Groth T, Falk H, Westgard J O
Scand J Clin Lab Invest Suppl. 1984;172:195-201.
A computer simulation program has been developed to aid clinical chemists in the design and theoretical evaluation of statistical control procedures. This "QC simulator" permits the user to study the effects of different parameters characterizing the measurement procedure (method standard deviation, components of variation, rounding of results) and parameters of the control procedure (decision criteria, control limits, number of observations). The performance of control procedures is characterized by the probability for rejection, estimated at several different levels of random and systematic error. Predictive values for reject and accept signals are also calculated at different error incident rates, given a specified error model. The relative performance of different control procedures can be compared based on these performance characteristics. Another important application of the program is the design of control procedures to assure that a specified level of analytical quality is achieved in routine analyses. Various optimization criteria may be applied, e.g., in terms of test yield and cost.
已开发出一种计算机模拟程序，以协助临床化学家进行统计控制程序的设计和理论评估。这个“质量控制模拟器”允许用户研究表征测量程序的不同参数（方法标准差、变异分量、结果的舍入）以及控制程序的参数（决策标准、控制限、观测次数）的影响。控制程序的性能由在几个不同水平的随机和系统误差下估计的拒收概率来表征。给定指定的误差模型，还会在不同的误差发生率下计算拒收和接受信号的预测值。可以基于这些性能特征比较不同控制程序的相对性能。该程序的另一个重要应用是设计控制程序，以确保在常规分析中达到指定的分析质量水平。可以应用各种优化标准，例如，从测试产量和成本方面考虑。
