Medial EarlySign – an Israeli health AI company has just announced the commercial availability of its first suite of machine learning-based predictive diabetes risk solutions. Expanding the company’s portfolio of clinical risk predictors, these new diabetes-focused AlgoMarkers are designed to help healthcare systems identify and engage patients at high risk for diabetes and downstream complications.
EarlySign’s Pre2D algorithm solution identifies prediabetic patients at highest risk of progressing to diabetes within twelve-months, while the Diabetes to CKD solution identifies type 2 diabetic patients at high risk of developing stage 2-4 chronic kidney disease within three years.
Machine Learning-Based Predictive Diabetes Risk Solutions
EarlySign’s Pre2D™ predictive solution applies advanced machine learning-based algorithms to identify “hidden signals” residing in existing, routine blood tests. Factoring in age, gender and BMI – and requiring no special patient preparation – it flags those prediabetic patients at high risk for progressing to diabetes in one (1) year or less. In a retrospective data study of 1.1 million prediabetic patients, the Pre2D AlgoMarker flagged the top 10% of the prediabetic population at risk and successfully identified 58.3% of patients who became diabetic within a 12-month period. This is a 14.7% increase over a logistic regression model that, by flagging 10% of the population, identified only 43.6% of future diabetics.
The Diabetes to CKD™ risk predictor uses basic demographic data, routine lab results, diagnostic codes, and medication information to flag type 2 diabetic patients most likely to develop stages 2-4 of chronic kidney disease in 3 years or less. In a retrospective data study of hundreds of thousands of diabetic patients, the algorithm was able to capture 25.5% of those most likely to progress to CKD within three years, by flagging only 3% of the diabetic population. This amounts to 77% more patients than would have been identified if the last eGFR value was used.
“In the U.S. alone, approximately 1.5 million prediabetic adults will become diabetic this year, while between 20% and 40% of diabetic patients worldwide suffer from diabetes-related kidney complications,” said Ori Geva, CEO of Medial Early Sign. “Our Pre2D™ and Diabetes to CKD™ solutions provide healthcare systems opportunities to identify and reach out to high-risk patients within an actionable timeframe, when preventative measures can be initiated, and resources allocated to potentially delay or prevent the onset of disease.”