Brenner mFIT includes all components of the standard Brenner FIT program and 6 mHealth components that are designed to target theoretically supported constructs (Table 1). The 6 mHealth components are (1) a mobile-enabled website, (2) dietary and physical activity tracking and a physical activity tracker, (3) caregiver podcasts (n=12), (4) animated videos (n=6) for adolescent patients, (5) interactive messaging (between the participants and clinical staff), and (6) tailored clinical feedback. All Brenner mFIT components are delivered by the clinical staff. The components are outlined individually as follows.

A mobile-enabled website (Multimedia Appendix 1) developed for the project and accessible via a telephone, tablet, or personal computer serves as a central hub for the materials. Podcasts, animated videos, tracking summaries (for the parent and child), goal setting, and messages are delivered via the website utilizing an application program interface that integrates data from a commercial app (Fitbit) with data entered by participants.

A wearable tracking device, chosen by the study team after consultation with caregivers and patients from previous work, is utilized for physical activity self-monitoring. Caregivers and adolescents are instructed to download the Fitbit app to their mobile devices, and adolescents are given a Fitbit Inspire HR activity tracker which synchronizes with the Fitbit app. Mobile phone apps have been shown to facilitate increased self-monitoring [25], which supports tracking of progress on behavioral goals (a component of Brenner FIT) by the youth and reporting to the caregiver and clinical team. Tracking app data are integrated into the mobile-enabled website.

Dietary behavioral tracking is supported by the mobile-enabled website. This strategy was chosen over the use of commercially available apps because no commercially available apps are congruent with the goals of the Brenner FIT program. The mobile-enabled website lets caregivers and youths track their goals in a manner that is appropriate for their role in the weight loss journey. For example, parents are asked to set goals related to providing a healthy food environment (eg, provide 5 family dinners this week), whereas youths are asked to set goals related to their behaviors (eg, eat 5 dinners with my family).

Caregiver podcasts tell a story of a caregiver of an adolescent Brenner FIT patient helping his/her child achieve a healthy weight. Through an engaging story, podcasts will help caregivers deal with the emotions that come with raising an adolescent in treatment for a medical condition (ie, obesity), provide strategies and encouragement to caregivers to help their adolescent build autonomy for healthy behaviors, engage in age-/ability-appropriate physical activity with their child, cook healthy family meals, and provide healthy snacks. Podcasts additionally focus on positive communication, interactions, and challenges often encountered in parenting. Podcasts have been used successfully in 4 previous studies by the team [26-30] and target several constructs from the self-determination theory and social cognitive theory, including (1) outcome expectations, (2) self-efficacy, (3) incentive motivation, and (4) autonomy support. Podcasts will be 5 to 10 mins long each and downloadable via the mobile-enabled website. Caregivers can download and listen to 1 podcast each week in weeks 1 to 12.

Animated videos were created based on feedback from adolescents previously enrolled in the Brenner FIT program. In this study, youths are given access to these animated episodic videos in six 30-second episodes. Animated videos are available to the youth biweekly for the first 12 weeks. Animated videos introduce scenarios encountered by adolescents who are part of a program like Brenner FIT. Each episode follows the youths as they deal with the negative and positive emotions of their weight loss journey and the difficult social situations in which they find themselves engaged. The stories and scenarios contain elements of humor and drama while targeting several constructs from the self-determination theory and social cognitive theory, including (1) observational learning, (2) outcome expectations, (3) autonomy, and (4) relatedness. Similar to the podcasts, the videos are accessible from the mobile-enabled website. The scenarios were developed based on feedback provided by the clinical staff with input from a panel of families enrolled in the Brenner FIT program. Audio for the videos is available in English with subtitles available in English and Spanish.

Peer support is associated with weight loss [31]. Therefore, opportunities are created for peer social support among caregivers and youths on the mobile-enabled website. As many adults do not have social media accounts, we also created a private, secure, moderated message board on the mobile-enabled website. Participants (caregivers and adolescents) are encouraged to (1) post questions to the clinical staff, (2) participate in group discussions and challenges (eg, post a picture of your healthiest meal today), (3) post about physical activities, and (4) share their success stories and challenges. Participants on social media are encouraged to follow each other as well as a curated list of health professionals on social media.

