APR 07, 2020 7:37 AM PDT

Computers Predict Diabetes with 94.9 Percent Accuracy

WRITTEN BY: Tara Fernandez

"Currently we do not have sufficient methods for predicting which generally healthy individuals will develop diabetes," says Akihiro Nomura of Kanazawa University, author of a new study that used artificial intelligence (AI) as a clinical tool to predict diabetes.

Diabetes mellitus is a chronic condition that affects how the body metabolizes glucose. Patients with Type 1 and Type 2 diabetes have too much sugar in their blood, a condition that can result in serious health complications. A CDC report estimates that over 100 million Americans have either diabetes or prediabetes. However, the problem, as Nomura points out, is that healthcare providers do not have a good way of foreseeing who to focus clinical and lifestyle interventions on to prevent the future onset of diabetes — until now.

Machine learning is a type of AI in which computers can “learn”, without programming, to make connections out of large datasets. Making sense of massive amounts of data is often too complicated and overwhelming for the human brain to achieve. Additionally, machine learning systems get “smarter” as they are exposed to more data, creating an ever-evolving system that is better equipped at recognizing subtle patterns. In this study, the researchers fed the computer system over half a million health checkup records from over 139,000 participants over a decade. About half of these individuals did not have diabetes, while around 5,000 did go on to be diagnosed with the disease.

According to Nomura, "Using machine learning, it could be possible to precisely identify high-risk groups of future diabetes patients better than using existing risk scores." 

"In addition, the rate of visits to healthcare providers might be improved to prevent the future onset of diabetes." Impressively, the machine learning system predicted the future development of diabetes with an accuracy of 94.9 percent.

 

 

The researchers’ next steps involve clinical trials to assess the robustness of the system among larger cohorts and how pharmaceutical interventions such as statin treatment can change patient outcomes.

 


Sources: Medical XPress, Journal of the Endocrine Society.

About the Author
  • Tara Fernandez has a PhD in Cell Biology and has spent over a decade uncovering the molecular basis of diseases ranging from skin cancer to obesity and diabetes. She currently works on developing and marketing disruptive new technologies in the biotechnology industry. Her areas of interest include innovation in molecular diagnostics, cell therapies, and immunology. She actively participates in various science communication and public engagement initiatives to promote STEM in the community.
You May Also Like
NOV 25, 2019
Clinical & Molecular DX
NOV 25, 2019
eRapid: molecular diagnostic power in the palm of your hand
We've heard lofty biotech promises in the news of being able to diagnose diseases from a single drop of blood. Yet, diag ...
FEB 15, 2020
Clinical & Molecular DX
FEB 15, 2020
FDA nod for AI-powered technology to detect strokes
The US Food and Drug Administration (FDA) has provided clearance for a novel technology that uses artificial intelligenc ...
APR 16, 2020
Genetics & Genomics
APR 16, 2020
A New Way to Quickly Test for COVID-19
Some of the reagents used in popular diagnostic tests may be in short supply, so other testing methods will be needed.
APR 22, 2020
Clinical & Molecular DX
APR 22, 2020
Microneedle Skin Patch Collects Bodily Fluids for On-patch Diagnostic Testing
The collection of blood for diagnostic tests is common practice in clinical settings. However, technicians often need to ...
MAY 21, 2020
Clinical & Molecular DX
MAY 21, 2020
Fidget Spinner Diagnoses Infections
The fidget spinner toy craze took the world by storm — a small, boomerang-shaped gadget that rotates hypnotically ...
MAY 23, 2020
Microbiology
MAY 23, 2020
The FDA Yanks Some COVID-19 Antibody Tests From the Market
The massive demand for diagnostic testing led the FDA to open a short window for many testing products to go to market w ...
Loading Comments...