VALIDITY OF LINEAR-REGRESSION IN METHOD COMPARISON STUDIES - IS IT LIMITED BY THE STATISTICAL-MODEL OR THE QUALITY OF THE ANALYTICAL INPUT DATA

We compared the application of ordinary linear regression, Deming regr ession, standardized principal component analysis, and Passing-Bablok regression to real-life method comparison studies to investigate wheth er the statistical model of regression or the analytical input data ha ve more influence on the validity of the regression estimates. We took measurements of serum potassium as an example for comparisons that co ver a narrow data range and measurements of serum estradiol-17 beta as an example for comparisons that cover a wide data range. We demonstra te that, in practice, it is not the statistical model but the quality of the analytical input data that is crucial for interpretation of met hod comparison studies. We show the usefulness of ordinary linear regr ession, in particular, because it gives a better estimate of the stand ard deviation of the residuals than the other procedures. The latter i s important for distinguishing whether the observed spread across the regression line is caused by the analytical imprecision alone or wheth er sample-related effects also contribute. We further demonstrate the usefulness of linear correlation analysis as a first screening test fo r the validity of linear regression data. When ordinary linear regress ion (in combination with correlation analysis) gives poor estimates, w e recommend investigating the analytical reason for the poor performan ce instead of assuming that other linear regression procedures add sub stantial value to the interpretation of the study. This investigation should address whether (a) the x and y data are linearly related; (b) the total analytical imprecision (s(a,tot)) is responsible for the poo r correlation; (c) sample-related effects are present (standard deviat ion of the residuals much greater than s(a,tot)); (d) the samples are adequately distributed over the investigated range; and (c) the number of samples used for the comparison is adequate.