View/Download PDF
Buy Reprints
PDF

Translate this page into:

Original Article
56 (
2
); 1-8

Colour Vision Testing for Selection of Civil and Military Pilots in India: A Comparative Study of Four Different Testing Methods

Senior Advisor Ophthal & Vitreo-retinal Surgery IAM IAF, Bangalore – 560017
Senior Advisor and Head Ophthal, CHAF Bangalore-7
Graded Specialist Ophthal, IAM IAF Bangalore - 17

Abstract

The Ishihara and the Martin Lantern Test (MLT) are the currently recommended tests for aviation colour vision testing in India. There are many anecdotal reports of conflicting results for these tests from different centres. This study was therefore conducted to compare these two tests and two other standard tests, the Farnsworth D15 and Heidelberg Multi-Colour anomaloscope(HMC) in assessing colour vision among 225 aspiring aviators for military flying and civil aviation prospectively over 20 months. Experienced colour normal pilots also tested as controls. All the four tests showed variable results in passing/failing same individuals with maximum variance attributable to the MLT and better correlation obtained on eliminating MLT. Comparing each test with HMC taken as gold standard, the best correlation was seen with the Ishihara and least with MLT. The D15 passed those with milder defects, failed those with severe anomaly and correlated well with Ishihara. The anomaly quotient of the anomalous was significantly different from colour normal controls. The study indicates that functional tests such as the MLT may not be reliable in assessing colour vision deficiency accurately and hence there is a need to include diagnostic tests such as the anomaloscope and the D15 in the testing protocol.

Keywords

colour vision
aviation
Ishihara
Martin Lantern test
Farnsworth D15
Heidelberg anomaloscope

Introduction

Colour coding is often used in aviation to make visual information more conspicuous. These include colour visual display in the cockpit, external visual cues such as airfield lighting, aircraft formation lights, coloured smoke or light signals used in military and colour coding for segmentation and grouping operation. It is estimated that approximately 4% of the human population (8-10 % males and 0.4% females) are congenitally colour deficient [1, 2].These individuals, who have no other visual deficit, may pass undetected unless tested for their colour perception. Though many tests for colour vision are available, there is no consensus on the ideal method, with different countries using different tests [3, 4].Studies comparing these tests have shown variable results for different tests [5, 6].Testing with a comprehensive battery of colour vision tests as is used in clinical settings [7, 8] cannot be done for occupational screening where we need simple, quick, and reliable tests. In India, the pseudoisochromatic plates (Ishihara and Tokyo Medical College) and the Martin Lantern test (MLT) are the colour vision tests [9], currently recommended for both military and civil aviation aspirants who are evaluated at medical centres of the Indian Air Force (IAF). There are manyanecdotal reports of conflicting results for these tests from different centres. This prospective study was therefore conducted to evaluate colour perception among applicants for military and civil aviation using these two tests and two other standard tests, namely the Heidelberg Multicolour Anomaloscope(HMC) and the Farnsworth D15 [10,11,12]. The anomaloscope is considered as the gold standard for colour vision testing in clinical research [10,11].Both these tests also test blue yellow perception for which presently there are no guidelines in the IAF. With the increasing use of blue and majenta in the modern day cockpits blue yellow testing may assume great significance in the near future, hence this study also evaluated blue yellow colour perception among the candidates.

Material and Methods

Individuals reporting to a large medical evaluation centre of the IAF for medical examination for selection to military flying or for grant of commercial pilot’s licence (CPL) over a period of 20 months were tested for colour perception with all four tests, after written informed consent. A group of controls consisting of colour normal experienced civil and military pilots were also evaluated with the 4 tests. Since the colour visions standards for selection are different for CPL and military flying in India, the subjects were grouped as below:

Test Group A (Military Pilots) -Current standard is a pass on Ishihara with no mistakes in reading the first 25 plates of the 38 plate Ishihara chart and pass on MLT by identifying correctly red, green and white light shown in pairs randomly through the smallest aperture of the Martin Lantern at 6 metres in a completely dark room, graded as Colour Perception standard One (CP-I).

Test Group B(CPL)- Current minimum standard is correct recognition of signal white ,green and red shown through the large aperture of Martin lantern at 1.5 m graded as CP-III or reads the requisite plates of Ishihara book as specified in IAF publication IAP 4303 4th edition para 2.11.24[9]. Those who fail are graded Colour Perception Standard IV (CP-IV), deemed unfit for selection.

Control Group A & B- Experienced colour normal Military Pilots and Airline Transport pilots less than 40 years of age. The age criterion of less than 40 years ensures a closer age match with the test population consisting of young individuals and minimises the effect of age related lenticular changes which can modify the spectrum of light reaching the fovea.

