Researchers at the University of Maryland have developed a diabetes risk tool that predicts the likelihood of complications in the near term, enabling doctors to target prevention more effectively and adapt care plans based on routine clinical data.
Researchers at the University of Maryland School of Medicine have built a diabetes risk calculator that aims to tell doctors not just what may happen years down the line, but which complications are most likely to emerge in the near term. The tool, known as the Diabetes Complications Risk Calculator, or DCRC, was led by Rozalina G. McCoy and published in Nature Communications. It uses routine clinical information already found in medical records to produce updated estimates as a patient’s condition changes.
The calculator draws on data from more than 400,000 adults in the United States who had recently been diagnosed with diabetes, combining insurance claims with electronic health records from ordinary care. According to the study, the model estimates the likelihood of nine acute and long-term complications, including cardiovascular disease, stroke, kidney disease, nerve damage and blood sugar emergencies. Unlike many existing tools that focus on a single outcome or distant future risk, the DCRC is designed to give encounter-specific predictions that can be refreshed over monthly intervals.
In validation testing, the models were assessed both in the original nationwide cohort and in an independent group of patients at Mayo Clinic. The researchers reported that the system generally showed good to strong ability to separate patients who went on to develop complications from those who did not. The study also found that complications can appear quickly after diagnosis: roughly one-third of patients had at least one complication within a year, and more than 40% had done so by two years.
The findings add to a growing body of work on personalised diabetes risk prediction. Earlier research has explored calculators for specific complications, including tools for microvascular disease in type 1 diabetes, while other models have used machine learning to estimate broader complication risk. The Maryland team said the DCRC could help clinicians and patients focus prevention efforts on the problems most likely to matter at a particular time, though they also noted clear limits. The study relied mainly on insured patients, prediction accuracy varied by complication, and the calculator still needs testing in everyday practice to show whether it improves decisions and outcomes.
Disclaimer: This content is for informational purposes only and is not intended to be a substitute for professional medical judgment, advice, diagnosis, or treatment.
