In recent years, depression has displaced many chronic diseases to be the leading cause of disability worldwide and the main contributor to the overall global burden of diseases [1]. Although depression can occur at any point in life, women and men are at increased risk of depression in the postpartum period [2,3]. The transition to parenthood, coupled with emotional, physiological, hormonal, and psychosocial changes experienced during childbirth, could have increased women’s susceptibility to postpartum depression [4-6]. The psychological state of women could influence their partners’ psychological well-being because maternal postpartum depression is a known predictor of paternal postpartum depression, affecting 24%-50% of all fathers [7]. Fathers also have the highest risk of developing paternal postpartum depression 3 to 6 months after childbirth [8,9]. Moreover, research has shown the ripple effects of postpartum depression adversely affecting an infant’s subsequent cognitive, physical, social-emotional, and behavioral development over the years [10,11]. Children of mothers who were depressed were found to score lower in intellectual and motor development, and they tend to display insecure attachment, behavioral problems, poor emotional regulation and academic performances, and are at an increased risk of developing psychiatric disorders [12-14]. In addition, a child’s internalizing and externalizing behavior problems are significantly related to paternal postpartum depression [15].

Given the potential widespread consequences of postpartum depression on the family unit, there is a strong impetus for couple-based interventions to curb postpartum depression. However, to date, the cause of postpartum depression among parents remains inconclusive, thus inhibiting the development of direct intervention. Instead, research has suggested a multifactorial etiology, as there has been evidence that various psychosocial variables, such as life stress, marital conflict, childcare stress, parenting self-efficacy, poor childcare knowledge, the lack of social support, and antenatal anxiety, are key predictors of postpartum depression [5,6]. Hence, a perinatal intervention targeting these predictors may help to alleviate feelings of stress and depression among parents. Reviews on previous interventions implemented during the perinatal period have indicated the need for more multimodal family-oriented educational programs that focus on parental physical, emotional, and social well-being [16-18]. Furthermore, theory-based educational interventions are scarce, and current interventions are often compartmentalized as antenatal or postpartum [16-19] and focus on specific groups of parents such as those with infants in the neonatal intensive care unit or with behavioral problems [16-18,20,21]. Therefore, there is an urgent need for a comprehensive perinatal educational intervention for parents with healthy infants to enhance parental psychological well-being and infant developmental outcomes.

Technology-enhanced health care and telemedicine have been gaining popularity because of their better accessibility, flexibility, cost-effectiveness, and ability to provide patient privacy and reduce stigmatization [22,23]. The rapid proliferation in technology-based health care is enabled by the increased use and advancements of mobile apps, wearable devices, live audiovisual communication, and telecommunication. The value of telehealth care and technology-based interventions has been further amplified during the ongoing COVID-19 pandemic as most households worldwide have faced lockdown restrictions, quarantine measures, and social isolation, which limit couples’ access to help and support [24]. Moreover, such technology-based interventions are especially crucial in a conservative society such as Singapore’s, where traditional views and postpartum practices (eg, home confinement) as well as the stigmatization of mental health disorders restrict parents from seeking help in the postpartum period [25,26].

Previous studies examining the needs of fathers and mothers during the perinatal period have also highlighted parents’ preferences for technology-based educational programs, especially mobile health (mHealth) apps [26,27]. mHealth apps have a relatively new edge in health care innovation that allows the delivery of health care and dissemination of educational information [27]. In 2010, more than 200 million mHealth apps were downloaded, and approximately 70% of the population worldwide has access to at least one mHealth app [28]. mHealth pregnancy-related apps are particularly popular among young parents, suggesting a shift in patient empowerment in terms of maternity care [29,30]. However, parents have expressed difficulty in finding pregnancy apps that provide localized and individualized advice and feedback from health care professionals [27,30]. Current literature has also revealed that most of the available mHealth apps are not validated or developed by health care professionals [31], thus calling into question the trustworthiness and credibility of the information provided.