All participants underwent detailed ophthalmic examination including refraction, orthoptic evaluation, slit lamp and fundus examination. Individuals with best corrected binocular vision less than 6/6, media opacity in the pupillary area/visual axis, evidence of macular disease/optic nerve disease including glaucoma were excluded from the study.

Test Protocol

The test protocol for Ishihara and MLT was guided by the Manual for Medical Examination and Medical Boards IAP 4303 4thedn [9]. Participantswere tested with full spectacle corrections.

Ishihara Test - The 38 plate Pseudoisochromatic Chart (Kanehara& Co Japan) was used with the book held at 75 cm under daylight illumination allowing 4 s for each plate. When tested indoor, illumination used was fluorescent lamp with daylight filter.

Martin Lantern Test (MLT) -The original Martin Lantern manufactured by Kelvin Bottomley and Baird Ltd Glassgow was used for the tests. All subjects were tested at 6 metres. Those who made mistakes at 6 m were tested at 1.5 m also, but pass/ fail recorded as per existing standard for each group.

Farnsworth D15 - The Farnsworth D15 manufactured by Richmond Inc USA with one reference and 15 coloured discsenclosed in plastic transparent case, was used for testing under fluorescent lamp with day light filter. Two or more major crossings on the scoring sheet were taken as an abnormal result.

Heidelberg Multi Colour Anomaloscope (HMC)- The Heidelberg Multicolour Anomaloscope (HMC) Type 47700 manufactured by Oculus Optikgerate GmbH Germany was used for the tests (Fig-1). The equipment uses Rayleigh equation for red green matching as in the Nagels anomaloscope and Moreland for blue yellow. It has automatic neutral adaptation to minimise accommodation by presenting a white target intermittently. The default settings for anomaly quotient were: 0 to <0.7 protanomalous; 0.7 to 1.4 normal and >1.4 to infinity deuteranomalous. The equipment software has a rapid screening test, a manual test and specific test protocol. The test results and the diagnosis appear on the screen at the end. All participants underwent the screening and manual testing and specific test was done only in those found abnormal. After explaining the tests subjects were asked to get a clear focus by adjusting the eyepiece. The first test was rejected so as to familiarise the examinees to the tests. Then each eye was tested sequentially, dominant eye first. For the manual tests a minimum of five matches were obtained. Statistical analysis was done using SPSS version 8. P value for significance was taken as 0.05.

Figure 1:: The Heidelberg Multicolour anomaloscope and the test screen .

Results

There were a total of 225 participants: - 115 in group A (102 males and 12 females) and 110 in group B (97 males and 13 females). The average age was 20.8 years (range 17-25) & 24.2 years (range 18-40) in Group A and B respectively. Control groups had 50 people each with an average age of 32.5 & 30.5 in group A & B respectively .The Group wise complete results are shown in table-1. Average testing time on HMC per individual for colour normals in Group A and B combined was 99.5 s and for those with deficient colour vision deficiency was 140.75 s and the difference was significant(P=0.03).

Table 1:: Test results Group wise
Total Ishihara MLT D15 HMC
Normal defective Normal defective Normal defective Normal Defective
Test Gp A 115 108
(CPII)
07
CPIII-04
CPIV-03
89
(CPI)
26
CPIII-21
CPIV-05
112 03
Protan-01
Deutan-02
115 10
Protan-02
Deutan-08
Test Gp B 110 101 (CPII) 09
CPIII-03
CPIV-06
98 12
CPIII-10
CPIV-02
104 06
Protan 01
Deutan-05
98 12
Protan-01
Deutan-11
Total A+ B 225 209 16 187 38 216 09 203 22

The comparison of the four tests in passing or failing the candidates in each group is presented in table-2. Similar result for fail/pass is seen in 79.1% and 88% in group A & B respectively which increases to 94% and 95% when MLT is disregarded, indicating that the MLT is at maximum variance. The agreement between Ishihara and HMC alone was 97%.

Table 2:: Comparison Pass and Fail Rates
Total All in Agreement Differing results
Only
MLT
fail
Only
MLT
Pass
Only
D15
pass
D15 & Ishihara Pass Only
HMC
failed
Total
Gp A 115 91 (79.1%) 17 0 04 02 01 24
Gp B 110 97(88%) 03 04 0 0 06 13
A+ B 226 189 20 04 04 02 06 37

The comparative analysis of each test versus the HMC is shown in table-3.The McNemar test was used for statistical significance. The results show that, when the best results of each test is compared with HMC as in group A, only the Ishihara mirrors the HMC closely (P>0.05) while in group B where the passing standards on MLT and Ishihara are lower (CP-III) test results for all three were significantly different (P<0.05) from the HMC.

