A total of 200 Canadian children who have overweight or obesity and one of their parents will participate in the trial, taking place in 2019-2020. The majority of the participants will be recruited for a study funded in part by the Public Health Agency of Canada with matched funds from partners. A portion of the participants (n=60) will be recruited as part of a study embedded within a Canadian Institutes of Health Research (CIHR) Team Grant in Bariatric Care (Team ABC 3) (27). After consent and baseline measurements, families will be randomized into one of two conditions: 1) intervention or 2) waitlist control. Families randomized into the intervention condition will receive the Aim2Be app with health coaching (i.e., receive tailored messages from a health coach with the option of scheduled and unscheduled text support). All participants will receive a package containing a scale, measuring tape, activity tracker (Fitbit), and a brochure with current health recommendations. At 3 months, families randomized to the waitlist control conditions will move to the intervention for months 3 to 6, but they will not be supported by the health coach. Data collection will occur at baseline, 3-months, and 6-months utilizing REDCap (Research Electronic Data Capture)(28, 29) hosted at BC Children’s Hospital Research Institute (BCCHRI). Staff involved in the data collection and analysis will not be involved in delivering the intervention. Ayogo Health Inc and the Childhood Obesity Foundation (COF) will deliver the Aim2Be intervention. The BCCHRI / School of Population of Public Health (SPPH) at the University of British Columbia (UBC) will collect and analyze the data. The study flow chart is shown in Figure 1.
(INSERT FIGURE 1 HERE)
Computer-generated (www.sealedenvelope.com) randomization schedule will be used to allocate participants in blocks of four, six, or eight participants with a randomization ratio of 1:1 (30). The allocation schedule will be concealed in the randomization module of REDCap and only assigned after informed consent and baseline assessments have been obtained. Research team members will not enter or modify the allocation schedule as it will be entirely computer generated. Participants will not be blinded to their allocation conditions as participants will know whether they receive the Aim2Be intervention immediately or in 3 months. Allocation assignment will be concealed to the researchers at the analysis stage.
Inclusion and exclusion criteria. Child participants must be 10-17 years old and literate in English, able to read at the grade 5 level or above, and have a mobile phone or a computer with Internet access at home. Child participants must have either overweight or obesity, based on the World Health Organization cut-offs for children and adolescents aged 5 to 19 (BMI > 85th percentile). Parent participants must be the caregiver with whom the child primarily lives, and must be literate in English.
Children must not have: diagnosis of any musculoskeletal, cardiovascular, pulmonary, or orthopedic problems; disabilities precluding the participant from being physically active; any other physical condition that precludes the participant from being physically active; diagnosis of anorexia nervosa or bulimia nervosa; diagnosis of type 1 diabetes; dietary restrictions or special diets that limit a participant’s ability to eat a variety of foods; simultaneous participation in another physical activity, nutrition, or weight management study/program; use of medication, nutritional supplements, or herbal preparations to help lose weight; pregnancy; or a history of psychiatric problems or substance abuse which would interfere with adherence to the study protocol.
Recruitment. Participants will be recruited using two main methods: clinical site referral and social media (i.e., Facebook advertisements with a link to study site). Seven clinical weight management programs across Canada are participating in this study and include: BC Children’s Hospital (Vancouver, BC), Alberta Children’s Hospital (Calgary, AB), Stollery Children’s Hospital (Edmonton, AB), Misericordia Community Hospital (Edmonton, AB), McMaster Children’s Hospital (Hamilton, ON), The Hospital for Sick Children (Toronto, ON), and Children’s Hospital of Eastern Ontario (CHEO) (Ottawa, ON). Clinical sites will use a combination of mailing and e-mailing invitation and information packages as well as telephone contact. All potential participants will be provided an invitational package that describes the study and includes copies of the consent and assent forms, as well as a link to the study website.
For both groups of potential participants, interested individuals will be asked to complete a secure, online consent to contact form in REDCap and provide their contact information. Individuals who express interest in the study will be screened via telephone for eligibility.