Given the benefits of mHealth apps and the paucity and limitations of current pregnancy-related mHealth apps, this study aims to develop the Supportive Parenting App (SPA), a theory-based perinatal educational mobile app for parents, and to subsequently examine its effectiveness in preventing postpartum depression and enhancing parenting self-efficacy, perceived social support, parent-child bonding, and physical and social-emotional development of an infant. According to the Singapore Infocomm Media Development Authority [32], 98% of households in Singapore in 2019 had internet access and 89% of individuals were internet users. In addition, the smartphone penetration rate of 82% in Singapore is one of the highest in the Asia Pacific region [33]. Therefore, technology is available in most households to support the implementation of the SPA during the perinatal period.

Detailed intervention development studies are limited because of research funding priorities that focused on efficacy or effectiveness studies [34]. However, a comprehensive, systematic, and transparent reporting of intervention development is necessary to help readers understand the challenges and benefits of different intervention development approaches and allow retrospective assessments to be performed to understand how different development approaches may affect the effectiveness of interventions. Therefore, this study aims to provide an insight into the SPA development process and highlight the challenges and lessons faced during the SPA development to improve future mHealth studies.

The SPA project was a multisite study that was conducted in tertiary public hospitals in Singapore. The intervention was delivered to both parents (fathers and mothers), and they were granted access to the SPA during the perinatal period (from the viability age of 24 weeks of gestation to 6 months after childbirth) when the chances of developing antenatal or postpartum depression were very high. The parents were able to access the mobile app anytime and anywhere. The SPA featured educational materials, a discussion forum, chat groups with peer volunteers, and advice from health care professionals. The inclusion criteria were as follows: (1) parents with a female partner who scored ≥9 on the Edinburgh Postnatal Depression Scale, (2) those aged ≥21 years, (3) those who were able to read and speak English, (4) those who had a low-risk pregnancy with more than 24 weeks of gestation (age of viability), and (5) those who had a smartphone with internet access. The exclusion criteria were as follows: parents who (1) had physical or mental disorders that would interfere with their ability to participate in the study; (2) had a high-risk pregnancy, including placenta previa major, preeclampsia, and pregnancy-induced hypertension; or (3) gave birth to a stillborn baby or a newborn with congenital anomalies or medical complications, including pathological jaundice that required special care in hospital. Mothers aged ≥21 years who had previously recovered from postpartum depression were recruited and trained by a psychiatrist to become peer volunteers to befriend and provide adequate web-based support to the parents.

The aims of this research project are to (1) develop a theory-based SPA intervention for parents across the perinatal period; (2) examine the intervention’s effectiveness in reducing parental postpartum depression and anxiety and encouraging parenting self-efficacy, help-seeking behavior, parent-infant bonding, parenting satisfaction, and infant developmental outcomes during the perinatal period up to 1 year after childbirth; (3) evaluate the SPA’s cost-effectiveness compared with standard perinatal care across major restructured hospitals; and (4) examine SPA user experiences. However, this paper only focuses on the SPA intervention development process, the challenges faced by the research team, and the lessons learned.

The information systems research (ISR) framework, which promotes an iterative, user-centered, nonlinear design process was used to guide the development of the SPA [35]. The ISR framework is commonly used in the creation of technological artifacts and comprises 3 research cycles (relevance cycle, design cycle, and rigor cycle) that link 3 research domains (knowledge base, design science research, and environment). The relevance cycle involves understanding the end users’ needs and requirements through feedback and focus groups. The design cycle is where the artifact is developed and evaluated, and the rigor cycle is where theories, past methods, and representations of the artifacts contribute to the current design and development of the new artifact [35]. This study’s adaptation of the ISR framework [35] is presented in Figure 1.

