Active Voice: How Accurate Are Wearable Activity Monitors?

By Jung-Min Lee, Ph.D.

Viewpoints presented in SMB commentaries reflect opinions of the authors and do not necessarily reflect positions or policies of ACSM.

Jung-Min Lee, Ph.D. completed his doctoral training at Iowa State University and currently is an assistant professor in the School of Health, Physical Education and Recreation at the University of Nebraska - Omaha. Dr. Lee’s research encompasses physical activity and health promotion, focusing on the development and validation of objective techniques to assess habitual physical activity among diverse audiences and, for the physical activity environment, using Geographical Information Systems (GIS).

This commentary presents Dr. Lee’s views on the topic of an article which he and his colleagues published in the July-August 2014 issue of ACSM’s Health & Fitness Journal (FIT).

Physical activity has been studied with various monitoring methods. It started with a variety of assessment tools such as activity logs, questionnaires and direct observation, then progressed to wearable monitors including pedometers, heart rate monitors and accelerometers. All of these methods have been tested both in the laboratory and free-living conditions. Among these, accelerometry-based activity monitors have been widely adapted as an objective assessment tool to measure individual’s regular physical activity.

Over the last several decades, tracking or monitoring physical activity was utilized primarily for research purposes - to examine the relationship between individuals’ daily physical activity level and health related outcomes. However, with significant advances in accelerometer technology and the public’s increased awareness of physical fitness, numerous companies have developed wearable monitors designed to help individuals track their personal activity patterns.

To date, the wearable monitors market has been growing rapidly, and tracking daily physical activity is getting popular among the general population. In addition, a growing number of companies and fitness facilities have adapted wearable monitors not only to promote employee wellness but to track client’s activity. This technology is able to provide web-based tailored physical activity advice through its website or smartphone application. However, there has been a lack of scientific research to evaluate the accuracy and practicality of these various wearable activity monitors.

In our recent research paper published in FIT, we formally evaluated the accuracy of energy expenditure in eight different wearable activity monitors (BodyMedia® Fit, DirectLife, Fitbit® One, Fitbit® Zip, Jawbone UP®, Nike+Fuel Band, Basis Band, and Actigraph GT3X+) under semi-structured, free-living conditions. The criterion measure for energy expenditure was a portable metabolic analyzer. Our results revealed that the performance in estimating energy expenditure of these wearable monitors is quite impressive, as most had mean absolute percentage error (MAPE) values between 10 and 15 percent. The performance is especially remarkable considering the diverse range of activities tested in the study. However, when it comes to measuring steps, the wrist-oriented wearable monitors have a tendency to overestimate compared to manual step counts because of the wearing position (unpublished data). But our pilot study indicated that clip-style wearable monitors do accurately measure daily steps (MAPE: 7-12%) compared to the research grade pedometer. Taken collectively, our results demonstrate good potential for almost all of the devices since the accuracy was generally similar to that of research grade activity monitors (i.e., Actigraph and New Lifestyles Pedometer).

This new class of wearable activity monitors offers considerable potential to advance the science and practice of physical activity promotion, but they should not be viewed as a complete solution or all-in-one package. The wearable monitors provide valuable feedback to help individuals monitor their activity behavior. However, training and support are needed to help users learn how to use and apply this information to achieve more active lifestyles. In addition, further research is still needed to systematically evaluate the reliability, validity, and utility of these tools for promoting physical activity behavior in the population.