Comparison of ordinary linear regression, orthogonal regression, standardized principal component analysis, Deming and Passing-Bablok approach for method validation in laboratory medicine
Comparison of ordinary linear regression, orthogonal regression, standardized principal component analysis, Deming and Passing-Bablok approach for method validation in laboratory medicine
复制标题
DOI：
10.1515/labmed-2013-0003
复制
发表时间：
2013-06-01
影响因子：
--
通讯作者：
Klauke, Rainer
中科院分区：
文献类型：
--
作者：
Haeckel, Rainer;Wosniok, Werner;Klauke, Rainer
方法比较研究是方法验证的一个公认工具。通常在散点图上评估结果，其中拟合线通过几种方法计算，例如普通（垂直）线性回归（OLR）、正交回归（OR）、德明回归（DR）、Passing-Bablok方法（PBR）或标准化主成分回归（SPCR）。DR以其一般形式（gDR）应用，要求至少两个不同数量的不精确信息，并作为简单DR（sDR）仅具有一个数量的不精确信息。根据这些概念计算的回归线方程因测量范围、分析变异和不精密度比（s（AY）/s（AX））而异。关于哪一个统计概念最适合验证目的，全球仍在辩论。对大小为100的各种配对随机样本进行5000次重复模拟，并使用不同的回归模型进行评估。在各种条件下比较回归线的斜率和截距的行为。两个极端的测量范围和几个方差比的存在和不存在偏倚进行了研究。结果清楚地表明，DR是唯一一种在方法比较研究中通常出现的条件下无需任何预防措施即可应用的模型，因此应首选用于实验室医学。其他型号需要限制测量范围和/或不精确轮廓。用两个DR的回归系数计算的测量间隔不同位置的浓度差异未偏离允许偏倚。因此，使用gDR的优点并不能证明其与SDR相比更大的缺点是合理的。
A well-accepted tool for method validation is a method comparison study. Results are usually assessed on a scatter plot of which the fitting line is calculated by several approaches, for example, ordinary (vertical) linear regression (OLR), orthogonal regression (OR), Deming regression (DR), Passing-Bablok method (PBR) or standardized principal component regression (SPCR). DR was applied in its general form (gDR), requiring information of the imprecision of at least two different quantities and as simple DR (sDR) with imprecision information of only one quantity. The equation of the regression line calculated by these concepts varies depending on range of measurement, analytical variation and on imprecision ratio (s(AY) /s(AX)). There is still a global debate about which statistical concept is the most adequate for validating purposes. Various paired random samples with a size of 100 were simulated in 5000 replicates and evaluated with different regression models. The behavior of the slope and intercept of the regression lines were compared under various conditions. Two extreme ranges of measurement and several variance ratios in the absence and presence of bias were studied. The results clearly demonstrated that DR is the only model which can be applied without any precautions under conditions which usually occur in method comparison studies, and therefore should be preferred in laboratory medicine. Other models require restrictions with regard to range of measurement and/or imprecision profile. Differences of the concentrations at different positions of the measurement interval calculated with regression coefficients of both DRs did not deviate more than the permissible bias. Therefore, the advantage of using gDR does not justify its greater disadvantages in comparison with sDR.