Activity data from a Fitbit can predict changes in blood sugar control for adults with prediabetes, a condition that affects around one in three adults in the United States, a new study shows. The findings point to a strategy that tech companies might use in their rumored efforts to build diabetes technology into wearable products.
“It kind of makes sense intuitively — more movement, more physical activity leads to overall better health, and better health is one of the factors behind improved glycemic control,” says Jessilyn Dunn, an assistant professor of biomedical engineering at Duke University who wasn’t involved with this study but has also done work on wearables and glucose monitoring.
People with prediabetes have elevated blood sugar levels, placing them at risk of developing diabetes. But most tools that predict whether someone with prediabetes will progress to develop diabetes look years in the future, says study author Mitesh Patel, an associate professor of medicine at the University of Pennsylvania and vice president for Clinical Transformation at hospital group Ascension.
“There were no good near-term models to say, in the next six months, whose blood sugar is going to increase and get worse versus whose is going to get better,” Patel says.
In the new study, published in NPJ Digital Medicine, Patel’s team built models that would use activity data collected from either wrist- or waist-worn Fitbits to predict both changes in average blood sugar and 5 percent improved or worsened blood sugar levels. Over the course of the six-month study, they found that they were able to make accurate predictions, and that predictions were more accurate using data collected from the wrist-worn devices.
“We know that people who are generally more active have better control of their blood sugar, and people who are less active have worse control,” Patel says. “But there are other hidden patterns in the daily information we’re getting — how many steps are fast steps versus slow steps, and other nuances — that we can get from this information.” Because the wearable can capture that additional data, it can give a more granular look at how activity drives changes in blood sugar.
Other studies have also found that data from wearable devices can help track and predict blood glucose levels. But that despite those relationships, activity and other wearable data is still only a proxy for blood sugar levels, Dunn says. It’s still important to directly monitor blood sugar through blood testing methods until there’s more evidence to show that the links are strong enough to make treatment decisions. This new study was in under 200 people, for example, and only looked at people with prediabetes — more work would be needed to test the relationship in more people and at varying levels of health.
Tech companies are looking toward diabetes and blood sugar management as the next set of tools to incorporate into consumer-facing products like smartwatches and smart rings. Apple has been looking into noninvasive glucose monitoring for years, Fitbit has a partnership with diabetes tech company LifeScan, and a new smart ring is looking to blood sugar monitoring as a future goal.
These companies want to find ways to measure blood sugar without breaking the skin — their attempts have focused on sweat, tears, breath, and the reflection of light off the skin. Those strategies could offer more direct measures of glucose than activity data, but most haven’t come to fruition yet, Dunn says.
Dunn says she could imagine that, in the short term, tech companies might build in features that use activity data to give people some type of information about their blood sugar. Then, if the other, noninvasive strategies that more directly measure glucose levels prove to work well, those might be the next-gen tools.
“And so, then the question becomes, what is the next technology?” Dunn says. “And I think that remains to be seen — but we can be sure there will be something.”