Researchers are developing machine learning models that accurately predict diabetes, the study says.

Researchers in Toronto-Canada have developed a machine learning model that accurately predicts diabetes in the population using health data collected on a regular basis, the new study says.

the study, Published in JAMA Network Open Journal, Tested new machine learning technology with regularly collected health data surveying the entire population of Ontario. This study was conducted by ICES’Non-Profit Data Institute.

Using linked administrative health data from Ontario from 2006 to 2016, researchers created a validated algorithm by training the model with information obtained from approximately 1.7 million patients.

Using collected validated data that includes variables such as body mass index and hypertension, the algorithm was 80% accurate in predicting people at risk for type 2 diabetes in the population over the years.

“There are effective preventive care and specific interventions to prevent type 2 diabetes, but often make sure that these approaches are targeted at those who need them most in the context of the healthcare system. It’s hard to do, “said ICES scientist and principal investigator Dr. Laura Rossella in the release. “Predictive models of type 2 diabetes are particularly useful in communicating more effective and efficient targeting of healthcare system interventions that support the prevention of type 2 diabetes.”

The role of machine learning models is not in the care of individual patients, but in the context of population health planning and management to predict diabetes and promote health fairness within the larger framework of the overall healthcare system. Used in.

Five years into the future, “Investments and targeted interventions to prevent diabetes, using artificial health planning tools that accurately distinguish between high-risk and low-risk groups by using machine learning and health data. We can guide you, “says the study.

“This model has been tested for a total of five years for accuracy in predicting the onset of diabetes among important demographic and socio-economic subgroups, confirming that it works well for the entire population. We did, “said Rossella.

According to the study, the number of people with diabetes in Ontario was estimated to be 785,000 in 2009, with associated costs of $ 3.5 billion, but seven years later in 2016, it was $ 1144,000 and $ 5.4 billion, respectively. Has increased.

Researchers found that the cohort of diabetics increased at an average annual rate of 51,800 new patients between 2009 and 2016, adding $ 242 million annually in associated costs.

The model also showed that patients with the highest risk of diabetes, according to the data, accounted for the largest portion of the associated medical costs. Medium-to-high-risk patients, whose population was surveyed, accounted for 26 percent of the total cost of associated diabetes.

According to this study, system-level barriers such as socio-economic status disparities, lack of access to healthy foods and medicines, and lack of access to health care “diabetes from individual patients to populations. It was difficult to “scale up prevention”.

The highest-risk patients averaged 58 years, included a larger cohort of immigrants, and those who were more likely to live in areas with low socio-economic indicators and high unemployment.

The new model is used by governments, health insurance providers, to identify those who are most in need of diabetes intervention and prevention, especially those who are older and come from “alienated areas in terms of ethnicity and material deprivation.” It aims to address the issues faced by public health planners.

Use Canadian Press files