Despite its frequent use, little is known about the ability of the erythrocyte sedimentation rate to discriminate between 'pathology' (inflammatory diseases and malignancies) and 'no pathology' in general practice. This has been studied by following 362 patients who presented to their general practitioner with a new complaint, for which the general practitioner considered determination of the erythrocyte sedimentation rate to be indicated. The test was performed at the local hospital laboratory and the patients were seen again after three months, in order to establish the follow-up diagnoses. By comparing the test results with the follow-up diagnoses, combined with receiver operating characteristic curves and regression analysis, the erythrocyte sedimentation rate was found to have a reasonable discriminating ability with respect to malignancies and inflammatory diseases (sensitivity 53%, specificity 94%, positive predictive value 48%, negative predictive value 91%, odds ratio 15.1). The upper limit for the normal erythrocyte sedimentation rate should be set at approximately 12 mm hour-1 for men and 28 mm hour-1 for women, and needs no correction for age. It is concluded that the erythrocyte sedimentation rate still deserves a place in the general practitioner's daily routine.