Although childhood obesity rates have been reported to stabilize in many developed countries [1], the prevalence of overweight and obesity in young children is still high. In general, 8% of Dutch children around the age of 2 years were overweight or obese in 2018 [2]. However, these rates were considerably higher among children from families with a lower socioeconomic position (SEP) [2]. In addition, a growing body of research indicates that plateauing prevalence rates are predominantly evident among groups with a higher SEP [3]. In groups with a lower SEP, the rates are still rising [4], implying an increase in socioeconomic health disparities. To reduce SEP disparities in childhood obesity, preventive intervention programs should promote healthy energy balance–related behaviors (EBRBs; ie, dietary intake, sleep, physical activity, and screen time) before children have developed fixed patterns of EBRBs [5]. As patterns of EBRBs, such as dietary intake and screen time, develop in the first years of life [6], infancy and preschool (0-4) years present a critical window for childhood obesity prevention.

Parents are considered the main agents of change in effective preventive intervention programs for childhood obesity, especially in the first years of life [7]. Parents’ behaviors, particularly their energy balance–related parenting practices (ie, energy balance–related, specific, discrete, and observable acts of parenting [8]), may significantly influence their children’s EBRBs and body weight [9,10]. Recent reviews have shown that promoting responsive feeding guidance to teach maternal awareness and attention to children’s hunger and satiety cues can support normal child body weight development [11]. Moreover, universal preventive intervention programs targeting early feeding and positive parenting skills, including programs that target sleep, show similar effects on the child’s body weight [12]. In general, current evidence supports starting with promoting responsive feeding and parenting during infancy and incorporating the promotion of structure and rule setting in early childhood [13]. For physical activity parenting, support has been found for the potential importance of parental support and parents’ own behaviors and role modeling [14,15]. However, it should be noted that only a few preventive intervention programs have targeted physical activity parenting [14]. More generally, preventive intervention programs for early childhood obesity that target parenting practices with respect to all relevant child EBRBs are scarce [12]. Moreover, parent-focused prevention programs have been mostly universal (ie, population based) in nature [12,16]. For instance, 3 of 4 parent-focused EPOCH (Early Prevention of Obesity in Children) trials that commenced by early infancy were universal programs [17]. Our parent-focused program to prevent early childhood obesity adds to the literature by addressing parenting practices with respect to all the important child EBRBs [12,18] while simultaneously applying selective prevention to the subgroup of parents with a lower SEP, who generally display more problematic energy balance–related parenting compared with other SEP groups [19].

In addition, our program uses an innovative approach to address 2 frequently reported limitations of traditional (parent focused) obesity prevention programs: the costly, time-intensive face-to-face setting [16,20] and difficulties in reaching people with a lower SEP [21]. To successfully target and reach people with a lower SEP, intervention programs should address both practical (eg, time constraints, lack of transportation, and inflexible working hours [22-24]) and attitudinal (eg, irrelevant and not engaging program elements [25]) barriers for participation [26]. Practical barriers for people with a lower SEP could (at least partly) be overcome by delivering interventions through smartphones, including mobile apps [21,27]. Furthermore, attitudinal barriers can also be overcome because app-based interventions allow for options to increase program engagement through the presentation of bite-sized information in plain language that is accompanied by appealing visuals and through the possibility of personalizing the intervention and the ability for users to monitor their behavior [28,29]. For these reasons, app-based intervention programs have the potential to be cost-effective and reduce the gap in socioeconomic health disparities [29,30]. The apps used in these programs need to be high quality (eg, in terms of engagement, esthetics, and information quality [31]) and encourage behavior change (eg, providing knowledge and information as well as prompting goal setting and planning [32]). To achieve this, app-based interventions should be based on evidence, grounded in behavior change theory, and incorporate the needs and wishes of the target audience through formative research [33].

Therefore, we developed the app-based prevention program Samen Happie! using the Intervention Mapping Protocol (IMP). The IMP is a widely used, standardized intervention planning format that helps intervention developers to incorporate empirical findings from the literature, effective behavioral change methods and their practical applications, and data collected in a representative population [34]. Although it is not specific to app-based program development, the IMP has been successfully applied to the development of digital (including mobile) interventions for youth health promotion [35-41]. The central goal of the app-based preventive parenting program Samen Happie! is to stimulate healthy energy balance–related parenting practices to prevent early childhood obesity among children of families with a lower SEP. In this paper, we inform readers about the development of the Samen Happie! program based on the 6 steps of intervention mapping.

