Facilitating Rapid & Cost-Effective Diagnosis Using a Data-Driven Approach
Facilitating Rapid & Cost-Effective Diagnosis Using a Data-Driven Approach
Liat Shabtay, M.Sc. student at School of Industrial & Intelligent Systems Engineering
Advisor: Dr. Reut Noham
Abstract:
The increasing use of diagnostic tests in hospitals has led to concerns about over-testing, resulting in higher costs, increased lab workload, and potential harm to patients. This research aims to enhance the diagnostic process by developing data-driven tools that support more efficient and accurate use of immunological tests. Using real-world clinical data from Ichilov Hospital, we apply Machine Learning (ML) as an intermediate step to enable the application of Explainable AI (XAI). We focus specifically on SHAP, a method that not only ensures transparency in model predictions but also helps identify the most relevant tests for distinguishing between immune and non-immune patients. As a preliminary study, this work demonstrates the potential of explainability to reduce unnecessary testing and lays the groundwork for a future Clinical Decision Support System (CDSS) to improve diagnostic efficiency and clinical trust.
Bio:
Liat Shabtay is an M.Sc. student at the School of Industrial Engineering and Intelligent Systems Engineering. She conducts her research in the HOPE Lab (Humanitarian and Healthcare Operations), under the supervision of Dr. Reut Noham. Her work, in collaboration with Ichilov Hospital and supported by a grant from the Alrov Center for Digital Medicine, focuses on enhancing diagnostic testing through data-driven insights. Liat holds a B.Sc. in Industrial Engineering and Management from Tel Aviv University.
