanaka, Hosaka, Miwa, Fuse, Yoshimine, Mashima, Shimizu, Ito and Shimizu2018); absence of a partner or spouse; pregnancy with someone other than a partner or spouse; concomitant medications; and any medical condition deemed inappropriate for participation by the research team.
Electronic informed consent was provided by all participants. The protocol of this research has been published previously (Toshishige et al., Reference Toshishige, Chatani, Kawasaki, Goto, Yoshida, Takahashi, Shimasaki, Okami, Sakaguchi, Furukawa, Mizushima and Akechi2025). The authors assert that all procedures contributing to this work comply with the ethical standards of the relevant national and institutional committees on human experimentation and with the Helsinki Declaration of 1975, as revised in 2013. All procedures involving human subjects were approved by the Asai Dermatology Institutional Review Board (approval number: NBP-24-01), an external ethics committee operating in accordance with the Japanese Ethical Guidelines for Medical and Health Research Involving Human Subjects. The trial was registered in the Japanese Clinical Trials Registry (UMIN000056014) on 1 November 2024. This study was reported in accordance with the CONSORT checklist (Moher et al., Reference Moher, Hopewell, Schulz, Montori, Gøtzsche, Devereaux, Elbourne, Egger and Altman2012), and the checklist is provided as a Supplementary File.
Procedures
Participants were recruited nationwide in Japan through online advertisements and research postcards distributed between 18 November 2024 and 6 December 2024. The materials provided a link to the webpage describing the study. They were also presented in a prescreening form that outlined the inclusion and exclusion criteria. Eligible respondents then received an email invitation for electronic consent. Individuals providing consent were invited to complete baseline assessments. After enrolment, participants completed assessments via Google Forms every 2 weeks for 2 months. The Client Satisfaction Questionnaire-8 (CSQ-8J) was administered at week 8. The PHQ-9 was administered at enrolment (baseline); at weeks 2, 4, 6, and 8; and at 1 month postpartum. Participants received ¥1,000 (approximately USD 6) for assessments at weeks 0 (baseline), 2, 4, and 6; ¥3,000 (approximately USD 18) for the 8-week assessment; and ¥5,000 (approximately USD 31) for the 1-month post-assessment.
Smartphone intervention informed by IPT
In collaboration with Shionogi, one author (YT), a certified trainer and supervisor authorized by the International Society of Interpersonal Psychotherapy, oversaw the design and functionality development of the smartphone intervention informed by IPT. The application provided perinatal mental health support through six integrated components: a mental health questionnaire; a feelings tracker; a log; an IPT-informed self-care program delivered via a preprogrammed, rule-based automated chatbot; columns; and a graphical depiction of interpersonal circles. The mental health questionnaire incorporated the EPDS as a self-monitoring tool, providing participants with feedback and advice based on their EPDS scores. The application also incorporated an automated safety function whereby participants reporting self-harm thoughts (EPDS item 10 ≥ 1) were directed to appropriate external public support services. The self-care program incorporated the feelings tracker (with entries automatically recorded in the log), columns, and the graphical depiction of interpersonal circles. Participants used these components during the lessons and were encouraged to continue using them outside the lessons. Detailed descriptions of the application components are presented in Supplementary Table S1, and representative screenshots are provided in Supplementary Figure S1. The IPT-informed self-care program, the core intervention component, is described in detail below.
The self-care program comprised eight sequential weekly lessons, each designed to be completed in approximately 5–10 min (maximum 20 min). The self-care program followed a structured weekly sequence in which participants completed one lesson each week and were required to complete each lesson before progressing to the next. The remaining application components could be accessed freely throughout the 8-week intervention period. Participants were encouraged to engage with the application daily for 5–10 min and to complete one self-care lesson each week throughout the 8-week intervention.
The content of the self-care program was as follows. Lesson 1 used linked columns within the application to provide psychoeducation on perinatal mental health and introduce core IPT-informed concepts, including the link between feelings and interpersonal interactions, as well as the role of social support in promoting well-being. It also introduced the feelings tracker and encouraged participants to use it to monitor the relationship between their feelings and daily experiences during the lesson and in daily life. Lesson 2 used linked columns within the application to introduce an interpersonal inventory and help participants identify people in their interpersonal network, reflect on their sources of social support and on how these relationships affected their feelings, and visually organize these relationships using the application’s graphical depiction of interpersonal circles. Lesson 3 provided psychoeducation on how relationships with significant others influence feelings and how social support can facilitate adjustment to the transition to parenthood. Participants were encouraged to reflect on their own relationships with significant others and their emotional responses to the transition to parenthood. Lesson 4 guided participants in reflecting on the interpersonal challenges they were currently experiencing, after which the application identified the interpersonal problem area most relevant to their responses. Lessons 5–7 provided tailored activities addressing either role transitions or interpersonal disputes. Participants addressing role transitions explored the positive and negative aspects of the transition to parenthood, identified associated emotions, and developed coping strategies suited to their individual needs. Participants addressing interpersonal disputes clarified their expectations of a significant other, reviewed their communication patterns, and considered more effective ways of communicating those expectations. Across these lessons, participants were encouraged to apply what they had worked on during the lessons in their daily lives, reflect on their experiences in the subsequent lesson, and review strategies for managing interpersonal challenges. Lesson 8 provided a comprehensive review of the program and reinforced the key learning points. After the 8-week intervention, participants could continue to use the application voluntarily until 1 month postpartum.