The Dutch title of the app-based parenting program (Samen Happie!) will be used throughout this paper. A possible translation is “Happy Together,” but this does not reflect the play on words the title indicates in Dutch.

The Samen Happie! program was developed through the 6 iterative and nonlinear steps of the IMP: (1) conducting a needs assessment; (2) preparing matrices of change objectives; (3) selecting theoretical methods and practical strategies; (4) developing the intervention program; (5) planning for adoption, implementation, and sustainability; and (6) planning the program evaluation [34]. We collaborated with a workgroup consisting of potential program implementers (ie, youth health care professionals of community health services) and users (ie, parents of young children with a lower SEP). Importantly, for the development and evaluation of the program, we asked only one parent-child dyad per family to participate, thereby focusing on the primary caregiver. In this section, the main tasks of each step of the IMP are explained and, when relevant, the role of the workgroup is exemplified. The outcomes of the IMP, including the choices and actions during each step, are described in the Results section.

The first step of the IMP was to conduct a needs assessment of our target group (ie, parents of children aged 0 to 4 years with a lower SEP) to build a logic model of the health problem [34]. Our needs assessment included a literature search, focus groups with parents with a lower SEP (N=16 mothers in total), and discussions with youth health care professionals (N=6 professionals in total). In the focus groups, the hindering and facilitating factors for healthy parenting and parenting practices in difficult parenting situations were discussed, with a focus on parenting practices regarding food and dietary intake. The focus groups, which were conducted until saturation was reached, were audio recorded, transcribed, and coded for themes and concepts using ATLAS.ti (ATLAS.ti Scientific Software Development GmbH). The discussions with youth health care professionals served to explore key parenting-related themes and issues that existed among the target group and as a sounding board for concrete questions during program development. For example, frequently reported parenting problems, (parental adherence to) national guidelines regarding child EBRBs, and effective, practical strategies to increase healthy child EBRBs through energy balance–related parenting practices were discussed. To build the logic model (based on the PRECEDE [Predisposing, Reinforcing, and Enabling Constructs in Educational Diagnosis and Evaluation] framework [42]), we identified the quality-of-life indicators associated with the health problem, behavioral and environmental risk behaviors for the health problem, and determinants related to these behaviors. On the basis of the knowledge derived from this assessment, we selected behavioral outcomes for the program and formulated the program goal.

In the second step of the IMP, the performance objectives were defined for the behavioral outcomes specified in step 1. These performance objectives constituted behaviors that are expected to contribute to achieving the program goal when performed by the target group [34]. By crossing the performance objectives with the determinants selected in step 1, the change objectives were specified. These change objectives indicate which actions are required to modify the determinants of the behavioral outcomes and reach the performance objectives [34].

The third step of the IMP evolved around the selection of theory- and evidence-based change methods to affect the determinants selected in step 1. We aimed to select a limited number of theoretical methods per determinant as interventions that use a small number of behavior change techniques are generally more effective for people with a lower SEP than interventions that use a larger number of techniques [43]. We then translated the selected methods into practical strategies through which they were delivered in the program [34].

The fourth step of the IMP involved building the program themes and components and drafting, pretesting, and producing the program materials based on the information gathered in the previous steps [34]. The development of the app-based parenting program consisted of 3 phases. In phase 1, qualitative user research was conducted to assess parents’ (N=16 mothers) wishes regarding the content and functionality of the app. In phase 2, a prototype of the app structure, functionalities, content, and visual design was constructed in continuous cocreation with parents (N=4 mothers) and youth health care professionals (N=3) and pretested by parents (N=16 mothers). Finally, phase 3 involved building the final version of the app. For the development of the app, we collaborated with Dio Agency, an agency specialized in deploying software design to facilitate behavioral change.

The fifth step of the IMP involved the identification of potential program users (eg, implementers or adopters) and the design of a program implementation plan [34]. We planned to make the app-based intervention available free of charge after program evaluation; nevertheless, the excess supply of health-related apps minimizes the chances that parents will spontaneously find and download the Samen Happie! app. This is one of the reasons that digital health programs in particular need a delivery system (ie, a program adopter) to get the program to its intended participants [34].

The sixth and last step of the IMP involved the development of a program evaluation plan [34]. We developed both process and effect evaluation plans to evaluate the quality of the implementation and the effectiveness of the preventive intervention program.

