| dc.description.abstract | Introduction: Measuring colour discrimination of people who carry out tasks where colour is used to convey information and accurate colour judgments are essential for safe and efficient performance of the task is important in order to ensure that they can carry out the tasks. Individuals with congenital colour vision deficiencies are at a greater risk in making an error in colour judgment and this is the primary reason for colour vision testing in industry. Today, there are a large number of colour vision tests to detect colour vision deficiencies and/or estimate one’s ability to discriminate colours.
Purpose: The purpose of this study is to determine the validity and repeatability of new colour vision test “Colour Vision Reaction Time” (CVRT) for screening for colour vision defect. The study will also determine the repeatability of a selection of clinical colour vision tests, which are currently in use.
Material and methods: The test series was administrated to 75 colour normal subjects and 47 participants with red-green defects. Colour vision was classified based on Nagel anomaloscope. In the pseudoisochromatic tests, the Hardy, Rand, Rittler 4th edition (HRR), Ishihara 38 plate edition and Pseudoisochromatic Plates Ishihara Compatible (PIPC) tests subjects are required to identify coloured figure within background of a different colour. For the Colour Vision Reaction Time (CVRT) test, subjects need to locate a coloured Square on a computer’s screen using a mouse. The Cone Contrast Sensitivity Test (CCST) requires individuals to identify coloured letters that may appear in a gray background on the computer’s screen. A prototype of the ColorDx (pColorDx) test is similar to the printed pseudoisochromatic plates except that the plates are displayed on a computer screen. The Farnsworth-Munsell D15 (D15) test requires subjects to arrange coloured caps in order according to colour starting from a fixed cap.
Results: The agreement of the printed pseudoisochromatic tests with anomaloscope in terms of screening for red-green defect was good with kappa (κ) coefficient of agreement value of 0.96 or more on all three tests. The repeatability of the three tests was good with κ coefficient of 0.96 or more on the three tests. Both HRR and PIPC tests can screen for blue-yellow defects. There were 2 deuteranomalous subjects at the first visit and a different deuteranomalous individual at the second visit who made a single blue-yellow error in the HRR test. In the PIPC test, only one deuteranomalous subject failed the blue-yellow screening plates at the first visit with two errors. In terms of the classification as either protan or deutan, the agreement with the Nagel anomaloscope was perfect with the HRR test and acceptable with the Ishihara, but only fair for the PIPC test. The agreement of the repeatability of the classification was perfect at the HRR test and good at the Ishihara test whereas it was reasonable at the PIPC test. The HRR test was designed to classify the severity of the defect and there was a reasonable correlation between the HRR severity and the Nagel anomaloscope matching ranges.
The agreement of the three computerized colour vision tests with anomaloscope was good with κ coefficients ≥0.91. The repeatability of these three tests was good with κ coefficients ≥0.98. All the three tests can screen for blue-yellow defects. In the CVRT test, the response times of most subjects to the blue-yellow test figures were within 1.0 standard deviation of the white control value. The single exception was deuteranomalous subject who did not fail any other blue-yellow screening test. Ten subjects failed the pColorDx blue-yellow test, whereas 3 subjects failed the CCST S-cone portion. The CCST and pColorDx computer test can classify the defect as protan or deutan. Both tests showed a good-to-perfect agreement with anomaloscope. The pColorDx test can grade the severity of the defect in terms of mild, moderate to severe. The Spearman rank correlation coefficient with the Nagel matching ranges was only moderate.
The Farnsworth D15 test was included to determine whether there was a difference in the pass rate using the results from the first trial or requiring the subjects to pass on 2 of 3 trials. There was a marginal improvement in the pass rate using the 2 out of 3 rule. The repeatability of the 2 out of 3 trails in the D 15 test showed that there was a good agreement between sessions with κ coefficient of 0.87. In terms of classifying the defect as protan or deutan, based on the visual inspection, there was a good agreement with κ coefficient = 0.83. However, based on the Colour Difference Vector (CDV) angle parameter, all the colour defective subjects was correctly classified. The repeatability of classifying the type of the defect based on the CDV showed perfect agreement between the first and second visit. The D15 can classify the defect as mild versus moderate-to-severe. As expected, the majority of individuals who failed the D15 were classified as having moderate to severe classification on the HRR and pColorDx tests.
Discussion and Conclusion: The current study confirms that the three pseudoisochromatic tests are effective in screening for red-green colour vision defect. The HRR test may be preferred over the Ishihara and PIPC because the sensitivity was marginally higher than the other two tests. Agreement of the diagnostic plates with the Nagel anomaloscope as to whether the colour vision defect was protan or deutan varied across tests. The results from this study agreed with the Birch’s (1997) results for the Ishihara in that approximately 85% of the colour defectives were classified correctly as either protan or deutan. However, HRR classification results were slightly better than Cole, et al’s. In terms of the severity, our results were similar to Cole et al in that there were a reasonable correlation between the HRR severity and the Nagel anomaloscope matching ranges.
The three computerized colour vision tests are effective in terms of screening for red-green defects. The CCST had the highest agreement with anomaloscope, but it was not significantly different from the other two tests. However, the pColorDx ability to grade the severity was moderate, but it was slightly lower than the HRR plates. All three tests are capable of screening tritan defects. Our results suggest that a small number of deutans are likely to fail this portion of these tests.
The D 15 test showed a reasonable repeatability on terms of pass/fail when we used 2 out of 3 rule and marginally better than performing only one trail on separate days. In terms of the repeatability of classification, the study showed that there was a good agreement between sessions based on the visual inspection and perfect agreement between sessions based on Colour Difference Vector parameters. | en |
