Abstract
Purpose: :
When no errors are allowed the Ishihara colour vision test is very sensitive at detecting loss of red / green (RG) chromatic sensitivity, but this is achieved at the expense of specificity. In practice, the number of errors is often taken as a measure of the subject’s severity of RG loss and colour "safe" certification within professional environments can require anything from zero to six or less errors. The purpose of this study was to derive a "severity" weight for each of the Ishihara pseudoisochromatic plates and to examine how the number of errors made relates to the subject’s loss of RG chromatic sensitivity.
Methods: :
742 subjects (236 normals, 340 deutans and 166 protans) were investigated using the full 38-plate edition of the Ishihara test. Errors were noted for each of the first 24 plates. In addition all subjects had their colour vision assessed using the CAD (Colour Assessment & Diagnosis) test and the RG thresholds measured.
Results: :
The Ishihara data were used to compute a plate specific "weight" to indicate its level of difficulty within each subject group. This approach made it possible to calculate a weighted score to take into account both the number and the difficulty of the plates failed. We label this the subject’s "severity index". The correlation between CAD thresholds and the subject’s corresponding severity index was then examined with and without the weights applied.
Conclusions: :
The results show very poor correlation between the subject’s CAD threshold and the number of plates failed. The relationship is highly non-linear and improves only marginally when the severity index takes is applied which accounts for both the number and the difficulty of the plates failed. The results demonstrate that even with appropriate weighting, the number of Ishihara plates failed cannot be used to describe the severity of RG colour vision loss.
Keywords: color vision • color pigments and opsins • clinical research methodology