Fig. 3
From: Validation of an equine serum amyloid A assay with an unusually broad working range
Bland-Altman plot of serum amyloid A (SAA) concentrations measured in a novel immunotubidometric assay developed for determination of SAA concentrations in horses (VET-SAA) and in an immunotubidometric assay previously validated for use in horses (LZ SAA). The difference between concentration readout in the two assays (y-axis) is plotted against the average SAA concentration (x-axis). The plot shows a large proportional and systematic bias, where the negative difference between concentrations measured in the two assays increases with increasing SAA concentration. The bias clearly exceeds the limits of the inherent combined imprecision of the assays (full lines), when concentrations exceeds 4000 mg/L (Data on LZ SAA imprecision derived from [13]; limits calculated as described in [16])