On supporting medical quality with intelligent data mining

The healthcare sector is currently facing both the economic necessity and the technical opportunity of a data based approach to quality management. Against this background, we introduce a process model for a data based medical quality management and apply intelligent data mining methods to patient data. Intelligent data mining incorporates advantages of both knowledge acquisition from data and from experts. We present the Knowledge Discovery Question Language (KDQL), a controlled language for business questions which abstracts from database and data mining terminology to allow high-level interaction. We use a knowledge-based measurement of relevant subjective interestingness facets like novelty, usefulness, and understandability which enables flexible ways to access the results of data mining. Questions asked in this project were targeted on diagnostic and therapeutic measures as well as the quality of documentation. For these issues in the field of medical quality management interesting results were found.