Intervention Studies

In the Netherlands, Slootmaker and colleagues [26] investigated the effectiveness of the PAM monitor in combination with PAM COACH for increasing the physical activity levels of adolescents (aged 13-17 years; n=87). Activity data from the PAM device were uploaded to PAM COACH, which was a Web-based system for self-monitoring activity levels and setting goals. At the conclusion of the 3-month intervention, girls in the intervention group increased their self-reported weekly moderate physical activity relative to controls (equated to ~59 min/day), though this was not evident at 8 months postintervention. No intervention effects were observed for boys after 3 months, though self-reported sedentary time was lower at 8 months relative to boys in the control group (~257 min/day). Greater adherence to the study (eg, log-in and upload frequency to PAM COACH) was not associated with greater physical activity. Minimal attrition was observed, with 78% and 91% of participants providing follow-up data at 3 and 8 months postintervention, respectively. Overall, the PAM was viewed positively by participants, and 65% reported regular wear, though monitor loss (12%) and damage (7%) may have influenced the results.

Hayes and Van Camp [25] used the Fitbit (model not reported) as a tool for increasing 6 grade 3 (8 years old) girls’ physical activity levels during school recess. After baseline data were collected from 7 recess periods, girls were provided with step goals (increments based on baseline data) for 7 further recess periods and were encouraged to self-monitor their steps against these goals. Following this, data were then collected for a further 7 periods (no step goals provided). The project culminated in a final intervention session where 3 goals were given, and a tangible reward (eg, a small toy) was provided based on the goal(s) achieved. The number of steps taken increased by 47% from baseline (1326 steps) to intervention (1956 steps; contribution of 18% to daily step recommendations [30]), while the proportion of time spent in MVPA increased from 4% to 25%, which equates to 5 minutes of MVPA during recess, or a contribution of 8% to daily recommendations [5]. Without the use of the Fitbit to self-monitor recess activity, steps taken and MVPA decreased to initial baseline levels. Basic process evaluation measures suggested that data were lost due to syncing issues, particularly during later recess sessions.

Hooke and colleagues [27] examined the efficacy of the Fitbit One to promote physical activity in clinical settings. A total of 16 children (mean age 8.7, SD 3.1 years) with acute lymphoblastic leukemia were provided with a Fitbit to wear for 17 days prior to and 5 days during a corticosteroid pulse. Monitoring over 3 days was used to identify baseline activity levels. Step goals were then tailored for each participant by a research nurse, and daily feedback was provided against these goals. No significant increases in daily steps were recorded, though there was a trend for steps to increase from week 1 to week 2 (average of 269 steps/day; 2% of daily step guidelines [30]) but to decrease from week 2 to week 3 (average of 307 steps/day). Process evaluation indicated that participants and their families were able to use the Fitbit One and the accompanying website, that they liked the Fitbit, and that data were available for 92% of measured days.

Intervention Effects

The limited intervention effects observed in studies included in this review may be attributable to several factors. First, the length of the interventions implemented ranged from 19 days to 3 months. It is possible that the shorter intervention periods were not sufficient to change behavior, a notion supported by a recent review that highlighted that behavioral interventions of longer durations (≥6 months) had greater success in changing physical activity levels [31]. There is a clear need for studies using longer measurement periods to examine the effectiveness of these devices in youth. Second, it is not clear whether the interventions were grounded on behavioral theories, which are critical for intervention effectiveness [32]. Third, 2 of the studies may have been underpowered to detect a change in physical activity due to the smaller number of children recruited [25,27]. Fourth, intervention effects were not examined using validated objective monitors (eg, accelerometers). While several studies used data from the wearable activity tracker, these devices have not been validated for assessing physical activity outcomes in youth to date [18]. This could be viewed as a limitation of these included studies. However, this may be less of an issue if the focus of an intervention study is to use the device as a tool to facilitate behavior change rather than to evaluate the outcome (ie, validity and reliability should be established prior to such use). In the only study to report significant effects, self-reported physical activity data were collected using a questionnaire with low validity [26]. There is a clear need for longer-term studies using randomized controlled trials that are grounded on behavioral theories to identify the effectiveness of wearable activity trackers for changing youth physical activity behavior.

Self-Monitoring Using Wearable Activity Trackers

A common feature of the identified intervention studies was that the wearable activity tracker was used to self-monitor physical activity in combination with different intervention approaches [17]. Unsurprisingly, these intervention approaches largely consisted of goal setting, identified as an effective behavior change technique [11] to enhance physical activity levels [33]. However, there was some variability in who set the goals. In the interventions with children, the researchers set the goals based on baseline values obtained, and then provided regular support [27] or rewards [25] in relation to reaching the goals. In contrast, the adolescents had an opportunity to set their own activity goal and then received tailored advice to achieve these based on their preferred activities [26]. While research has shown that assigned goals are as effective as self-set goals, provided the reason for the goal is given [34], it is unclear whether the included studies provided this information to participants. It is important for future research to establish how youth engage with wearable activity trackers (eg, is the frequency of self-monitoring mediated by the activity goal source?), as this will provide critical insights into how these devices can be integrated into future interventions and strategies for engaging children and adolescents in the behavior change process.

