NORMAN, OKLA. –Talayeh Razzaghi, an assistant professor in the School of Industrial and Systems Engineering, University of Oklahoma Gallogly College of Engineering, is leading a project using machine learning and artificial intelligence techniques to predict when pregnant women may have increased risk of preeclampsia.
“Preeclampsia is a subtype of hypertension (high blood pressure) developed during pregnancy that can lead to serious, even fatal, complications for both the mother and the fetus,” Razzaghi said. “Our central hypothesis is that machine learning-based models can fundamentally transform clinicians’ existing decision support tools for detection and monitoring preeclampsia for minority groups by addressing the key issues specific to preeclampsia datasets. This approach assists clinicians in the prognosis of adverse delivery outcomes. In particular, the research methodology in this study addresses biases and outcome health delivery disparities among Hispanic and Native populations in Oklahoma and Texas.”
“I’d like to see that what we are doing is influential on human lives, particularly on women and minority groups,” Razzaghi added. “Racial and ethnical health disparities can be attributed to various aspects of inequities including patients’ socioeconomic status and lack of universal healthcare policies.”
The Office of the Vice President for Research and Partnerships on the OU Norman campus recently funded 11 short-term projects that position OU faculty and their collaborators to effectively compete for significant external funding opportunities related to the impact of social inequities on knowledge creation and dissemination.
“This seed grant program illustrates how OU is committed to solving global challenges through research that provides a real impact to society,” said OU associate vice president for research and partnerships Ann West. “Dr. Razzaghi’s project applies data science and engineering ideas to improve public health and ultimately will advance our understanding of underlying social and policy dynamics.”
As an engineer, Razzaghi hopes to bring a unique perspective and data science tools and techniques to public health problems.
“We are going to study Texas research data files representing a massive-size patient discharge database,” Razzaghi said. “We plan to apply modern machine learning algorithms to this dataset in order to detect patients who are at risk of preeclampsia early in the clinical setting. Those findings will then ultimately inform our development of a physician support system to help clinicians identify these heath inequalities and disparities to better ensure appropriate treatments for these groups.”
Razzaghi notes that a major challenge lies in the complexity of health care data.
“We are going to be challenged to create an accurate, interpretable model to identify these patients at early stages,” Razzaghi said. “We are working to come up with an interpretable and robust technique using a deep learning algorithm to identify these patients as early as possible.”
Razzaghi is working with Zuber Mulla on the study. Mulla is a professor in the Department of Obstetrics and Gynecology, Paul L. Foster School of Medicine, at Texas Tech University Health Sciences Center El Paso.
“Our study is bridging the gap between the fields of medicine and engineering to advance the health of mothers,” Mulla said. “Collaboration among engineers trained in artificial intelligence methods and epidemiologists will no doubt be a boon to population health.”