CLEF eHealth 2017 Multilingual Information Extraction task Overview: ICD10 Coding of Death Certificates in English and French.
Résumé
This paper reports on Task 1 of the 2017 CLEF eHealth evaluation lab which extended the previous information extraction tasks of ShARe/CLEF eHealth evaluation labs. The task continued with coding
of death certificates, as introduced in CLEF eHealth 2016. This large-scale classification task consisted of extracting causes of death as coded in the International Classification of Diseases, tenth revision (ICD10). The languages offered for the task this year were English and French. Participant systems were evaluated against a blind reference standard
of 31,690 death certificates in the French dataset and 6,665 certificates in the English dataset using Precision, Recall and F-measure. In total, eleven teams participated: 10 teams submitted runs for the English dataset and 9 for the French dataset. Five teams submitted their systems
to the reproducibility track. For death certificate coding, the highest performance was 0.8674 F-measure for French and 0.8501 for English.