European Clinical Case Corpus

AbstractInterpreting information in medical documents has become one of the most relevant application areas for language technologies. However, despite the fact that huge amounts of medical documents (e. g., medical examination reports, hospital discharge letters, digital medical records) are produced, their availability for research purposes is still limited, due to strict data protection regulations. Aiming at fostering advanced information extraction technologies for medical applications, we present E3C, a corpus of clinical case narratives fully based on freely licensed documents. E3C (European Clinical Case Corpus) contains a vast selection of clinical cases (i. e., narratives presenting a patient’s history) that cover different medical areas, are based on different styles and produced in different languages. A portion of the corpus has been manually annotated to be used for training and testing purposes, while a larger set of documents has been automatically tagged to serve as a baseline for future research in information extraction.