tailieunhanh - Báo cáo khoa học: "Automatic Analysis of Patient History Episodes in Bulgarian Hospital Discharge Letters"

This demo presents Information Extraction from discharge letters in Bulgarian language. The Patient history section is automatically split into episodes (clauses between two temporal markers); then drugs, diagnoses and conditions are recognised within the episodes with accuracy higher than 90%. The temporal markers, which refer to absolute or relative moments of time, are identified with precision 87% and recall 68%. The direction of time for the episode starting point: backwards or forward (with respect to certain moment orienting the episode) is recognised with precision . . | Automatic Analysis of Patient History Episodes in Bulgarian Hospital Discharge Letters Svetla Boytcheva State University of Library Studies and Information Technologies and IICT-BAS Galia Angelova Ivelina Nikolova Institute of Information and Communication Technologies IICT Bulgarian Academy of Sciences BAS galia iva @ Abstract This demo presents Information Extraction from discharge letters in Bulgarian language. The Patient history section is automatically split into episodes clauses between two temporal markers then drugs diagnoses and conditions are recognised within the episodes with accuracy higher than 90 . The temporal markers which refer to absolute or relative moments of time are identified with precision 87 and recall 68 . The direction of time for the episode starting point backwards or forward with respect to certain moment orienting the episode is recognised with precision . 1 Introduction Temporal information processing is a challenge in medical informatics Zhou and Hripcsak 2007 and Hripcsak et al. 2005 . There is no agreement about the features of the temporal models which might be extracted automatically from free texts. Some sophisticated approaches aim at the adaptation of TimeML-based tags to clinically-important entities Savova et al. 2009 while others identify dates and prepositional phrases containing temporal expressions Angelova and Boytcheva 2011 . Most NLP prototypes for automatic temporal analysis of clinical narratives deal with discharge letters. This demo presents a prototype for automatic splitting of the Patient history into episodes and extraction of important patient-related events that occur within these episodes. We process Electronic Health Records EHRs of diabetic patients. In Bulgaria due to centralised regulations on medical documentation which date back to the 60 s of the last century hospital discharge letters have a predefined structure Agreement 2005 . Using the section headers