Author(s)

A Riha, V Svátek, P Nemec & J Zvárová

Abstract

We investigate the possibility of two-step approach to electronic healthcare record mining, in the context of analysing the compliance of healthcare practice with standards formulated in medical guidelines. Non-compliance patterns detected in the process of guideline–based data pre–processing provide additional attributes for subsequent association rule mining. The approach has been preliminarily tested on databases of hypertensive patients from different Czech hospitals. It should help reveal causes of frequent non–compliance; its sensitivity however depends on the quality of guideline formalisation, on the eligibility of patients for the given guideline, and on the coverage of datasets. 1 Introduction From the point of view of data mining, the medical domain hosts the nearly full variety of known forms of data: relational as well as texts and images. An ubiquitous type of data is the electronic healthcare record (EHR) maintained by physicians for their patients. Although the most advanced EHR systems accommodate even multimedia data, most are still written as free text. However, there is an increasing pressure on (and even interest of) the physicians to move towards the structured, database representation of the real EHRs. This gradually makes the application of the mainstream data mining techniques possible. The generic analysis of the EHRs themselves can reveal useful ‘nuggets’. We are, however, interested in a more focused analysis, which also takes into account

Keywords