A randomized trial of a fitness tracker was conducted over 3 months. Participants were randomized (1:1) using computer block randomization techniques into 1 of 2 groups: (1) adolescent group with the adolescent receiving a fitness tracker or (2) dyad group with the adolescent and their caregiver receiving a fitness tracker (Figure 1). Adolescents in both groups also received counseling regarding physical activity data as part of standard tertiary care weight management treatment, as described in detail below. At the time of trial implementation, the study did not meet requirements for registration on clinicaltrial.gov because of the sample size, lack of a control group naïve to an intervention, and intent to test the feasibility of the study procedures.

Participants recruited for the trial were adolescents (aged 13-17 years) who were new patients to 1 of 2 tertiary care weight management clinics in a children’s health system. Patients were referred by their primary care provider to the clinic for having a body mass index percentile ≥85th percentile for age and having failed attempts at weight management in the primary care setting. One of the adolescent’s caregivers was also enrolled as a participant in the study. Participants were excluded if the adolescent had a primary genetic or endocrine syndrome associated with obesity or was taking a medication that would predispose them to weight gain; if the caregiver was not the adolescent’s legal guardian, did not reside in the same household, or was of limited English proficiency; or if either the adolescent or their caregiver was unable to understand how to use the fitness tracker, had used a fitness tracker before, did not have a smartphone or tablet computer, or had a condition limiting physical activity. Participants were also excluded from the study if they were unable to attend monthly weight management clinic visits for logistical or insurance reasons. This study was approved by the institutional review board, with both adolescents and their caregivers signing consent forms at the time of enrollment.

The primary aim of this study was to assess the differential impact of an adolescent using the fitness tracker alone or in conjunction with their caregiver on physical activity patterns. As this is a novel comparison of the intervention, the effects are unknown. To detect a clinically meaningful difference in mean daily steps of 2000 (equating to approximately 1 mile of additional walking) between groups with a power of 80% and a level of significance of .05, assuming a 30% attrition rate for this 3-month study, we estimated that a sample size of 90 participants (45 in each experimental group) would be needed.

Each of the 2 tertiary care weight management clinics was located inside a tertiary care children’s hospital. One clinic was located in the mid-Atlantic, and 1 clinic was located in the South Atlantic. Each clinic provides individualized, interdisciplinary weight management treatment to patients aged younger than 18 years (please see below section on Standard Weight Management Treatment for more details). Participants in the South Atlantic clinic were more likely to be of Hispanic background (48% [14/29] vs 14% [8/59] in the mid-Atlantic, P<.001), less likely to be of non-Hispanic white background (21% [6/29] vs 44% [26/59], P=.03), and more likely to have Medicaid insurance (55% [16/29] vs 29% [17/59], P=.02). There were no significant differences between participants at the 2 clinical sites in terms of gender, age, and weight status.

Adolescent demographics were summarized and compared between the adolescent and dyad groups. To describe fitness tracker satisfaction, responses to the Fitness Tracker Utility questionnaire were summarized. Data collected from the fitness tracker were summarized to describe fitness tracker utilization and physical activity patterns. To explore the impact of providing caregivers with a fitness tracker as part of the intervention, intent-to-treat analysis was used to compare adolescent fitness tracker satisfaction, fitness tracker utilization, and physical activity patterns between the adolescent and dyad groups. Due to differences in season of enrollment and differences in demographics between the 2 clinical sites, we also tested to see if there were any differences in adolescent fitness tracker utility responses, fitness tracker utilization, and physical activity patterns between adolescents based on season of enrollment and clinical site. Due to high fitness tracker discontinuation rates, we also conducted exploratory analyses comparing demographic and physical activity patterns between adolescents who discontinued use of the fitness tracker during the study and those who did not. Chi-square or Fisher exact analyses were used to compare categorical variables, and 2-sample t tests were used to compare continuous variables. All tests were 2 tailed, with an overall level of significance of .05. Statistical software R was used.

A total of 90 adolescents were recruited to receive a fitness tracker, with random assignment of 45 to the adolescent group and 45 to the dyad group (Figure 1). Two adolescents were excluded from the dyad group in the first month of the study because of insurance problems and diagnosis of insulin-dependent type 2 diabetes. The 88 remaining adolescents were representative of the patient population with the majority being female (61/88, 69%), having severe obesity (55/88, 63%), and representing a diversity of racial and ethnic backgrounds (non-Hispanic black [32/88, 36%], non-Hispanic white [26/88, 29.5%], and Hispanic [20/88, 23%]). There were no significant differences in baseline demographics between the adolescent and dyad groups (Table 1).

