This article is more than 5 years old. Information may no longer be current.
App predicts blood glucose levels in type 2 diabetes
Click Here to Manage Email Alerts
Researchers at Columbia University are developing a personalized algorithm that forecasts what impact certain foods will have on an individual’s glucose levels, according to a press release.
The algorithm has been built into the app Glucoracle, which aims to assist patients with type 2 diabetes in achieving a better grasp on their glucose levels.
The algorithm uses data assimilation and is frequently updated to improve the model’s predictions.
Glucoracle collects data by allowing users to upload fingerstick glucose measurements and a photo of their meal directly to the app. The user must also provide an estimate of the nutritional content of the meal, to which the app responds with an intermediate prediction of post-meal glucose levels.
The app can begin generating predictions after a week of use.
"While we know the general effect of different types of food on blood glucose, the detailed effects can vary widely from one person to another and for the same person over time," David Albers, PhD, associate research scientist in Biomedical Informatics at CUMC, said in the release. "Even with expert guidance, it's difficult for people to understand the true impact of their dietary choices, particularly on a meal-to-meal basis.”
In a study, five individuals (three with type 2 diabetes and two without diabetes) used the app. The predictions produced by the app were compared with real post-meal blood glucose measurements and the predictions of certified diabetes educators.
For the individuals without diabetes, the app’s forecasts were comparable to the actual measurements. However, the app’s predictions for the people with diabetes were less accurate but still comparable to the predictions made by the diabetes educators. This disparity in accuracy could have been due to parameter error or the fluctuating physiology of diabetes patients, according to the release.
The researchers are planning a larger clinical trial to better understand the app’s efficacy. The team estimates the app could be ready for widespread use in as little as 2 years.