S Hanson1, S K Krishnan, J Phillips. 1. Department of Optometry and Vision Science, The University of Auckland, New Zealand.
Abstract
BACKGROUND: The validity of the Cup:Disc (C:D) ratio as a clinical measure depends on the level of agreement among observers. In this study, we investigated the effect of observer experience on interobserver and intraobserver agreement in C:D ratio estimation. METHODS: Nine 3rd-year (Yr3) and nine 4th-year (Yr4) optometry students and nine optometrists (Opt), determined C:D ratios from 26 optic disc photographs viewed monoscopically and stereoscopically. Observer agreement was investigated using weighted kappa (Kw) analysis. RESULTS: Interobserver agreement was higher (p < 0.01) among optometrists than among 4th-year students and higher (p < 0.01) among 4th-year than among 3rd-year students (mean Kw: Opt = 0.61; Yr4 = 0.51; Yr3 = 0.39). Stereoscopic estimates of C:D ratios were higher (p < 0.0001) than monoscopic estimates, and interobserver agreement was higher (p < 0.001) when evaluating the vertical rather than the horizontal C:D ratio. Intraobserver agreement (mean Kw: Opt = 0.67; Yr4 = 0.66; Yr3 = 0.62) did not vary significantly (p = 0.78) with experience. CONCLUSIONS: The extent to which an observer reproducibly assesses the C:D ratio from optic disc photographs (intraobserver agreement) does not increase with experience. However, the agreement between different observers in assessing the C:D ratio (interobserver agreement) increases with experience.
BACKGROUND: The validity of the Cup:Disc (C:D) ratio as a clinical measure depends on the level of agreement among observers. In this study, we investigated the effect of observer experience on interobserver and intraobserver agreement in C:D ratio estimation. METHODS: Nine 3rd-year (Yr3) and nine 4th-year (Yr4) optometry students and nine optometrists (Opt), determined C:D ratios from 26 optic disc photographs viewed monoscopically and stereoscopically. Observer agreement was investigated using weighted kappa (Kw) analysis. RESULTS: Interobserver agreement was higher (p < 0.01) among optometrists than among 4th-year students and higher (p < 0.01) among 4th-year than among 3rd-year students (mean Kw: Opt = 0.61; Yr4 = 0.51; Yr3 = 0.39). Stereoscopic estimates of C:D ratios were higher (p < 0.0001) than monoscopic estimates, and interobserver agreement was higher (p < 0.001) when evaluating the vertical rather than the horizontal C:D ratio. Intraobserver agreement (mean Kw: Opt = 0.67; Yr4 = 0.66; Yr3 = 0.62) did not vary significantly (p = 0.78) with experience. CONCLUSIONS: The extent to which an observer reproducibly assesses the C:D ratio from optic disc photographs (intraobserver agreement) does not increase with experience. However, the agreement between different observers in assessing the C:D ratio (interobserver agreement) increases with experience.
Authors: Maged S Habib; James A Lowell; Nick S Holliman; Andrew Hunter; Daniella Vaideanu; Anthony Hildreth; David Hw Steel Journal: BMC Ophthalmol Date: 2008-07-23 Impact factor: 2.209