Application-use days were defined as the number of days on which the application was opened at least once. Usage of each application component was determined based on prespecified in-application events. For the self-care program, each completed lesson was counted as one use. Detailed operational definitions for all application components are provided in Supplementary Table S2.
Outcome measures
The primary outcomes were the acceptability and efficacy of the application at 8 weeks after enrolment, prespecified as co-primary outcomes. In self-guided digital interventions, actual use and continued engagement served as prerequisites for efficacy, and acceptability is a key determinant of engagement with the intervention (Perski & Short, Reference Perski and Short2021). Accordingly, acceptability was defined as a CSQ-8J score of ≥17 of a possible score of 32 (i.e. in the upper 50%) (Larsen, Attkisson, Hargreaves, & Nguyen, Reference Larsen, Attkisson, Hargreaves and Nguyen1979; Nguyen, Attkisson, & Stegner, Reference Nguyen, Attkisson and Stegner1983; Tachimori & Ito, Reference Tachimori and Ito1999). The CSQ-8 is commonly utilized to evaluate patient satisfaction within psychiatric care settings (Miglietta, Belessiotis-Richards, Ruggeri, & Priebe, Reference Miglietta, Belessiotis-Richards, Ruggeri and Priebe2018; Woodward, Berry, & Bucci, Reference Woodward, Berry and Bucci2017), and it has established reliability and validity (Tachimori & Ito, Reference Tachimori and Ito1999). Efficacy was assessed as the change in the PHQ-9 score from baseline to 8 weeks after enrolment. This score is widely used to assess depression severity in both general and perinatal populations across clinical and research settings (Kroenke et al., Reference Kroenke, Spitzer and Williams2001), and the reliability and validity of the Japanese version have been confirmed (Muramatsu et al., Reference Muramatsu, Miyaoka, Kamijima, Muramatsu, Tanaka, Hosaka, Miwa, Fuse, Yoshimine, Mashima, Shimizu, Ito and Shimizu2018). PHQ-9 scores of ≥5 indicate the presence of at least mild depressive symptoms (Kroenke et al., Reference Kroenke, Spitzer and Williams2001). PHQ-9 was selected as the primary outcome measure because its widespread use in depression research facilitates comparison of findings with the broader depression literature beyond the perinatal field. Furthermore, the PHQ-9 is recommended by the International Consortium for Health Outcomes Measurement (ICHOM) for the assessment of adult depression outcomes, supporting standardized outcome assessment and facilitating international comparability of outcomes (Obbarius et al., Reference Obbarius, van Maasakkers, Baer, Clark, Crocker, de Beurs, Emmelkamp, Furukawa, Hedman-Lagerlöf, Kangas, Langford, Lesage, Mwesigire, Nolte, Patel, Pilkonis, Pincus, Reis, Rojas and Rose2017). Changes in PHQ-9 scores from baseline to weeks 2, 4, and 6, as well as at 1 month postpartum, were prespecified as secondary outcomes. Changes in EPDS scores from baseline to week 8 and at 1 month postpartum were also prespecified as secondary outcomes. The EPDS is one of the most widely used self-report measures for screening and assessing depressive symptoms during the perinatal period. By emphasizing cognitive and affective symptoms while minimizing somatic symptoms that may be attributed to pregnancy and childbirth, the EPDS is considered particularly suitable for screening perinatal depression (Cox, Holden, & Sagovsky, Reference Cox, Holden and Sagovsky1987). The instrument comprises 10 items rated on a four-point scale, yielding a total score ranging from 0 to 30, with higher scores indicating greater depressive symptom severity. The Japanese version of the EPDS has demonstrated good reliability and validity (Okano, Reference Okano1996). In addition to serving as a secondary outcome measure, the EPDS was integrated into the smartphone intervention as a self-monitoring tool because it was specifically designed for use during the perinatal period. Prespecified subgroup analyses according to baseline depressive symptom severity (PHQ-9 score of ≤4 or ≥5) were also conducted to examine changes in PHQ-9 scores from baseline to week 8 and from baseline to 1 month postpartum.