Tailored feedback is provided at regularly scheduled clinic visits by the clinical staff based on a web-based dashboard generated from self-monitoring data. The clinical staff have access to the web-based dashboard that provides a summary of the level of self-monitoring, level of physical activity, and progress on behavioral/weight loss goals for the adolescent and their caregiver. The staff use this information to engage with dyads via regularly scheduled face-to-face meetings and emails to give feedback based on behavioral progress, encouragement to engage in greater self-monitoring, and/or theoretically informed messages to promote self-efficacy, positive outcome expectations, and self-regulatory behaviors.

Acceptability is measured using a brief survey from caregivers and youths to gain insight into their reception of the 6 mHealth components following the completion of the 6-month assessment (ie, completion of the study). Surveys include the system usability scale [32] to inquire about the usability of the mobile-enabled website, tracking apps, Fitbit tracking device, podcast format and content, animated video format and content, and web-based interactions with other participants and the clinical staff.

Feasibility is assessed consistent with the recommendations of Leon et al [33]. Specifically, the team examines screening, recruitment, randomization, retention, adherence, fidelity, and the assessment process. The number of patients screened per month, the number enrolled per month, the proportion of those enrolled who are eligible, and the number who remain enrolled in the study by condition are all tracked. Retention rates (completion of at least 75% of monthly sessions through 6 months) are tracked for each intervention condition. The time needed for assessments is monitored, and feedback on participant burden is recorded. Examples of clinical staff data include increases in clinical staff time interacting with patients and families, costs of delivering clinical staff training (eg, orientation to the technology), time spent in technical support with families related to mHealth components, and hosting/maintenance of the mobile-enabled website. The team developed a brief survey for the clinical staff to gain insight into the benefits and challenges of using the mHealth components and suggestions for improvements. Previous publication documents from the study group have demonstrated good engagement, retention, feasibility, and acceptability of a mobile tracking intervention in youths and their families [34].

Implementation fidelity is assessed to inform a future implementation-effectiveness study. Specifically, activity by the clinical staff and participants is compared against what has been directed (staff) or prescribed (participants). For example, the number of downloads of podcasts/videos, frequency of dietary and physical activity self-monitoring, goal setting, and open rates for web-based messaging from the clinical staff are recorded (Figure 1). The logs of participant engagement with website components are utilized rather than using self-reporting to remove recall bias. In addition, the participant retention rate is considered as a measure of fidelity (the criterion for retention rate has been detailed in the Trial Sample Size section). These data give indicators of fidelity and dose and highlight areas for improvement in future studies.

Accelerometer data are collected at a 7-day interval at baseline and again at 6 months using ActiGraph GT3X+ accelerometers. The accelerometers are set to collect data in the raw format and analyzed by converting the data to 1-second epochs to account for the intermittent and sporadic nature of children’s physical activity [35] and to improve the ability to capture the transitory physical activity patterns of children. During each 7-day data collection period, these monitors are affixed to a wrist strap on the participant’s nondominant hand. Research by our team and others suggest that compliance is greater when the accelerometer is worn on the wrist, and this placement also allows for comparison with the National Health and Nutrition Examination Survey normative data collected at the wrist beginning in 2011 [36-38]. The cut points are distilled in a manner consistent with emerging research that allows for comparability with waist-based estimates [39]. Participants are instructed to wear the monitor continuously over the next 7 days except during bathing and sleeping. Once an accelerometer is received, data from the accelerometer are downloaded using the manufacturer’s software and a USB computer interface. The accelerometer data are reduced, cleaned, and processed using custom scripts in Stata (StataCorp). The accelerometers are initialized and set to record beginning at 5 AM the day following their distribution. This provides 7 days of data for analysis, from which 4 or more complete (>10 hours) days will be extracted. Participant’s accelerometer data are included in the analyses if 4 days with 10 hours of data are available.

