Nat Metabol: New algorithms could use artificial intelligence to help manage people with type 1 diabetes

, June 16, 2020 /PRNewswire/ -- In a recent study published in the international journalNature Metabolism, scientists from institutions such as Oregon Health and Science University have developed a new method that could help patients with type 1diabetesbetter manage their blood sugar levels using artificial intelligence and automated monitoringPicture Source: Wikipedia
Researcher Dr Nichole Tyler said: 'Our system design is so unique that we use special mathematical simulators to design artificial intelligence algorithms, but when this algorithm is validated on the real data of type 1diabetespatients, the recommendations are highly similar to those suggested by endocrinologists.' This is important becausepatients with diabetesoften take 3-6 months to make an appointment and appointment with an endocrinology doctorduring this time, if the patient's blood sugar level scane is too high or too low, there may be a risk of dangerous complications, type 1diabetespatients themselves can not produce enough insulin, so they have to use insulin pump or multiple insulin injections per day to control blood sugar, the new algorithm developed by the researchers in this study can use data collected from continuous glucose monitoring equipment and wireless insulin pen to guide patients' treatmentWhen used with a mobile phone called DailyDose
app , the algorithm's recommendations showed that it was in line with doctors 67.9 percent of the time study, researchers monitored 16 people with type 1 diabetes over a four-week period, and the new model showed that the new model could help reduce low blood sugar levels, which, if left untreated, can lead to coma and even death In this regard, there are other algorithms developed by the researchers, but not many related clinical studies, and few have shown statistically relevant results, and most do not compare the algorithm recommendations with those of doctors, so in addition to showing that blood sugar levels are under control, the recommendations produced by the new algorithm are very highly correlated with the recommendations of doctors, with more than 99 percent of the algorithms recommended by clinicians in 100-week patient tests being considered safe now researchers want to continue to delve into improving the technology, and they plan to conduct several larger trials over the next few years , ranging from 8 weeks to 12 weeks , while comparing daily dose differences with other insulin treatment strategies , including automated insulin-to-drug devices (BioValley Bioon.com) original origins: Tyler, N.S., Mosquera-Lopez, C.M., Wilson, L.M et al.
An artificial intelligence fod for the management of the type 1 diabetes .
Nat Metab (2020) doi: 10.1038/s42255-020-0212-y