Extraction and Exploration of Correlations in Patient Status Data

The paper discusses an Information Extraction approach, which is applied for the automatic processing of hospital Patient Records (PRs) in Bulgarian language. The main task reported here is retrieval of status descriptions related to anatomical organs. Due to the specific telegraphic PR style, the approach is focused on shallow analysis. Missing text descriptions and default values are another obstacle. To overcome it, we propose an algorithm for exploring the correlations between patient status data and the corresponding diagnosis. Rules for interdependencies of the patient status data are generated by clustering according to chosen metrics. In this way it becomes possible to fill in status templates for each patient when explicit descriptions are unavailable in the text. The article summarises evaluation results which concern the performance of the current IE prototype.