Colloquia Archive
Advanced Methods for Adverse Drug Event Detection in Clinical Notes
March 20, 2007
Host: Matthew Palakal
AbstractAdverse drug event (ADE) detection is vital to the patient safety efforts of healthcare organizations. To prevent ADE, computerized surveillance techniques use structured clinical data to generate alerts and reminders in clinical settings. Unfortunately, these techniques fail to detect textual signals in the electronic medical record. This presentation describes the development of methods to detect ADEs in free-text clinical notes based on approaches from cognitive science, artificial intelligence, and linguistics. These approaches may be combined in an ADE detection expert system. The accuracy of different text-processing methods at detecting ADEs is compared.
BiographyShobha Phansalkar received her Bachelors in Pharmacy from the University of Pune in India and a Masters in Medical Informatics from the University of Utah. Her research interests lie in the area of design, implementation and evaluation of clinical information systems. Key research topics include understanding clinician decision making, knowledge acquisition and engineering, patient safety and development of intelligent clinical information systems. Shobha’s doctoral dissertation focuses on developing methods for the detection of adverse drug events in clinical notes. She has published and given several presentations on the implementation of computerized protocols, use of natural language processing for clinical event detection and evaluation of clinical information systems. She co-authored a chapter on the impact of integrative functions in Electronic Health Records for the Agency for Healthcare Research and Quality. She was recently awarded the John D. Morgan Award for excellence in research, academic work and service.
