Using clinical tests of colour vision to predict the ability of colour vision deficient patients to name surface colours

Purpose: To determine the predictive power of commonly used tests for abnormal colour vision to identify patients who can or cannot name surface colours without error. Methods: The colour vision of 99 subjects with colour vision deficiency (CVD) was assessed using the Ishihara, the Richmond HRR (2002), the Farnsworth D15, the Medmont C100 and the Nagel anomaloscope. They named 10 surface colours (red, orange, brown, yellow, green, blue, purple, white, grey and black), which were presented in two shapes (lines and dots) and three sizes. The surface colours were also named by an age-matched group of 20 subjects with normal colour vision. The performance of the clinical tests to predict the CVD subjects who made no colour naming errors and those who made errors is expressed in terms of the predictive value of a pass P(p) and the predictive value of a fail P-(F). Results: The P-(p) values of the tests were between 0.59 and 0.70 and P-(F) values were between 0.77 and 1.00. Conclusions: A 'mild' classification with the Richmond HRR test, especially if no more than two errors are made on the HRR diagnostic plates, identifies patients with abnormal colour vision who are able to name surface colour codes without error or only the occasional error. A pass of the Farnsworth D15 test identifies patients who will make no or few (up to 6%) errors with a 10 colour code, but who will be able to name the colours of a seven colour code that does not include orange, brown and purple. If protans are excluded, the predictive value for a pass P(p) for the Farnsworth D15 is improved from 0.59 to 0.70. The anomaloscope is not an especially good predictor of those who can recognise surface colour codes. However, an anomaloscope range >35 units identifies those who have difficulty in recognising surface colour codes, as does a fail at the Farnsworth D15 test.