To establish treatment fidelity, the strategies proposed by Bellg et al [36] and Eaton et al [37] for intervention design, training of providers, intervention delivery, and receipt of intervention were adopted. During the intervention design phase, intervention fidelity was ensured by using a theoretical framework (social cognitive theory) as the foundation of the intervention and by preparing a protocol for the intervention. In anticipation of provider and peer volunteer dropout, backup providers and volunteers with the requisite skills were on constant standby. To ensure consistency, delivery of intervention materials (eg, knowledge-based content and expert advice) was standardized by providing briefings and issuing providers and peer volunteers a handbook on the intervention procedures. In terms of receipt of intervention, participants’ frequency of access to the app and the use of audio and video materials and peer discussion forum were monitored and recorded through a back-end admin portal. Weekly reminders were sent to parents to reinforce the use of the intervention by accessing the SPA.

As the SPA development process involved stakeholders across various departments, immense collaborative effort and a high-performing team were crucial to ensure consistent progress and adherence to the project timeline. The Tuckman model of team development [38] highlighted different phases of team development—forming (initial meeting and beginning of collective work), storming (dealing with confusion and conflict over decision-making and roles), norming (accepting goals, roles, and work positively), performing (focusing on achieving goals and obtaining personal growth), and adjourning (recognizing and celebrating achievements). Although all phases are necessary for successful team building, problem solving, work planning, and delivering of results, the team leader plays a critical role in regulating the team dynamics and ensuring effective workflow by applying appropriate conflict resolution strategies as recommended by Tuckman [38].

The end-user environment and requirements were obtained through a focus group session with new parents. In all, 3 pairs of parents representing each major ethnic group in Singapore (Chinese, Malay, and Indian) were interviewed on their informational and perinatal educational needs and their expectations of the SPA. These parents were not included in the subsequent main study. In addition, because the objectives and components of the SPA were similar to those of the research team’s previous research projects—Home-but-not-Alone (an mHealth app-based postnatal educational program) and the Peer-Support Intervention Program—user requirements were obtained from the projects’ focus groups and interviews [39,40]. Previous local research among Singaporean mothers who only received standard routine hospital care indicated a need for more support after childbirth, such as having peer volunteers to talk to, more follow-up appointments with clinicians, and having an educational parenting mobile app catered to parents in Singapore [40]. Mothers who received the intervention requested the inclusion of individualized push notifications to mark important baby developmental milestones and special parenting tips. A study that investigated the use of a technology-based supportive intervention among multiracial Singaporean mothers during the postpartum period suggested the extension of the parenting support programs beyond 1 month after childbirth, introducing the parenting app early in pregnancy, and including a greater variety of educational topics [39].

To meet the goals of the rigor cycle, well-established theoretical frameworks were incorporated into the development process of the SPA. The research team referred to the mHealth user engagement pyramid proposed by Singh et al [41] to identify features that could be incorporated into the SPA. Providing educational information is the most fundamental way to engage a patient, and the level of engagement increases with add-on features such as reminders and notifications, records of health information, display and summary of health information, availability of individualized guidance, means of communication between clinicians and family members or caregivers, provision of support through social networks, and incentivization of behavioral changes [41].

Other theoretical frameworks that guided the development of the SPA’s features were the social cognitive theory proposed by Bandura [42,43] and the attachment theory proposed by Bowlby [44]. The social cognitive theory posits that an individual’s knowledge acquisition depends on the interaction among the environment, cognitive behavior, and personal factors [43]. An important component of the social cognitive theory is self-efficacy, which is defined as one’s belief in one’s capability to accomplish a task effectively [42]. Therefore, to ensure successful parenting, Bandura [42,45] emphasized that parents need to possess the self-confidence to perform specific skills and believe that they can achieve their desired outcome. As evidence has shown low levels of parenting self-efficacy as a predictor of postpartum depression [46], the concepts of the self-efficacy theory proposed by Bandura [42,45] were incorporated into the development of the SPA. Mastery of experiences, vicarious experiences, verbal persuasion, and improving physical and emotional states are ways in which self-efficacy can be attained [42]. In addition, Bandura [43] emphasized the importance of social support in achieving self-regulation to aid learning. There are different types of social support (emotional, instrumental, informational, and appraisal) [43]. As the SPA is a mobile app intervention, providing instrumental support was not possible; hence, it focused primarily on emotional and informational support. Appraisal support, which involves the provision of feedback regarding one’s performance or qualities, was also not feasible in the context of the SPA. However, specific feedback requested by parents was provided for their individual needs.