Our logic model of the health problem is presented in Multimedia Appendix 1. The model displays the quality-of-life indicators (eg, cardiovascular diseases, depression, and risk of obesity in adulthood [44-46]) as correlates of health problems (ie, early childhood obesity). Moreover, it shows behavioral (eg, child dietary intake) and environmental (eg, unhealthy home environment) risk factors for health problems and determinants related to these factors (ie, parents’ knowledge about healthy dietary intake). The parental determinants impact child risk behaviors through energy balance–related parenting practices. The child energy balance–related risk behaviors, role of parents, and determinants of healthy parenting are described hereafter.

Table 1 presents 3 examples of performance and change objectives for each behavioral outcome (ie, children have a healthy dietary intake, sufficient sleep, and restricted screen time and sufficient physical activity). Multimedia Appendix 2 [10] presents an overview of all performance objectives and change objectives specified for the app-based preventive intervention program.

The theoretical methods we selected for each of the determinants (ie, knowledge, attitudes, self-efficacy, and habits) were derived from the taxonomies described by Kok et al [104] and Michie et al [105]. Table 2 presents an overview of the methods and examples of their practical application. This table shows that we included both methods involving the provision of information (eg, consciousness raising and persuasive communication) and those that target more automatic processes (eg, implementation intentions and self-nudging). Automatic, nonconscious methods might be particularly effective for people with a lower SEP as these methods are less dependent on literacy capabilities [106].

In addition to potentially important theoretical methods, the form in which an intervention is delivered is also a key ingredient in behavior change interventions [107] and might be even more crucial for groups with a lower SEP in terms of comprehensibility and engagement [108]. Considering the literacy capabilities of participants with a lower SEP, research has shown that people with literacy problems can remember written texts more easily when they are supported by (audio)visual aids [109,110]. Therefore, we included both written texts and supporting icons, images, videos, and voice-overs of important information in the app-based prevention program. Moreover, to ensure that the app was comprehensible for parents with varied literacy skills, all textual components were revised by a specialized organization to match language levels A2-B1 according to the Common European Framework of Reference [111]. In addition, the technological features of the app allowed us to tailor the program to individual participants, which might positively impact user engagement [112] and subsequent intervention retention. Illustratively, the app uses data about the name, sex, and birth date of the child (entered by the parent upon registration) to personalize texts with respect to names and pronouns (eg, “Set a good example for Maria. Try to eat healthy when she is around”). Finally, the data on children’s age were used to present parents with developmentally appropriate information via age-based modules (see Step 4: Developing the Intervention Program), which could enhance the perceived relevance of the information, thereby increasing engagement [113].

In phase 1 of the app development process, the results from our focus groups and discussions with youth health care professionals informed the development of the prototype of the app and the way in which we tailored the materials to our target group. For instance, based on parents’ negative affective attitudes toward water consumption, we asked them whether they would provide their children tea (without sugar) as an alternative, and they affirmed that they would do so. Therefore, we drafted the app content with “water or tea,” instead of focusing entirely on water consumption. Other examples of content that we based on the suggestions of parents from our focus groups included providing information about saving money on groceries and tips to stimulate vegetable consumption.

In phase 2, the content of the preventive intervention program was drafted based on national health care guidelines, relevant literature on parenting practices in relation to child EBRBs, and previous intervention projects [103,114]. For example, youth health care professionals suggested using videoclips that were specially designed by the Dutch Child and Family Center (Centrum voor Jeugd en Gezin) to stimulate health literacy among young (low literate) parents. The pretest of the prototype of the app yielded, among others, the following suggestions: the option to enlarge images, the possibility of replaying the instruction clip, a summary of the most important information per level, and the option to reread a lesson. These suggestions were incorporated into the final version of the app.

The final version of the app (phase 3) was launched in September 2018. Figure 1 presents 6 screenshots of the app (in Dutch). Screenshot 1 shows the main menu of the app. From here, parents could navigate to their profile, app settings, and the actual app content. The content of the app was divided into 5 age-based modules ranging from 7 to 12 months (module 1) to 24 to 48 months (module 5). On the basis of their child’s birth date (entered upon registration), parents were granted access to the appropriate modules (eg, parents of children aged 14 months were granted access to the first 2 modules). When the child reached the minimum age of the subsequent module, the new module became unlocked. The content of each age-based module was divided into 6 themes reflecting all relevant child EBRBs (ie, eating, drinking, sleep, and physical activity and screen time) and themes on the well-being of the parent and temper tantrums (this last theme was only included in the modules for children aged 18 months and older).

