Statistical Process Control in Biochemical and Hematological Quality Control Data
Nien F. Zhang, P Winkel
Daily quality control (QC) measurements of common biochemical and hematological quantities were recorded during several months while methods and analyzers showed no signs of malfunctioning.Usually it is assumed that QC data may be described as i.i.d. In this case an X chart and/or an EWMA chart are the proper control charts to use. When autocorrelation is presented, the traditional control charts may be inefficient. An alternative control chart, the EWMAST chart proposed in Zhang (1998) has been developed for stationary process data. The EWMA and the EWMAST chart were applied to each of the 11 QC data series. In six of the 11 series, significant process autocorrelations were demonstrated. The results show that the conventional EWMA chart may give false alarms in the presence of autocorrelation while the EWMAST chart gave few false alarms.
Journal of the American Statistical Association
autocorrelation, control chart, stationary process, time series
and Winkel, P.
Statistical Process Control in Biochemical and Hematological Quality Control Data, Journal of the American Statistical Association
(Accessed February 26, 2024)