Comparison of subjective grading of limbal redness by eyecare practitioners in three countries

To compare grading of a biomicroscopy sign between eye care practitioners in three countries of widely different geographies-the United States, Australia, and Germany-and to test the accuracy of grading the sign within each country.

Four eye care practitioners licensed in the practice of optometry in each country were sent sets of limbal redness images that were printed from the Efron Grading Scale morphing program. Each set consisted of 18 images randomly selected over the range of 0 to 4, inclusive. Each image was printed on a separate page with the grade scale from 0 to 4 with 0.5 steps. Each practitioner received the same set of images that were packaged in random order. Seventy-two (72) grades were collected from each country, which exceeded the sample size estimates for both the between and within country objectives.

No statistically significant differences were found between countries for pairwise comparisons (P<0.02, McNemar test, adjusted for multiple comparisons). No statistically significant differences were found within each country between observed and expected values for grades from Germany and Australia (P>0.05, chi test); however, a statistically significant difference was found for the United States (P<0.05, chi test).

These findings support the generalization of results of limbal redness grades from trials that are conducted in several countries or in one country and used in another, i.e., practitioners may expect similar findings for this sign among their patients no matter which of their colleagues are investigators. They also indicate that different strategies for grading limbal redness may exist among practitioners in the countries tested. Study sponsors may want to include standardized training for investigators.