Each of the 5 modules consisted of 2 types of activities: lessons and challenges. Screenshot 2 in Figure 1 shows the lessons (indicated by the larger icons) and challenges (indicated by the smaller icons with the thunderstroke) of the themes sleep and temper tantrums in module 5. Upon completion of a lesson or challenge, the icon of that lesson or challenge became filled (see screenshot 2 in Figure 1). The lessons and challenges were constructed using different types of cards. Every lesson or challenge started with an introduction card that showed the title and length of the lesson or challenge (in minutes; see screenshot 3 in Figure 1). The lessons consisted of multiple information cards that presented information on that specific theme in an engaging and easy-to-comprehend manner, for instance, through the use of facts (“Did you know...?”; see screenshot 4 in Figure 1), practical examples, tips, and quizzes and supported by icons or pictures. The challenges consisted of exercises that prompted parents to apply the information from the lessons. By employing techniques that tackle (unhealthy) automatic behaviors, parents were encouraged to implement (new) parenting skills as habits. The challenges were similar in their design. Each challenge started with providing information and examples of specific parenting target behaviors. Next, parents were asked whether they thought performing that behavior was important (ie, attitude) and whether they felt capable (ie, self-efficacy) of performing that behavior. On a slider card, they could indicate their responses on a scale from 0 (low importance or self-efficacy) to 5 (high importance or self-efficacy). Appropriate feedback on their responses was provided through a pop-up notification. Finally, parents created a personal goal or action plan in the context of the target behavior on a fill-in card (see screenshot 5 in Figure 1). For each goal or action plan, parents could set reminders to receive a notification at a date and time of their choice to help fulfill that goal (see screenshot 6 in Figure 1). An illustrated example of a lesson and the accompanying challenge within the theme Sleep can be found in Multimedia Appendix 3. The example includes screenshots of different types of cards within the lesson or challenge, a translation of the original Dutch text to English, and a reference to the theoretical method that formed the basis for that card. Moreover, an overview of the 5 modules and the corresponding lessons and challenges (specified per theme) is presented in Multimedia Appendix 4.

We identified national community health services and child day care centers as potential adopters of the app-based program, as representatives of these organizations were involved in the development and execution of the program. The program implementers will be youth health care professionals and pedagogical staff working at these organizations, and their tasks are to bring the app to the attention of parents (whose children they perceive as being at high risk for obesity) and motivate them to use the app. The implementation of the app-based program can support the daily practice of the program implementers in 2 ways. First, the app could function as an addition to usual care, given that youth health care professionals indicated that standard consultations are generally too brief to give parents elaborate, well-rounded advice. In this sense, it is advantageous that the app is an easy-to-use product that does not require detailed instructions from a health care provider. Second, the app could function as an educational material that creates a legitimate opening to discuss topics such as food parenting and body weight [115], which health professionals often find to be difficult topics to address [67]. In addition, the active role of youth health care professionals in the referral of high-risk groups (in terms of obesity and other health problems) and a primary focus on prevention is in line with the national prevention agreement that the Dutch government issued in the first half of 2019 [116]. Moreover, the agreement calls for the inclusion of (potentially) effective (preventive) interventions for childhood obesity to be registered at the Healthy Living Desk, an intervention database of the Dutch ministry of Public Health, Welfare and Sports. To facilitate nationwide dissemination and aid in intervention sustainability, the Samen Happie! program was submitted to this database and accepted in September 2020.

The process and effects of the app-based preventive intervention program were evaluated in 2 separate trials, the designs, eligibility criteria, procedures, and measures of which are explained hereafter.

The need for effective preventive intervention programs for childhood obesity is high, particularly among families with a lower SEP. The app-based parenting program Samen Happie! was developed primarily for these families and aims to stimulate healthy energy balance–related parenting practices from early childhood, before unhealthy energy balance–related habits have been established. More specifically, the program promotes both structured and autonomy-supportive practices and limits coercive controlling parenting practices with respect to all relevant energy balance–related determinants of childhood obesity (ie, dietary intake, sleep, physical activity, and screen time), with the ultimate goal of preventing children aged 0 to 4 years old from being overweight and obese. The successful development of the program was aided by the use of the IMP. The process of stepwise decision making made this large-scale and complex project manageable and contributed to thorough considerations. Regarding the selection of eminent EBRBs, for instance, we initially selected only parenting practices related to the children’s dietary intake but we decided to also include practices relating to sleep, physical activity, and screen time after an extensive literature search and discussions with the target group. Moreover, insights from recent literature reviews facilitated the selection of the most promising energy balance–related parenting practices (eg, parental support and modeling for physical activity). By facilitating a collaboration with experts and the target group, the IMP assured that the intended end users of the program were involved in multiple stages of program development. Overall, by integrating theory, empirical studies, professional knowledge, and the needs and preferences of the target group through continuous cocreation, we increased the chances of developing an effective preventive intervention program for childhood obesity [34].

