Federal Grant Award Summary

The National Institute of Environmental Health Sciences (NIEHS) awarded The Feinstein Institutes for Medical Research a Project Grant of $139,836 under the Medical Library Assistance program (CFDA 93.879) to support the development of artificial intelligence-driven software tools for early prediction of clinical deterioration in hospitalized patients. Initiated July 1, 2026, and concluding June 30, 2029, this award funds a Research Software Engineer position to advance Aim 2 of the parent NIH Grant R01NR020774. The primary deliverable involves developing next-generation, personalized deterioration prediction models utilizing generative foundation models and multimodal learning techniques that integrate continuous vital signs from wearable devices with electronic health record (EHR) data.

The project will employ self-supervised pre-training methodologies on unlabeled physiological time series data, followed by fine-tuning on labeled deterioration events, with validation across both short-term (4-hour) and mid-term (24-hour) prediction windows. To address data scarcity challenges, the initiative will investigate synthetic data generation techniques, including Generative Adversarial Networks (GANs) and optimal transport-based methods. A secondary aim involves creating and publicly releasing robust software tools and resources to disseminate these clinical wearable foundation models, directly supporting the Research Software Engineer's career development while enabling broader healthcare applications aimed at improving patient outcomes and reducing costs associated with acute deterioration events in non-critical care settings.