Abstract

Purpose: To observe the inter-rater reliability between the observers for visual acuity, refractive error, corneal reflex, and inter-pupillary distance.


Methods: A random subset of 50 students from primary (1st to 4th grade), higher primary (5th to 7th grade) and secondary school (8th to 10th grade), for a total of 150 students were selected by stratified sampling from the prevalence study VARES (Visual acuity, refractive error, and squint) performed among 1752 school children of Udupi district, India in 2013-2014. The Institutional Ethical Committee & District Health Administration board permission was acquired. There were two examiners: Examiner 1 (JT), a qualified optometrist with two years of experience, and Examiner 2 (MT), an optometry student. Examiners 1 and 2 performed a single measurement for visual acuity using the COMPlog computerized logMAR vision chart at 3 metres; Landolt C or Sloan letter optotypes used. Refractive error, corneal reflex, and inter-pupillary distance was measured by Plusoptix Power-refractor twice by Examiner 1 and thrice by Examiner 2; the average of each value was considered for analysis. Data was tested for normality using the Kolmogorov-Smirnov test. Statistical significance was considered p < 0.05. The agreement between the observers for the two procedures was traced by a Bland-Altman plot.


Results: The mean age of the students was 10.21 years (± 2.80 SD). The measurement of agreement for visual acuity (COMPlog) between the examiners was tested using Kappa statistics with 0.79 for OD, p < 0.05 and 1.0 for OS, p < 0.05 supported by Wilcoxon sign rank test showing 0.067 for OD and 0.564 for OS between the observers. The Bland-Altman plot also showed good agreement between the observers. Paired t-test for refractive error showed 0.117 (CI: -0.11­–0.10) for OD & 0.067 (CI: -0.005–0.157) for OS, with Bland-Altman plot showing good agreement between the observers. Paired t-test showed 0.323 for corneal reflex and 0.520 for inter-pupillary distance.


Conclusion: From the results we report that there is no interobserver variability for visual acuity, refractive error, corneal reflex and inter-pupillary distance in school screening with COMPlog and Plusoptix A09.