A machine-learning approach to identify women with ANA-related connective tissue diseases

Women are disproportionately affected by autoimmune diseases, 80% of individuals affected by autoimmune diseases are women and this includes conditions such as systemic lupus erythematosus, Sjogren’s syndrome, scleroderma, and immune-mediated myositis (Anti-nuclear antibody related connective tissue diseases/ARCTD). Anti-nuclear antibody (ANA) is a biomarker traditionally used to screen for autoimmune diseases and is a helpful screening biomarker among individuals who displays signs and symptoms of such conditions. Unfortunately, the value of ANA as a diagnostic aid is low and it is prone to false positivity. This results in delayed care among women with actual ARCTD.

We will develop a machine learning using multimodal data from electronic health records, including medication utilization patterns, within the Yale New Haven Health Hospital System to accurately identify ANA-positive individuals who would develop ARCTD.