Polygenic Scores Help Reduce Racial Disparities in Predictive Accuracy of Automated Type 1 Diabetes Classification Algorithms

In a study, led by CGM PI, Miriam Udler, and her colleagues they implemented an automated clinical algorithm and a type 1 diabetes polygenic score to identify individuals with type 1 diabetes in two large biobanks: MGB Biobank and BioMe (Mt. Sinai). They assessed the accuracy of the clinical algorithm compared to a gold standard of clinician diagnosis on chart review. The authors found that the clinical algorithm more accurately predicted type 1 diabetes status for self-reported White individuals, compared to other race/ethnicity groups. However, after updating the clinical algorithm to incorporate type 1 diabetes polygenic scores, the accuracy improved for all individuals, and the racial/ethnic disparity was reduced. These results demonstrate the potential for polygenic scores to aid in clinical phenotyping and to help reduce health disparities.

Read more in Diabetes Care and a podcast here.

February 17, 2023


CGM Primary Investigator

Miriam Udler