1999 Costenbader Lecture. Outcome study in amblyopia: treatment and practice pattern variations

This study retrospectively evaluates the results of amblyopia therapy and suggests hypotheses for future study. We address the various methods of treatment and evaluate the results from the most common therapy techniques. Practice pattern variations were analyzed in addition to the analysis of overall results. For ophthalmologists, there is a need to determine whether actual medical practice approaches the established standard of care, if it exists at all. How often are medical procedures, thought to be appropriate, based on anecdotal observation (case reports) rather than hard data (clinical trials)? The 3 types of vision loss evaluated were strabismic, anisometropic, and deprivation amblyopia. The methods of treatment studied were full-time patching part-time occlusion, penalization, and occlusion of the contact lens. Nine centers, thought to have private as well as indigent patients, were recruited to participate in this study. The centers responded by filling out an extensive questionnaire and sending the information through the World Wide Web for inclusion in a spreadsheet. This information was then collated, and various statistical programs tabulated the results. Although trends, as a consequence of therapy, are suggested from our retrospective analysis, concrete results can only arise from a randomized prospective study. The study included 279 patients. There were a similar number of male and female patients. Only 77% of the patients without fusion before treatment had either single binocular vision or peripheral fusion at the conclusion of therapy. The log improvement of vision was significant in each group. Factors that potentially influenced the results were severity of distance acuity in the amblyopic eye before treatment, duration of treatment, and length of daily patching. The paper suggests that worse vision, not better vision, at the beginning, predicts better outcome in terms of improvement of visual acuity. For example, visual acuity less than 20/70 at the initiation of treatment led to better visual results of geometric log improvement. Surprisingly, among the 9 centers studied, there was a statistically significant difference in many of the areas related to practice patterns. Patient compliance, which directly affects outcome, was highly variable and is a factor that may be readily influenced by the treating physician.