According to the attachment theory proposed by Bowlby [44], parenting self-efficacy, social support, and parental emotional well-being are crucial for early parent-infant bonding and infant attachment styles, and these serve as a foundation for the development of positive social relationships in infants. Bowlby [44] postulated that a healthy parent-infant bond may lead to secure attachment, whereas insecurely attached infants may take their insecure patterns (avoidant, anxious ambivalent, or disorganized) with them into adulthood. Congruent to the attachment theory proposed by Bowlby [44], Bandura [45] postulated that self-efficacy will consequently influence parent-infant bonding, which is an important determinant of parental emotional well-being and infant development. Therefore, it is necessary to examine the relationship among parenting self-efficacy, parental outcomes, and infant outcomes. The theoretical framework for this project is presented in Figure 2 [42,43].

The development process of the SPA, which included requirement identification, the use of theories and past artifacts, the design of artifacts, and user testing, was well documented using the ISR framework. Although the ISR framework was able to capture the design technicalities and rigor of the project, it failed to highlight interpersonal and real-life challenges. Apart from technical and user issues, the design cycle faced prominent issues such as logistical limitations and also in terms of teamwork and communication because of the effects of the COVID-19 outbreak, which affected the process timeline.

With respect to the FAQ and expert advice section, the research team ideally wanted to provide a readily available hotline or personalized help for parents in need. However, because of limited resources and other work commitments, it was not feasible for the research team to be on constant standby to respond to parents’ queries immediately. Hence, a decision was made to have a maternity unit nurse or midwife to check and respond to queries posted in the FAQ section once a day and respond within 24 hours of the queries raised by the parents. Parents with urgent queries were recommended to call the hospital or visit the physician immediately. Despite the team’s best efforts to provide parents with a personalized, fast, trusted, and reliable information source, the strain on logistics and manpower led the research team to consider the future incorporation of live chatbots in the mobile app. The use of chatbots or conversational agents in the health care setting has been on the rise, especially with smartphone apps, in the areas of patient treatment and monitoring, health care service support, and patient education [47]. Given the novelty of the use of chatbots, there is a paucity of randomized controlled trials evaluating their efficacy in providing knowledge-based information to parents in the perinatal period. However, promising user test results from a few development and usability studies [48-50] indicate the viability of chatbots, thus paving the way for future research.

Second, the recruitment and training of peer volunteers posed a challenge because peer volunteers were also mothers who had work and family commitments; hence, it was difficult to recruit dedicated peer volunteers to take on this yearlong project. In addition, the conflicting schedules of peer volunteers resulted in a few postponements of face-to-face training sessions until the COVID-19 outbreak led to their conversion to web-based sessions instead. Team members in charge of conducting the sessions had to rapidly adapt to the situation and convert the face-to-face training to a web-based session. Research has shown that the efficacy of web-based interpersonal skills (ie, communication, counseling, and interview) training on skills acquisition is on par with face-to-face training [51-53]. Furthermore, participants have also commended web-based courses for their better accessibility and the ease of remote learning [54,55]; thus, a web-based session may better cater to volunteers who are working mothers. However, Doo [56] has argued that cognitive understanding of interpersonal skills does not translate into successful skills execution and interpersonal skills can only be acquired through real-life practice, which is difficult in web-based learning environments. Although web-based training may not be as beneficial as in-person training in terms of building rapport among peer volunteers and conducting role-plays, this was the optimal resolution in times of a COVID-19 outbreak. We are waiting to interview the peer volunteers at the completion of this project about their perceptions of receiving the web-based training. The interviews may provide true evaluation and recommendations on enhancing such web-based training in the future.