Consent/assent process. All participants will provide informed consent or assent prior to taking part in this study. All parent participants, with the exception of parents of children 16 years of age or older recruited from CHEO, will be asked to sign a consent form consenting to their own and their child’s participation in the study. Parents of children 16 years of age or older recruited from CHEO will be asked to sign a consent form agreeing to their own participation in the study. All child participants, with the exception of those 16 years of age or older recruited from CHEO, will be asked to sign an assent form agreeing to their own participation in the study. Child participants 16 years of age or older recruited from CHEO will be asked to sign a consent form agreeing to their own participation in the study.
Sample size calculation. Based on previously published data (31), the RCT will have 80% power at an alpha of .05 to detect a .5 decrease in BMI z-score when the sample size in each group is 60. BMI z-score was used to calculate sample size as it is the most difficult variable to change, and requires the largest sample size of all of the primary outcomes. Based on this and accounting for both missing data and attrition, we project needing at least 80 families in each condition. However, to be able to detect a 20% difference in adherence (e.g., secondary aims outcome) between the two intervention groups (odds ratio of 2.33) at an alpha of .05 with 80% power using a one-sided t-test, 77 families are needed in each group. As the waitlist control group will receive the intervention at 3 months, we require 100 participants enrolled in each group (this accounts for 15% attrition from baseline to 3-month and that about 8% of the families will not download the Aim2Be app). Power calculations were conducted in nQuery (Statsols, USA). Based on these computations, we aim to enroll 200 participants in the RCT.
Development and description of Aim2Be
Aim2Be development. Aim2Be, which will be evaluated in this RCT, was built on the foundational knowledge learned in the first generation of the program called LiGHT (Living Green Healthy and Thrifty) and the second generation of the program, called Aim2Be version 1 (v1). LiGHT was an 11-module online program that integrated lifestyle behaviour modification principles with environmental and financial concerns to address childhood obesity among 10-17 year old children and their families (32). Modest and non-significant improvements in readiness to change were initially observed for both the child and parents, likely due to ceiling effects on the measure and small sample size (N=17 child-parent dyads). However, the qualitative feedback (N=10 child-parent dyads) suggested that the LiGHT approach held promise. Specifically, both children and parents indicated that they related strongly to the foundational approach of the LiGHT intervention, meaning they liked having behaviours linked with health, environmental, and financial concerns. The evaluation also uncovered key areas that needed further improvement including making the program more visually appealing, adding more interactivity, and creating a sense of community for the users.
The second generation of the intervention transitioned from a web-based intervention to a mobile app (iOs and Android) and was renamed Aim2Be. For those who do not have a smartphone, Aim2Be has remained accessible as a web-based platform via the internet. Aim2Be became a gamified app that supports children and their families to initiate sustainable behaviours in four primary areas: healthy eating, active living, reducing screen time, and healthy sleeping habits. It retained its focus on linking behaviours with health, and and living green and economically as well as adding emphasis on healthy body image and strong self-esteem. While Aim2Be retained some core elements from the LiGHT program, it: 1) became strongly grounded in theories of behaviour change (i.e., being family-based; supporting the development of self-regulatory skills at both the individual and familial levels; and enhancing self-efficacy through graded tasks)(10, 11, 19, 33); 2) integrated principles of maintenance of health behaviours and as such it included self-regulatory processes that support intrinsic motivation (34-38); 3) integrated gamification practices – designed to maximize for both enjoyment and motivation (39); 4) recognized the importance of body image and self-confidence in lifestyle management, 5) used the mHealth context to support self-regulatory processes (40); 6) aligned with best clinical practices and guidelines in three areas: a) the treatment of childhood obesity in Canada as it aligns with the curriculum of Canadian Weight Management programs (e.g., being family based, multi-behavioural, and focus on supporting skill development at the individual and familial levels); b) clinical guidelines for the management of childhood obesity as it integrates the central elements that need to be incorporated in these interventions (e.g., family focused, focused on improving lifestyle behaviours, skill based support) (41); and c) Canadian health recommendations including the Canadian 24 Hour Movement Guidelines (42) and Canada’s Food Guide (43); and 7) was designed specifically with and for children and their families (see details below).
