Clinician Decision Support Dashboard: Extracting value from Electronic Medical Records using Text Mining Iccha Sethi and Harold R. Garner (email@example.com and firstname.lastname@example.org) Virginia Bioinformatics Institute, Virginia Tech
Objective• To leverage EMRs to enhance patient care.• To realize a system that will analyze a patient’s evolving EMR: • in context with all available biomedical knowledge. • using the accumulated experience recorded in the EMRs of other patients. • with the help of interactive, automated, actionable text mining tools. • using multidisciplinary approach. Computer Science • Text Mining Marketing • HCI Medicine • Biomedical research Clinician Decision Support Dashboard
Flow Chart of the CDS Dashboard The CDS Dashboard, in a secure network, will help physicians find de-identified electronic medical records similar to their patients medical record, relevant medical literature, recent research findings, and clinical trials thereby aiding them in diagnosis, treatment, prognosis and outcomes. Patient EMR Clinician Input INPUT Query Assembly and Conditioning SIMILARITY eTBlast ENGINE De-OUTPUT Related Case Clinical Customized identified Literature Reports Trials Features EMRsFigure 1: Flow chart of CDS (The Patient EMR is processed by the CDS and then comparedto various databases by eTBlast)
ChallengesSeveral challenges exist in creating a CDS Dashboard that can be universally usedincluding: Variation in entries in the EMRs; each doctor uses their own set of acronyms and shortcuts. The presence of billing and CPT codes. Combinations of unstructured and structured laboratory data. Presence of lexical variants and synonyms.
GoalsOur goal in building the CDS Dashboard was the ensure : platform portability independence can layer on top of existing electronic modularity medical record management softwareDiscussion• The tool can be used in multiple ways: • Teaching aid for medical students • Exploratory tool for doctors and nurses