Table 3:: Comparative analysis each test Vs HMC
Group A HMC P value Group B HMC P Value
Variables Fail Pass Fail Pass
Ishihara Ishihara
Fail 7 (70.0) 0 (0.0) 0.250 Fail 6 (50.0) 0 (0.0) 0.031
Pass 3 (30.0) 105 (100.0) Pass 6 (50.0) 98 (100.0)
MLT MLT
Fail 9 (90.0) 17 (16.2) <0.00 Fail 2 (16.7) 3 (3.1) 0.029
Pass 1 (10.0) 88 (83.8) Pass 10 (83.3) 95 (96.9)
D15 D15
Fail 3 (30.0) 0 (0.0) 0.016 Fail 6 (50.0) 0 (0.0) 0.031
Pass 7 (70.0) 105 (100.0) Pass 6 (50.0) 98 (100.0)

The sensitivity, specificity and the predictive value of each test as compared with the anomaloscope taken as the gold standard are presented in table-4. Consistent results for Ishihara and D15 in both groups were seen while the MLT results were different in the two groups. In group A it had the highest sensitivity and lowest specificity but in group B it had the poorest sensitivity as well as specificity among the three .It passed some of those failed by Ishihara (CPIV) and D15. The overall accuracy was highest for the Ishihara and somewhat low for the MLT.

Table 4:: Sensitivity, Specificity and Predictive value
True True False False Sensi- Speci- PPV NPV Accuracy
+ ve - ve +ve -ve tivity ficity
Group A
Ishihara 07 105 0 03 70% 100% 100% 97% 97%
MLT 09 88 17 01 90% 84% 35% 99% 84%
D15 03 105 0 07 30% 100% 100% 94% 94%
Group B
Ishihara 06 98 0 06 50% 100% 100% 94% 95%
MLT 02 95 03 10 17 % 97% 40% 90% 88%
D15 06 98 0 06 50% 100% 100% 94% 95%

Table-5 shows the comparison of anomaly quotient (AQ).The AQ for the colour normals was taken as the average of all the values and for the anomalous the value farthest from the mean normal. Mann Whitney U test was used to compare medians. Anomaly quotient was significantly different only for the anomalous versus normal controls. For individuals found to be CP-III and CPIV by MLT and Ishihara there was no significant difference between the anomaly quotients of CPIII versus CP-IV.

Table 5: Comparative analysis of anomaly quotient
Test A Control A Test B Control B Anomalous(A&B)
All CPIII CPIV
Mean AQ 1.20 1.00 1.08 1.01 3.03 3.06 2.22
SD 0.92 0.10 0.49 0.09 2.79 0.57 1.37
Median 1.005 1.000 1.00 1.00 2.65 2.96 2.96
25,75 percentile 0.99.1.07 0.90,1.09 0.94,1.03 0.95,1.05 1.06,2.64 2.96,2.96 0.74,2.96
P Value 0.07
(Test Vs Control )
0.75
Test Vs Control
<0.001
Anomalous Vs Controls
0.408
CPIII Vs CPIV

Table-6 shows the results in terms of the CP classification currently followed in IAF. The following patterns could be discerned: All but one of the candidates (99.5%) declared MLT CPI (186/ 187) passed the HMC .The one who failed had only one match in each eye slightly outside range (0.4 Mild protanomaly). On the other hand 61.3% (19/31) of MLT-III and 14.3% (1/7) of MLT-IV also passed the anomaloscope. The corresponding anomaloscope pass rate for Ishihara CP-II, CPIII, and CP-IV is, 97 %, 0%, 0%. All MLT CP-I were normal on D15, 84% (26/31) of MLT-III were D15 normal and 16 % D15 fail, and 43 %(3/7) of MLT- IV passed D15 and 57% (4/7) failed D15. All Ishihara CP-II /CP-III passed the D15 test and all CP-IV failed the D15. The Farnsworth D15 passed candidates with CP-III or better on Ishihara (anomalous trichromats) and invariably failed those declared CPIV on Ishihara.

