Selection of controls for clinical research studies in ophthalmology

Ophthalmologists are becoming increasingly aware of the need for well-designed, controlled studies to investigate etiology and treatment of ocular conditions. While patients with the conditions of interest, or with characteristics that place them at high risk of eventual development of such conditions, may be readily identified in the ophthalmology clinic, identification of appropriate controls to be used for comparison may be quite difficult. In this article, epidemiologic principles for selection of controls for retrospective (case-control) and prospective (cohort) studies are reviewed. Two related studies, a case-control study in which multiple control groups were employed and a 15-year cohort study of all cases and controls, provided an opportunity to compare controls selected at random from an ophthalmology practice with controls selected at random from the general population. Participation rates were higher among office patients selected for study, both in the case-control study and in the cohort study. In the cohort study, office controls were more likely to lose visual acuity and to develop new degenerative eye conditions. These differences between control groups selected from two different sources emphasize the need for careful evaluation of potential groups of controls with respect to biases that each may bring to interpretation of clinical research findings.