June 17, 2026

Canadian researchers develop tool to identify kidney disease risk earlier

Kidney disease can often be slowed—or even prevented—when it is identified early. But many people do not know they are at risk until kidney function has already declined.  

To help address this at a population level, researchers in Manitoba and Ontario have developed a model that uses anonymized health data to predict which people may be at high risk of developing chronic kidney disease (CKD).  

The goal is to help identify people earlier so they can receive monitoring, treatment, and support sooner—potentially delaying or preventing the need for dialysis or a kidney transplant.  

Ottawa-based researcher Dr. Manish M. Sood and colleagues developed the prediction model using anonymized health records from more than 413,000 people in Manitoba. The model uses six factors that are commonly found in health records and are linked to kidney disease risk: age, sex, baseline eGFR (a measure of kidney function), hemoglobin, high blood pressure (hypertension), and diabetes.  

The model can also include a seventh factor—the albumin-to-creatinine ratio (ACR), a urine test related to kidney health—if that information is available in a person’s health record.  

To test how accurate the model was, the researchers then validated it using a second database containing health records from 7.7 million adults in Ontario.  

The researchers found that the model was very good at predicting whether someone would develop CKD within the next five years, and it continued to provide strong predictions for up to nine years.  

The researchers say this type of tool could eventually help health systems identify large numbers of people at risk of kidney disease earlier, allowing more people to access medications and care that may help slow disease progression.