Using wearable device data to predict changes in glycemic control


About 90 million adults in the United States have prediabetes and are at risk of developing diabetes. Lifestyle modifications such as physical activity and weight loss can reduce the likelihood of diabetes, but it is often challenging to determine patient engagement with these behavior changes and their impact on glycemic control. Wearable devices can remotely transmit data on activity levels, but different types of wearable devices have not been compared for their ability to enhance risk prediction models.


We designed a clinical trial and randomly assigned prediabetic patients to use either a wrist-worn or waist-worn wearable device for six months.  The wrist-worn device is more expensive but tracks physical activity, sleep patterns, and heart rate.  The waist-worn device is less expensive but only tracks physical activity. We will compare these data collection methods head-to-head to enhance prediction models.


This trial is currently in progress and results are expected to be reported in 2019.


Pennsylvania Department of Health


Work with us


3400 Civic Center Boulevard
14th Floor, South Pavilion
Philadelphia, PA 19104