Of the 72 adolescents who filled out the questionnaire at the 1-month follow-up visit, the majority reported that the fitness tracker was helping them meet their healthy lifestyle goals (50/72, 69%), be more physically active (49/72, 68%), and be more motivated to achieve a healthy weight (48/72, 66%). Of the 41 adolescents who completed the questionnaire at the 3-month follow-up visit, an even greater majority reported that the fitness tracker was helping them meet their healthy lifestyle goals (32/41, 78%), be more physically active (34/41, 82%), and be more motivated to achieve a healthy weight (33/41, 80%). Every month, more than 75% of adolescents reported accessing their mobile app weekly, but less than 45% of adolescents reported accessing their dashboard on the consumer website weekly. One-third of all adolescents used virtual rewards, and one-third of adolescents in the dyad group engaged in challenges with their caregiver. Food and activity logs were used rarely. There were no significant differences on any Fitness Tracker Utility questionnaire responses between the adolescent and dyad groups, between participants based on clinical site, or between participants based on the season of enrollment.

Our study is the first to describe the use of a fitness tracker as an adjunct to family-based weight management treatment among a cohort of adolescents with obesity, a historically difficult to engage patient population with lower levels of physical activity and worse health consequences, and to describe the impact of providing fitness trackers to caregivers in support of their adolescents. Although the majority of adolescents in this study reported that the fitness tracker was helpful in their weight management efforts, discontinuation rates were high. In addition, there was no impact of providing caregivers with fitness trackers on adolescent physical activity levels; caregiver fitness tracker discontinuation rates were also high and associated with adolescent fitness tracker discontinuation. Adolescents who did not discontinue use of the fitness tracker by the end of the study were able to maintain physical activity levels, whereas those who discontinued use of their fitness tracker during the study had a decline in physical activity levels over time.

The majority of adolescents in this study reported that the fitness tracker was helpful in their weight management efforts and reported accessing their physical activity data weekly through a mobile app, supporting the utility of physical activity feedback especially when easily accessible via mobile technology. This is consistent with other studies that have demonstrated high levels of acceptability of fitness trackers among children and adolescents [38-40]. For example, the majority of adolescents in a study by Slootmaker et al found value in the physical activity assessments provided by fitness trackers [40]. In addition, qualitative studies by Schaefer et al found that pre-adolescents enjoyed the feedback feature of fitness trackers and that these features promoted self-reflection about their physical activity [38,39].

Despite high acceptability rates, the majority of adolescents discontinued use of the fitness tracker by the end of the study. Even among adolescents who continued to use the device for the entire study, the number of days the fitness tracker was used declined each month. Adherence to fitness trackers and discontinuation of fitness trackers have also been a major challenge in other studies [28,38,40,41]. A systematic review of randomized trials using accelerometers to measure physical activity in children found that one-third of children were not adherent to using accelerometers. In trials of commercial fitness trackers as the actual intervention, adherence is even worse. For example, only 24% of adolescents in Slootmaker et al’s study provided fitness tracker data for the entire study [40], and among pre-adolescents in Schaefer et al’s study, there was a 50% decrease in the use of the fitness tracker in the first month [38].

There may be several reasons for fitness tracker discontinuation over time. Seasonality appears to play a role, with participants who enrolled in the summer being more likely to continue use of their fitness tracker. This is likely because adolescents have more time when not in school, and the weather is more conducive in the summer for engaging in physical activity and use of the fitness tracker. There may also be other reasons that are more psychological in nature. For example, a recent mixed-methods study by Kerner and Goodyear found that adolescents lost motivation to use a fitness tracker over time because of feelings of incompetence when not achieving their goal [42], and a review by Sullivan and Lachman noted that inactive individuals struggle to come up with a plan for how to be physically active and overcome obstacles to being physically active [43], barriers that are not addressed with current fitness trackers.

In yet another study, Schaefer et al noted that utilization of the fitness tracker was highest during times where there was contact with research staff, suggesting that in-person support may still be a necessary component to engagement despite the automatic support inherent in the fitness tracker itself [38]. Similarly, utilization of the fitness tracker was highest in our study during the first month when contact with research staff was more frequent. At the same time, only about one-third of adolescents in our study used the automated functions of the fitness tracker that provide positive reinforcement such as rewards and challenges, which are important for behavior change, especially among inactive individuals [43], suggesting that more directive guidance on how to use these functions may also help to increase utilization of the fitness tracker and promote physical activity.

