[Everything (or almost everything) about the Kappa coefficient]

Sensitivity, specificity, and predictive value are the standard parameters used to evaluate the efficacy of diagnostic tests. However all three parameters assume the existence of a gold standard test. Thus the problem arises as to what method to use in the absence of a benchmark. The solution involves assessment of agreement between two raters. This article based on an example drawn from an African setting describes the value and computation of the most widely used tool for assessing agreement and reproducibility of measures, i.e., the kappa coefficient.