چکیده (انگلیسی):

This research contributes to the domain of long-term care by exploring knowledge discovery techniques based on a large dataset and guided by representative information needs to better manage both quality of care and financial spendings, as a next step towards more mature healthcare business intelligence in long-term care. We structure this exploratory research according to the steps of the CRoss Industry Standard Process for Data Mining (CRISP-DM) process. Firstly, we interview 22 experts to determine the information needs in long-term care which we, secondly, translate into 25 data mining goals. Thirdly, we perform a single case study at a Dutch long-term care institution with around 850 clients in five locations. We analyze the institution’s database which contains information from April 2008 to April 2012 to identify patterns in incident information, patterns in risk assessment information, the relationship between risk assessments and incident information, patterns in the average duration of stay, and we identify and predict Care Intensity Package (ZZP) combinations. Fourth and finally, we position all data mining goals in a two-by-two matrix to visualize the relative importance of each goal in relation to both quality of care and financial state of care institutions.