After publication of the basic findings (Kahn et al., <it>Am J Epidemiol</it> 106:17–32, 1977) in the Framingham Eye Study (FES), the National Eye Institute began preparation of a more extensive FES Monograph. During supplementary data editing for the Monograph, the coding of visual field tests was found to contain errors. The changes in glaucoma prevalence resulting from corrections of these coding errors, together with prevalence changes related to a modified reference population, alternate disease definitions, and editing changes other than those related to coding visual field tests, all contributed to a difference in prevalence between the Monograph and the original report. The present paper identifies and evaluates these differences and concludes for glaucoma that: 1) the only important differences in glaucoma prevalence relate to the visual field coding errors and the definition of disease; 2) the lower glaucoma prevalences of 1.9%, and 1.2% excluding blind spot enlargement, reported in the Monograph for ages 52–85, probably underestimate the true prevalence in this Framingham population because a) glaucoma suspects who did not complete the definitive exams are probably not all disease free; b) persons screened who did not meet criteria for glaucoma suspects are probably not all disease free; and c) the incompletely recorded Framingham visual fields, recoded from definite to doubtful, probably include persons who would be classified as having definite defects had the fields been properly recorded. An analogous study of differences between original report and Monograph in prevalence of diabetic retinopathy indicated that change in definition of disease was the only contributor to significant difference.
No takes yet. Share an insight, caveat, or question.
Kahn et al. (1980) studied this question.