Text Mining to Support Knowledge Discovery from Electronic Health Records

The use of electronic health records (EHRs) has grown rapidly in the last decade. The EHRs are no longer being used only for storing information for clinical purposes but the secondary use of the data in the healthcare research has increased rapidly as well. The data in EHRs are recorded in a structured manner as much as possible, however, many EHRs often also contain large amount of unstructured free‐text. The structured and unstructured clinical data presents several challenges to the researchers since the data are not primarily collected for research purposes. The issues related to structured data can be missing data, noise, and inconsistency. The unstructured free-text is even more challenging to use since they often have no fixed format and may vary from clinician to clinician and from database to database. Text and data mining techniques are increasingly being used to effectively and efficiently process large EHRs for research purposes.
Most of the methods developed for this purpose deal with English‐language EHRs and cannot simply be applied to non‐English EHRs. This thesis concerns the use of data mining and natural language processing techniques to process unstructured Dutch‐language EHRs. We present all methods and approaches in this thesis in a wider and formal framework of knowledge discovery. We described several data-mining and data-preparation techniques for automated processing of Dutch electronic health records.