Adverse events
The data on AEs were gathered through participant self-reports during the study. All reported events were recorded and assessed by study psychiatrists or obstetricians for their causal relationship to the application as per the AE definitions and procedures specified in the study protocol (Toshishige et al., Reference Toshishige, Chatani, Kawasaki, Goto, Yoshida, Takahashi, Shimasaki, Okami, Sakaguchi, Furukawa, Mizushima and Akechi2025).
Sample size estimation
For acceptability, a one-sided test for proportions (null hypothesis of 0.5; assumed proportion of 0.65; α = 0.025; 80% power), adjusted for 21.1% attrition, indicated that 108 participants were required in the intervention arm. For efficacy, using effect estimates derived from a prior RCT (mean difference = −1.19; SD = 3.40), a two-sided t-test (α = 0.05; 80% power) with 1:1 allocation and accounting for 21.1% attrition yielded a target of 165 participants per group (Suharwardy et al., Reference Suharwardy, Ramachandran, Leonard, Gunaseelan, Lyell, Darcy, Robinson and Judy2023). Therefore, we planned to randomly allocate 330 women into the study groups. Detailed assumptions are provided in Supplementary Material S1.
Randomization, allocation and blinding
Participants entered their personal information via an online Google Form, and Integrity Healthcare Inc. verified eligibility and completed formal registration. After baseline evaluations, eligible participants were randomized 1:1 using the stratified block randomization method based on PHQ-9 scores (≤4 or ≥5).
The random allocation sequence was generated in Microsoft Excel by Integrity Healthcare staff, who were not involved in recruitment or assessment, and the allocation list was stored and managed by Integrity Healthcare to prevent any access by study staff or participants. Group allocation was also performed by Integrity Healthcare. According to the open-label design, neither participants nor study staff were blinded to group assignments. All outcomes were self-reported, and analysts at Emeraid Inc. were also not blinded. Further details on randomization, allocation procedures, and data management are available in the published study protocol (Toshishige et al., Reference Toshishige, Chatani, Kawasaki, Goto, Yoshida, Takahashi, Shimasaki, Okami, Sakaguchi, Furukawa, Mizushima and Akechi2025).
Statistical analysis
The performed statistical analyses were based on available data only. Missing values were not imputed, excluding the sensitivity analysis of the primary outcome (acceptability). To control for multiple evaluations, acceptability was tested first, and efficacy was assessed only if acceptability reached significance. Acceptability was defined as the proportion of women in the intervention group with CSQ-8 scores of ≥17 at week 8 and analyzed using a one-sided test for proportions following normal approximation (α = 0.025). Efficacy analyses were conducted in the full analysis set (FAS; participants with ≥1 post-baseline assessment) using mixed-effects models for repeated measures (MMRMs). The models covered fixed effects for treatment group (intervention vs. control), time point (week), the treatment-by-time point interaction, and baseline PHQ-9 scores as covariates. The covariance structure was specified as unstructured. Degrees of freedom were calculated using the Kenward–Roger method. Hypothesis testing for PHQ-9 comparisons was based on the t-statistic produced by the MMRMs. Two-sided tests at α = 0.05 were applied, and p-values and 95% confidence intervals (CIs) were reported.
Secondary outcomes and prespecified subgroup analyses by baseline PHQ-9 scores (≤4 or ≥5) were evaluated similarly. Because the study was not sufficiently powered to detect differential treatment effects between subgroups and interaction testing was not prespecified in the statistical analysis plan, these subgroup analyses were considered exploratory.
As a post-hoc analysis that was not prespecified in the protocol, standardized mean differences (SMDs) were calculated for PHQ-9 and EPDS outcomes at week 8 and 1 month postpartum. SMDs were also calculated for the prespecified PHQ-9 subgroup analyses at both time points. SMDs were calculated by dividing the MMRM-estimated adjusted between-group difference at each time point by the observed pooled standard deviation of the endpoint scores. No interim analyses were planned. Criteria for participant discontinuation and trial termination were prespecified in the study protocol. Additional details are displayed in Supplementary Material S1.
Patient and public involvement
An early prototype of the application was tested by a group of pregnant and postpartum women, external to the research team, who provided feedback on usability and content clarity. Their input contributed to refining the intervention prior to the trial. No participants or members of the public were involved in the conduct or reporting of the trial.
Results
Participants
From 18 November 2024 to 6 December 2024, 1,179 women were assessed for eligibility (Figure 1). Of these, 957 met the inclusion criteria, and 352 consented to participate; of these, two withdrew consent before randomization. Therefore, 350 women were allocated to the intervention (n = 176) or control (n = 174) group. Two participants were excluded from the FAS, including one who did not install the application or complete any questionnaires and one with a criteria deviation. One additional participant in the intervention group withdrew consent because of the burden of study participation but permitted the use of her questionnaire data, permitting her retention in the FAS. Two participants in the control group did not complete the postpartum questionnaires, but they consented to the use of previously collected data. Accordingly, the FAS comprised 348 participants (intervention, n = 175; control, n = 173).
