Hitachi and University of Utah Health develop new AI method for diabetes patients needing complex drug treatment

Hitachi has collaborated with the University of Utah Health and the non-profit health informatics research organisation Regenstrief Institute to develop and test a new AI method that supports care for patients with type 2 diabetes who require complex drug treatment.

WHAT IT’S ABOUT

In a study, the novel AI method analysed EHR data across Utah and Indiana in the United States and learned generalisable treatment patterns and clinical outcomes among type 2 diabetes patients with similar characteristics. It then further organised patients into disease state groups and predicted the range of potential health outcomes depending on treatment options. 

According to a press statement, the algorithm was able to support medication selection for over 83% of patients, even when two or more medications were used.

Some of the findings from this study have been published in the peer-reviewed medical journal, Journal of Biomedical Informatics.

WHY IT MATTERS

More than 500 million people in the world are at risk of getting type 2 diabetes, according to the latest estimate by the International Diabetes Federation. A small portion of them, Hitachi noted, require multiple medications to control blood sugar and avoid serious complications like vision loss and kidney disease. Their doctors may also have limited experience in evidence-based guidance for choosing drug combinations and giving the right treatment decision.

The patterns learned by the new AI method developed by Hitachi and its partners can be used to assist clinicians in determining an optimal drug regimen for a specific patient. 

In a statement, the research team behind this technology said they wanted to help diabetes patients requiring complex treatment to receive the right treatment plan from their doctors after checking the efficacy of various drug combinations. “This will lead not only to better management of diabetes but increased patient engagement, compliance, and quality of life,” they added.

The partners said they will continue evaluating and improving the effectiveness of their AI method and accelerate its practical application by partnering with healthcare IT businesses and R&D groups.

THE LARGER TREND

Hitachi has worked with U of U Health since 2018 on the development of a medication selection system for treating diabetes. The system they earlier built, however, cannot always accurately predict more complex and less prevalent treatment patterns due to insufficient data. This is why they have partnered with Regenstrief, a research institution associated with Indiana University, to tap into a wealth of health data.

In recent years, more AI-enabled tools for diabetes management have been approved for commercial rollout. Bigfoot Biomedical, a medtech startup in California, was cleared last year by the US Food and Drug Administration for its diabetes management system that provides insulin dose recommendations for diabetic users who are under multiple-dose injection therapy. 

Israeli startup DreaMed also scored an FDA clearance for its AI-enabled diabetes care platform catering to type 2 diabetes patients. The Advisor Pro Al Clinical Decision Support System helps providers manage their patients’ insulin needs and monitor their blood sugar levels using continuous glucose monitors.