I-Corps: Translation potential of machine learning algorithms for early postpartum depression detection

Award ID: 2628652

This I-Corps project is based on the development of machine learning tools for the early detection of postpartum depression (PPD) in new mothers. Postpartum depression affects 10 to 20 percent of women who give birth in the United States each year, yet up to half of all cases go undiagnosed, creating substantial human and economic cost. Untreated perinatal mood disorders cost the U.S. healthcare system an estimated $14 billion annually, with costs falling disproportionately on Medicaid programs serving the highest risk populations. This technology uses a shortened clinical screening tool requiring minimal provider time, and a passive monitoring approach using consumer wearable devices that continuously assess depression risk without any effort from the patient. This may expand PPD detection to millions of women who would otherwise go unscreened, reduce burden on healthcare providers, generate cost savings for payers, and improve outcomes for mothers and children across health systems, community health centers, and Medicaid managed care programs. This I-Corps project utilizes experiential learning coupled with first-hand investigation of the industry ecosystem to assess the translation potential of two machine learning algorithms for early postpartum depression (PPD) detection. Currently, there is a surveillance gap with new mothers where after the postpartum clinical visit, traditional health data stops precisely when depression risk peaks. This technology is designed to analyze passive signals from consumer wearable devices, which generates continuous risk stratification without active patient participation. In early validation studies, this demonstrated strong predictive accuracy in identifying patients with PPD. In addition, a key scientific advance is the use of within-person deviation modeling, which tracks changes from an individual's personal baseline rather than population averages, improving performance across all populations tested. This technology may provide a new tool for PPD detection that is easier to use by clinicians and may improve patient outcomes. This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.

NSF Program Director: Ruth Shuman