Sustainability of Wearable Activity Trackers Over Time

Arguably, one of the biggest concerns regarding wearable activity trackers is whether individuals sustain their engagement with the technology over time [16]. Research has suggested that approximately one-third of US adults stop using their wearable activity tracker after 6 months [16], with expectation mismatch (ie, the technology doesn’t do what was expected) a commonly cited reason [35]. While it is difficult to draw any conclusions about this due to the small number of studies identified and the variability in the length of time the wearable activity trackers were worn, there is some preliminary evidence to suggest that youth might regularly use the technology to self-monitor their physical activity levels when the technology is integrated into an intervention [26,27], but sustained use may not be observed in the longer term when these devices are simply provided to youth to use [28]. Interestingly, Slootmaker and colleagues found that the frequency that adolescents uploaded their data to PAM COACH was not associated with physical activity, yet physical activity levels were found to increase in girls at 3 months and sedentary time decreased in boys at 8 months [26]. These results could be explained by the lower activity levels of the girls in the study compared with boys; therefore, girls could achieve greater gains in activity levels [26]. However, it is also possible that there are different degrees of engagement, ranging from brief glances at the device [36], to tracking activity across the day using the monitor, to using specific functions within the accompanying app (eg, trends data), which may mediate the efficacy of the device on activity levels. Research is needed to provide further evidence about how youth engage with these devices and the accompanying apps over time (eg, attrition rates), whether this differs by population subgroups (eg, sex, age), whether the engagement with different features may have differential effects, and reasons for potential changes in use over time. Such research will be critical for identifying how to incorporate wearable activity trackers into interventions, and for informing best practice in future physical activity interventions and health promotion practice.

Feasibility of Wearable Activity Trackers in Youth

This review identified few studies that have examined the feasibility of using wearable activity trackers in children and adolescents, either as part of an intervention or as a stand-alone study. While these technologies offer significant promise for increasing physical activity levels and benefiting health promotion practice [14] or, potentially, clinically relevant outcomes in healthy and clinical populations [27], it is important to establish whether wearable activity trackers are a practical tool for youth to use regardless of the setting. Overall, the results from this review suggest that such devices were viewed positively by youth and their parents [26-29], and that they appreciated the devices’ range of functions, which included the tracking of physical activity. Ease of use, comfort, and aesthetics were important to the participants [28,29], which is an interesting point to note given that such devices are unlikely to have been developed with youth in mind [35]. Such factors have been previously identified as important and potential barriers to use in adults [37,38]. Interestingly, while there was some evidence that youth used their wearable activity tracker to compete with (boys) or support (girls) each other [28], which has also been observed in adults [37], few studies noted concerns over the accuracy of the devices. Schaefer and colleagues found that, while some children did test the accuracy of the monitors [28], this did not appear to influence use. Of potentially greater concern, access to technology was identified as a potential issue in low-socioeconomic-status youth, such as the ability to sync and access data at home [28]. This supports a previous study that found that the use of a mobile phone app that enabled boys from low-socioeconomic areas to track their goals and behavior was moderate, due to their prioritizing data for entertainment over the app [39]. As such, we recommend that researchers examine the feasibility and acceptability of different devices, and we suggest that youth are interested in using wearable activity trackers and that these devices are feasible for use in both clinical and healthy populations. However, it must be noted that there are issues concerning the age required (≥13 years old) to hold an account associated with wearable activity trackers that may preclude their use with children, unless alternative feedback mechanisms are employed (eg, researcher-led feedback [25,27]). Identifying how best to integrate the functions and features of wearable activity trackers into physical activity interventions, in order to maximize their potential within the population of interest, should be a future area of research.

Limitations

There are some limitations in this review that warrant attention. First, this review highlights the surprisingly small number of studies that have used wearable activity trackers in youth despite their widespread prevalence in daily life. This makes it difficult to establish any firm conclusions. Given the pervasiveness of these technologies and widespread appeal, further research is needed to explore these issues in youth in order to inform interventions and public health guidance. In addition, research is needed to establish whether differential effects are observed based on the age of the participants. Second, the quality of the included studies was low, with only 1 study using a randomized controlled design and evaluating intervention effects over a 3-month period. There is a need for more rigorous and robust intervention designs to evaluate the longer-term effectiveness of these devices on youth physical activity levels. Third, while a range of monitors were used in the included studies, some monitors have since been discontinued (eg, PAM, SenseWear) or largely usurped by new models (eg, Fitbit). In addition, the validity and reliability of these devices in youth has not been established.