In her lecture ‘Using Activity Trackers in Research Studies’ Dr. Kelly Evenson discussed the brief history of tracking devices in physical activity studies and the potential of these devices to aid in public health surveillance of physical activity.
Personal activity trackers are a relatively new phenomenon, gaining much popularity with brands like Nike and Garmin offering commercial products for individuals to keep track of their health behaviors. There are currently over 400 activity trackers on the market and devices like Fitbit have been used to track physical activity in over 140 research studies worldwide. The advantages of using such devices in research studies include the fact that they are low-cost, passive, and low-burden ways to assess physical activity levels.
In order to be confident in the results gained from studies using activity trackers, we must first assess their reliability and validity. Dr. Evenson and her research team completed a systematic review on the topic to determine the reliability and validity of these devices in five specific measures: steps, distance, physical activity, energy expenditure, and sleep. The results of this work indicated that trackers are relatively accurate at capturing the number of steps taken, but less accurate at estimating energy expenditure and sleep. In addition, the authors indicate that more studies are needed in order to determine the validity of trackers to provide accurate information on distance and overall physical activity. Luckily, Dr. Evenson shared that since this review was undertaken, there have been a multitude of new studies using these devices so her group is beginning to plan another review of the literature to assess these topics and update their conclusions.
With the popularity of these devices and the increasing amount of research being conducted on their accuracy, it is only logical to wonder if activity trackers could have a role in public health surveillance of physical activity on a national level. Currently, physical activity at this level is assessed by survey and sometimes accelerometry in nationwide surveillance systems. Dr. Evenson and her colleagues performed a study to assess the agreement between Fitbit data and that from the Behavioral Risk Factor Surveillance System and unfortunately at this time, there was no correlation between the two.
In closing her discussion, Dr. Evenson mentioned a few pieces of advice for different audiences regarding use of activity trackers. For individuals, she recommends making sure to wear the device in the same position every day, enter and update personal characteristics (i.e. height, weight, and age), and make sure the device is set to receive updates so its algorithms remain accurate. For researchers, Dr. Evenson urged more studies focused on assessing reliability and validity of these instruments and making sure that they are tested for such in both free-living and lab-based conditions. Finally, she recommends that device manufacturers publicize information about updates to their software (apps), allow users to have access to data in smaller time intervals, and that they establish standards for rigor in what they measure.
It was a truly interesting lecture! For those who are interested in running future research about physical activity tracking, Dr. Evenson mentioned that there is a lack of evidence regarding the accuracy of trackers in a variety of populations including youth, older adults, participants with higher body mass index, and people with disabilities. She also recommended considering issues of privacy and informed consent in further studies because of long-term data storage.
Ginny – considering that the fitness industry is currently being flooded with a wide assortment of physical activity trackers, I certainly believe that tracking devices and fitness monitoring will serve as a cornerstone of physical activity research in the near future as access continues to increase. However, once this technology is deemed valid and reliable, it seems that it might be difficult to normalize data across multiple platforms and companies to ascertain that data is collected and interpreted correctly and without confusion. Data collected by one device brand may not be consistent with data collected by another, thus increasing the difficulty of establishing widespread public health recommendations based on activity trackers. One other issue to consider is the inability of researchers to blind their participants to the use of physical activity monitors. While it is ideal that subjects continue to participate in their regular amount of physical activity free of outside influence, some may wish to increase their energy expenditure because of the presence of the device while others could decrease or discontinue their physical activity due to discouraging results communicated by their tracker.
Ginny, great summary on activity trackers for physical activity research. The discussion on using objective measures such as these come up regularly in my other classes. Especially in regards to the reliability and validity of these instruments. The possibly using an activity tracking in future research studies is tempting because they, appear feasible, user-friendly, and low burden on participants. Do you currently have a Fitbit or other device that you are using to track your activity?
Hey Ginny, great write up! I did not get to attend this lecture although the topic was of interest to me personally as I am very passionate about health and fitness. I am also a personal trainer and this information is useful to me since clients use wrist-worn activity trackers that measure heart rate and activity levels. Initially skeptical of these tracker’s accuracy, I did some of my own research; a recent study in the Journal of Personalized Medicine looks at “accuracy in wrist-worn, sensor-based measurements of heart rate and energy expenditure” of seven major fitness trackers (Shcherbina et al., 2017). The studied showed that heart rate tracking was relatively accurate with three models under 5% median error, while calorie tracking was very inaccurate with a median error ranging from (27-98%). I would be interested to see what the next literature review Dr. Kelly Evenson completes on more recent data reveals on fitness trackers accuracy. As for the future of fitness trackers, I could definitely see them being utilized in public health studies on physical activity as the technology is perfected and accuracy increases. Very interesting!