CONSORT diagram. The intervention and control groups were evaluated at week 8 (primary outcome). Note: Some participants met more than one exclusion criterion. Thus, the numbers denote the total number of reasons for exclusion and not the number of excluded participants. One participant in each group was excluded from the full analysis set (FAS). The participant in the intervention group was excluded because of a failure to install the application and lack of data, and the participant in the control group was excluded because of a protocol deviation. Therefore, the FAS comprised 175 participants in the intervention group and 173 participants in the control group.
Figure 1. Long description
The flowchart begins at the top center and moves downward through the following stages:
* Enrollment Phase:
* Assessed for eligibility n equals 1179. An arrow points right to Excluded n equals 222, with reasons including gestational age criteria n equals 155, prior childbirths n equals 33, age criteria n equals 16, P H Q 9 score criteria n equals 16, and other reasons n equals 37.
* Included n equals 957. An arrow points right to Excluded n equals 605 for not providing informed consent.
* Consented n equals 352. An arrow points right to Excluded n equals 2 for withdrawal of consent.
* Randomised n equals 350.
* Allocation and Follow-up (Two parallel columns):
* Left Column (Intervention Group):
* Allocated to the intervention group n equals 176. 175 received intervention; 1 did not start because the app was not installed.
* Lost to follow-up at 8 weeks n equals 1 due to withdrawal of consent.
* Analysed F A S n equals 175. Excluded from analysis n equals 0.
* Lost to follow-up at 1 month post-partum n equals 0.
* Final Analysed F A S n equals 175.
* Right Column (Control Group):
* Allocated to the control group n equals 174.
* Lost to follow-up at 8 weeks n equals 1.
* Analysed F A S n equals 173. Excluded from analysis n equals 0.
* Lost to follow-up at 1 month post-partum n equals 2.
* Final Analysed F A S n equals 173.
Baseline characteristics were balanced between the groups (Table 1). Participants had a mean age of approximately 30 years and a mean gestation time of 21 weeks. Most women were employed, and prior use of mental health applications was uncommon.
Baseline characteristics of participants in the full analysis set
Table 1. Long description
The table consists of three columns: Characteristic, Intervention group (N = 175), and Control group (N = 173).
Key data points include:
* Age: The mean age is 31.12 (S D 3.98) for the intervention group and 30.95 (S D 4.04) for the control group. The largest age bracket for both is 30 to 34 years.
* Gestational age: Mean weeks are 21.75 (S D 3.53) for intervention and 21.50 (S D 3.57) for control.
* Education: Over half of both groups hold a four-year university degree (55.43 percent intervention, 58.38 percent control).
* Employment: 53.14 percent of the intervention group and 65.90 percent of the control group work full-time.
* Pregnancy History: Previous miscarriages were reported by 16.00 percent of the intervention group and 16.76 percent of the control group.
* Pregnancy Method: Natural pregnancy accounts for 73.14 percent (intervention) and 67.63 percent (control). Infertility treatments include I V F / I C S I, artificial insemination, and ovulation inducers.
* Mental Health: A P H Q-9 score greater than or equal to 10 was found in 10.29 percent of the intervention group and 14.45 percent of the control group. History of psychiatric visits was 17.14 percent and 13.87 percent respectively.
Note: SD, standard deviation; IVF, in vitro fertilization; ICSI, intra-cytoplasmic sperm injection; PHQ-9, Patient Health Questionnaire-9.
a Based on the participants’ self-report.
The trial was ended as planned, and the 1-month postpartum follow-up observation was completed by 27 June, 2025. Postpartum assessments were scheduled 1 month after the expected date of delivery, although the actual timing varied. Among participants with available data (intervention, n = 174; control, n = 171), the mean assessment timing was 4.75 ± 1.78 and 4.42 ± 1.67 weeks in the intervention and control groups, respectively.
Outcomes
In the intervention group, 174 of 176 participants (98.9%) completed the week 8 evaluation, whereas one woman withdrew consent without completing CSQ-8J at week 8 (Figure 1). Among the 175 participants in the intervention group of the FAS, all used the application at least once during the 8-week intervention period. During this period, participants used the application for a mean of 35.85 days (SD = 16.24), engaged with a mean of 5.56 of the six integrated components (SD = 0.80), and completed a mean of 7.45 self-care lessons (SD = 2.06). After the intervention period, 96% of the women used the application at least once, with a mean usage period of 14.06 days (SD = 12.18) until 1 month postpartum. Intervention use stratified by various components is summarized in Supplementary Table S2.
Among these 174 participants, 157 (90.2%, 95% CI = 84.8–94.2) achieved CSQ-8J scores of ≥17, exceeding the prespecified threshold of 50% (p < 0.001).
