Evaluation of the Third edition of the City University Colour Vision Test

The 3rd edition of the City University Colour Vision Test (CUT) was originally based on the Farnsworth-Munsell D-15 test (D15). The first part of the test is for detecting a defect, and the second part is used to diagnose the type and severity of the defect. This study evaluates the CUT 3rd edition relative to the Ishihara and the D15 colour vision tests.

Fifty nine colour vision normal subjects and 60 subjects with a congenital red-green colour vision defect were recruited. Subjects were tested with the Ishihara and CUT tests. Subjects who failed the Ishihara also performed the D15 test.

The agreement between the Ishihara and CUT screening plates was marginally higher when using the CUT failure criterion of >1 error compared with using >2 errors. If the diagnostic plates were included with the screening plates in determining the pass/fail outcomes, the agreement between the Ishihara and CUT was high, with a first-order agreement coefficient (AC1) of 0.90. The AC1 coefficient agreement between the D15 and CUT diagnostic plates in terms of pass/fail was 0.81 when using the D15 failure criteria of >1 or >0 crossing.

The level of agreement between the 3rd edition of the CUT and D15 was lower than the 2nd edition of the CUT. The primary reason for the lower agreement of the 3rd edition of the CUT was that it had a lower specificity relative to the D15 compared to the 2nd edition. Although the CUT predictive value for failing the D15 is over 90%, the predictive value for passing shows that 19%-25% of patients who pass the 3rd edition of the CUT test will fail the D15. The 3rd edition tends to misclassify protans as deutans or cannot classify the type of defect relative to the D15 and Ishihara.