As this was intended to be a multisite study, extensive collaboration among health care professionals from various institutions was required. However, a key challenge faced during the Tuckman model’s forming phase of team development [38] was the assembling and scheduling of face-to-face meetings with different experts from various health care institutions, given their hectic schedules. As face-to-face meetings were crucial at the initial stage for the research team to get acquainted and share their expertise and viewpoints, meeting dates were scheduled when there was a likelihood of the highest possible attendance. Team members who were unable to attend received the meeting materials and minutes, and any further feedback was communicated through email. After the delegation of the main roles and responsibilities, subsequent follow-up meetings, updates, and discussions were conducted through web-based meetings with respective parties. At the storming phase, given each team member’s field of expertise and the dearth of essential information, it was difficult for the team to agree on the finalized topics and knowledge-based content to be included in the app. However, the team managed to ascend to the norming phase through resolution by constructive negotiation and eventual mutual agreement to include less essential information in the FAQ section than in the knowledge-based content section. During the performing phase, the team managed to work together effectively and cohesively and successfully rolled out the SPA for user acceptability testing. Overall, the Tuckman model of team development highlighted key characteristics and conflict resolution strategies that enabled optimal team development, which future mobile app developmental studies can use.

Another prominent issue concerned the communication between the app developers and the health care research team. Although the research team comprised experts in content development, they were unfamiliar with technical programming terms and were unable to communicate effectively with the app developers on their app preferences such as cloud hosting, data analytics function, and security mechanisms of the app. This communication difficulty resulted in different expectations for the app and led to last-minute enhancements that incurred additional costs that were not initially considered. Therefore, both parties should be careful with the use of technical jargon to avoid confusion and to constantly clarify any uncertainty or details. Alternatively, visual diagrams (eg, mind maps, drawings, and sample screenshots) could be useful in visualizing and conveying expectations of specific mobile app features.

Much can be learned from the intervention development process of the SPA based on the aforementioned challenges. A major setback was the sudden COVID-19 outbreak, which interrupted the development of the intervention and delayed the project. The flexibility and quick adaptability of the team in finding solutions was crucial in ensuring the continuity of the project. Although face-to-face meetings were impossible, the team was able to conduct meetings through web-based video calls and follow-up emails. Face-to-face training for peer volunteers was also revised and conducted on the web. Team cohesiveness and adaptability were crucial in effectively managing unforeseen circumstances.

Team cohesion is dependent on open and clear communication among team members to ensure that everyone’s expectations are aligned, uncertainties are clarified, and there is transparency for group decision-making. This is especially crucial between the research team and the app developers because their areas of expertise differ and miscommunication in expectations could lead to unfavorable outcomes. However, the interdisciplinary, team-based approach ensured the success of the app development in this study. In addition, a structured and effective workflow was implemented through strict planning of timelines, delegation of responsibilities to team members, close updates and timely feedback among team members, and the charting of project milestones using the Gantt chart. The physical visualization of the project timeline kept everyone in check and allowed better handling of delays.

Another critical lesson learned was to ensure sufficient financial and time budgeting for last-minute enhancements of the intervention. Because of project delays during the COVID-19 pandemic, the team could not spend more time and money for potential feature enhancements that may influence the detailed data analytics. As the basic infrastructure for the parenting app has already been established and the team members have acquired sufficient experience, incorporating future improvements such as adding additional features to the platform would require much less time than that needed for the initial project. Future studies developed as part of this SPA project could assess actual app use, cost-effectiveness, impacts on longitudinal parental and infant outcomes, and qualitative user experience. The use of chatbots to answer parental queries on a timely basis could also be incorporated and evaluated in future research on supportive apps for parents.

This study presents a multistage iterative development process of an mHealth SPA using the ISR framework. Rooted in social cognitive theory and attachment theory, the SPA offers parents a holistic means of obtaining informational and emotional support from knowledge-based content, parents with similar backgrounds, trained peer volunteers, and health care professionals. However, the development process had its challenges because the design cycle involved other logistical and team communication issues that were not taken into consideration in the ISR framework. These issues can potentially be resolved with adequate time and financial budgeting and by developing team cohesion.