We investigated the ability of the leucocyte count to discriminate between "inflammatory diseases and malignancies" on the one hand and other, often harmless and self-limiting diseases on the other, and compared this ability with that of the erythrocyte sedimentation rate (ESR) in general practice. In cooperation with nine general practitioners (GP) we prospectively followed 151 patients who were seen by their GP because of a new complaint for which the GP wanted to know their ESR. ESR and one leucocyte count measurement (Technicon H6000) were performed at the local hospital laboratory. The patients were seen again after three months, to determine the follow-up diagnoses. By comparing the test results with the follow-up diagnoses, ESR was found to have a better discriminating ability than the leucocyte count test (ESR: sensitivity = 53%, specificity = 84%, positive predictive value = 29%, negative predictive value = 93%, Odds Ratio = 5.73). Using Receiver Operating Characteristic analysis, this conclusion could be shown to be independent of the chosen reference values. We conclude that the determination of the leucocyte count is not a clinically interesting alternative to ESR.