Various brands of fitness trackers were identified on the basis of listings from 4 websites (CNET, TechRadar, Wareable, and AllThingsWaterproof). A list of their corresponding webpages about fitness trackers was created on January 9, 2018. The official product websites for each brand of fitness trackers were reviewed to identify the various models of fitness trackers pertaining to each brand. Each identified fitness tracker was assessed for the following inclusion and exclusion criteria through reviewing the specifications of each fitness tracker on its official product website.

Inclusion criteria for the fitness trackers were the following: (1) be able to track or monitor activity (eg, number of steps taken, minutes spent in swimming), (2) be swim proof, (3) be compatible with both Android and Apple operating systems, and (4) cost below Aus $150.

A total of two raters, the first author (GC) and an independent researcher (PP), wore each of the fitness trackers for a week and downloaded and used companion apps of each fitness tracker on a smartphone. The fitness trackers and their companion apps were coded using the BCTTv1 coding manual, which contains labels, descriptions, and examples of each BCT. The raters also read the manual of each fitness tracker to ensure that no functions of the fitness tracker were left unnoticed and unassessed. Each BCT was coded using a dichotomous score of either 0 (not present) or 1 (present). Any coding disagreements were discussed between the two raters until an agreement was reached. To support study objectives, BCTs were only coded when they were applied to target behaviors of increasing physical activity or reducing sedentary behavior. BCTs that targeted other behaviors (eg, diet or sleep) were not coded. Before assessing the fitness trackers and their companion apps, the main author completed the Web-based BCTTv1 training through the official website and trained the other rater to use the BCTTv1 taxonomy. After this, a test for calibration was conducted, where the mobile app Runkeeper was coded using the BCTTv1 taxonomy. This app was chosen as it was found to contain the highest number of BCTs in a review of apps to promote physical activity in adults [56]. Furthermore, by using an app, the two raters were able to access the app and conduct the calibration process concurrently. Any ambiguous descriptors or definitions from the BCTTv1 taxonomy were discussed between the two raters until an agreement was reached during the calibration process.

The 6 fitness trackers that were included in this study were the Fitbit Flex 2 (Fitbit Inc), Huawei Band 2 Pro (Huawei Technologies), Misfit Shine 2 (Fossil Group), Moov Now (Moov Inc), Nokia Go (Withings), and Polar A300 (Polar Electro). The companion app of each fitness tracker was downloaded to an Apple iPhone 8 and a Samsung J5 Pro smartphone. Record forms were used to document the results from the data analysis of each fitness tracker. Each form included a table containing labels, definitions, and examples of each BCT from the BCTTv1 coding manual, as well as a 0 and 1 against each BCT for the raters to mark the absence or presence of the BCT.

Descriptive statistics were used to summarize the BCTTv1 ratings of the fitness trackers. Interrater reliability was calculated by dividing the number of agreements plus disagreements and multiplying by a hundred. To assess the relationship between the cost of fitness trackers and the number of BCTs incorporated, a Pearson correlation coefficient was computed via SPSS Version 26 (IBM), using an alpha of .05 to determine statistical significance.

A total of 39 fitness trackers were identified across 14 brands as shown in Figure 1. These 39 fitness trackers were assessed by reviewing their specifications on their official product website, and 12 of them were considered eligible for inclusion. Of the eligible 12 fitness trackers, 6 fitness trackers were variants of others from similar brands. As such, only the latest model of each brand of fitness tracker was retained, leaving a total of 6 fitness trackers for BCT analysis.

The definition of each BCT is provided in Multimedia Appendix 1. The number and type of BCTs included in each fitness tracker were evaluated, and these are summarized in Tables 1 and 2. The interrater reliability for evaluating the presence of BCTs measured by percent of agreement was 100%. Disagreements were resolved through discussion. The cost, number of BCT clusters as identified by the BCTTv1 classification, and the number of self-management and nonself-management BCTs incorporated in each fitness tracker are outlined in Table 1.

