Assessing accuracy of non–eye care professionals as trainee vision screeners for children

OBJECTIVE: To determine the level of agreement between non-eye care trainees and a trainer (ophthalmologist) in a vision screening program. DESIGN: Prospective, observational study carried out in 3 phases (Phase I-III). PARTICIPANTS: Study population included 1228 children, aged 6-14 years, at 5 elementary schools in the city of Hamilton. METHODS: In Phase I, 1228 children were screened by the trainee screeners, of which 273 children failed the vision testing. Of these 273 children, 170 consented to enrolment into Phase II and were examined by an ophthalmologist, who confirmed that 105 of these children were true positives. On retesting (Phase III), the ophthalmologist passed 158 of the 163 randomly selected children who passed in Phase I. RESULTS: Overall, trainee screeners had a sample sensitivity of 95.5% and sample specificity of 70.8% in detecting children who should fail vision screening. When we used the positive and negative prediction values obtained, 198 of the 1228 children had vision impairment-providing an estimated prevalence of 16.1%, or 161 children per 1000 population. CONCLUSIONS: Non-eye care professionals can be trained to an acceptable degree of accuracy to perform certain vision screening tests on children. Such screening methods may be a useful approach to address existing gaps in provision of eye care for many Canadian children, thereby ensuring that all children receive timely vision screening.