Brown Simon, Cooke Delwyn G, Blackwell Leonard F
Deviot Institute, Deviot , Tasmania , Australia.
Scand J Clin Lab Invest. 2015 Sep;75(5):367-73. doi: 10.3109/00365513.2015.1031688. Epub 2015 May 11.
Ligand-binding assays, such as immunoassays, are usually analysed using standard curves based on the four-parameter and five-parameter logistic models. An estimate of the uncertainty of an analyte concentration obtained from such curves is needed for confidence intervals or precision profiles. Using a numerical simulation approach, it is shown that the uncertainty of the analyte concentration estimate becomes significant at the extremes of the concentration range and that this is affected significantly by the steepness of the standard curve. We also provide expressions for the coefficient of variation of the analyte concentration estimate from which confidence intervals and the precision profile can be obtained. Using three examples, we show that the expressions perform well.
配体结合分析，如免疫分析，通常使用基于四参数和五参数逻辑模型的标准曲线进行分析。为了得到置信区间或精密度曲线，需要对从这些曲线获得的分析物浓度的不确定性进行估计。通过数值模拟方法表明，分析物浓度估计的不确定性在浓度范围的极值处变得显著，并且这受到标准曲线斜率的显著影响。我们还给出了分析物浓度估计变异系数的表达式，从中可以得到置信区间和精密度曲线。通过三个例子，我们表明这些表达式表现良好。