Fitness tracker discontinuation rates of adolescents in the dyad group were strongly associated with those of caregivers in the dyad group, suggesting that there is a strong influence of caregiver actions on adolescents. This is not surprising, given the importance of caregivers as role models to their children and the importance of the home environment to a child’s lifestyle habits [29-32]. However, caregivers in the dyad group also demonstrated low fitness tracker utilization rates, which may explain why adolescents in the dyad group did not demonstrate better fitness tracker utilization and physical activity patterns than adolescents in the adolescent group. Indeed, these findings suggest that more attention to increasing the physical activity of caregivers may be necessary to indirectly promoting physical activity in their adolescents.

At the same time, the lack of difference in outcomes between the adolescents in the adolescent and dyad groups may also speak to the need to focus efforts on peer-based interventions, especially among this age group that is more responsive to peer influence and support [44-46]. For example, Schaefer et al found that pre-adolescents in their study engaged with their fitness tracker most when competing with peers [38]. This concept is supported by studies demonstrating the positive effect of exergaming on physical activity among children, especially when peer socialization elements are incorporated [47-49]. Social media trends such as Pokemon GO further exemplify the role of social gaming in motivating individuals to be physically active [50,51]. Incorporating peer socialization as a component of fitness tracker interventions may also enhance physical activity outside of the school setting, which is important given that our study and others have observed significantly lower physical activity levels after school hours than during school hours [52].

Physical activity levels did not improve over the course of the study and were lower than evidence-based recommendations of 10,000 steps or 60 min of MVPA every day [53]. This lack of improvement in physical activity is in contrast to other studies reporting positive effects of fitness trackers on physical activity levels [20,28] but is consistent with studies reporting generally low levels of MVPA among children in this country, especially those with obesity and of adolescent age, who are a very difficult group to motivate to be more active [52,54,55] for many of the reasons that Sullivan and Lachman note in their study on inactive adults [43]. Notably, adolescents who continued to use the fitness tracker for the entire study period maintained their physical activity levels, whereas adolescents who discontinued use of the fitness tracker by the end of the study demonstrated a decline in their physical activity levels over the course of the study. This suggests that fitness tracker use may at least promote maintenance of physical activity levels among adolescents, even if it does not encourage an increase in physical activity levels. Because higher physical activity levels are associated with long-term weight loss and health benefits [12,13,56], this finding supports the importance of using fitness trackers with adolescents who are most likely to benefit from them or finding ways to increase adolescent engagement with fitness trackers.

There were several limitations to this study. Although our study was larger than other fitness tracker trials in pediatric populations, the sample size was still small, and the duration of our study was short. As this was a pilot study, we also did not include a control group of adolescents receiving standard treatment in the weight management clinic. Because we wanted to evaluate a commercial fitness tracker as an adjunct to care, leveraging the automated functions of the fitness tracker, we provided minimal intervention or incentives beyond counseling around the fitness tracker data at routine clinic visits, and use of the intervention may have been limited by functions specific to the fitness tracker, including the need to sync and charge the device. Also, the majority of adolescents in our study reported that the fitness tracker was helpful, but this may have been subject to self-report bias. Collecting qualitative data about why adolescents discontinued use of their fitness tracker may have provided meaningful information to inform future interventions. In addition, fitness tracker discontinuation rates were high, limiting the data that were available for analysis of physical activity, potentially leading to under- or overestimation of physical activity among participants.

Despite high levels of satisfaction with the fitness trackers among adolescents enrolled in a weight management clinic, fitness tracker discontinuation rates were high, suggesting that more guidance and support may be needed beyond what is provided by the fitness tracker alone. This includes the support of caregivers, among whom fitness tracker discontinuation rates were also high and associated with adolescent fitness tracker discontinuation. Adolescents who continued use of the fitness tracker for the entire study demonstrated sustained physical activity levels compared with the decline in physical activity levels seen among adolescents who discontinued use of the fitness tracker during the study, suggesting that fitness tracker use may promote maintenance of physical activity levels over time. We believe the findings of our study provide valid and valuable insight into how adolescents with obesity use fitness trackers and underscores the importance of continued research to identify which subgroups will benefit most from this type of intervention while identifying innovative ways to engage adolescents in use of fitness trackers over time, including enhanced caregiver or peer involvement and features that address barriers to physical activity specific to obese patients.