Machine learning helping to predict diabetesThursday, 2 April 2020
A special supplement about to be published in the Journal of the Endocrine Society, reports that a type of Artificial Intelligence (AI) called machine learning may help predict who will develop diabetes.
Lead author of the study Akihiro Nomura, said “We currently don’t have sufficient methods for predicting which generally healthy individuals will develop diabetes.”
What is machine learning?
The researchers investigated how machine learning could help with the diagnosis of diabetes.
Machine learning enables computers to learn without being explicitly programmed. With each exposure to new data, a machine-learning algorithm grows increasingly better at recognizing patterns over time.
“Using machine learning, it may be possible to identify people at high risk of developing diabetes with more accuracy than current risk scores or calculators. In addition, analysing visits to healthcare providers might help predict the future onset of diabetes and improve prevention,” said Dr Nomura.
The data set
Nomura and his colleagues analysed over half a million Japanese national annual health records from 139,225 participants over a period of ten years. This included 65,505 participants who did not have diabetes.
The data included physical exams, blood and urine tests and participant questionnaires. People without diabetes at the beginning of the study who underwent more than two annual health check-ups during this period were included. New cases of diabetes were recorded during patients’ check-ups.
The researchers identified a total of 4,696 new diabetes patients (7.2%) in the study period. Their trained computer model predicted the future incidence of diabetes with an overall accuracy of 94.9%.
The next phase of research
The phase of the study is to perform clinical trials to assess the effectiveness of preventative treatments on those identified by the machine learning model as being at high risk of developing diabetes.