Repeated automatic visual field screening in a cohort study.

During 1977 and 1978 automatic visual field screening was performed in 1511 persons born 1907-1921 and taking part in a population study. Three years later automatic visual field screening was repeated in 2565 eyes of 1293 persons. "Irrelevant" defects were common and had a limited repeatability. "Relevant" defects were easily detected at the first and reproduced at the second survey. From a clinical point of view, the yield of the visual field screening at the second survey was restricted. We concluded that repeated visual field screening by automatic perimetry is quite feasible but has little reward.