Diagnostic Variation Caused by Differences in Viewbox Illumination and Visual Ability

Two pairs of bitewing radiographs containing verified carious lesions at various depths in the enamel and dentin were presented to twelve clinical dental teachers for diagnosis. The teachers used viewboxes with, unknown to them, either high, x̄: 5,900 lux, medium, x̄: 3,600 lux or low, x̄: 1,800 lux illumination. A line pattern test was used as a measure of the observers visual acuity. Ranking of ability in the viewing test was compared with and tested against diagnostic ability for sensitivity of total actual caries and dentin caries, overall quality of caries depth estimation (using Cohen's weighted Kappa, 1968), and the number of unreadable surfaces scored. There was no significant association between ranking in the viewing test and the various measures of diagnostic accuracy (Kendall's coefficient r > 0.22). Illumination level was found to be significantly associated with caries diagnostic accuracy although inter-examiner diagnostic variation was responsible for most of the variance for dentin caries sensitivity (p < 0.001, teachers variance 77%, illumination variance 15%, mean sensitivity 53%). Variance due to the differences between radiographs was estimated at < 2%. The range of visual ability of the teachers did not significantly effect caries diagnostic accuracy. The illumination range had a small but significant effect.