Applied AI Summit
Free online conference | October 14-16, 2025Smoking Out the Missing Data: Extracting Patient Insights from Provider Notes
Much of the information needed for clinical research lives in unstructured provider notes, not structured EHR fields. At TriNetX, we use John Snow Labs’ NLP to extract insights such as smoking status—rarely captured in structured data and often missing for non-smokers.
This talk will outline our six-step annotation lifecycle, show how we scale pipelines from one-to-many sites, and share results from fine-tuning models to overcome site-level variation. We’ll also discuss how we create structured, harmonized labels that bring consistency across networks, enabling richer research and better outcomes.
About the speaker
Mike Temple
Lead Healthcare Data Scientist at TriNetX
Mike is a physician-data scientist who has expertise in predictive modeling using advanced AI methodologies. He has clinical experience as a pediatrician in private practice and at Nationwide Children’s Hospital where he served as a hospitalist in the Neonatal ICU. He has been on faculty at Vanderbilt University in the Department of Biomedical Informatics and was previously a clinical data scientist at Cardinal Health and the Hospital Corporation of America. Dr. Temple brings both strong technical and clinical experience.
Zuzanna Drebert
Data Scientist at TriNetX
Zuzanna Drebert is a data specialist with over seven years of experience working with real-world
medical data. With a PhD in Experimental Cancer Research from Ghent University in Belgium, she
combines her biomedical expertise with technical skills to help solve problems and drive
advancements in healthcare and research. As a Data Scientist at TriNetX, Zuzanna supports both
internal and external research projects by leveraging her experience in data analysis, machine
learning, data visualization, and domain knowledge. She also contributes to experiments
exploring various methods for extracting data from free-text clinical notes. Zuzanna is passionate
about learning, experimenting with new ideas and techniques, and continuously improving her
work to achieve meaningful outcomes.