Aim2Be was iteratively and incrementally developed. The first version of Aim2Be (v1) was field tested for 4.5 months among 301 teens 14-17 year and 315 parents. The quantitative evaluation revealed that teens who were moderately and/or highly engaged in the app (>30 min of app usage) as compared to those with low engagement (≤30 min of app usage) significantly increased their motivation and self-efficacy to improve their dietary habits (i.e., limit sugar-sweetened beverages and increase intake of fruits and vegetables) and sedentary behaviours (i.e., limit screen time). At 4.5 months, children significantly increased their previous day intake of fruits and vegetables, decreased their consumption of 100% fruit juice, and reduced their screen time. In addition, parents who were more engaged with Aim2Be v1 reported significant improvements in their dietary behaviours (i.e., increased intake of fruits and vegetables and decreased intake of sugar-sweetened beverages). Additionally, multiple rounds of qualitative evaluations were conducted among 36 teens/pre-teenss and 24 parents. The qualitative evaluations included both focus groups and 2-week prototype testing followed by semi-structured interviews. The results of these evaluations led to numerous improvements in Aim2Be including: clarifying the overall purpose of Aim2Be, supplementing the tracking and check-in sections, adding more engaging features, and syncing the app with physical activity monitoring (i.e., Fitbit).
The conceptual framework of Aim2Be is shown in Figure 2. At its core, the behaviour change techniques (BCT) incorporated in Aim2Be are rooted in: 1) Social Cognitive Theory (SCT) (44); 2) the Player Experience and Need Satisfaction (PENS) model – an extension of the Self Determination Theory model for the gamified context which incorporates enjoyment to support intrinsic motivation and the basic psychological needs that promote intrinsic motivation (autonomy, relatedness, and competence/self-efficacy) (39, 45); and 3) the ACUDO framework (46). The ACUDO framework, is a best practice framework developed by Ayogo Health Inc. (46) to promote engagement and enjoyment by: a) supporting Agency, b) incorporating Challenges, c) infusing Uncertainty, d) supporting self-Discovery and exploration, and e) adding fun Outcomes such as rewards. Specifically, Aim2Be includes strategies that target: 1) Gamified mediators of behaviour change to ensure the experience is both enjoyable and engaging; 2) Behavioural mediators that aim to activate self-regulatory skills and support the development of intrinsic motivation and increase self-efficacy as a way to support behaviour change and ultimately support healthy body weights; and 3) Environmental mediators as it is recognized that behaviour change needs to be physically and emotionally supported in the familial environment and that social support from both peers and/or coach is also important to changing health behaviours (see Figure 2).
(INSERT FIGURE 2 HERE)
Aim2Be features. Aim2Be is not a prescriptive app, but instead it utilizes strategies to activate self-regulatory skills through a self-guided journey of self-discovery of current health behaviours and autonomy to select which health behaviours the user wants to work on. Strategies employed by the various app features are grounded in Michie’s behaviour change taxonomy which specifies the “active ingredients” of behaviour change interventions (47). Children will be subdivided into two versions of the app based on their age: teens (13-17 year olds) and pre-teens (10-12 year olds) and will be able to select which Aims they wish to address. Once an Aim is selected, teens start a stage and pre-teens start the Aim as a whole, and then are provided tasks to help them set incremental goals, plan, and self-monitor their behaviours. In the application environment, the user progresses along their journey by completing games, tasks and activities including quizzes. The user receives currency that they can use to unlock collectibles and ‘choose your own adventure’ stories. Users are provided with tools within the app to further support their journey including Check-Ins, Quick Wins, Articles, Quizzes, and a moderated Social Wall with daily topics to help engage users. Features that are incorporated in the pre-teen, teen, and parent apps, as well as the BCT (47) targeted by these features, are summarized in Table 1.
(INSERT TABLE 1 HERE)
Intervention condition participants will have access to the full version of the Aim2Be app (as described in Table 1) for six months. In addition, participants (teens and parents) will have the option to access a health coach who can support them in health behavior change using motivational interviewing techniques. The health coach has training in motivational interviewing and expertise in lifestyle management and working with families. The health coach will communicate with participants through the in-app text feature and participants have the option to schedule a live text session if they wish to. The health coach will send an initial contact to all participants, and follow-up supportive messages. The exact messaging and frequency of messaging that the health coach will provide to participants cannot be predicted, as this coaching will be individually tailored to the questions and concerns participants raise with the health coach, and the frequency with which they reach out to the health coach. This is in line with the autonomy support model featured throughout Aim2Be, where participants can choose which health behaviours to focus on.
