From metabolic profiles to clinical interpretation: multivariate approaches to population-based and personalized reference intervals and reference change values

Interpretation of laboratory test results is a comparative process that requires reference data. Such data are derived for each analyte separately, without accounting for, the interrelationships among analytes. Physicians use test panels containing multiple analytes to enhance clinical significance and improve the accuracy of decision-making. However, current interpretation practices apply reference intervals and reference change values in a univariate manner - that is, each analyte in the panel is interpreted independently and no reference data are available to interpret the panel as a whole. Yet, metabolism is a network of biomolecules, each of which is related to others. Therefore, a multivariate approach - based on the correlations among biomolecules - can provide a more informative reference than univariate approaches and can be used more effectively in the interpretation of laboratory data. This concept can be summarized by a motto: <i>Combine single tests into meaningful groups, but interpret the group as a single clinical entity.</i> In this opinion paper, we present a practical approach for obtaining reference data for both reference intervals and reference change values to interpret laboratory test panels composed of related analytes.