Table 6:: Clinical correlation of the Tests and the CP standards
Total No Ishihara MLT D15 HMC
CPII CPIII CPIV CPI CPIII CPIV N Anom N Anom
MLT CPI187 187 0 0 187 0 0 187 0 186 1(AQ=0.4-1)
Mild Protanomaly
Ishihara CPII209 209 0 0 187 21 0 209 0 203 06
03 Protanomaly
03 Deuteranomaly
Ishihara CPIII07 0 07 0 0 05 02 07 0 0 07Deuteranomaly
Ishihara CPIV09 0 0 09 0 05 04 0 09 0 09
06 Deuteranomaly
03 Deuteronopia
MLT CPIII31 20 05 06 0 31 0 26 05 19 12
10 Deuteranomaly
02 Deuternopia
MLT CPIV07 0 03 04 0 0 07 03 04 01 06
05 Deuteranomaly
01 Deuternopia

Test results for blue yellow perception are presented in table-7.The HMC did not give consistent results for blue yellow. Only 22.6% and 11.5% in Group A& B respectively had a perfect match. The D15 reported 2 out of the total 225 tested in both groups as tritanope who were also found abnormal on the anomaloscope.

Table 7: Blue Yellow HMC and D15
HMC D15
Normal Indeterminate Anomalous Normal Anomalous
Test Gp A 26(22.6%) 89 (60%) 0 115(100%) 0
Test Gp B 12(11.5%) 93(88%) 5(0.47%) 108((8.2%) 02(1.8%)
Total 38 187 5 223 02

Discussion

The Ishihara and the Martin lantern, the two tests currently recommended in India, have been compared with the Farnsworth D15 and the Heidelberg Anomaloscope (HMC), in terms of efficacy and level of agreement. HMC was taken as the gold standard. Majority of the earlier studies have been done on the Nagels anomaloscope [10, 11&13].

Unlike the Nagels, which is technically difficult to administer, the Heidelberg Multicolour Anomaloscope (HMC) is a microprocessor controlled computerised test based on same principle as Nagels, which is easy to understand.It generates results automatically, thus eliminating the need for a skilled technician. Our results show that the HMC enables comprehensive testing in a short time. The colour deficient participantstook a significantly longer time, but even in them, the average time was less than two and half minutes. Our study showed that there is no perfect agreement between the various tests in passing or failing the same individuals which is similar to the findings of Squire et al who compared three types of lanterns, the Ishihara and the anomaloscope [13]. However in our study, the correlation improved substantially when Martin Lantern Test (MLT) was disregarded. Comparison of each test with HMC reveals the best correlation with the Ishihara and the least with MLT which shows high sensitivity only in Group A, but at the cost of lower specificity. The sensitivity of Ishihara was rather low (70% in Group A and 50% in group B). Birch et al reported 98.7% sensitivity and 94.1% specificity for Ishihara compared to the Nagels anomaloscope using 471 colour normals and 401 colour deficient [14]. Ours was a smaller study with only 22 colour abnormal overall. In group B, the standards for civil aviation being lower allowedparticipants to pass with specified mistakes and hence a lower sensitivity was expected. Moreover our study design was different in the sense that we tested actual aspiring aviators whose colour perception status was not known prior to testing. Despite random presentation of plates, passing by memorisation by candidates who have practised the test in order to get selected, could not be ruled out completely. Nevertheless, it is safe to conclude that though the overall results suggest that Ishihara matched the HMC closely, it allowed some HMC abnormals to pass even when a strict criterion of no mistakes was used. The D15 had the poorest sensitivity amongst the three but it was 100% specific. The D15 consistently passed individuals classified CP-III by Ishihara and failed those who were CP-IV indicating that the D15 failed only individuals with severe defects. This was on expected lines as studies that have evaluated the D15 have shown that it effectively detected only the severely anomalous trichromats and the dichromats while mild to moderate anomalous trichromatswere likely to pass this test [15, 16].

The study results showed the overall accuracy and reliability of the MLT was questionable especially when testing was done at 1.5 m. The Martin Lantern is of 1939 vintage initially designed to simulate the lights of a ship in different weather conditions [17]. As an occupational screening test in aviation there are no studies validating the MLT. The lanterns which are being used more commonly and for which operational trials are available are the Farnsworth, the Holmes Wright and the Beynes lanterns [3]. Studies evaluating these lantern tests have noted many disparities and conflicting results [13, 18, 19, 20&21]. When MLT was eliminated we found that the level of correlation between the other three tests became high and much more acceptable. The study highlights the lacunae in using functional tests such as the MLT in its existing form, as the sole or final test for colour vision deficiency. The results of MLT should be read in conjunction with other diagnostic tests for colour vision such as the anomaloscope and the Farnsworth D15 and hence a review of the existing protocols for testing colour vision in the IAF is suggested, which should include these diagnostic tests.

