The first aim is to examine the feasibility and acceptability of a 12-week FMS intervention delivered through a mobile app to parents and children.

The second aim is to test the hypothesis that a 12-week FMS intervention delivered through a mobile app will improve children’s FMS, compared to the unstructured PA (UPA) app comparator group.

The first exploratory aim is to test the hypothesis that a 12-week FMS intervention delivered through a mobile app will improve children’s PA levels, perceived movement competence, and academic readiness (ie, self-regulation skills), compared to the UPA app comparator group.

The second exploratory aim is to test the potential mediating or moderating effect of FMS on changes in exploratory outcomes.

The third exploratory aim is to test the hypothesis that the effects of the FMS intervention will be sustained through week 24.

Parents received detailed information about the study and provided written informed consent. Eligibility measures were also completed. Parents completed a weekly availability worksheet to identify available times and opportunities to practice lessons with their children using the mobile app at least 5 days per week. The preschool child performed the TGMD-3. TGMD-3 assessments were scored by a trained coder before the baseline visit to determine eligibility. The child was fitted for a hip-worn accelerometer (ActiGraph GT3X+BT), and parents were asked to have their child wear the accelerometer for 24 hours/day for 7 days while the parent completed a wear-time log. The parent was present when the child was fitted for their accelerometer to assist in appropriate placement while at home; parents were also given informational handouts on how to complete the parental log and when and how to take the monitors off.

The baseline visit was scheduled within 2-3 weeks of the screening visit. Parents returned the accelerometer at the baseline visit. If the child did not have adequate accelerometer wear during the time period, the parent was asked to have the child wear the accelerometer for an additional 7 days. Parents completed questionnaires at the baseline visit, which included parent and child demographics and child self-regulation. A trained research assistant measured the child’s height and weight and conducted the Pictorial Scale of Perceived Movement Skill Competence for Young Children (PMSC) [38,39].

After the collection of baseline data, the trained research assistant revealed randomization from an opaque envelope provided by the statistician. The trained research assistant helped download the PLAY study mobile app onto the parent’s smartphone and selected the FMS version or UPA version based on randomization. The trained research assistant familiarized the parent with the PLAY study mobile app by using the first weekly lesson as an example, showing the videos and explaining the points system.

The parent and child returned for 2 follow-up visits at 12 weeks (end of intervention) and 24 weeks (follow-up) after baseline. The parent was mailed the accelerometer and wear-time log 2 weeks ahead of these visits to provide adequate time for the child to wear the accelerometer for 7 days. Parents returned the accelerometer in-person for each visit. At both follow-up visits, parents completed questionnaires including child self-regulation and acceptability. Anthropometry, PMSC, and TGMD-3 were conducted with the preschool child and in that specific order. A trained research assistant deleted the PLAY study mobile app from the parent’s smartphone at the week 24 visit.

The PLAY app was developed by research scientists with expertise in FMS, developmental psychology, and behavior change and programmed by CyberFision, an app development company. Because the app is web-based, the app can function on a smartphone, tablet, or other mobile internet-enabled device. Parents in both conditions downloaded the PLAY study mobile app. To standardize the appearance and usability of the mobile app across the two conditions, one mobile app was created to house all components of the intervention, with specific features turned on for each condition (ie, parents of the FMS intervention did not see the UPA intervention components and vice versa).

Parents in both conditions were instructed to first read the lesson each week, which detailed the purpose and goals for the activity breaks, and then to have their child perform the respective activity break (which was either the FMS or UPA suggested activities). Parents were asked to have their child engage in the activity breaks for at least 12 minutes/day, 5 days a week, for 12 weeks. This resulted in a total of 720 minutes of time directed toward either the FMS or UPA activity breaks. If the child was unable to obtain 12 minutes in one bout, the parent was asked to obtain that total amount within the day (eg, two 6-minute sessions).

One-way SMS or text messages were sent 5 times each week via the mobile app to prompt the parent to read each week’s lesson (1 time/week) and to prompt the parent and child to engage in the activity break (5 times/week). Parents selected times at their screening visit of when they wanted to receive these reminders. Additionally, a reinforcement schedule was built into the PLAY study mobile app by a point system in which the child pressed a star on the screen for each day that they performed the activity break (5 stars available for each week). Parents were encouraged to reward their child with non-food–based rewards such as small toys or non-tangible items such as a high-five. A new lesson was available every 7 days for 12 weeks for 12 lessons total. Lessons were unlocked regardless of the number of activity breaks completed; parents had access to previous lessons but were unable to access future lessons until the respective week. Research assistants also monitored parents’ usage of the app weekly by monitoring whether the parents reported that the child engaged in at least one activity break through the star reporting system. The research assistants telephoned the parent when this information was missing to inquire if there were technical difficulties accessing the app.

