Blum A S
Clin Chem. 1985 Feb;31(2):206-12.
I describe a program for definitive comparison of different quality-control statistical procedures. A microcomputer simulates quality-control results generated by repetitive analytical runs. It applies various statistical rules to each result, tabulating rule breaks to evaluate rules as routinely applied by the analyst. The process repeats with increasing amounts of random and systematic error. Rate of false rejection and true error detection for currently popular statistical procedures were comparatively evaluated together with a new multirule procedure described here. The nature of the analyst's response to out-of-control signals was also evaluated. A single-rule protocol that is as effective as the multirule protocol of Westgard et al. (Clin Chem 27:493, 1981) is reported.
我描述了一个用于不同质量控制统计程序的确定性比较的程序。一台微型计算机模拟由重复分析运行产生的质量控制结果。它对每个结果应用各种统计规则，将违反规则的情况制成表格，以评估分析师常规应用的规则。该过程随着随机误差和系统误差的增加而重复。对当前流行的统计程序的错误拒绝率和真误差检测率以及这里描述的一种新的多规则程序进行了比较评估。还评估了分析师对失控信号的反应性质。报告了一种与Westgard等人（《临床化学》27:493，1981）的多规则方案同样有效的单规则方案。
