Use of alternative rules (other than the 1(2)s) for evaluating interlaboratory performance data.

Abstract

Previous studies have documented the ineffectiveness of using either the group mean +/- 2 group standard deviations (SD) or the 1(2)s rule as the standard of acceptable performance in evaluating interlaboratory proficiency testing (PT) data. Using computer simulation of PT data, we evaluated the efficiency of 244 alternatives to the 1(2)s rule, all based on the PT population's mean and SD. Using the traditional interlaboratory PT format, we determined the ability of each rule to correctly identify both good and deficient intralaboratory performance. The rules are based on results from one to five PT samples "analyzed" at the same time. Because the effectiveness of the criteria set for acceptable performance in a PT program is influenced by the population SD, each rule's capabilities were examined for PT populations with interlaboratory SDs ranging from 1% through 10% of the population mean value. All rules achieve their maximum efficiency over a narrow range of interlaboratory SDs. For PT evaluations of intralaboratory performance to be optimally effective, selection of the rule must be based on the SD of the PT population.