Besides providing diagnostic or prognostic accuracy, a biomarker must be effective to ensure its adequate use in routine clinical implementation. Effectiveness can be defined as an individual’s potential to benefit from using the biomarker. This benefit includes the prevention of undesirable clinical outcomes and improvement in an individual’s quality of life.

Assessment of a biomarker’s effectiveness should include two sequential criteria. First, the biomarker should add prognostic value in relation to basic clinical and laboratory data. Second, the information provided by the biomarker should promote changes in medical conduct in a way that ultimately benefits the patient. Beginning with the analysis of prognostic value, a new biomarker must have a prognostic value independent of conventional risk markers. However, this criterion of statistical significance is not enough to guarantee clinical significance. Once the statistical significance is confirmed by multivariate analysis, it is necessary to proceed and evaluate whether the biomarker increases our ability to identify individuals who will present the undesirable outcome. This is done by incremental analysis of the C-statistic and net reclassification analysis. For example, although high sensitivity C-reactive protein has an independent association with cardiovascular risk, it only slightly increases the discriminatory capacity in Framingham score. The coronary calcium score, on the other hand, can increase the C-statistic of Framingham score, in addition to correctly reclassifying a proportion of patients. This incremental value is an essential condition for a biomarker to be effective, because only then can its result correctly modify clinical decisions.

[…]