Machine Learning Predicts Risk of Heart Failure in Patients with Diabetes

A machine-learning model accurately predicted the risk of future heart failure among patients with diabetes, according to new research from Brigham and Women’s Hospital and UT Southwestern Medical Center.

Researchers created the WATCH-DM risk score for five-year incidence based off the top performing predictors of heart failure. The predictors included weight, age, hypertension, creatine, high-density lipoprotein cholesterol, diabetes control, QRS duration, myocardial infarction and coronary artery bypass grafting.

Patients who developed heart failure were older, more commonly men and had a higher body mass index. These patients also had higher frequencies of prevalent atherosclerotic cardiovascular disease and had longer average durations of Type 2 diabetes mellitus, hypertension and hyperlipidemia.

“Our risk score provides a novel prediction tool to identify patients who face heart failure risk in the next five years,” said co-first author Matthew Segar, M.D., a resident physician at UT Southwestern. “By not requiring specific clinical cardiovascular biomarkers or advanced imaging, this risk score is readily integrable into bedside practice or electronic health record systems and may identify patients who would benefit from therapeutic interventions.”

The research team used data from 8,756 patients with diabetes who were enrolled in the Action to Control Cardiovascular Risk in Diabetes trial. Data included 147 variables, including demographics, clinical information and laboratory data.

Researchers used machine-learning methods to determine the top-performing predictors of heart failure. Over the course of fewer than five years, 3.6% of the patients developed heart failure.

Patients with the highest risk scores faced a five-year risk of heart failure near 20%.

Each one-unit increment of the risk score was associated with a 24% higher risk of heart failure over the next five years.

The WATCH-DM risk score is available as an online tool for clinicians to use.

“We hope that this risk score can be useful to clinicians on the ground — primary care physicians, endocrinologists, nephrologists and cardiologists — who are caring for patients with diabetes and thinking about what strategies can be used to help them,” said co-first author Muthiah Vaduganathan, M.D., MPH, a cardiologist at Brigham and Women’s.

Patients can look at the predictors to better manage their health, said Vaduganathan. The research highlights ways for patients to intervene through lifestyle changes and through the use of SGLT2 inhibitors to delay or prevent heart failure, he added.

“This risk tool is an important step in the right direction to promote prevention of heart failure in patients with Type 2 diabetes,” said senior author Ambarish Pandey, M.D., a preventive cardiologist at UT Southwestern. “It can be readily used as part of clinical care of patients with Type 2 diabetes and integrated with the electronic (health) records to inform physicians about the risk of heart failure in their patients and guide use of effective preventive strategies.”

The research team is currently working to integrate the risk score into the electronic health record systems at their respective hospitals.

More studies will be needed to develop specific risk scores to predict heart failure with preserved ejection fraction among the general population and among patients with diabetes, the authors noted.

Get the best insights in digital health directly to your inbox.

Related

AI Model IDs Congestive Heart Failure from Single Heartbeat

How Machine Learning Can Identify Patients at Risk of Diabetes

Unnecessary Testing Proves Costly for Patients With Type 2 Diabetes