In the FAS analysis set, no significant between-group difference in PHQ-9 scores was detected at week 8 (group difference = 0.41, 95% CI = −0.24 to 1.06; p = 0.214; SMD = 0.11). However, at 1 month postpartum, the intervention group had significantly lower PHQ-9 scores than the control group (group difference = 1.01, 95% CI = 0.29–1.73; p = 0.006; SMD = 0.28; Figure 2, Table 2). The results for weeks 2, 4, and 6 are displayed in Table 2.
Changes in PHQ-9 scores from baseline up to 1 month postpartum (FAS participants with available PHQ-9 data). Note: FAS, ‘full analysis set’; PHQ-9, ‘Patient Health Questionnaire-9’.
Figure 2. Long description
The x-axis is labeled Time with intervals for Baseline, Week 2, Week 4, Week 6, Week 8, and 1-month postpartum. The y-axis is labeled Mean 95 percent C I, ranging from 4.0 to 6.0.
* The Control Group (blue line with circular markers) starts at approximately 5.05 at baseline, rises to 5.5 at Week 2, dips slightly at Week 4, then rises steadily to a peak of 5.75 at Week 8 before ending at 5.2 at 1-month postpartum.
* The Intervention Group (orange line with circular markers) starts at 5.0 at baseline, rises to 5.4 at Week 2, remains stable through Week 8, and then shows a sharp decline to approximately 4.2 at 1-month postpartum.
* Vertical error bars representing the 95 percent C I are present for all data points except baseline, with the Intervention Group showing a significantly lower mean and non-overlapping error bar compared to the Control Group at the final 1-month postpartum time point.
Primary and secondary PHQ-9 outcomes of the smartphone intervention informed by IPT
Table 2. Long description
The table compares unadjusted and adjusted P H Q 9 scores between an intervention group and a control group.
Columns include:
- Time point: Baseline, Week 2, Week 4, Week 6, Week 8, and One month postpartum.
- Unadjusted data: n, Mean score (S D), and Mean score change (S D) for both groups, plus Group difference (S E).
- Adjusted data: Mean score change (S E) for both groups, Group difference (S E), 95% C I, p-value, and S M D.
Key data points:
- Baseline: Intervention n = 175, Mean 5.02; Control n = 173, Mean 5.06.
- Week 8 (Primary Outcome): Unadjusted mean score change is 0.34 for intervention and 0.72 for control. Adjusted group difference is 0.41 (S E 0.33) with a p-value of 0.214.
- One month postpartum: Shows a significant difference. Unadjusted mean score change is -0.84 for intervention and 0.15 for control. Adjusted group difference is 1.01 (S E 0.36) with a 95% C I of 0.29 to 1.73 and a p-value of 0.006.
Note: P H Q 9 stands for Patient Health Questionnaire 9. S D is Standard Deviation, S E is Standard Error, and S M D is Standardized Mean Difference.
Note: The assessment included participants from the full analysis set with PHQ-9 data, including 175 in the intervention group and 173 in the control group. The primary outcome was assessed at week 8, with additional secondary outcomes evaluated at other time points. Adjusted results were estimated with a mixed-effects model with repeated measures. The mean score change from baseline might not reflect the difference between the baseline and time point means because of missing data at a particular time point. IPT, interpersonal psychotherapy; PHQ-9, Patient Health Questionnaire-9; SD, standard deviation; SE, standard error; SMD, standardized mean difference.
a Number of respondents.
A similar pattern was observed for EPDS. Although no significant between-group difference was found at week 8 (group difference: 0.61, 95% CI −0.10 to 1.33; P = 0.093; SMD = 0.14), EPDS scores were significantly lower in the intervention group at 1 month postpartum (group difference: 1.51, 95% CI 0.65 to 2.38; P < 0.001; SMD = 0.35; Table 3).
Secondary EPDS outcomes of the smartphone intervention informed by IPT
Table 3. Long description
The table compares an intervention group and a control group across three time points: Baseline, Week 8, and One month postpartum.
Header Structure:
- Unadjusted data includes n, Mean score (S D), Mean score change (S D) for both groups, and Group difference (S E).
- Adjusted data includes Mean score change (S E) for both groups, Group difference (S E), 95% C I, p-value, and S M D.
Data Rows:
- Baseline: Intervention group (n = 175) mean score 6.23 (4.55); Control group (n = 173) mean score 6.50 (4.45).
- Week 8: Intervention mean score 5.15 (4.14), change -1.11; Control mean score 5.89 (4.75), change -0.61. Adjusted group difference is 0.61 (S E 0.36) with a p-value of 0.093 and S M D of 0.14.
- One month postpartum: Intervention mean score 4.79 (4.01), change -1.48; Control mean score 6.35 (4.58), change -0.04. Adjusted group difference is 1.51 (S E 0.44) with a p-value less than 0.001 and S M D of 0.35.
Note: E P D S stands for Edinburgh Postnatal Depression Scale. Adjusted estimates were obtained using a mixed-effects model with repeated measures.
