Accuracy of vision measurement by ancillary personnel among rural Chinese secondary school children: The X–PRES Study

Purpose: To evaluate the accuracy of teachers and government-run program as compared to screening by vision professionals in detecting low vision among school children in rural China.

Methods: After training, 32 teachers assessed uncorrected and presenting vision on a random cluster sample of 1892 children from all secondary schools in Xichang, China. Vision was independently assessed by professionals masked to teacher results. Annual government screening by health technicians also evaluated presenting vision of all children. Factors predicting accurate screening results were examined.

Results: Compared to vision professionals, sensitivity (93.5%) and specificity (91.2%) of teachers in identifying low vision without refractive correction (uncorrected vision <=/12 in either eye) were better than teacher values for low presenting vision (sensitivity 85.2%, specificity 84.8%). Failure of teachers to identify and correctly test children who owned but were not wearing glasses, and apparent teacher bias towards better vision measurement among spectacle wearing children, seemed to explain worse results for screening of presenting vision. Among teachers, 87.5% (n=28) thought teacher vision screening was beneficial. Wearing spectacles was the student factor most strongly predictive (P<0.0001) of inaccurate teacher screening. Government screening compared to study personnel had low sensitivity (86.7%) and specificity (28.7%).

Conclusion: Teacher vision screening after training can achieve good diagnostic accuracy in this setting. There appears to be support among teachers for this approach. Screening of uncorrected rather than presenting vision is recommended in similar settings with high prevalence of corrected and uncorrected myopia. The government program was of little benefit due to low specificity.