Statistical measures of clinical agreement.

Disagreements over clinical findings, diagnoses and management decisions are unavoidable in medicine. It is, however, necessary to ascertain whether they agree 'sufficiently'. Such investigations are often analysed inappropriately using Pearson's product-moment correlation coefficient. In this article, we discuss these as well as other methods, such as the kappa statistic, intra-class correlation coefficient and the graphical procedure to measure agreement using clinical examples.