Abstract
Abstract: :
Purpose: The MTI PhotoScreenerTM has been widely utilized to screen preschool children for amblyogenic factors with drastically varying predictive value positive outcomes due to inconsistent grading criteria for the evaluation of screen fails. The purpose of this study is to compare exam outcomes of children referred from each of two sets of criteria; those instituted by Tong et al (Ophthalmology, 2000) and those instituted by Donahue and Johnson (Ophthalmology, 2000). Methods: Between 8/17/00 and 6/1/01, 17,309 preschool children were screened with the MTI PhotoScreenerTM in childcare centers by trained volunteers. Photo interpretation was completed by the Vanderbilt Ophthalmic Imaging Center using criteria established by Tong et al and Donahue and Johnson (SD/TJ), with oversight of all referrals by Donahue (pediatric ophthalmologist) and Johnson. If children met the referral criteria for either set of criteria, they were sent to local optometrists or ophthalmologists for formal examination. Results: 786 children were referred (4.5%) of which 638 had completed exams (81.2% follow-up). Results of eye exams were divided into four categories: 1) Failed both Tong and SD/TJ criteria for hypermetropia (105 refer, 97 exams). Results showed exam fail=82, predictive value positive (PVP)=90.1%; 2) Tong criteria fail, SD/TJ fail for reason other than hyperopia (52 refer, 44 exams). Exam fail=33, PVP= 76.7%; 3) Tong criteria fail, SD/TJ pass (48 refer, 41 exams). Exam fail=16, PVP=42.1%; and 4) Tong criteria pass, SD/TJ fail for reason other than hyperopia (581 refer, 456 exams) Exam fail=307, PVP=71.2%. The SD/TJ criteria missed 13 children with amblyogenic factors that Tong criteria identified at the cost of 22 over-referrals (false positives). The highest hypermetropia in this group was +7.25. These 13 children represent less than 0.1% of the total children screened. Conclusion: These results suggest that the Tennessee referral criteria could be relaxed to avoid missing some children. But, this would substantially lower the predictive value positive with very few additional children identified. This may not be desirable in a screening setting led primarily by volunteers.
Keywords: 575 screening for ambylopia and strabismus • 313 amblyopia