Besides the use of the IMP, the app-based parenting program has several other notable strengths. First, previous digital preventive intervention programs for childhood obesity focused solely on the sociocognitive determinants of energy balance–related parenting [36]. A unique aspect of the Samen Happie! program is its focus on both sociocognitive (ie, knowledge, attitudes, and self-efficacy) and automatic (ie, habits) determinants, bridging the well-known gap between health parenting intentions and behaviors [88]. Furthermore, formulating long-term goals is potentially too proximal for parents with a lower SEP as their focus lies primarily on surviving in the here and now [119]. By using a self-regulatory planning approach with personally tailored prompts (eg, through implementation intentions), we facilitated the fulfillment of short-term goals. In addition, by assessing both general parenting style and parental mental well-being as potential moderators of the app-based prevention program, we might be able to identify groups of parents who might particularly benefit from the program, which could contribute to more personalized app usage. Finally, the app is an easy-to-use, stand-alone product that can potentially have a significant reach through its opportunities for widespread implementation.

One of the challenges of mobile health interventions is to keep users engaged for longer periods, which is particularly important for interventions targeting behavior change maintenance [120]. Although the preliminary results of the process evaluation indicated that parents appreciated the functionality, design, and content of the app to a reasonable degree, most parents who downloaded the app did not continue their app use over the course of several months. This might indicate that parents’ information needs were fulfilled and new behavior patterns had been developed, explaining that further app use was no longer needed, but it could also suggest that the user engagement of the app should be increased. This is something that future research should consider. We will further develop the prevention program based on the results of the process evaluation and the input of potential program implementers (eg, youth health care professionals of community health services). On the basis of parents’ suggestions for app improvement, the Samen Happie! website [121] has been developed (available in Dutch and English). This website contains more elaborate and structured parenting information and includes a search function, personal accounts of parents, and a forum on which parenting experiences can be exchanged. Future research should examine whether parents’ evaluations of the Samen Happie! program improve by offering them access to both the website and the app.

The app-based prevention program also has some limitations. With respect to the design of the trials, it was not possible to blind both participants and investigators to the allocation of conditions (ie, double blinding). As we used a waitlist control condition, the participating parents knew they would eventually receive an app about healthy parenting. However, neither the participants nor the investigators knew which trial condition the participants would be allocated to before randomization took place. Although double blinding in RCTs is generally recommended, methodological studies have shown that adequate allocation concealment is most important in minimizing bias [122,123]. Moreover, regarding our effect evaluations, we were unable to include in-depth measures of the 4 determinants that were targeted in the program (ie, knowledge, attitudes, self-efficacy, and habits), as it was imperative to keep our questionnaires short for our target group of parents with a lower SEP. Nevertheless, we included some questions that could serve as proxies for parental attitudes and self-efficacy. Even though this will give us some indication of the degree to which our preventive intervention program successfully targeted the selected determinants, future research should aim to include detailed measures of its program determinants to be able to examine the working mechanisms of the program.

In addition, future preventive intervention programs for childhood obesity should consider involving both caregivers. Although we intentionally targeted only primary caregivers (who turned out to be primarily mothers) for the recruitment, program materials, and questionnaires of our program, recent research has indicated that parents within a family differ in the energy balance–related parenting practices they perform [124]. This highlights the need for the inclusion of both parents in future preventive intervention programs targeting energy balance–related parenting. Finally, app-based parenting support might not be sufficient for the needs of parents of preschoolers entering their terrible twos and food neophobic phase [125,126], especially in the case of parents (with a lower SEP) who already experience parenting problems. To address the more extensive needs of these parents, future research should explore a combination of online tools with additional offline (group based [127]) counseling, which could provide a promising approach to change parenting attitudes and behaviors [128].

In conclusion, the IMP allowed for effective development of the app-based parenting program Samen Happie! to promote healthy energy balance–related parenting practices among parents of infants and preschoolers. By applying the IMP, including continued cocreation, the program specifically addressed the needs of parents with a lower SEP through a tailored program content and through theory-based behavior change techniques. This increases the potential of the program to prevent the development of obesity in early childhood among families with a lower SEP.