Suffering in Silence. Population-Level Detection and Risk-Prediction of Women’s Reproductive Disease from Claims Data

Abstract

We present a new approach to population health where data-driven predictive models are learned for various stigmatized, underserved, and neglected chronic conditions that affect women. Our approach enables early detection within large populations at low cost from readily available medical claims datasets. The model uncovers early and late stage risk factors up to two years in advance of onset that can be used as guides for early interventions for endometriosis, polycystic ovarian syndrome, and infertility, allowing women suffering from these conditions to get on treatment plans earlier.