For blue yellow colour vision the study showed widely varying results with the anomaloscope. Blue yellow colour perception can vary in normal population due to physiological variations in macular pigment densities and lens density, which selectively absorbs the lower wavelengths [22, 23]. A study designed specifically for blue yellow colour vision testing which takes into account these variations, may be able to give better results. An abnormal result on the Farnsworth D15 usually indicates a severe defect in blue yellow perception [24]. However, for formulating guidelines for blue yellow screening the results of present study were insufficient.

Conclusion

There is a need to supplement the existing colour vision tests for military pilots in India, with more objective, diagnostic tests such as the anomaloscope. Since this was the first study in the Indian Armed Forces on this kind of equipment, it needs to be replicated across all major medical evaluation centres and service hospitals as part of a multicenter trial.

References

  1. . UnderstandingColour,Normal and Defective colour vision. TrendCogn Sci. 2003;7:434-436.
    [Google Scholar]
  2. , . In Human colour vision (2nd ed). Washington, DC: Optical Society of America; .
    [Google Scholar]
  3. , , . A Review of NATO Aviation Colour Vision Testing Chapter Ftp.rta.nato.int/public/PubFulltext/RTO/ TR/RTO...///TR-16-06.
    [Google Scholar]
  4. . Selection of colour vision tests for army air forces: A summary of study at the army air forces school of aviation medicine. Arch Ophthal. 1946;36(3):263-283.
    [Google Scholar]
  5. , . Colour vision standards in aviation. ADF Health. 2005;6:62-66.
    [Google Scholar]
  6. . Comparative study of several diagnostic tests of colour vision used for measuring types and degrees of congenital red-green defects. ActaOphthalmolScand. 1972;115:1-64.
    [Google Scholar]
  7. Report of Working Group 41: Procedures for testing colour vision. In: Committee on Vision, Assembly of Behavioural and Social Sciences, National Research Council. Washington, DC: National Academy Press; .
    [Google Scholar]
  8. . A practical guide for colour-vision examination: report of the standardization committee of the International Research Group on colour-vision deficiencies. Ophthalmic Physiol Opt. 1985;5:265-85.
    [Google Scholar]
  9. Manual of Medical Examinations and Medical Boards IAP 4303. (4th ed). Air Headquarters New Delhi
    [Google Scholar]
  10. . Nagel’s Anomaloscope for Testing Colour Vision. Trans Am Ophthalmol Soc. 1915;14(Pt1):161-5.
    [Google Scholar]
  11. . Diagnosis of defective colour vision using the Nagel anomaloscope. DocumOphthalProc Series. 1982;33:231-5.
    [Google Scholar]
  12. . The Farnsworth Munsell 100 hue test for the examination of colour discrimination. Manual for MunsellColour.Macbeth Corp.
    [Google Scholar]
  13. , , , . Colour vision tests for aviation:comparison of the anomaloscope and three lantern types. Aviation Space Environ Med. 2005;76(5):421-9.
    [Google Scholar]
  14. . Efficiency of the Ishihara plates for identifying red/green colour deficiency. OphthalPhysiol Opt. 1997;17:403-8.
    [Google Scholar]
  15. . Pass rates for the Farnsworth D15 colour vision test. OphthalmicPhysiolOpt. 2008;28(3):259-64.
    [Google Scholar]
  16. , . Does the Farnsworth D15 predict the ability to name colours? ClinExpOptom. 2003;86:221-229.
    [Google Scholar]
  17. . A standardized lantern for testing colour vision. Br J Ophthalmol. 1939;23(1):1-20.
    [Google Scholar]
  18. , . Survey and evaluation of lantern tests of colour vision. AmJOptomphysiolOpt. 1982;59:346-74.
    [Google Scholar]
  19. , . Validation of the Holmes–Wright lanterns for testing colour vision. Ophthalmic Physiol Opt. 1983;3(2):137-52.
    [Google Scholar]
  20. , . Performance of Red-green colour deficient subjects on the Farnsworth Lantern (FALANT) Aviat Space Environ Med. 1999;70(1):62-7.
    [Google Scholar]
  21. , . Colour vision testing by Farnsworth lantern and ability to identify approach-path signal colors. Aviat Space Environ Med. 2008;79(6):585-90.
    [Google Scholar]
  22. , , , . Analysis of the macular pigment by HPLC: retinal distribution and age study. Invest Ophthalmol Vis Sci. 1988;29:843-55.
    [Google Scholar]
  23. . Short-wavelength-sensitive cone acuity: individual differences and clinical application. Appl Optics. 1989;28:1151-7.
    [Google Scholar]
  24. , . On acquired deficiency of colour vision with special reference to its detection and classification by means of the tests of Farnsworth. Vision Res. 1961;1:201-19.
    [Google Scholar]
Show Sections