The primary aim of this study was to evaluate how continuous self-monitoring using the SWA affected weight loss and waist circumference reduction both alone and in combination with GWL in sedentary overweight or obese adults. We demonstrated that a 9-month lifestyle intervention with the SWA produced significant weight loss and reduction in waist circumference. Thus, the SWA approaches, including real-time feedback and self-monitoring of energy balance, appeared to be beneficial tools in weight loss intervention. To our knowledge, only one other study  has evaluated weight loss in overweight and obese adults using a lifestyle intervention based on self-monitoring through the SWA.
The goal behind the design and development of the SWA was to create a wearable device that could quantitatively assess energy balance, sleep, and physical activity in free-living environments more efficiently and effectively than current alternatives. With regard to weight loss programs, SWA offers valuable assistance with the goal to simply self-monitor the physical activity and caloric intake during an intervention. In fact, the adherence to armband wear was excellent. Among the 76 participants assigned to an Armband group (SWA-alone and GWL+SWA) and who completed month 4 assessments, 75% wore the armband more than 75% of days. Those who wore the armband more than 75% of days experienced significant weight loss compared with those who wore it less than 75% of days . Self-monitoring, a "cornerstone" of behavioral treatment, has been found to be correlated significantly with weekly weight loss . However, due to the variability in adherence measurement methods, it is difficult to compare adherence across studies . Burke and colleagues reported a median 55% adherence to standard self-monitoring with a paper diary [24, 25]. In our study, adherence to armband wear was higher than 55%, suggesting that weight loss participants may better adhere to self-monitoring protocols that use technology, compared to standard protocols.
Our finding that self-monitoring of diet and physical activity with the SWA was related to successful weight loss supports previous studies that have identified the value of self-monitoring in weight loss [23, 26]. A recent paper  presents a systematic review of the literature on three components of self-monitoring in behavioral weight loss studies: diet, exercise, and self-weighing. The use of technology, which included the internet, personal digital assistants, and electronic digital scales, was reported in five of the 22 identified studies. A significant association between self-monitoring and weight loss was consistently identified. One unique aspect of our study was that the self-monitoring was continuous and in "real-time". Recent systematic reviews of randomized controlled trials of weight loss [27, 28] have concluded that weight loss interventions can be effectively delivered over the Internet. Successful online obesity treatment programs have targeted reduced energy intake, increased physical activity, and cognitive-behavioral strategies including personalized feedback, self-monitoring, and social support. However, limitations of previous studies include no intention-to-treat (ITT) analysis, no assessor blinding, follow-up measures based only on participants' self-report, moderate retention rates, and insufficient follow-up.
In a recent study the armband was used as a real-time self-monitoring device in conjunction with a GWL . This small study (n = 57) reported improved weight loss over 3 months when the armband was worn in conjunction with a GWL beyond that which was accomplished with a GWL alone . Polzien and colleagues' focus was to evaluate the efficacy of providing real-time feedback on energy balance as part of a weight reduction program, and to evaluate whether this approach enhances standard weight loss methods. Research suggests that individuals who use self-monitoring strategies (i.e., frequent weight checks, monitoring physical activity, monitoring caloric intake) experience improved weight loss [4, 29]. Our study also showed a significant weight loss at month 9 for participants who wore a SWA compared with those who did not. These findings further confirm the importance of self-monitoring, which is indeed a consistent predictor of successful weight loss in technology-assisted weight management programs [29, 30].
Due to the public health importance of overweight and obesity, weight loss interventions must be effective, available and accessible to the public. The Internet and technology devices provide unique opportunities for developing and implementing of lifestyle interventions that promote self-monitoring [31–33]. One study reported that overweight participants who had access to an Internet behavioral weight loss program for 6 months showed greater weight loss and reductions in waist circumference than participants given access to a weight loss education material only .
Our study had several strengths: a randomized design, primary and secondary outcomes assessed, including objective measures of adiposity, outcomes assessed by researchers blinded to group assignment, a participant population comparable to South Carolina's racial profile (67% white) and follow-up assessments at months 4 and 9. The significant weight loss results from SWA-alone group are likely to be applicable outside the research setting because participants in this group received minimal face-to-face intervention, prepared all their own meals, and established their own physical activity regimens.
There were also several limitations in the present study. First, there was a large attrition rate, particularly from the Standard Care group, where only 52% of the initial sample had complete data at month 9. Although the attrition rate is disappointing, it does not diminish our findings. Those lost to follow-up were similar to those who completed the study with the exception of a difference in education levels. Moreover, since we assumed no weight loss occurred in individuals lost to follow-up (initial weights carried forward), attrition biases our results toward finding no effect rather than overstating the effects of our interventions. Future studies are warranted to confirm or reject the findings reported here. Second, the participant sample was mostly female (82%) and highly educated (77%) and therefore may not be generalizable to the general populations, but is representative of individuals typically seeking weight loss treatment[35–37]. Third, this was a short intervention with significant weight loss, and therefore, it is unknown whether the weight loss would be maintained or continued over long term. Lastly, the GWL did not perform well compared with the published studies. One possibility is that participants were disappointed with group assignment. Another possible cause is the different characteristics of participants recruited during different waves. The first wave of recruitment was mainly from the University employees and students. We later found out that students were the most unreliable group in this study, and their adherence was especially poor for homework assignments and other assignments.