The PLAY app provided parents in the FMS condition with brief instructional lessons, peer modeling videos of FMS, and peer modeling videos of activity breaks to deliver targeted, structured FMS instruction time. The FMS condition was built on a curriculum focused on 6 key FMS (3 locomotor, 3 object control) that were selected to be challenging but developmentally appropriate (hop, throw, slide, kick, jump, catch; each repeated twice). Activity breaks were developed and tested previously in a preschool-age population, demonstrating effectiveness to improve children’s motor skill competence with good adherence [40,41]. Each week’s lesson included 6 videos; 5 videos (approximately 5 seconds in length) showed preschool-aged children performing different aspects of an FMS (eg, stepping with opposition, bending knees). These series of videos were intended to help the parent select components of the FMS to focus on to shape the child’s skill towards mastery [42]. A final video each week (<1 minute in length) showed children demonstrating activities for the activity break, which included activities that reinforced the week’s selected skill. This final video was narrated by children.

For example, on Week 1 Day 1, the parent received a notification to open the app and access the first themed lesson, “Hop.” The parent was asked to read a brief instructional lesson about the targeted FMS (including a description of a proficient “hop”), view the brief videos of the demonstrated FMS and suggested activity break activities, and then engage in a 12-minute activity break designed to help the parent model the skill and provide the preschool child with practice for the targeted FMS. Given the complexity of each FMS, parents were given 5 videos of the subcomponents of completing each FMS and asked to practice the progression that fit the child. For example, for overhand throwing, the child would practice winding up, stepping forward with the opposite foot, following through across their body, and aiming either for distance or accuracy. Therefore, the child and parent had video examples of child peers completing these FMS subcomponents. The goal was to obtain a total exposure of 720 minutes of directed instruction over 12 weeks, a dosage that aligns with prior interventions that effectively improved children’s FMS [35,43].

Behavioral scaffolding [44] and social cognitive theory [44,45] informed the FMS approach. Behavioral scaffolding is a cognitive learning approach to problem solving that allows children to master skills beyond his or her current ability [46,47] and was implemented through video segments of children performing motor skills that increase in complexity. According to social cognitive theory, for modeling to effectively elicit behavior change, the child must undergo a process involving attention, retention, reproduction, and motivation [44,45]. This was accomplished through viewing of peers performing and narrating skills (attention), repeated practice and reminders (retention), replication of modeled activities (reproduction), and perceived competence and video and text files providing intrinsic and extrinsic motivation, respectively (motivation).

Parents in the UPA (defined as child-led free play) condition had access to the UPA lessons and UPA activity breaks to promote the equivalent amount of UPA time for the child (ie, 720 minutes over 12 weeks). The lessons were adapted to be developmentally appropriate for preschool children using a curriculum previously developed and tested based on social cognitive theory [30]. This comparator arm was selected as it has shown to increase children’s PA levels [30] but does not explicitly target FMS or provide structured lessons to parents on how to model these skills. The following 6 topics were covered: setting goals, making time for child’s free play, being active indoors and outdoors (eg, dancing), reinforcing PA, reducing sedentary behaviors (eg, screen-time), and parental co-participation. The activity breaks provided specific strategies to encourage the child’s UPA (eg, take your child to the park or outside, use your phone alarm to remind your child to be physically active) with an accompanying video narrated by an adult that read aloud the lesson and featured photos depicting images that aligned with the lesson (eg, a photo of a park). There were no children featured in the photos so that the UPA group did not receive any peer modeling. Features of each condition may be found in Table 1.

