Researchers at the University of Maryland School of Medicine (UMSOM) have developed and validated a new risk calculator that can estimate an individual patient’s short-term risk of developing a range of diabetes-related complications, using information already collected during routine medical care.
Results were published in the journal Nature Communications.
The study, led by Rozalina G. McCoy, MD, MS, Associate Professor of Medicine in the Division of Endocrinology, Diabetes, and Nutrition, analyzed health data from more than 400,000 adults newly diagnosed with diabetes across the United States. She and her colleagues created a set of prediction models—called the Diabetes Complications Risk Calculator (DCRC)—that can estimate a patient’s likelihood of developing several common complications and update those estimates as new clinical information becomes available.
People with diabetes are at risk for a wide range of complications, including heart disease, kidney disease, nerve damage, eye disease, and emergencies caused by very high or very low blood sugar. While prediction tools exist, most focus on just one complication at a time or predict the risk of complications over a much longer period of time. The current models also usually rely on data from specialized research groups rather than real-world care settings.
The new risk calculator was designed to address those gaps. It can estimate risk for nine different acute and chronic complications, including cardiovascular disease, stroke, kidney disease, nerve damage, and blood sugar crises—all at once and over short time intervals that may help improve clinical decisions and patient care.
“Our goal was to create a tool that reflects the reality clinicians face, where patients often have multiple concurrent and competing risks,” said Dr. McCoy who is also Director of the Precision Medicine and Population Health Program at the University of Maryland Institute for Health Computing. “By looking at these risks together and updating them over time, we can better understand what complication or complications our patients are most likely to experience, which can ultimately support more informed and actionable conversations between patients and their clinicians.”
The researchers used machine learning — a type of statistical method that can identify patterns in large datasets — to analyze insurance claims and electronic health record data. The models incorporate commonly available information such as age, existing health conditions, medications, and laboratory tests.
Unlike traditional models that provide a single long-term estimate, the DCRC produces monthly, encounter-level risk estimates that change as a patient’s health status evolves.
In testing, the models showed good to strong accuracy in predicting whether patients would develop specific complications, both in the original nationwide dataset and in an independent group of patients treated at Mayo Clinic.
Over time, diabetes complications were common in the study population. Within one year of diagnosis, about one-third of patients had experienced at least one complication, and that number rose to more than 40 percent after two years. The models also identified factors linked to higher risk including:
- Older age
- High blood pressure and related complications
- Longer duration of diabetes
- Kidney function measures
- Coexisting health conditions
Importantly, risk varied from person to person and could change over time—sometimes rising or falling as health conditions and treatments changed.
“Machine learning and other AI-based methods agentic systems can scan routine clinical and laboratory data in our electronic health records to identify patients who are at high risk for developing irreversible complications of diabetes,” said UMSOM Dean Mark T. Gladwin, MD. “This diabetes risk calculator demonstrates the power of this approach by continuously updating individual risk estimates as a patient’s health evolves using routine clinical data that already exists in electronic health records.”
Dr. McCoy emphasized that the calculator is not intended to replace clinical judgment. Instead, it is designed to support care decisions, helping clinicians and patients weigh risks and prioritize prevention strategies. For example, the tool could help identify patients who may benefit from closer monitoring or earlier interventions, or guide discussions about treatment choices.
“While our results are encouraging, these models should be used cautiously and in combination with clinical expertise,” she said. “We need to do more testing to understand how the tool performs when used in everyday clinical practice.”
Limitations of the tool include the use of data only from insured patients, which may not fully reflect patients without consistent access to medical care. In addition, some of its predictions were less accurate for certain complications.
The research team plans to evaluate how the calculator performs when integrated into real-world clinical workflows and whether it can improve shared decision-making and long-term health outcomes.
Study funding was provided by the National Institute of Diabetes and Digestive and Kidney Diseases (grant number K23DK114497), the National Institute on Aging (NIA) (grant number P30AG097158), the Diane Deshong Family Fund for Artificial Intelligence in Healthcare Delivery, and the Mayo Clinic Robert D. and Patricia E. Kern Center for the Science of Health Care Delivery.
Original release: https://www.medschool.umaryland.edu/news/2026/new-clinical-tool-helps-predict-short-term-risk-of-diabetes-complications-using-routine-health-data.html