Waitlist Control condition participants will be put on a waitlist for three months during which time they receive a brochure with Canadian health recommendations about physical activity, diet, screen time, and sleep habits (this represents the duration in which they are part of the efficacy trial - see Figure 1). Once their 3-month assessment is complete, they will have access to Aim2Be app with no health coach for the following three months. Participants will have access to all features of the Aim2Be app except the health coach and will be assessed three months after they were given access to the Aim2App with no health coach. This three-month period will be compared to the three-month period of the intervention condition to determine differences in adherence in Aim2Be with and without health coach support.
Data collection occurs at baseline, 3-months, and 6-months and at each time point the participating parent and their teen will complete an online survey administered using REDCap. The teen will also complete three 24 hour dietary recalls using the Waterloo Eating Behaviour Questionnaire (WEBQ) and parents will measure their teen’s height and weight following the protocol and tools provided by the research team. Finally, the study team will extract 14 days of Fitbit Flex 2 (Fitbit Inc., San Francisco, CA) wear for each teen at each assessment time point using Fitabase (Small Steps Lab LLC, San Diego, CA, USA), an online platform designed specifically for research using Fitbits. Participants will receive their Fitbit at baseline. See the Figure 3 for the SPIRIT Figure (48, 49) to see when data collection occurs and for a detailed list of data that are collected to address the primary and secondary aims of this study. The complete SPIRIT checklist can be found in Additional File 1.
(INSERT FIGURE 3 HERE)
Change in weight outcomes. Height and weight will be measured by parents using the Centers for Disease Control and Prevention home protocol (50) and used to calculate Body Mass Index (BMI) kg/m2 which is a valid and reliable method to assess BMI in children (51). Each participant will be mailed a digital scale (Active Era) and tape measure (HDX) so that all measurements are taken using the same instruments. This procedure has been validated to assess height and weight at home among adolescents and found to provide valid assessment (51). Teen BMI z-scores will be computed using a Stata macro developed by the World Health Organization whereby a BMI z-score > 1 and ≤ 2 Standard Deviation = overweight and >2 Standard Deviation = obese) (52).
Change in physical activity behaviours will be assessed using a Fitbit Flex 2, which can be immersed underwater (53) (Fitbit Inc., San Francisco, CA) and with a questionnaire. Each participant will be mailed their Fitbit Flex 2 at baseline. From the Fitbit Flex 2, total steps per day for a period of 14 days will be extracted using the Fitabase platform. The child physical activity questions were modelled after the International Physical Activity and the Environment Network questions (54) and ask about participation in physical education at school in past week, invovlement in team sport, and number of days of moderate-vigorous physical activity in the previous week.
Change in dietary habits will be measured using the WEBQ, a 24 hour dietary recall web-questionnaire developed by the University of Waterloo which has been validated in children (55), and with a questionnaire. Three 24 hour dietary recall WEBQs (including 1 weekend day) will be analyzed using the Food Processor Software (ESHA Research, Salem, OR) using the 2015 Canadian Nutrient data file. An index of diet quality using the Healthy Eating Index adapted for Canadian food recommendations will be computed. The index computes a score from 0-100, where <- 50 is categorized as poor diet; 50-80 as needing improvement, and >= 80 as being good (56, 57). In addition, seven dietary questions were included into the child’s questionnaire and these questions were taken from the 2016 Canada Community Health Survey (58) with some questions originating from the Behavioral Risk Factors Surveillance System (59) and Centers for Disease Control and Prevention National Youth Physical Activity and Nutrition Study (60). Questions asked about previous week and previous day consumption of fruits and vegetables, 100% fruit juices, and sugar-sweetened beverages.
Change in sedentary behaviours will be measured with an adapted version of French et al.’s assessement of screen time which has been found to be sensitive to intervention-mediated change (61). Two questions ask the amount of time children spend in front of screen in their free time on weekdays and weekend day.
