A novel method for identifying continuity of care in hospital discharge summaries

Abstract

Discharge summary is one of the most important clinical narratives because it provides essential information about the patient, including the continuity of care. Due to its free-form writing and the lack of consensus on its essential content, the identification of data contained on the text becomes a difficult task. In this project, we propose a rule-based method to verify the presence of information related to the continuity of care in Brazilian Portuguese texts, applying Natural Language Processing (NLP) techniques with an opensource tool named CoGrOO, and based on an annotated medical corpus. After the experiments, four rules were defined and applied on 200 summaries to identify the existence of the content "continuity of care". This process had resulted in a Precision value of 83%, Recall value of 69% and F-Measure value of 76% related to algorithm evaluation.