Preschool vision screening: rationale, methodology and outcome

Abstract

Although population outcome studies support the utility of preschool screening for reducing the prevalence of amblyopia, fundamental questions remain about how best to do such screening. Infant photoscreening to detect refractive risk factors prior to onset of esotropia and amblyopia seems promising, but our current understanding of the natural history of these conditions is limited, thus limiting the prophylactic potential of early screening. Screening for strabismic, refractive and ocular disease conditions directly associated with amblyopia is more clearly proven, but the diversity of equipment, methods and subject populations studied make it difficult to draw precise summary conclusions at this point about the efficacy of photoscreening. Sensory-based testing of preschool-age children exhibits a similar combination of promise and limitations. The visual acuity tests most widely used for this purpose are prone to problems of testability and false negatives. Moreover, the utility of random-dot stereograms has been confused by misapplication, and new small-target binocularity tests, while attractive, are as yet inadequately field-proven. The evaluation standard for any screening modality is treatment outcome. However, variables in amblyopia classification and quantitative definition differences, timing of presentation, nonequivalent treatment comparisons, and compliance variability have been uncontrolled in virtually all extant studies of amblyopia treatment outcome, making it difficult or impossible to evaluate either the relative efficacy of different treatment regimens for amblyopia or the effects of age on treatment outcome within the preschool age range. The latter issue is a central one, since existence of such an age effect is the primary rationale for screening at younger rather than older preschool ages. The relatively low prevalence of amblyopia makes it difficult to achieve a high screening yield in terms of predictive value, but functionally increasing prevalence by selective screening of high risk populations causes further problems. Unless a "supertest" can be devised, with very high sensitivity and specificity, health policy decisions will be required to determine which of these two characteristics should be emphasized in screening programs. Performance of screening tests can be optimized, however, with adequate training, perhaps via instructional videotapes.