Please join Wayne State University's Center for Leadership in Environmental Awareness and Research (CLEAR) and Center for Urban Responses to Environmental Stressors (CURES) for their upcoming hybrid seminar on November 5, 2026 from 12:30 to 1:30 p.m. The seminar is free; registration is required. The seminar will be held in the Integrative Bioscience Center at Wayne State University and via Zoom.

The guest speaker will be Dr. Melike Yildirim, assistant professor of industrial and systems engineering, family medicine and public health sciences, and in the Institute for AI and Data Science (AIDaS) at Wayne State University. Dr. Yildirim will present, "Who Is at Rising Risk? Explainable AI Linking Environmental Exposures, Genomics, and Social Context to Asthma Outcomes."

ABSTRACT: Asthma remains a leading cause of preventable emergency visits and hospitalizations, and its burden falls hardest on urban communities like Detroit, where residents face overlapping environmental stressors: traffic-related air pollution, industrial emissions, aging housing, and lead. Yet most risk assessment tools rely on static clinical snapshots and examine environmental, social, and genetic factors in isolation.

This talk presents an integrative, explainable AI framework that links individual-level electronic health record, survey, and genomic data with neighborhood-level social determinants and outdoor environmental exposures from AHRQ, PHOENIX and EPA. In preliminary analyses of about 7,600 Michigan adults with persistent asthma, our models predicted uncontrolled asthma at 3-, 6, and 12-month horizons with strong discrimination. Patient clustering revealed sharp geographic contrasts. Detroit-based subgroups showed poor asthma control alongside high poverty, housing vacancy, lead exposure, and traffic pollution, while suburban subgroups showed substantially better control and controller adherence.

I will also discuss where prediction falls short. Prior emergency utilization dominates model performance, so the harder and more consequential goal is identifying patients at rising risk before their first acute event. Finally, I will describe our partnership with the Asthma and Allergy Foundation of America–Michigan Chapter. Through that partnership, we are building a public community-level dashboard and testing whether this information changes where community organizations focus outreach and remediation resources.

BIOGRAPHY: Melike Yildirim, PhD, is an Assistant Professor of Industrial and Systems Engineering at Wayne State University, with a secondary appointment in Family Medicine and Public Health Sciences. Her research uses large-scale community-based data, predictive analytics, simulation, and optimization to promote equitable access to healthcare and to guide intervention and policy decisions. Before joining Wayne State, she was a CDC Prevention Effectiveness Fellow at the Agency for Healthcare Research and Quality (AHRQ) and a postdoctoral fellow at Massachusetts General Hospital/Harvard. She has developed decision-support tools for the CDC, AHRQ, the Georgia Department of Public Health, and UNICEF. These include BREATHE, which estimates the returns of investment for asthma home visits and self-management education. She received her PhD in Industrial and Systems Engineering from Georgia Tech.