Note: The analysis included participants in the full analysis set with available EPDS data, including 175 participants in the intervention group and 173 participants in the control group. Adjusted estimates were obtained using a mixed-effects model with repeated measures. Because of missing data at individual time points, the mean change from baseline may not correspond exactly to the arithmetic difference between the baseline and follow-up mean scores. IPT, interpersonal psychotherapy; EPDS, Edinburgh Postnatal Depression Scale; SD, standard deviation; SE, standard error; SMD, standardized mean difference.
a Number of respondents.
In the subgroup analysis, the proportion of participants with baseline PHQ-9 scores of ≤4 did not differ between the groups (week 8: group difference = −0.29; 95% CI = −1.04 to 0.47; p = 0.455; SMD = −0.11; 1 month postpartum: group difference = 0.44; 95% CI = −0.42 to 1.30; p = 0.314; SMD = 0.14). Among the participants baseline scores of ≥5, the intervention group exhibited significantly more pronounced reductions in PHQ-9 scores at week 8 (group difference = 1.20; 95% CI = 0.11–2.29; p = 0.031; SMD = 0.30) and at 1 month postpartum (group difference = 1.63; 95% CI = 0.46–2.79; p = 0.006; SMD = 0.42; Supplementary Table S3).
AEs
Several AEs occurred during the trial, including a small number of hospitalizations; however, none was found to be related to the intervention. No participant discontinued the intervention because of an AE attributable to the application.
Discussion
To the best of our knowledge, this is among the first RCTs to evaluate the acceptability and efficacy of a smartphone intervention informed by IPT aiming to maintain and promote mental well-being during pregnancy in a predominantly nonclinical population. In our cohort of primiparous women with partners, the intervention achieved a high acceptability rate. However, no significant difference in overall efficacy, defined as changes in PHQ-9 scores, was detected between the intervention and control groups at the end of the 8-week prenatal intervention period. During the prespecified secondary postpartum analysis, more favorable changes in PHQ-9 scores were noticed in the intervention group. Similarly, no significant between-group difference in EPDS scores was observed at week 8, whereas more favorable changes in EPDS scores were observed in the intervention group at 1 month postpartum.
Face-to-face IPT is an effective evidence-based psychotherapy for perinatal depression (Sockol, Reference Sockol2018), but there are many barriers to such services (Goodman, Reference Goodman2009; Kim et al., Reference Kim, La Porte, Corcoran, Magasi, Batza and Silver2010). Conversely, smartphone applications incorporating elements of IPT have relatively few barriers to access, and they might help meet the psychological needs of pregnant women (Goodman, Reference Goodman2009; Johansen, Robakis, Williams, & Rasgon, Reference Johansen, Robakis, Williams and Rasgon2019; O’Mahen & Flynn, Reference O’Mahen and Flynn2008). However, the acceptability and efficacy of mobile health applications delivering IPT-informed interventions have not been rigorously assessed in large populations of pregnant women.
Bright et al. reported the feasibility and acceptability of web-based IPT modules with low-intensity guided support for prenatal mental health and suggested delivery via a smartphone application (Bright et al., Reference Bright, Stuart, Mcneil, Murray and Kingston2022), although the study only included a small qualitative sample (n = 15). Suharwardy et al. (Reference Suharwardy, Ramachandran, Leonard, Gunaseelan, Lyell, Darcy, Robinson and Judy2023) investigated a smartphone application incorporating CBT and IPT techniques in postpartum women, with 62 of 68 (91%) reporting at least a moderate level of satisfaction. Meanwhile, our study recorded high acceptability rates for a smartphone intervention informed by IPT among pregnant women.
Fisher et al. (Reference Fisher, Davey, Wong, Morgan-Trimmer, Howard, Sharp, Atmore, Brook, Collins, Domoney, Makinde, McCree and O’Mahen2024) identified four pillars of engagement in mental healthcare: perinatal competence, relationship building, accurate reassurance, and reliability. Users of the CBT/IPT chatbot developed by Suharwardy et al. (Reference Suharwardy, Ramachandran, Leonard, Gunaseelan, Lyell, Darcy, Robinson and Judy2023) reported forming affective bonds with the chatbot. Similarly, our application featured a friendly character that guided users and encouraged the application of learned strategies in daily life. Such features might have facilitated engagement with mental health care. The application might also have reflected other engagement pillars, as it was co-developed by psychiatrists, an obstetrician and a pharmaceutical company, thus potentially enhancing users’ perceptions of reliability and reassurance.
The lack of significant between-group differences in PHQ-9 scores at the end of the 8-week prenatal intervention period might be attributable to the generally low baseline PHQ-9 scores and less significant than expected between-group differences. More than half of the participants had baseline PHQ-9 scores of ≤4, and the mean scores were extremely low (2.56 ± 1.22 in the intervention group and 2.13 ± 1.38 in the control group). These low baseline scores might have produced a floor effect, limiting the potential for measurable changes. In addition, the between-group difference was less pronounced than anticipated, potentially indicating insufficient statistical power.
