Diagnosis: Method Validation
• A method correlation study between your laboratory’s proposed method and a gold standard or reference method should always be part of a method validation for a high-complexity, laboratory-developed test (subject to the availability of a reference method).
• Classically, the reference method is plotted on the x-axis and your laboratory’s method on the y-axis.
• If the plot is classically drawn, a slope greater than 1 implies a positive and proportional bias of your method relative to the reference method.
• If plotted classically, a y-intercept term greater than 0 implies a constant positive bias.
• Both constant and proportional biases may coexist in a method comparison study.
• An R2 value approaching 1 indicates that the linear regression line drawn between the two methods is an excellent fit. However, high coefficient of determination does not rule out bias in your method.