Statistics from Altmetric.com
Clinicians must continue to refine as well as develop new visual acuity tests for preschool children
A ccurate measurement of visual acuity is fundamental to ophthalmological care for all ages. This measure is used for diagnosis, for documentation of clinical outcomes, for qualification for special education or disability programmes, and for permission to drive a motor vehicle. More than 20 years ago the importance of ensuring that such a measure be scientifically rigorous was recognised with the development of the Bailey-Lovie visual acuity test1 and the Early Treatment of Diabetic Retinopathy Study (ETDRS) test derived from it.2 The ETDRS test incorporated important design elements including letter selection, proportional optotype spacing, standard illumination, methods of administration, and scoring. Despite the quality of this test for school age children and adult patients, it is too difficult for preschool children.
Clinicians testing visual acuity of preschool children have not yet found a single best test. An accurate, reproducible, and highly testable measure is needed by the paediatric eye care community for screening and office testing. Practitioners have utilised an assortment of test chart configurations with letter, number, and picture optotypes. Letter optotypes have generally been too difficult for preschool children, while traditional picture optotypes, like the Allen pictures, though much more testable, substantially reduce the sensitivity of the test for the detection of amblyopia.3,4 This is a critical drawback since amblyopia is the most common reason to test visual acuity in this age group. In addition, the practitioner has individualised most testing with plenty of room for the art of administration, including the judicious use of second chances.
An ideal test of visual acuity in early childhood should include as many of the design features of the ETDRS as possible. Using the same test optotypes and presentation strategy throughout childhood would eliminate the measurement variability that is introduced by using a different test at each stage of therapy as the child ages. The ideal test should be constructed in a regular logMAR progression to facilitate detection of change at all levels of acuity. The test should utilise single optotypes, which increase testability in children.5 However, single optotypes are not as sensitive to amblyopia as linear optotypes, owing to the absence of contour interaction with the single optotypes. Crowding bars or even a surround box should be placed 0.5 letter width away from the test optotype to improve the test sensitivity.6,7 The test should incorporate letter optotypes, which are more sensitive to amblyopia than the pictures often used (for example, Allen), unless a picture test that is as sensitive to amblyopia can be devised.3 The single surrounded HOTV is a good example.8 The surround bars combined with single letters produce a result that closely approximates line letter acuity.7 The Glasgow acuity cards are a test which uses letter optotypes (X V O H U Y) presented as groupings of four in a surround box ensuring contour interaction.9–,11 Random presentation of optotypes is preferred to avoid memorisation. Commercially available computerised display systems would seem to be ideally suited for this task compared to printed charts.
An accurate, reproducible, and highly testable measure is needed by the paediatric eye care community for screening
A specific and reproducible testing protocol should be part of the test as it is with the ETDRS. The protocol will probably use a modified staircase, rather than asking the preschool child to identify every optotype in the test, to reach threshold more quickly and improve overall testability.10 One such strategy was developed and implemented for the Amblyopia Treatment Study (ATS).8 This measurement protocol used single surrounded HOTV optotypes. It had high testability between age 4 and 7 years (92%), though it could be performed by only about 50% of children between 2 and 4 years of age. It is this reduced testability of even the simplest crowded letter optotypes in young children that has kept clinicians searching for a picture based optotype test for use in younger preschool children.12
In this issue of the BJO (513) Becker and his colleagues extend their work in preschool visual acuity testing with the Lea symbols. Dr Lea Hyvärinen designed a set of tests based on picture optotypes for use in children.13 They were designed to have contours like the Landolt C. These tests make use of common pictures felt to improve testability among young children and eliminate cultural biases.13–,15 They have become very popular among clinicians. It is hoped that these optotypes will exhibit sensitivity for amblyopia equivalent to letter optotypes.
Becker and colleagues report on the results of two projects. In the first, the single Lea symbols distance test was used in a paediatric practice on unselected children from 21 to 93 months of age (median 47 months). Overall, the testability was a disappointing 54%. The testability after age 4 years was quite good (93%), but these older children are in the age range when letter based tests can be successfully administered.8,10,11 More troublesome was the poor testability between 21 and 48 months, when only 31% of children completed the testing. This is just the age when a sensitive picture based optotype test would be revolutionary. Though the performance of the Lea tests in this setting was disappointing, other authors have found better testability with Lea symbols, both in isolated and line versions, in children between 3 and 4 years of age.14,16
The second study in this report compared the single Lea symbols to the crowded Lea symbols as well as to single and crowded Landolt C tests. The crowded Landolt C test was designed with fixed rather than proportional interoptotype spacing, which may impact the sensitivity of the test. The study was performed in an ophthalmology clinic setting and included normal children from 23 to 70 months of age. Single Lea picture testability was better in this clinic setting, 77%, though it was only 56% before 48 months of age. The crowded Lea symbols were more difficult to successfully complete, with only 57% of studied children able to complete the test. As might be predicted both versions of the Lea symbols test had higher testability than the Landolt C tests. However, the surrounded Lea tests overestimated acuity by 1.9 lines compared to the crowded Landolt C in normal eyes.
Reproducibility was not directly assessed. However, it is possible to infer such a rate from the intereye difference of this visually normal population by assuming the acuities of each eye are equal. The measurements for the two eyes were within one logMAR line for the single Lea symbols, either better or worse, in only 84% of the children screened in paediatric offices. This means that when a one line change is observed, 16% of the time it does not represent actual change. This imprecision was noted for all ages tested. Better test reproducibility was noted during ophthalmological office testing, with about 90% of patients within one line on retest for both the single and surrounded Lea symbols. Similar rates of 95% and 93% have been reported for the Glasgow acuity cards10 and the single surrounded HOTV (ATS protocol) respectively.8
These studies did not address the crucial issue of the sensitivity of the Lea symbols in single or surrounded format for the detection of amblyopia. The authors have previously reported on the sensitivity of Lea symbols in older strabismic amblyopes.17 In that study the Lea test with crowded symbols overestimated visual acuity by about two logMAR lines compared to a crowded Landolt C.
It would seem logical for future testing that the surrounded versions of the Lea test should be preferred based on the recognised impact of contour interaction on amblyopia sensitivity with other tests. Additional studies comparing the sensitivity of the Lea tests to line and surrounded letter acuity in amblyopic patients, such as has been done for the Glasgow test,10 would be helpful.
How should we currently be testing preschool visual acuity? It would seem that the best visual acuity measurement strategy for children over age 4 years would be to use a single surrounded letter optotype in a logMAR progression18 or the Glasgow test.8–,10 For the developmentally delayed and children unable to perform these letter tests, the use of the surrounded Lea symbols test is a reasonable alternative.18 Another option would be the “O” test, which had a testability of 90% in a group of 62 mildly delayed 4–10 year old children.19,20 Practitioners should have more than one test available, as not one of these tests can be used successfully in all young children. However, many young children will not be testable with any of the currently available tests.
Picture based test optotypes are clearly necessary to quantitatively measure the visual acuity of young children. Clinicians must continue to refine as well as develop new visual acuity tests for preschool children, which combine picture optotypes, which have high testability, with crowding features ensuring sufficient sensitivity to detect amblyopia.