Reliability of vision screening tests for school children.

PURPOSE: Estimate the reliability of the E-chart as used with Israeli school children. DESIGN: Cross-sectional, population-based study conducted among 751 Israeli students of the Northern District, aged 6- and 7-year-olds and 13- and 14-years-old in 30 schools in 2003. METHODS: Each student was screened separately by two public health nurses using the illiterate E-chart. Collected data included the students' vision and demographic characteristics, the nurses' professional background, and whether they referred students for medical testing. The reliabilities of vision testing and of the recommendations were determined using total, positive, and negative percentages of agreement and Kappa coefficients. FINDINGS: Total percentage of agreement on vision (combined findings for both eyes) was 78.2% (Kappa 0.47, 95%CI 0.41-0.53). Logistic regression models to predict agreement on vision abnormality showed a higher percentage of agreement among females and 13- and 14-year-old students than among males and 6- and 7-year old students. Total agreement of 85.8% was found in referral recommendations (Kappa 0.58, 95%CI 0.51-0.65). Significant relationships were noted with student age, ethnicity, subdistrict of residence, nurse seniority, and agreement on vision findings. CONCLUSIONS: Improvement in school vision-screening reliability is needed, especially among 6- and 7-year-old students. To this end, the determinants of fair reliability should be investigated and training programs planned. Reasons for differences in the reliability of nurses' recommendations detected among subdistricts must be further studied, together with careful supervision, to ensure better performance and adherence to PHS guidelines. Implications for nurses and nursing should be considered. CLINICAL RELEVANCE: Demographic characteristics were found to predict reliability, which can guide nurses in selecting students who need more careful attention or closer supervision during vision testing.