A new algorithm accurately predicts whether a diabetes patient is at high risk for developing chronic kidney disease, a frequent long-term complication of diabetes, according to study results published in Nature Medicine.
IBM and Roche created the algorithm with data from half a million people with diabetes. The goal was to determine how an algorithm derived from real-world patient data would compare to previously released algorithms that relied on clinical trial data.
The algorithm from IBM and Roche outperformed previously released algorithms, according to the study results.
The study was a collaboration between IBM, Roche Diabetes Care, Eli Lilly and Co., the Regenstrief Institute in Indianapolis and the Indiana Bioscience Research Institute, also based in Indianapolis.
The Regenstrief Institute provided the research team with a data set of longitudinal, real-world medical data for the effort, according to the institute's president and CEO Peter Embi, MD.
"This is a great example of how by working together, partnerships like these can empower breakthrough discoveries that ultimately will benefit patients and healthcare systems throughout the world," he said in a news release.