QC chart mining: extracting systematic error patterns from quality control charts

This paper presents a novel method; "QC Chart Mining", which extracts systematic error patterns from quality control charts in order to manage clinical test data at a medical laboratory. In this paper we describe the basic principle of a time series decomposition mechanism for QC Chart Mining. QC Chart Mining is used to recognize quality problems such as long-term trends and/or daily cyclic variations in analytical processes of clinical tests, then to improve the quality level over clinical laboratory medicine. Intensive experiments from both actual quality-control data and artificial data have revealed the validity of the proposed method. Our results have shown that the proposed method is useful and effective for quality management in a medical laboratory.