Jun 15 2020
At Oregon Health & Science University (OHSU), scientists and physicians have used automated monitoring and artificial intelligence to develop a technique to assist people with type 1 diabetes to better regulate their glucose levels.
The study was reported in the Nature Metabolism journal.
Our system design is unique. 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.
Nichole Tyler, MD, PhD, Study Lead Author, School of Medicine, Oregon Health & Science University
The study is crucial as people suffering from diabetes normally visit their endocrinologist after a period of three to six months.
During that time, people could face the risk of having harmful complications if their blood glucose levels increase too high or fall too low. People having type 1 diabetes cannot produce their own insulin, and hence, they must take it constantly throughout the day through an insulin pump or via several daily injections.
The algorithm designed by researchers at OHSU makes use of the data gathered from a continuous glucose monitor and wireless insulin pens to offer advice on adjustments.
When coupled with a smartphone app known as DailyDose, the suggestions from the algorithm were demonstrated to be in accordance with that of physicians 67.9% of the time.
As part of the new study, 16 people with type 1 diabetes were monitored for four weeks, demonstrating that the model can be useful in reducing hypoglycemia—low glucose. Hypoglycemia could lead to coma or even death if it is not treated.
The algorithm was designed jointly with the OHSU Harold Schnitzer Diabetes Health Center and the Artificial Intelligence for Medical Systems Lab headed by Peter Jacobs.
There are other published algorithms on this, but not a lot of clinical studies. Very few have shown a statistically relevant outcome–and most do not compare algorithm recommendations with those of a physician.
Peter Jacobs, PhD, Study Senior Author and Associate Professor, Biomedical Engineering, School of Medicine, Oregon Health & Science University
Jacobs continued, “In addition to showing improvement in glucose control, our algorithm-generated recommendations that had very high correlation with physician recommendations with over 99% of the algorithm’s recommendations delivered across 100 weeks of patient testing considered safe by physicians.”
OHSU aims to continue to further develop the technology.
We have plans over the next several years to run several larger trials over eight and then 12 weeks and to compare DailyDose with other insulin treatment strategies, including automated insulin delivery.
Jessica Castle, MD, Study Co-Author and Associate Professor of Medicine (endocrinology, diabetes and clinical nutrition), School of Medicine, Oregon Health & Science University
This study was financially supported by the National Institutes of Health/National Institute of Diabetes and Digestive and Kidney Diseases (grant 1 R01DK120367-01), The Leona M. and Harry B. Helmsley Charitable Trust (grant 2018PG-T1D001) and a Dexcom grant.
Tyler, N. S., et al. (2020) An artificial intelligence decision support system for the management of type 1 diabetes. Nature Metabolism. doi.org/10.1038/s42255-020-0212-y.