Biomarker development: current issues and perspectives

International audienceWe have read the letter of Ho et al. in response to our work with great interest. We agree with their general comment regarding the interest of evaluating agreement between two multiplex immunoassays. Ho et al. reminds us that correlation analysis suffers from limitations in assessing agreement between methods. Such agreement should be assessed using Bland-Altman plots. This method is largely known as the original 1986 paper of Bland and Altman has been the most frequently cited article ever to appear in the Lancet and is one of the ten most frequently cited statistical articles ever (Bland and Altman, 2012). Nevertheless, we decided to use Spearman’s rank correlation, since our original intent was to compare methods and biological fluids, which is not exactly the same as assessing the degree of agreement. Ultimately, several conditions are required to evaluate agreement with Bland-Altman graphics