Federal Project Grant Award Summary

The University of Pennsylvania received a $1.45M project grant from the National Institute of Environmental Health Sciences under the Medical Library Assistance program (CFDA 93.879) effective June 1, 2026 through May 31, 2030. The award funds the development of statistical methodologies and predictive modeling frameworks designed to integrate irregularly collected longitudinal multi-modal health data—including electronic health records (EHR), patient-reported outcomes (PRO), and mobile health (mHealth) data—into clinically actionable prediction models. The research specifically addresses three critical gaps in current biomedical informatics approaches: incorporating the predictive signal of informative missingness patterns, preserving fine-grained intra-day detail from mHealth sensors and wearables rather than reducing data to summary statistics, and developing population-flexible models that generalize effectively to underrepresented groups and individual patients with limited data availability.

This research initiative directly supports the Medical Library Assistance program's mission to advance biomedical informatics and data science methodologies for improved human health discovery and clinical outcomes. The methodological framework will enhance early intervention capabilities for chronic disease management, including diabetes, heart failure, cancer, and obesity, by enabling more accurate prediction of symptom escalation, acute care utilization, and mortality risk. The project employs distributional learning with low-dimensional embeddings and develops both univariate and multivariate longitudinal models that account for delays between predictors and outcomes, delivering generalizable computational tools and validated statistical approaches for the biomedical research community.