Feasibility was measured as adherence to the mobile app intervention, including objective data on the number of lessons, videos, and activity breaks accessed, and self-reported frequency of interaction with the app through the reinforcement schedule (ie, star reporting system). These data were provided on an ongoing basis by CyberFision. Acceptability was captured at weeks 4, 8, and 12 with an in-app parent survey that assessed satisfaction with the intervention in 4 domains (overall satisfaction, helpfulness, ease of use, and perceived change in child’s motor skills) using a Likert-type scale. The System Usability Scale [48] and a satisfaction scale [49] were completed by the parents at the week 12 and 24 assessment visits. Additional usage data were captured via weekly parent surveys embedded within the app: the context of the activity breaks including the parent’s location, the child’s location, and who participated (weeks 1, 5, 9); barriers to performing the activity breaks and the use of equipment during the breaks (weeks 2, 6, 10); and engagement with the mobile app (weeks 3, 7, 11).

FMS were assessed using the TGMD-3, a direct observation assessment used with children aged 3-10 years [37]. The TGMD-3 is a process-oriented and product-oriented assessment to evaluate FMS performance in 2 subscales: locomotor (run, gallop, one-legged hop, skip, jump, and slide) and ball skills (two-hand strike, one-hand strike, catch, kick, dribble, overhand throw, and underhand throw). By design, these include skills targeted in the intervention. Previous research has demonstrated that the TGMD-3 is a valid and reliable assessment tool [50,51]. Assessments were filmed in an open space and coded by trained research assistants blinded to the purpose of this project. An FMS expert (EKW), who previously attained 99% reliability coding with the author of the TGMD-3 assessment, coded the administrations and was blinded to the experimental condition of the participants.

To measure PA, the child was asked to wear an Actigraph GT3X+BT accelerometer for 7 days on the right hip, which has been previously validated in preschool children [52]. Minimal wear time was accepted as 4 days at ≥10 hours/day (≥1 weekend day) and 15-second epoch length [53] were used. Cutpoints by Pate et al [52] were used to classify moderate and vigorous PA, and cutpoints by Evenson et al [54] were used to classify sedentary time.

The PMSC [38] was used to examine the child’s perceived movement competence on the 13 skills assessed with the TGMD-3 and took approximately 10 minutes for the child to complete. Previous research has established validity and reliability with this scale in this age range [39].

Self-regulation skills were reported by the parent using the Devereux Early Childhood Assessment, 2^nd Edition, which is a 38-item proxy report with good validity and reliability to measure self-regulation and behavioral concerns in children aged 3-5 years [55].

For anthropometry, height and weight were measured using a stadiometer and portable scale, respectively, without shoes, and recorded to the nearest 1.0 cm and 0.1 kg, respectively. A third measurement was taken if the 2 measurements differed by more than 0.5 units. BMI z-score was calculated [56].

To collect sociodemographic data, parents reported information on child and parent age, sex, race/ethnicity, parental education, family structure, childcare or away-from-home care, home environment [57], and household income. To collect health behavior information, parents reported the child’s screen use, sleep, and prior experience of the preschool child and parent with mobile apps. The parent also completed a short questionnaire related to child diet, including intake frequency of specific food groups (high fat foods and sugary drinks), that has been previously validated in this age range [58].

Initial results for feasibility, broken down separately for number of lessons and videos accessed, activity breaks accessed, and frequency of interaction, will be expressed using contingency tables. Rates, based on a generalized linear model with a Poisson distribution, will be used to both adjust for covariates and compare the 2 versions of the app (FMS vs UPA). Each of the 4 domains of acceptability will be treated independently and considered as a continuous response, analyzed with a linear model for covariate and app effects.

FMS results will be estimated using a linear mixed effect model with the baseline and follow-up scores at week 12 (end of intervention) and week 24 (end of study) as the outcome. This model will use random effects to account for the correlation in a participant over time. Results from the model will be reported as least square means, with P values based on t tests.

PA levels, perceived movement competence, and self-regulation will be analyzed similarly as for Aim 2. FMS will be tested as a potential mediator and moderator. Mediations will investigate how the covariates affect the relationship between the condition and each outcome variable (for the exploratory analyses, these include PA levels, perceived movement competence, and academic readiness or self-regulation). Mediation analyses will be applied using both structural mean models and principal stratification [59].

Sustained effects will be estimated using a similar linear mixed model as described in an earlier section. However, time will be considered a continuous covariate in order to estimate the rate of sustained effect. This allows the interaction of app and time to test for rate differences between the 2 apps on outcomes of interest through the follow-up at weeks 12 and 24 (separate models conducted for each dependent variable measured continuously over time: FMS, PA levels, perceived movement competence, and self-regulation).