Singer J M, Edberg S C, Markowitz R L, Glickman J D, Miller L, Marchitelli R
Am J Clin Pathol. 1979 Oct;72(4):597-603. doi: 10.1093/ajcp/72.4.597.
Ten latex-fixation kits (both slide and tube) were evaluated by use of a reference serum containing 250 IU of rheumatoid factor (Communicable Disease Center). All tests were performed with serial dilution of serum; adjacent tubes differed by a factor of 0.10 log10. Geometric titers, means, standard deviations, and variance analysis were used to evaluate commercial kits. Three levels of variance were applied: (1) 0.18 or less for intralaboratory variation when the same serum is tested with the same kit; (2) 0.18-0.35 when various lots or batches of the same serum were used; (3) 1.0 or less when various commercial kits were tested with the same serum. The sensitivity of each commercial kit was determined by assigning to it a value (the minimum detectable unit) arrived at by dividing the number of units in the standard by the titer obtained. The technic presented could be applicable to other serologic tests, and thus provide a general method for their standardization and quality control.
使用含有250国际单位类风湿因子的参考血清（传染病中心）对10套乳胶固定试剂盒（玻片法和试管法）进行了评估。所有检测均采用血清系列稀释法；相邻试管的稀释倍数相差0.10 log10。采用几何滴度、均值、标准差和方差分析来评估商用试剂盒。应用了三个方差水平：（1）使用同一试剂盒检测同一血清时，实验室内变异为0.18或更低；（2）使用同一血清的不同批次时，变异为0.18 - 0.35；（3）使用同一血清检测不同商用试剂盒时，变异为1.0或更低。通过将标准品中的单位数除以所获得的滴度来确定每个商用试剂盒的灵敏度（最小可检测单位）。所介绍的技术可能适用于其他血清学检测，从而为其标准化和质量控制提供一种通用方法。