Diet is assessed using the National Cancer Institutes’ automated self-administered 24-hour (ASA24) dietary assessment tool (version: ASA24-2016) on 3 nonconsecutive days (including 1 weekend day). The ASA24-2016 is available on both computers and mobile devices and takes approximately 25 min to complete. All youth participants complete one ASA24 at their baseline orientation and at 6-month follow-up office visits to encourage compliance while allowing youths and caregivers to ask questions and gain comfort with using the system. The project coordinator contacts participants within 6 days (on a randomly selected day) following the office visits to inform participants that they should complete the second/third dietary recall and offers to conduct the recall over the phone if the youths need assistance (the ASA24 has been shown to perform equally well as a self-administered or interviewer-administered questionnaire) [40,41]. A series of follow-up reminder messages and/or phone calls are made by the coordinator to encourage completion of the second/third of the two recalls if not completed within 24-hours following the first contact after the clinic visit. Dietary data are collected at baseline and 6-month follow-up.

Family and individual (youth and caregiver) constructs are assessed to identify potential mediators or moderators of observed effects of the intervention; these are incorporated into a larger trial in the future. Brenner FIT uses the family systems theory as a guiding model to address child and parent behaviors within the context of their family [42] and presently uses a number of scales to assess families participating in Brenner FIT. Specifically, Brenner FIT families complete the Family Assessment Device General Functioning subscale (capturing family function) [43], Olson’s Family Communication Scale [44], perceived stress [45,46], self-efficacy for physical activity, impulsivity, and health behaviors of the family [47]. In addition, the team administers scales to capture social cognitive theory and self-determination theory constructs targeted by the intervention that are not captured as part of the intake process. These include outcome expectations, autonomy, autonomy support, and relatedness [48]. All psychosocial data are collected at baseline (before randomization), at 3 months, and at 6 months. Details regarding the variables collected, their theoretical rationale, and measurement instrument are presented in Table 2.

The caregiver and youth participants’ age, sex, race, and ethnicity are collected via a parental self-report upon enrollment in Brenner FIT. The youth participants will also self-report their gender, race, and ethnicity during baseline data collection. The weight status of caregivers and youths are quantified through the calculation of BMI derived from the measurement of height and weight at the intake and follow-up visits. Both height (SD 0.1 cm) and weight (SD 0.5 kg) are recorded twice, and the values are averaged to produce the final value. BMI is calculated as kg/m². Height and weight are measured without shoes in normal clothing. The BMI z-score and percent of the 95th percentile BMI are calculated using the Centers for Disease Control and Prevention growth charts.

Sample size was determined relying on guidance from Eldridge et al [59] and Whitehead et al [60]. Specifically, we based our sample on a feasibility outcome (retention) and the ability to estimate parameters that will inform future trial power calculations. Given a sample size of 40 (per group), there is an anticipated range of possible precision for the retention probability estimates (CIs) as follows: a true retention rate of 60% yields a 95% CI of 55% to 75%; for a retention rate of 90%, the CI is 81% to 99%. The observed CIs provide a plausible range for the true retention rate in our population and trial conditions. If the CIs exclude values <80%, this would be an acceptable retention rate to plan a large-scale trial of Brenner mFIT; regardless, these estimates (CIs) are weighed heavily and, together with many other components of the trial experience, are used to determine overall acceptability and feasibility, and along with estimates of effect size, these estimates will assist with the design aspects of any future trial.

This study will be informed by emerging literature on the scalability of health promotion programs [61,62]. Specifically, the several scalability considerations identified by Milat et al [61] are evaluated, such as (1) reach and adoption, (2) organizational resources required (including costs), (3) intervention delivery (eg, acceptability, fidelity), (4) contextual factors (eg, interaction of the intervention with organizational contexts), and (5) evaluation approach. For example, we will characterize the features of the clinic and its operations (eg, equipment, staffing, costs, reimbursement) to identify areas of potential variability to inform future rollout. The data collected provide answers to crucial questions suggested by Klingner et al [63]: (1) Under what conditions and with whom does Brenner mFIT work? (2) What is necessary to support clinical staff implementation of Brenner mFIT? (3) What is necessary to enhance the capacity of clinics to support staff in implementing Brenner mFIT under different conditions? (4) What is necessary to support broad, deep, sustained implementation of Brenner mFIT? This information is crucial in helping us to scale the intervention.