Serguei V Pakhomov, Ph.D. Research Associate Division of Biomedical Informatics Mayo Clinic http://www.tc.umn.edu/~pakh0002/homepage/ (personal page) http://mayoresearch.mayo.edu/mayo/research/bmi/index.cfm (division page) Monday, March 7, 2005 Kirby Plaza 309 12:00 - 1:20pm Title ====== Natural Language Processing and Electronic Health Record at the Mayo Clinic Abstract ======== Electronic Health Record (EHR) (a.k.a. Electronic Medical Record (EMR) and Electronic Patient Record (EPR)) are getting rapidly introduced into medical practice and health science research across the country. EHR typically contains all information available for a given patient including but not limited to his/her lab test results, radiology images and various records of encounters with physicians. Much of the EHR exists in the form of unrestricted text and needs to be structured in order to be accessible by health care professionals. Due to the massive amount of such unstructured textual data, there is a demand for Natural Language Processing methods and systems that are flexible and scalable. In this presentation, I will give a brief overview of the EHR at the Mayo Clinic, the challenges it poses for NLP and the efforts of our Division of Biomedical Informatics to address some of these challenges. As an example of a challenging issue, I will give a more in-depth overview of a recent collaborative project with the UMD Computer Science department focused on developing measures of semantic relatedness between terms used in the biomedical domain. Biography ========= Dr. Pakhomov has a Doctorate degree in Linguistics from the University of Minnesota, Twin Cities and is currently a Research Associate and an Instructor in Medical Informatics at the Mayo Clinic in Rochester, MN. He has been involved in research and development of text analysis and NLP technology in the Division of Biomedical Informatics at the Mayo Clinic since 2001. Currently he is the scientific lead on the Text Analysis project of the Mayo Clinic Life Science System. Prior to Mayo Clinic he was a Research Scientist at Lerhout and Hauspie, Inc. where he did research in automatic speech recognition for medical dictations. Dr. Pakhomov's interests in NLP include automatic term extraction, automatic classification of medical records, abbreviation detection and disambiguation, semantic relatedness and frame-based concept representation. Dr. Pakhomov is an alumnus of UMD, having received his M.A. in English Studies in 1994.