Can AI Predict Who Will Develop Diabetes?
TUESDAY, March 31, 2020 (HealthDay News) -- Artificial intelligence (AI) may be able to identify people who will develop type 2 diabetes, researchers say.
For the study, the researchers used machine learning AI to analyze more than 509,000 annual health checkup records of more than 139,000 people in Japan from 2008 to 2018. They included more than 65,000 who did not have diabetes in 2008.
The data included information from physical exams, blood and urine tests, and patient questionnaires.
During the study period, nearly 4,700 cases of diabetes were diagnosed. Machine learning AI was 94.9% accurate in predicting cases, according to the study published March 31 in a special supplemental section of the Journal of the Endocrine Society.
Diabetes is associated with an increased risk of severe health problems, including heart disease and cancer, so preventing it is a key to reducing the risk of illness and premature death.
"Currently we do not have sufficient methods for predicting which generally healthy individuals will develop diabetes," said lead author Dr. Akihiro Nomura, a cardiologist at Kanazawa University Graduate School of Medical Sciences.
"Using machine learning, it could be possible to precisely identify high-risk groups of future diabetes patients better than using existing risk scores," he said in a news release from the Endocrine Society. "In addition, the rate of visits to health care providers might be improved to prevent future onset of diabetes."
In machine learning, computers learn without being explicitly programmed. With each exposure to new data, a machine-learning formula gradually becomes better at recognizing patterns.
The next step in this research is conducting clinical trials to assess the effectiveness of using statins to treat patients identified by the machine learning model as being at high risk for diabetes, Nomura said.
The U.S. National Institute of Diabetes and Digestive and Kidney Diseases explains how to prevent type 2 diabetes.