The average number of BCT clusters included in each fitness tracker was 12, ranging from 6 to 9. The Fitbit Flex 2 and Nokia Go had the highest number of BCT clusters (n=9), followed by Huawei Band 2 Pro (n=8). Misfit Shine 2 had the lowest number of BCT clusters (n=6). The most common BCT cluster incorporated across all fitness trackers was goals and planning (mean BCTs incorporated=25%), followed by feedback and monitoring (mean BCTs incorporated=15.5%), antecedents (mean BCTs incorporated=14.3%), and reward and threat (mean BCTs incorporated=13.1%), as shown in Figure 2. The average number of BCTs per fitness tracker was 14, ranging from 11 to 20. The Fitbit Flex 2 had the highest number of BCTs (n=20), followed closely by Nokia Go (n=19). The Huawei Band 2 Pro and Misfit Shine 2 had the lowest number of BCTs (n=11). The Fitbit Flex 2 incorporated the most number of BCTs related to self-management strategies (n=14), followed by Nokia Go (n=12). Huawei Band 2 Pro and Polar A300 contained the least number of BCTs related to self-management strategies (n=9).

The names and types of BCTs incorporated in each fitness tracker are outlined in Table 2.

A total of 6 BCTs out of the total 93 were present in every fitness tracker. These 6 BCTs were discrepancies between current behavior and goal, feedback on behavior, self-monitoring of behavior, feedback on outcomes of behavior,adding objects to the environment, and body changes. A total of 8 BCTs were present in 50% or more of the fitness trackers. These were goal setting of behavior (n=5), goal setting of outcomes (n=3), self-monitoring of outcomes of behavior (n=3), information about health consequences (n=4), social comparison (n=5), prompts and cues (n=5), social reward (n=4), and social incentive (n=3). The BCTs that were present in fewer than 50% of the fitness trackers were action planning (n=1), social support unspecified (n=1), instruction on how to perform behavior (n=2), monitoring of emotional consequences (n=1), information about emotional consequences (n=2), demonstration of behavior (n=2), behavioral practice and rehearsal (n=1), habit formation (n=1), graded tasks (n=1), credible source (n=2), incentive outcome (n=1), and reward outcome (n=2). A Pearson correlation coefficient was computed to assess the relationship between the cost of fitness trackers and the number of BCTs incorporated. There was a nonsignificant correlation (r=0.21, n=6, P=.69) between cost of fitness trackers and the number of BCTs incorporated.

All fitness trackers incorporated components of self-management strategies to target participation in physical activity. These self-management components included goal setting, self-monitoring, and self-evaluation of activity in relation to goals, providing feedback of progress toward goals, and providing rewards on meeting goals.

A total of 11 BCTs that were not related to self-management strategies were found to be incorporated across the fitness trackers. A total of 4 fitness trackers (Fibit Flex 2, Huawei Band 2 Pro, Nokia Go, and Polar A300) provided information about health consequences through textual information on the app (eg, “put on your running shoes…work up a sweat…it can improve cardiovascular and respiratory health”). A total of 2 fitness trackers (Fitbit Flex 2 and Nokia Go) incorporated information about emotional consequences, presented as textual information on the app (eg, “…boost self-esteem and emotional state”). A total of 2 fitness trackers (Fitbit Flex 2 and Moov Now) incorporated the BCT instruction on how to perform behavior and demonstration of behavior, which were presented through videos on the app showing movements and describing the steps to undertake the physical activity. A total of 2 fitness trackers (Fitbit Flex 2 and Huawei Band 2 Pro) incorporated credible source, presented as textual information on the app (eg, “the American heart association recommends…”). All fitness trackers, except the Huawei Band 2 Pro, incorporated social comparison, using leaderboards of steps accumulated. Only the Nokia Go incorporated the BCTs behavioral practice and rehearsal, action planning, habit formation, and social support by setting alerts as prompts for action. Only the Polar A300 incorporated monitoring of emotional consequences through rating of emotions using emoticons on the app.

All 6 fitness trackers incorporated BCTs to increase physical activity, but only 2 fitness trackers (Fitbit Flex 2 and Moov Now) aligned with public health physical activity guidelines by including active minutes as a physical activity goal. The Huawei Band 2 Pro, Misfit shine 2, Nokia Go, and Polar A300 centered on the number of steps achieved daily as physical activity goals. A total of 2 fitness trackers (Fitbit Flex 2 and Polar A300) incorporated BCTs to reduce sedentary behavior, which was implemented through the BCT information about health consequences (eg, “sitting for long periods is bad for your blood circulation, especially in your legs” and “moving regularly breaks up sedentary time and can improve your well-being”). The Huawei Band 2 Pro, Misfit shine 2, Moov Now, and Nokia Go did not specifically target sedentary behavior.

