Introduction of z-score arrow ranges (ZSAR) in proficiency testing

Abstract

Most proficiency testing (PT) providers do not calculate the performance of participants that report results that are below the limit of quantification, so-called truncated results ("< results"). The ISO FDIS 13528 standard describes three approaches for the analysis of censored data. DUCARES leaves out truncated results from the statistical evaluation. Although these results are removed from the statistical dataset, it is possible, however, to give a performance indication with z-score ranges. For this, DUCARES has developed the graphical z-score arrow range (ZSAR) method which considers the full interval of potentially possible results as described by the truncated value <x reported by a laboratory, i.e., from the limit value x down to zero. The ZSAR method translates this into a z-score interval and displays it graphically with a double-headed arrow in the distribution plot. This presentation gives the participating laboratory a good indication whether its reported truncated value potentially includes satisfactory scores |z| ≤ 2 or not, and to decide on corrective actions. In this paper, we present the ZSAR method and discuss its advantages in PT.