Active Voice: Measuring Children’s Physical Activity with Accelerometers: Apples and Pears

By Stuart J. Fairclough, Ph.D. 

Children are the most active segment of the population, but concerns exist about their declining levels of physical activity – a problem that becomes progressively more pronounced as children advance through adolescence. Accurate measurement of children’s physical activity (PA) is challenging because their activity behaviors typically include high tempo bouts of moderate and vigorous PA, interspersed with periods of light PA and rest. Accelerometers are the instruments of choice to measure children’s free-living PA. However, variation in accelerometer brands, location that the device is worn on the body, output metrics and data processing procedures all combine to limit the degree to which children’s PA data are comparable between studies. 

Most previous studies have used hip-mounted Actigraph accelerometers to record children’s PA, based on conversion of the manufacturer’s proprietary ‘counts’ to time spent in PA at different intensities. Recently, PA intensity thresholds derived from raw accelerations have been developed for the ActiGraph GT3X+ and GENEActiv accelerometers. Basing PA data on raw accelerations provides an opportunity to improve comparability between studies using different devices, and promote transparency and consistency of post-data collection analytical processes. 

In our study, recently reported in MSSE, we examined children’s compliance to wearing hip and wrist-worn accelerometers, compared children’s PA derived from raw acceleration signals from these devices. Using these data, we investigated differences in PA estimated from raw data with that from counts data. For seven consecutive days during waking hours, 129 healthy children (9-10 years old) wore an Actigraph GT3X+ accelerometer (AGhip) on the right hip and a GENEActiv accelerometer (GAwrist) on the non-dominant wrist. Data were downloaded and analysed based on raw acceleration data (AGhip and GAwrist) and counts data (AGhip only). 

We found that children wore the GAwrist for longer each day and for more days over the week than the AGhip. Children recorded more time being physically active based on GAwrist raw accelerations compared to those from the AGhip. In fact, 86.9 percent of children achieved current PA guidelines based on their GAwrist data, compared to 19 percent according to raw AGhip data. Comparison of PA from AGhip raw and counts data revealed a seven min/d difference in the category of moderate PA (raw data greater than counts data) and a 30 min/d difference in vigorous PA (counts data greater than raw). 

In children, it would appear that accelerometer wrist placement promotes superior compliance compared to the hip. This reflects the more recent NHANES accelerometer data collection cycles that have employed wrist-worn accelerometry (see: Freedson & John, MSSE 2013, for related comment). Using raw accelerations would seem to offer greater potential for comparability of PA outcomes between devices than previous counts-based approaches based on manufacturers’ proprietary algorithms. However, we found substantial differences in PA that were possibly due to accelerometer placement location and technical differences between devices. From a health promotion perspective, current PA guidelines are based mainly on self-report questionnaires and, to a lesser extent, on data from hip-worn accelerometer counts. As the use of raw acceleration data increases, examination of activity-health relationships using raw data from wrist-worn devices is warranted.

Viewpoints presented on the ACSM blog reflect opinions of the authors and do not necessarily reflect positions or policies of ACSM. 

Stuart J. Fairclough, Ph.D., is a professor of physical activity education at Edge Hill University in the United Kingdom. His research focuses on youth physical activity and health, with an emphasis on physical activity correlates, school-based interventions and physical activity measurement. 

This commentary presents Professor Fairclough’s views on the topic of a research article which he and his colleagues published in the February 2016 issue of Medicine & Science in Sports & Exercise® (MSSE). 

Your Comments