Federal Grant Award Summary

The University of Pennsylvania received a $723,841 Project Grant award from the National Institute on Drug Abuse, administered under the Mental Health Research Grants program (CFDA 93.242), with an award date of May 18, 2026, and completion date of February 28, 2031. The project, "Advancing Multimodal AI/ML to Enhance HIV Clinical Care," will develop artificial intelligence and machine learning (AI/ML) models that integrate multiple data modalities encountered in HIV clinical practice, including structured and unstructured electronic health records (EHRs), claims data, pharmacy records, imaging data, and omics/molecular biomarker data. The research aims to deliver novel, accurate, and unbiased multimodal AI models for HIV care and prevention while addressing significant technical challenges related to data heterogeneity, missing data, and bias inherent in healthcare datasets.

The project's deliverables are organized across three specific aims: developing advanced multimodal AI models for HIV applications; adapting and creating causal knowledge graphs to enhance model interpretability for HIV care and prevention; and synergistically integrating knowledge graphs with multimodal AI models to improve precision and clinical usability. These products will address critical barriers in AI implementation for HIV applications, particularly the scarcity of publicly available multimodal HIV data and the complexity of training robust models on heterogeneous, incomplete clinical datasets. The research will advance clinical decision-support capabilities and contribute to improved HIV prevention and treatment outcomes through enhanced personalized medicine approaches.