Lei Wang1,2, Nathan Congdon3,4,5, Ruth E Hogg3,4,5, Siqi Zhang2, Mengjie Li1,2, Yaojiang Shi2, Ling Jin4, Fei He6, Huan Wang2, Matthew Boswell7, Mony Iyer8. 1. International Business School, Shaanxi Normal University, Shaanxi, China. 2. Center for Experimental Economics in Education, Shaanxi Normal University, Shaanxi, China. 3. Centre for Public Health, Queen's University Belfast, Belfast, UK. 4. State Key Laboratory of Ophthalmology, Zhongshan Ophthalmic Center, Sun Yat-sen University, Guangzhou, China. 5. Orbis International, New York City, New York, USA. 6. Food and Resource'Economics Department, University of Florida, Gainesville, Florida, USA. 7. Freeman Spogli Institute of International Studies, Stanford University, Stanford, California, USA. 8. OneSight, Mason, Ohio, USA.
Abstract
PURPOSE: To explore the accuracy and cost-effectiveness of three vision screening models among preschool children in rural China. METHODS: Vision screening was carried out among children aged 4-5 years in 65 preschools in two counties in Northwest China, using Crowded Single Lea Symbols to test visual acuity. Children were assigned randomly by school to one of three screening models: screening by teachers (15 schools, 1835 children), local optometrists (30 schools, 1718 children) or volunteers (20 schools, 2183 children). Children identifying ≥2 symbols incorrectly in either eye failed screening. Accuracy of screening was compared with screenings executed by experienced optometrists among 141 children selected randomly from the three screening models. Direct and indirect costs for each model were assessed. Costs to detect a true case failed screening were estimated. RESULTS: The sensitivity for three models ranged from 76.9% to 87.5%, specificity from 84.9% to 86.7% and standardized positive predictive value from 83.7% to 85.7%. None differed significantly between models. The costs per case detected were $37.53, $59.14 and $52.19 for the teachers, local optometrists and volunteers. In producing the cost estimates for teacher screening and local optometrist screening models, we used a salary payment that was identical for both models (with the salary being equal to that of the optometrist). The teacher screening model was the most cost-effective. CONCLUSION: Accuracy of screening by teachers, local optometrists and volunteers was the same in this setting, but the use of teachers was most cost-effective, reducing the cost per case detected by almost 40%.
PURPOSE: To explore the accuracy and cost-effectiveness of three vision screening models among preschool children in rural China. METHODS: Vision screening was carried out among children aged 4-5 years in 65 preschools in two counties in Northwest China, using Crowded Single Lea Symbols to test visual acuity. Children were assigned randomly by school to one of three screening models: screening by teachers (15 schools, 1835 children), local optometrists (30 schools, 1718 children) or volunteers (20 schools, 2183 children). Children identifying ≥2 symbols incorrectly in either eye failed screening. Accuracy of screening was compared with screenings executed by experienced optometrists among 141 children selected randomly from the three screening models. Direct and indirect costs for each model were assessed. Costs to detect a true case failed screening were estimated. RESULTS: The sensitivity for three models ranged from 76.9% to 87.5%, specificity from 84.9% to 86.7% and standardized positive predictive value from 83.7% to 85.7%. None differed significantly between models. The costs per case detected were $37.53, $59.14 and $52.19 for the teachers, local optometrists and volunteers. In producing the cost estimates for teacher screening and local optometrist screening models, we used a salary payment that was identical for both models (with the salary being equal to that of the optometrist). The teacher screening model was the most cost-effective. CONCLUSION: Accuracy of screening by teachers, local optometrists and volunteers was the same in this setting, but the use of teachers was most cost-effective, reducing the cost per case detected by almost 40%.