This study aimed to evaluate the use of BCTs in fitness trackers that are swim proof and cost less than Aus $150 (US$ 105). Overall, all fitness trackers incorporated more than 3 BCT clusters, which has been shown to produce significant effects in physical activity interventions [51]. The Fitbit Flex 2 and Nokia Go incorporated the most BCT clusters. There was a nonsignificant correlation between the cost of fitness trackers and the number of BCTs incorporated. The Fitbit Flex 2, which costs Aus $149.95 (US$ 105, had the most BCTs (coded at 20), followed by the Nokia Go, at a cost of Aus $89.95 (US$ 63) with 19 BCTs. In comparison, the Misfit Shine 2 Pro, which costs Aus $107.77 (US$ 75), incorporated the least number of BCT clusters and BCTs, at 6 and 11, respectively. The findings indicated that the cost of fitness trackers does not necessarily reflect or is associated with the number of BCTs incorporated in them. The BCTs that were frequently incorporated across the 6 fitness trackers were mostly related to self-management strategies, such as goal setting, self-monitoring, self-evaluation (eg, feedback, provision of discrepancies between current behavior and goal), prompts and cues, and rewards (eg, social reward) [57]. This finding is similar to those from previous studies of fitness trackers [52-54] and apps [56,58] targeting physical activity, and they also overlap with reviews of BCTs that were rated as important by users of fitness trackers [55]. Notably, all 6 fitness trackers in this study incorporated a combination of 9 or more BCTs that were related to self-management. However, similar to the study by Lyons et al [52], this study also found that other effective BCTs were rarely incorporated in fitness trackers. Only Nokia Go contained the BCTs behavior practice and rehearsal and social support, through the use of prompts, that have been found to be effective in increasing physical activity for adults with obesity [24] and cardiovascular disease [45]. This indicates that certain fitness trackers may be more effective for particular population groups because of the types of BCTs incorporated in them. The combinations of informational and instructional BCTs that have been found to support maintenance of change in physical activity were only observed in the Fitbit Flex 2 fitness tracker. The Fitbit Flex 2 companion app provided links to a separate app called Fitbit Coach that contained a library of videos to instruct and model the way to perform various workouts. The Fitbit Flex 2 companion app also provided textual information about health and emotional benefits of physical activity and sedentary behavior. It seems that manufacturers of fitness trackers are more focused on functional features (eg, recording data, providing prompts) compared with informational and instructional features. It could be that certain BCTs, such as features that support self-management through recording behavioral data, evaluating and providing feedback about behavior and providing prompts, are better suited to delivery via wearable technology. There is some support for this contention in this study in the finding that informational and instructional BCTs were often presented on the companion apps instead of the device. A recent study also found that providing instructions was one of the most frequently implemented BCTs (17 of 25 apps) in mobile apps that aimed to improve physical activity and reduce sedentary behavior [59]. These findings are not surprising, as longer information and instructions might be more suited to technologies with larger reading screens or audio outputs, such as computers, tablets, and smartphones. Taken together, these findings highlight the importance of delivering BCTs through appropriate technological channels and leveraging and integrating different technologies to create a more holistic intervention for promoting physical activity.

