Abstract
Abstract: :
Purpose: To determine testability and inter-tester reliability of preschool children tested with the Random Dot E (RDE) stereoacuity test. Methods: Subjects were 1133 3- to 5-year-old children who were participants in Head Start programs in 5 communities. The sample was over-weighted with children who had failed the routine Head Start vision screening (58% failures; 42% non-failures). Each child had stereoacuity measurement attempted by two licensed eye care practitioners (optometrists and pediatric ophthalmologists) using the RDE test, first in a screening setting and later (median 3 days) as part of a comprehensive eye examination. Test distance for the non-stereo E test plate was 50 cm, and for the stereo E test plates 50 cm (504 arc sec disparity), 100 cm (252 arc sec disparity), and 150 cm (168 arc sec disparity). Children who discriminated between the non-stereo E and blank cards on 4 of 4 or 4 of 5 presentations were considered testable. The extent of agreement beyond chance between the two stereoacuity test results was assessed using the weighted Kappa statistic (Kw). Results: Testability increased with year of age at both the first (86% of 3-, 89% of 4-, and 93% of 5-year olds; p=.016) and second (90% of 3-, 94% of 4-, and 98% of 5-year olds; p<.0001) test sessions. Overall, the stereoacuity score was better at the second test session and agreement between sessions was moderate (Kw=0.43; 59% identical scores). Although the percentage with identical scores increased with age (53% of 3-, 59% of 4-, and 63% of 5-year-olds), Kw statistics did not differ with age (p=0.49). Conclusions: Testability of the RDE test increased with age of the preschool child. Test-retest agreement was moderate and did not differ by age of the child.
Keywords: binocular vision/stereopsis • clinical (human) or epidemiologic studies: sys • clinical (human) or epidemiologic studies: tre