Dr. Zhe He, an Associate Professor in the iSchool, recently had his Knowledge Discovery and Data Mining Work Group (KDDM WG) pre-symposium workshop accepted by the American Medical Informatics Association (AMIA) for the 2023 Informatics Summit, as well as a student paper, and a podium presentation by several students from the iSchool and other departments.

The AMIA Informatics Summit will be held in Seattle, WA, from March 13-16, 2023, and will have content dedicated to translational bioinformatics, precision medicine, clinical research informatics, data science, and artificial intelligence.

The pre-symposium workshop, “Which Comes First: High Quality Clinical Data or Reliable AI-based Applications?”, focuses on four guidance documents that were published by the Food and Drug Administration (FDA) in 2021 regarding how to use Electronic Health Records such as medical claims and patient registry data to support regulatory decision-making for drug and biological products, with a center of attention on data collection and data quality. In the workshop, world-leading experts from machine learning, biomedical informatics, healthcare delivery system, and life science industry will discuss real-world data (RWD) quality and its impact on building reliable, AI-based applications that can be adopted for patient care and regulatory decision making.

Dr. Zhe He says, “This event will cover in-depth topics such as missing data, data traceability and documentation, data bias assessment and mitigation, AI assisted data collection and abstraction, and deployment of AI systems in real world settings. These topics synergize tightly with the broader informatics interests of the AMIA attendees and help to raise their awareness of the crosstalk and opportunities in AI and healthcare.”

iSchool doctoral student Forhan Bin Emdad will be presenting his student paper, “Towards Interpretable Multimodal Predictive Models for Early Mortality Prediction of Hemorrhagic Stroke Patients”, that presents an ensemble of deep learning framework to predict early mortality among ICU patients with hemorrhagic stroke using patients’ clinical data extracted from a well-known clinical database called MIMIC III. This paper follows the MINIMAR (Minimum Information for Medical AI Reporting) standard, presenting an important step towards building trust among the AI system and clinicians. Other authors on this paper include: Shubo Tian (Statistics), Eshan Nandy (Computer Science), and Dr. Karim Hanna (USF Health).

The podium presentation, “Towards Semi-Automated Construction of Laboratory Test Result Comprehension Knowledgebase for a Patient-Facing Application”, is about building a robust test result comprehension knowledge base for a tool Dr. He and his team plan to develop, called LabGenie, which aims to help patients with their understanding of lab results data in patient portals.

In the project for the podium presentation, four undergraduate students in the FSU Undergraduate Research Opportunity Program (UROP) helped annotate the data: Jessica Valyou, Shawntah Thomas, Maddy Dupuis, and David Garner. As well as two PhD students, Arslan Erdengasileng and Shubo Tian, who helped with the curation of the annotated data and evaluated the named entity recognition models.

“Having these important studies accepted by AMIA Informatics Summit is a strong indication of the significance of these work and motivated us to keep working hard to push these projects forward,” says Dr. He. “It’s great to have students leading or involved in these studies, as they can get first-hand training in health informatics research.”