Software support for ICD-10 classification of hospital patient data

Défense de mémoire de Sarah Kumwimba

Since a couple of years, the ICD-10 BE (the International Classification of Diseases Belgian revision) is used by Belgian hospitals to code the do􀀀to􀀁s􀀂 practises on patients in order to claim financial resources from the Belgian government. However, the INAMI codes (National Institute of Health and Disability Insurance) are widely used by doctors unlike the ICD-10 BE codes. Therefore, it is up to specialised hospital teams to translate these INAMI codes into ICD-10 BE codes that will be reported to the SPF finances (Federal public service of finances). However, this task is considered laborious for these teams as they have an average knowledge in medical vocabularies and practises.

In this thesis, the issue of translating codes from a classification to another one is discussed and an introduction to a tool aimed at helping to improve this process is proposed. The tests with a first prototype show that it is a challenging problem.

Keywords: Medical classification, terminology translation, medical vocabulary, ICD10, Machine learning, UMLS, INAMI

Contact : Isabelle Daelman - 4966 - isabelle.daelman@unamur.be
Télecharger : vCal