At the prespecified 1-month postpartum follow-up, more favorable changes in PHQ-9 scores were observed in the intervention group. The magnitude of the effect (SMD = 0.28) was compatible with the pooled estimates from meta-analyses of digital interventions for perinatal depression (SMD = 0.29) (Anyanwu & Jenkins, Reference Anyanwu and Jenkins2024), suggesting high consistency with the outcomes in the existing literature. The detected differences in the postpartum follow-up, despite the absence of effects noticed immediately after the intervention, might reflect a delayed manifestation of the intervention’s impact. One possible explanation is that knowledge and coping skills related to mental health acquired during the intervention were internalized and further applied beyond the intervention period. Such a self-care psychotherapy application has also been demonstrated to allow users to acquire therapeutic skills during an active intervention phase, which might be sustained and applied beyond this period (Furukawa et al., Reference Furukawa, Tajika, Toyomoto, Sakata, Luo, Horikoshi and Noma2025). Interpersonal skills introduced through the smartphone intervention informed by IPT during pregnancy might have been maintained and applied after the intervention, potentially contributing to the observed between-group differences in depressive symptoms in the postpartum follow-up, particularly in the context of postpartum role transitions and emerging interpersonal difficulties with close others. Some degree of engagement with the application after the intervention period might have played a role in the development of these effects; however, the extent to which such engagement influenced outcomes remains unclear.
Exploratory subgroup analyses suggested more favorable changes in PHQ-9 scores in participants with baseline PHQ-9 scores of ≥5; however, these findings should be interpreted cautiously.
Women who experience depressive symptoms during pregnancy are known to have elevated risks of postpartum depression (O’Hara & McCabe, Reference O’Hara and McCabe2013; Xu et al., Reference Xu, Luo, Huang, Liu and Ding2026). Previous studies demonstrated that face-to-face IPT group interventions during pregnancy can reduce the incidence of postpartum depression in high-risk women (Zlotnick et al., Reference Zlotnick, Tzilos, Miller, Seifer and Stout2016). Clinical guidelines also recommend preventive IPT for pregnant and postpartum women at increased risk of perinatal depression (Motrico et al., Reference Motrico, Moreno-Peral, Uriko, Hancheva, Brekalo, Ajaz, Apter, Bramante, Conejo-Cerón, Christoforou, Dikmen-Yildiz, Evagorou, Fonseca, Lupattelli, Radoš, Al Maach, Rodriguez-Muñoz, Žutić and Lambregtse-Van Den Berg2022; O’Connor et al., Reference O’Connor, Senger, Henninger, Coppola and Gaynes2019; US Preventive Services Task Force et al., Reference Curry, Krist, Owens, Barry, Caughey, Davidson, Doubeni, Epling, Grossman, Kemper, Kubik, Landefeld, Mangione, Silverstein, Simon, Tseng and Wong2019). These findings suggest that interpersonal support delivered via a mobile application during pregnancy can help women with depressive symptoms address interpersonal challenges associated with the transition to parenthood, including role changes, partner-related conflict, and limited social support (O’Hara & McCabe, Reference O’Hara and McCabe2013; Spinelli, Reference Spinelli2017; Xu et al., Reference Xu, Luo, Huang, Liu and Ding2026).
The present study had some limitations. First, the low baseline PHQ-9 scores may have limited the potential for observable changes due to a floor effect. Second, several high-risk groups for perinatal depression were excluded (Brito, Alves, Ludermir, & Araújo, Reference Brito, Alves, Ludermir and Araújo2015; Muraca & Joseph, Reference Muraca and Joseph2014; Qiu et al., Reference Qiu, Zhang, Sun, Li and Wang2020; Siegel & Brandon, Reference Siegel and Brandon2014) (women aged <18 or >39 years, those without partners, and women with unintended or nonmarital pregnancies), therefore reflecting the original design constraints of the application. Third, only primiparous women were enrolled, limiting the applicability of the findings to multiparous women. Fourth, daily usage time and day-to-day usage patterns were not directly assessed, limiting detailed evaluation of intervention adherence. Fifth, the participants were informed that the application had been developed by psychiatrists, an obstetrician, and a pharmaceutical company, which might have increased expectancy effects. Sixth, postpartum outcomes were assessed only up to 1 month after delivery. Because postpartum depression commonly occurs within the first 6 months postpartum (Bai et al., Reference Bai, Li, Cheng, Caine, Tong, Wu and Gong2023), longer follow-up is needed to determine whether these findings are relevant.
