Baltimore, Maryland-based mobile health company WellDoc has developed a model that can predict hypoglycemic events in patients with Type 2 diabetes 91 percent of the time based on once-a-day patient blood glucose testing.
Predicting hypoglycemic events, a term for when people with diabetes are dangerously low blood sugar, up to 24 hours before they occur -- as WellDoc says its model can do -- could potentially be very valuable in diabetes management. But most of the work that's been done on predicting diabetes uses continuous glucose monitoring, which doesn't reflect the way people with Type 2 diabetes actually monitor their blood glucose.
"People respond best to information that is personal, timely, and actionable," Robin Anthony-Kouyate, director of behavioral science at WellDoc, said in a statement. "Once a patient is armed with the knowledge that a hypo event is likely to occur in the next 24 hours they may become more receptive to learning about the consequence of the event and what they can do to avoid going hypo."
WellDoc tested the model using more than 10,000 data points from the International Diabetes Center and WellDoc’s proprietary databases. This also enabled them to include a wide variety of patients using different medications and on different insurance plans. They found that the algorithm "correctly predicted hypoglycemic incidents 42 percent more often than expert human endocrinologists, with low rates of false positives and false negatives," according to lead researcher Bharath Sudharsan.
WellDoc plans to incorporate the model into its BlueStar prescription diabetes management system, pending clearance from the FDA. The technology could improve the system's mobile coaching element, reducing the instance of hypoglycemic events in BlueStar users.