Bland and Altman plot method is a graphical plot approach that compares related data sets, supporting the eventual replacement of a measurement method for another one. Perhaps due to its graphical easy output it had been widely applied, however often misinterpreted. We provide three nested tests: accuracy, precision and agreement, as a means to reach statistical support for the equivalence of measurements. These were based on structural regressions added to the method converting it on inferential statistical criteria, verifying mean equality (accuracy), homoscedasticity (precision), and concordance with a bisector line (agreement). A graphical output illustrating these three tests were added to follow Bland and Altman's principles. Five pairs of data sets from previously published articles that applied the Bland and Altman's principles illustrate this statistical approach. In one case it was demonstrated strict equivalence, three cases showed partial equivalence, and there was one case without equivalence. Here we show a statistical approach added to the graphical outputs that turns the Bland-Altman otherwise graphical subjective interpretation into a clear and objective result and with significance value for a reliable and better communicable decision.
翻译:Bland 和 Altman 绘图方法是一种图形绘图方法,比较相关数据集,支持最终将测量方法替换为另一个数据集。也许由于该方法的图形简单,它被广泛应用,但往往被误解。我们提供了三种嵌套测试:精确、精确和一致,作为达到统计对等测量支持的手段。这些测试基于结构回归,在转换方法中添加了推断统计标准,核实平均平等(准确性)、同质性(精度)和与双部门线(协议)的一致性。在Bland 和 Altman 原则中添加了显示这三种测试的图形输出。从先前发表的应用Bland 和 Altman 原则的文章中得出的五对数据集说明了这一统计方法。在一种情况下,它表现为严格对等,三个案例显示部分对等,有一个案例没有对等性。我们在这里展示了一种统计方法,图形输出将Bland-Altman 的图形化主观解释转化为清晰和客观的结果,对于可靠和更好的传染性决定具有重大意义。