Hamilton Martin A, Hamilton Gordon Cord, Goeres Darla M, Parker Albert E
Big Sky Statistical Analysts LLC, 309 South Sixth Ave, Bozeman, MT 59715, USA.
J AOAC Int. 2013 Sep-Oct;96(5):1138-51. doi: 10.5740/jaoacint.12-217.
This paper presents statistical techniques suitable for analyzing a collaborative study (multilaboratory study or ring trial) of a laboratory disinfectant product performance test (DPPT) method. Emphasis is on the assessment of the repeatability, reproducibility, resemblance, and responsiveness of the DPPT method. The suggested statistical techniques are easily modified for application to a single laboratory study. The presentation includes descriptions of the plots and tables that should be constructed during initial examination of the data, including a discussion of outliers and QA checks. The statistical recommendations deal with evaluations of prevailing types of DPPTs, including both quantitative and semiquantitative tests. The presentation emphasizes tests in which the disinfectant treatment is applied to surface-associated microbes and the outcome is a viable cell count; however, the statistical guidelines are appropriate for suspension tests and other test systems. The recommendations also are suitable for disinfectant tests using any microbe (vegetative bacteria, virus, spores, etc.) or any disinfectant treatment. The descriptions of the statistical techniques include either examples of calculations based on published data or citations to published calculations. Computer code is provided in an appendix.
本文介绍了适用于分析实验室消毒产品性能测试（DPPT）方法的协作研究（多实验室研究或循环试验）的统计技术。重点在于评估DPPT方法的重复性、再现性、相似性和响应性。所建议的统计技术易于修改以应用于单个实验室研究。内容包括对数据初步检查期间应构建的图表的描述，包括对异常值和质量保证检查的讨论。统计建议涉及对现有类型DPPT的评估，包括定量和半定量测试。内容强调将消毒剂处理应用于与表面相关的微生物且结果为活菌计数的测试；然而，统计指南适用于悬液测试和其他测试系统。这些建议也适用于使用任何微生物（营养细菌、病毒、孢子等）或任何消毒剂处理的消毒测试。统计技术的描述包括基于已发表数据的计算示例或对已发表计算的引用。附录中提供了计算机代码。
