OBJECTIVE Despite early interest in contrast sensitivity as a screening test for ophthalmic disease, most published opinion suggests that there is no benefit over conventional measurement of visual acuity. Taking a primary care perspective of screening, the authors evaluated the ability to discriminate those with any diagnosed ophthalmic disease in a large sample representative of the general population. METHODS Retrospective analysis of a clinical, cross-sectional survey. Snellen visual acuity, contrast sensitivity (Arden plates, American Optical contrast sensitivity test), and ophthalmic diagnosis were reported previously. METHODS A sample of 3283 subjects, all aged at least 50 years, were selected randomly from residents of a health district in Sydney, Australia. Ophthalmologic diagnosis (ophthalmic disease presence/absence) had been confirmed for 2522 of these subjects. METHODS Signal detection techniques (the receiver-operating characteristics function [ROC], quality ROC [QROC], and weighted kappa coefficient of association [kappa(r)]) were used to evaluate test discriminability. RESULTS Because analyses of right and left eyes were almost identical, only right eye results are presented. Advantages of kappa(r) over ROC were shown. Discrimination of those with diagnosed ophthalmic disease from those without ophthalmic disease was best with Arden plate 7 (kappa0.5 = 0.93) and was better than distance Snellen visual acuity (kappa0.5 = 0.59). Arden plate 7 (6.4 cyc/deg) correctly assigned 96% of subjects at its optimal pass-fail criterion. CONCLUSIONS In the primary care setting, a person older than 50 years of age with reduced contrast sensitivity, as determined by Arden plate 7, requires extra care in subsequent examinations because this person is likely to have an ophthalmic disease.