Adherence will be measure with web-analytic data that is internally collected within the Aim2Be app for children. Adherence will be first operationalized as time spent using Aim2Be, number of days using Aim2Be, and number of tasks completed. Given that web-analytic data tracks all features that children are using and the content utilized, if variability in the data allow profiles of users will be used as outcomes.
Mediators that will be examined with respect to adherece are linked to the theoretical framework of Aim2Be (see Figure 2). The list of mediators are included in Appendix 1 and include: 1) assessment of satisfaction with Aim2Be for gamification mediators; 2) measures of motivation, self-efficacy, and psychological autonomy for behavioural mediators; and 3) assessments of the household environment (family communication and functioning), parenting styles and practices, access/opportunities (cell phone, TV in bedroom, co-participation in physical activity), modeling healthy behaviours (e.g., parent own behaviours), and parental support for environmental mediators.
Moderators that will be examined will be taken from the demographic questions (i.e. age, gender), as well as assessment of enjoyment and preferences for engaging in the behaviours targeted by Aim2Be (5 items assessing food and sedentary behaviours preferences and enjoyment for physical activity), knowledge of current health recommendations (8 Yes/No items asking knowledge about the physical activity, diet, screen time, and sleep recommendations), weight concerns (2 items), self-esteem/optimism (6 items) (62), and anxiety/worries (3 items (63).
To address the primary aims, multilevel mixed-effect linear and logistic models, which account for the repeated nature of the data, will test whether zBMI, dietary quality, fruit and vegetable intake, sugar-sweetened beverage intake, physical activity, and screen time behaviours at 3-months differ between Aim2Be and control condition participants. To address the secondary aim that examines whether health coaching influences adherence to the Aim2Be intervention, linear regressions will be used to compare time spent using Aim2Be, number of days using Aim2Be, and number of tasks completed. Three analyses will be run, one for each adherence variables (i.e., time spent in app, numbers of days in app, tasks completed). As the data are expected to be skewed, the dependent variables will be log transformed or utilize the optimal transformation to deal with the skewness of the data. Finally, examination of the mediators and moderators of adherence will use path analyses and relevant socio-demographic covariates and explore whether recruitment routes influenced adherence (e.g., self-referral or healthcare facility referral). Multiple imputations techniques will be used to deal with the missing data. Analyses will be conducted in Stata (StataCorp LLC, College Station, TX, USA) and MPlus (Muthen & Muthen, Los Angeles, CA, USA).
Quality assurance and monitoring
Operating procedures will be documented in a Standard Operating Procedure manual to standardize administration of trial conditions, data collection methods, tracking procedures, and checking programming into REDCap (e.g., randomization, administration of tools within specified timeline). Any adverse events will be reported to both the Childhood Obesity Foundation and the University of British Columbia Children’s and Women’s Research Ethics Board and accordingly added to the trial registry.
Study participants will be provided access to the app, and any updates to the app, for as long as the app is supported by the providers. At the end of the evaluation period, participants will receive emails and a push notification alerting them that the study is complete, at which time they will need to sign a new agreement with the providers if they wish to continue using the Aim2Be app. If a user does not agree to continue using the app their user account will be deactivated and a reactivation code will be emailed to them. If they chose to return to the app, they will need to agree to the provider’s agreement. Participants are given the option to continue using the app after the study ends, as there are ethical concerns associated with discontinuing access once the evaluation window is over, particularly if the participant found the program useful.
All questionnaire data will be collected in REDCap hosted at BCCHRI. REDCap is a secure, web-based software platform designed to support data capture for research studies, providing 1) an intuitive interface for validated data capture; 2) audit trails for tracking data manipulation and export procedures; 3) automated export procedures for seamless data downloads to common statistical packages; and 4) procedures for data integration and interoperability with external sources (28, 29). Personal identifiable information about participants will be kept separate from the main dataset and will not be shared. The main dataset will be deidentified. All data will be stored securely on site at BCCHRI, University of British Columbia with access limited to authorized study personnel.
The trial is scheduled to be completed by November 30, 2020. Recruitment began on January 2, 2019 and will be completed by approximately October 31, 2019.
Study results will be reported according to CONSORT guidelines. Study findings will be submitted to a peer-reviewed scientific journal for publication and be presented at academic conferences. The study results will also be reported in the ClinicalTrials.gov registry.