With regard to future implementation, the application is intended to be made available for direct download and use by perinatal women. Although the specific implementation strategy, including its potential interface with clinical practice, has not yet been determined, the application has the potential to support self-care among perinatal women and to be used alongside clinical practice. To enhance user safety, the application incorporates an automated safety monitoring function. When users report a score of ≥1 on EPDS item 10 (self-harm thoughts), an alert banner is displayed with a link to a webpage provided by the Ministry of Health, Labour and Welfare. The webpage provides telephone hotline numbers and information on SNS-based counseling services for users requiring immediate support.
In conclusion, this RCT demonstrated the high acceptability of a smartphone intervention informed by IPT during pregnancy among primiparous women with partners. Although no overall difference in depressive symptoms was detected at the end of the prenatal intervention period, more favorable changes were observed at 1 month postpartum. These findings suggest that an accessible and self-paced smartphone application incorporating interpersonal support represents a promising approach to perinatal mental health support. However, larger trials with longer follow-up and more diverse populations are required to validate these findings.
Supplementary material
The supplementary material for this article can be found at http://doi.org/10.1017/S0033291726105868.
Data availability statement
The data supporting the findings of this study are not publicly available because they contain confidential patient information, and participants did not provide consent for their data to be shared outside the research team. Requests for limited access to deidentified data cannot be accommodated due to these consent restrictions and privacy considerations. Additionally, related study documents, including the study protocol, statistical analysis plan, and informed consent form, are not available due to intellectual property constraints. The analytic code used in this study cannot be made publicly available because it is derived from the confidential study data, and participants did not provide consent for any data-related materials to be shared outside the research team. However, the code can be submitted to the journal’s editorial office upon reasonable request for the purpose of verifying the study’s findings.
Acknowledgments
During the preparation of this work, the authors used ChatGPT (OpenAI, GPT-5) for English editing purposes (i.e. to improve grammar, wording, and clarity). After using this tool, the authors reviewed and edited the content, as needed, and take full responsibility for the content of the published article. We thank Enago (www.enago.jp) for their assistance with English language editing.
Author contribution
All authors contributed to the concept and implementation of this study. Y.T. (Yuko Toshishige), M.S., N.C., S.K., S.G., T.O., and H.M. (Hiroko Mizushima) took part in the development of the smartphone application. This study design involved the participation of Y.T. (Yuko Toshishige), N.C., S.K., S.G., Y.Y., Y.T. (Yusaku Takahashi), T.O., Y.M., T.A.F., and T.A. The manuscript was drafted by Y.T. (Yuko Toshishige), M.S., N.C., S.K., Y.T. (Yusaku Takahashi), T.O., E.T., H.M. (Hiroko Maekawa), and T.A., and all authors contributed to its review and approved the final manuscript.
Funding statement
This study was financially supported by Shionogi.
Competing interests
Y.T. has received a grant from Shionogi and is a co-inventor of a pending patent (2025-128748) (Institute) related to the submitted work. Y.T. has received personal fees from Viatris, Takeda, Seiwa Shoten, Igaku-Shoin, and Ishiyaku Publishers, outside the submitted work. N.C. and S.K. are employees of Shionogi and are co-inventors of a pending patent (2025-128748) (Company). S.G. has received a grant from Shionogi. M.S. has received financial support from Shionogi and is a co-inventor of a pending patent (2025-128748) (Company) related to the submitted work. T.A.F. reports personal fees from Daiichi-Sankyo, DT Axis, Micron, Shionogi, SONY, and UpToDate, and a grant from DT Axis, outside the submitted work. In addition, T.A.F. has a patent 7448125 and a pending patent 2024-521973, and has licensed intellectual properties for Kokoro-app to DT Axis. T.A. has received a grant from Shionogi and is a co-inventor of a pending patent (2025-128748) (Institute) related to the submitted work. T.A. received lecture fees from AstraZeneca, Chugai, Daiichi-Sankyo, Eisai, Eli-Lilly, Kowa, Lundbeck, MSD, Meiji-Seika Pharma, Merck, Otsuka, Pfizer, Shionogi, Sumitomo Pharma, Takeda, Tsumura, UCB, and Viatris, as well as royalties from Igaku-Shoin, outside the submitted work. T.A. serves as the representative director of the General Incorporated Association NCU CRESS and receives compensation as an adviser to Snom Inc. T.A. has pending patents (2020-135195, 2024-516288) (Institute) and a granted patent (7313617) (Institute). All other authors declare no conflicts of interest.
Transparency declaration
The lead author affirms that this manuscript is an honest, accurate, and transparent account of the study being reported; that no important aspects have been omitted; and that any deviations from the planned study procedures have been fully disclosed.
Research material availability
The research materials associated with this study, including the application content informed by IPT and related instructional materials, are not publicly available because they contain proprietary content developed in collaboration with Shionogi & Co., Ltd., and the materials cannot be shared outside the research team. These materials can therefore not be provided to third parties.
Open access
