Researchers at DePaul University and Sinai Chicago have received a $1.5 million grant from the National Science Foundation to expand their SCHOLAR model, which integrates artificial intelligence with community health workers to reduce hospital re-admissions.

The SCHOLAR model pairs AI with human insights to identify high-risk patients and match them with appropriate support systems. This approach not only aims to reduce emergency department re-admissions but also has potential applications in other areas.

“The responsible and ethical use of AI moving forward will depend on the integration of experts outside the technology domain,” said Jacob Furst, a professor of visual computing at the Jarvis College of Computing and Digital Media. “The SCHOLAR model recognizes the important role of users in shaping AI for good.”

Community health workers are crucial in this initiative, administering surveys and using their experience to determine patient needs based on factors like food insecurity, housing, and transportation access.

“We’re drawing on these professionals and their experiences, asking community health workers what they feel are the most influential determinants in their encounters with patients. This allows us to prioritize and define those for the system,” said John Mazzeo, associate professor of health sciences in the College of Science and Health.

Improved patient care leads to fewer emergency re-admissions, lowering costs for health systems like Sinai and enhancing access to government grants. DePaul researchers are leveraging data collection to provide better insights for community health care workers.

“This important and innovative work highlights the crucial role that community health workers play in connecting with patients on a personal level," said Kelly McCabe, Director of Community Health Interventions & Research at Sinai Urban Health Institute.

DePaul computer scientist Roselyne Tchoua conducted an initial study involving over 1,300 patients at Chicago’s Sinai hospital system in 2023. The study showed that combining AI models with community health worker data predicted readmission rates with 5% more accuracy.

The NSF grant will enable DePaul researchers to expand their research by adding more patient data. “We’ll be able to expand our current work by adding more patient data,” said Daniela Stan-Raicu, associate provost for research at DePaul. “Our prototype model will become more accurate and can be applied to other communities with this expansion.”

A user-friendly dashboard is being developed to streamline analysis and note-taking processes for community health workers, reducing time spent on assessments while ensuring that tools cater specifically to their needs.

This interdisciplinary project involves both undergraduate and graduate students participating in system modeling, data input, and interviews with health workers.

“The models that we are building can be applied beyond the healthcare domain to other fields like education,” said Stan-Raicu. “Rather than identifying patients, social workers and other student support staff could modify the tool to identify students who may not graduate.”