Given that the purpose of fitness trackers is to promote physical activity, it is remarkable that not all fitness trackers assessed in this study aligned with public health physical activity guidelines. Physical activity guidelines for adults include engagement of at least 150 min of moderate-intensity physical activity per week, performed in bouts of at least 10 min duration [4,14,60]. Only the Fitbit Flex 2 and Moov Now aligned with public health physical activity guidelines by including active minutes as a physical activity goal. Active minutes were calculated only if activities above 3 METs were detected continuously for 10 min at a time. Other fitness trackers focused on the number of steps achieved daily. The tracking of steps is a prominent part of data collection for most of the fitness trackers in this study. This was expected, as steps are a fundamental part of daily living, which might be more easily measured. However, the sole use of steps to measure physical activity may be insufficient. The tracking of step counts does not provide information about the intensity of the activity, which has been indicated to be more beneficial to health than number of steps taken [61]. Moreover, there has been a change in the fitness culture over the past decade, in which the trend for physical activity has been moving away from step-based activities to spinning (indoor cycling), dance workouts (eg, Zumba), and body weight training [62-64], as well as water sports, such as stand-up paddle boarding and kitesurfing. Although most fitness trackers allow for self-monitoring of physical activity that is not based on steps, these data have not been directly incorporated into physical activity goals. Given the findings from previous studies that goal setting is an effective BCT to increase physical activity [25,48] and reduce sedentary behavior [65], it would be optimal to integrate non-step-based activities into physical activity goals, as has been achieved by Fitbit Flex 2 and Moov Now, through the inclusion of active minutes as a goal. Regarding guidelines to minimize sedentary behavior by minimizing the amount of prolonged sitting [14-16], only the Fitbit Flex 2 and Polar A300 provided BCTs that directly targeted sedentary behavior in the form of providing information about health consequences. The Fitbit Flex 2, Huawei Band 2 Pro, Misfit Shine 2, and Polar A300 provided prompts to move after detecting inactivity over 1- or 2-hour periods. However, these prompts were based on a lack of steps detected within that period, instead of detecting sedentary behavior that has been defined as engaging in a sitting or reclining posture, with an energy expenditure less than or equal to 1.5 METs [8]. The authors of a previous study have coded the prompts of inactivity as targeting a reduction of sedentary behavior [54]. However, this study adhered to the definition of sedentary behavior by the Sedentary Behaviour Research Network [8] and took into account the distinction between physical inactivity and sedentary behavior in the coding of BCTs. Furthermore, studies have noted that interventions, which promote physical activity, do not necessarily have an effect on reducing sedentary behavior, and techniques that primarily aimed to change sedentary behavior were more effective in reducing sedentary behavior [65]. Despite literature showing that both physical activity and sedentary behavior are important to good health, it seemed that most developers of fitness trackers assessed in this study have focused on achieving sufficient amount of physical activity, but they have placed less emphasis on directly reducing sedentary behavior. It could be that the technology required to detect sitting and reclining postures is too costly to be incorporated into fitness trackers within the price range of Aus $150 and below and that a more cost-effective workaround to target sedentary behavior is to frequently prompt users to take more steps. Overall, the findings from this study show that fitness trackers below Aus $150 contain evidence-based BCTs known to promote physical activity and reduce sedentary behavior. The majority of BCTs implemented targeted physical activity, and the most frequently used BCTs overlapped with self-management strategies. This suggests that fitness trackers offer the prospect for physical activity interventions that are cost effective and easily accessed by a wide population. Opportunities to improve the effectiveness of fitness trackers include incorporating BCTs that facilitate maintenance of behavior change in physical activity, such as instruction or demonstration of behavior, and providing information on the consequences of behavior [24], as well as incorporating BCTs to target sedentary behavior. Furthermore, public health physical activity guidelines ought to be used to inform the type of goals that are incorporated into fitness trackers.

Although this study highlights that low cost fitness trackers contain evidence-based BCTs known to promote physical activity, it is worth noting that only 1 fitness tracker contained BCTs effective for special population groups, such as adults with obesity and chronic illnesses. Further work is needed to determine if these low cost fitness trackers are sufficient as standalone interventions, especially for these special population groups.

The most recent taxonomy of behavioral change techniques was used to assess the content of the fitness trackers and their companion apps, and it compared the BCTs to those that have been found to be successful at improving physical activity, thus adding to the evidence base for the use of fitness trackers to support interventions to promote physical activity. This study focused on fitness trackers that are swim proof and cost less than Aus $150 to broaden the reach of fitness trackers to the community in terms of preferred and accessible physical activity, as well as affordability. A limitation is that the ratings may not be representative of the most current version of the fitness trackers assessed in this study because of frequent updates of fitness trackers and their associated apps. A further limitation is that this study only included fitness trackers under Aus $150 and did not cover the entire range of fitness trackers available. The small number of fitness trackers places a limitation on the calculation of Pearson correlation coefficient to assess the relationship between the cost of fitness trackers and the number of BCTs incorporated.