Figure 1.  Diabetes Onset Prediction Performance. A Calibration is assessed...
Figure 1. Diabetes Onset Prediction Performance. A Calibration is assessed visually with a calibration curve composed of 20 population bins of equal size. B Precision and recall curves are displayed. The left y-axis corresponds to precision, and the right y-axis to recall. The test area under the receiver operating curve is 80.26.
Source: DOI: 10.1001/jamanetworkopen.2021.11315

AI predicts risk of developing type 2 diabetes

Artificial intelligence could be used to predict who is at risk of developing type 2 diabetes—information that could be used to improve the lives of millions of Canadians.

Researchers at the University of Toronto used a machine learning model to analyze health data, collected between 2006 to 2016, of 2.1 million people living in Ontario. They found that they were able to use the model to accurately predict the number of people who would develop type 2 diabetes within a five-year time period. The machine learning model was also able to analyze different factors that would influence whether people were high or low risk to develop the disease.

"We know that identifying people who are at risk of developing type 2 diabetes is really important because there are things we can do to prevent the onset of the disease," says senior author Laura Rosella, an associate professor in the University of Toronto's Temerty Faculty of Medicine and the Dalla Lana School of Public Health. "This machine learning model can help with managing one of the biggest chronic disease challenges in North American society. There is a demonstrated advantage to intervening early when people are at risk of type 2 diabetes."

Rosella, who is the education lead for the Temerty Centre for AI Research and Education in Medicine (T-CAIREM), says the study's findings could help inform larger health system strategies to decrease the number of people who develop type 2 diabetes. "The model is about 80 percent accurate when it comes to predicting who will develop type 2 diabetes," she says. "By using this information in a proactive way, we can plan health systems better and help prevent what can be a serious, burdensome condition."

The use of a machine learning model is important, says Rosella, because it shows how routinely collected data can be used to address complex health problems in a more effective way.

Preventing type 2 diabetes means looking at larger structural factors like food insecurity and access to primary care physicians, Rosella adds. "We know diabetes can be prevented or delayed. We know there are effective ways we can prevent the onset of a chronic disease. This study offers a way to start thinking about how to identify who is at risk of type 2 diabetes, and then start implementing strategies to stop the onset of a debilitating, lifelong condition."

Vinyas Harish, an MD-Ph.D. candidate at the Temerty Faculty of Medicine and learner co-lead at T-CAIREM, says the research illuminates how scrutinizing social determinants of health have an important impact on stopping the spread of type 2 diabetes. "It helps us think about what we can do to get a health system to intervene on larger, more structural factors," he says.

Rosella says medical research that incorporates artificial intelligence requires a team approach. "You need a multi-disciplinary group of people that include a group of really good computer scientists, people that understand data and how to use it, and people with a health-system perspective and a clinical perspective," she says. "This is needed to make sure that you're coming up with algorithms that are actually going to be used and have an impact."

The results of the study were published in JAMA Network Open.

Subscribe to our newsletter

Related articles

'Deepfaking the mind' to improve brain-computer interfaces

'Deepfaking the mind' to improve brain-computer interfaces

Researchers are using generative adversarial networks to improve brain-computer interfaces for people with disabilities.

ReSkin helps to discover a sense of touch

ReSkin helps to discover a sense of touch

Carnegie Mellon University and Meta AI (formerly Facebook AI) want to increase the sense of touch in robotics, wearables, smart clothing and AI.

AI Eve augments genetic tests

AI Eve augments genetic tests

AI model called EVE shows remarkable capacity to interpret the meaning of gene variants in humans as benign or disease-causing.

An AI that ‘thinks’ like humans

An AI that ‘thinks’ like humans

Creating human-like AI is about more than mimicking human behaviour – technology must also be able to process information, or ‘think’, if it is to be fully relied upon.

Enabling AI-driven advances without sacrificing privacy

Enabling AI-driven advances without sacrificing privacy

Secure AI Labs is expanding access to encrypted health care data to advance AI-driven innovation in the field.

Monitoring mental health in cancer survivors with "FAITH"

Monitoring mental health in cancer survivors with "FAITH"

AI-based solution FAITH is designed to monitor the mental health status of people who have undergone cancer treatment.

Researchers psychoanalyse artificial intelligence

Researchers psychoanalyse artificial intelligence

We can run tests and experiments, but we cannot always predict and understand why AI does what it does.

The progress and risks of artificial intelligence

The progress and risks of artificial intelligence

Artificial intelligence has reached a critical turning point in its evolution, according to an international panel of experts.

Machine learning helps diagnose leukemia

Machine learning helps diagnose leukemia

Researchers at the University of Bonn show how artificial intelligence improves the evaluation of blood analysis data.

Popular articles

Subscribe to Newsletter