A new study released online in the peer-reviewed journal Nature Metabolism unveiled an algorithm using artificial intelligence effectual for type 1 diabetes (T1D) maintenance. The study was conducted by a research group at Oregon Health & Science University.
The Oregon research group placed their focus on 16 participants diagnosed with type 1 diabetes over a one month period. The engine used in the study was constructed by the Artificial Intelligence for Medical Systems Lab, in conjunction with the OHSU Harold Schnitzer Diabetes Health Center.
In the study, the AI-based algorithm was used alongside a mobile application known as DailyDose.
According to their findings, the Oregon research group showed that the new AI-based algorithm was effectual at recommendations, with high similarities to that of endocrinologists. Such attributes are crucial given that patients with diabetes experience a three to six month wait time to consult with their endocrinologist.
“Over 40% of people with T1D manage their glucose through multiple injections of long-acting basal and short-acting bolus insulin, so-called multiple daily injections. Errors in dosing can lead to life-threatening hypoglycaemia events and hyperglycaemia, increasing the risk of retinopathy, neuropathy, and nephropathy,” the co-authors explained in their findings.
“Machine learning (artificial intelligence) approaches are being harnessed to incorporate decision support into many medical specialties.”
“We designed the AI algorithm entirely using a mathematical simulator, and yet when the algorithm was validated on real-world data from people with type 1 diabetes at OHSU, it generated recommendations that were highly similar to recommendations from endocrinologists,” the co-authors determined.
The Oregon research group is set to conduct more research using DailyDose in hopes of finding more efficient ways of managing insulin treatment for diabetic patients.
