SEARCH

SEARCH BY CITATION

Keywords:

  • amblyopia;
  • crowding;
  • HVOT chart;
  • KM chart;
  • Monoyer chart;
  • preschool vision testing;
  • visual acuity

Abstract.

  1. Top of page
  2. Abstract.
  3. Introduction
  4. KM chart design
  5. Material
  6. Methods
  7. Results
  8. Further investigation of the data
  9. Discussion
  10. Acknowledgements
  11. References

Purpose: To evaluate a new letter matching visual acuity (VA) chart (the KM chart) for children aged 5–7 years, designed as a Monoyer-based chart, in order to obtain a better consistency between school and preschool VA recordings.

Methods: Visual acuities were assessed using three methods: the HVOT, KM and Monoyer charts. Comparisons were made between the KM method versus the HVOT and Monoyer methods, respectively. Children with normal vision and with different degrees of amblyopia were investigated.

Results: Visual acuity levels appeared significantly higher in children with amblyopia when tested with the HVOT chart than when tested with the KM chart. Visual acuities obtained with the Monoyer and KM methods were comparable. The difference between the Monoyer and KM methods on the one hand, and the HVOT method on the other, can be explained by the fact that the HVOT chart elicits less crowding effect than the other two charts.

Conclusions: Visual acuity in children with amblyopia might be overestimated if the HVOT test alone is used to assess vision. Use of the HVOT chart, therefore, should be restricted to the 3.5−4.5 years age group, for whom the KM chart is somewhat too difficult. In our opinion the KM chart should be preferred for use with older preschool children because it shows good consistency with the Monoyer chart.


Introduction

  1. Top of page
  2. Abstract.
  3. Introduction
  4. KM chart design
  5. Material
  6. Methods
  7. Results
  8. Further investigation of the data
  9. Discussion
  10. Acknowledgements
  11. References

The diagnosis of functional amblyopia can easily be missed if visual acuity (VA) is tested with single optotypes only. Therefore, it is generally accepted that VA testing in preschool children should be performed with the optotypes in rows (line acuity measures) as soon as the child is mature enough to co-operate. In Sweden, the most commonly used method for testing line acuity in the preschool age group is the HVOT chart, one of the numerous modifications of a matching test originally described by Sheridan (1963). Later improvements of the original chart design included reduced interletter distances in order to improve the efficacy of detection of amblyopia by an increased crowding phenomenon (Hedin et al. 1980). Based on these charts, Hedin et al. (1980) developed the HVOT test, where the interletter separations are equal to the letter width.

The HVOT test has proved very popular because it is easy to apply and is well accepted for use with children from the age of 3.5 years. The accuracy of the test has been considered to be very high. However, especially among orthoptists, the opinion has arisen that VA measured with the HVOT chart is overestimated compared to that measured with the Monoyer chart, especially in children who have amblyopia or are undergoing treatment for amblyopia. The Monoyer chart is the test most commonly used in Sweden among school children and adults.

The differences in VA measurements obtained with the HVOT chart compared to those obtained with the Monoyer chart might be explained by the differences in letter shapes and interletter separations between the two charts. The letters used on the HVOT chart are of the Sloan 5 × 5 unit type, in contrast to the 4 × 5 unit type used on the Monoyer chart's letters. The letter separation on both charts is equal to the width of one letter. As letter widths on the HVOT chart are larger than on the Monoyer chart, the interletter separations are also wider. The smaller separation distances between the Monoyer letters implies greater separation difficulties or crowding phenomena and this effect is likely to be most pronounced during examination of the VA of children with amblyopia (Stuart & Burian 1962).

In order to develop a letter matching test that would elicit the same degree of crowding as the Monoyer chart, one of the authors (KM) designed a letter matching chart (designated the KM chart in this study) based on the Monoyer standard. The aim of this study was to make a clinical and statistical evaluation of the KM chart compared to the HVOT chart and the Monoyer test, respectively, with the focus on each chart's ability to detect amblyopia.

KM chart design

  1. Top of page
  2. Abstract.
  3. Introduction
  4. KM chart design
  5. Material
  6. Methods
  7. Results
  8. Further investigation of the data
  9. Discussion
  10. Acknowledgements
  11. References

An earlier evaluation carried out by Hedin & Olsson (1984) indicated that 10 letters have similar legibility. Seven of these were selected for the KM chart: C, D, E, F, K, N and V. The letter size used is 4 units wide and 5 units high, as in the Monoyer chart, and the interletter space is equal to one letter width. The letters are of the Monoyer type without serif. We chose a progression more similar to that of the HVOT chart than the Monoyer chart for easy comparison of results with HVOT chart results (Fig. 1). The progression is geometric, with a size increment varying between 1.25 and 1.5 to fit the visual acuities in a scale of decimal regression (1.0, 0.8, 0.65, 0.5, 0.4, 0.3 0.2, 0.15 and 0.1) (Fig. 1), except in the case of the 0.9 line. We used the 0.9 line when comparing the Monoyer chart and the KM chart. The Monoyer chart is arranged in rows with an arithmetical progression of VA recordings in tenths from 0.1, 0.2… to 1.0 and is designed for use at a testing distance of 5 metres. Twenty letters are used.

image

Figure 1. The HVOT and KM letter charts are used at a distance of 3 metres and the Monoyer chart at a distance of 5 metres. The HVOT is a letter matching chart; the KM chart can be used as either a matching or a naming chart. The Monoyer chart is used in schools and for adults.

Download figure to PowerPoint

The KM chart is designed for use at a testing distance of 3 metres, a distance that has been shown to give maximum testability for children aged 3–6 years (Lippmann 1969).

The chart is 26 × 60 cm in size and fits all standard illuminated light boxes in Sweden. It is made of white PVC that can be cleaned with a mild detergent to maintain high contrast between the black letters and the white background. The letters are arranged so that no words or acronyms can be spelled out.

Material

  1. Top of page
  2. Abstract.
  3. Introduction
  4. KM chart design
  5. Material
  6. Methods
  7. Results
  8. Further investigation of the data
  9. Discussion
  10. Acknowledgements
  11. References

Two groups of children with mild or moderate amblyopia were assessed. Group A (HVOT versus KM) included 162 children with a mean age of 6.2 ± 1.0 years. Group B (Monoyer versus KM) included 120 children with a mean age of 8.2 ± 1.2 years. Strabismic, anisometropic and ametropic types of amblyopia were represented. Some of the children were already undergoing treatment for amblyopia and all children with significant refractive errors wore full correction. At the time of the study the corrected VAs of the amblyopic eyes were recorded in the range of 0.9–0.3.

Methods

  1. Top of page
  2. Abstract.
  3. Introduction
  4. KM chart design
  5. Material
  6. Methods
  7. Results
  8. Further investigation of the data
  9. Discussion
  10. Acknowledgements
  11. References

Monocular VA was assessed in each child in group A with both the KM and HVOT charts and in each child in group B with both the KM and Monoyer charts. Although the Monoyer chart lacks some of the lines of the HVOT and KM charts (0.15 and 0.65), we chose to use this chart because it is the most popular chart for testing VA in children who have outgrown the HVOT chart.

The original HVOT chart (the LM chart), designed by Hedin et al. (1980), was used (Fig. 1).

Monocular VA was tested, with the subjects wearing glasses if these had been previously prescribed, at 3 metres for group A (HVOT versus KM), and at 3 metres and 5 metres for group B (KM versus Monoyer). A standard illuminated box was used. The assessments were carried out in well lit rooms. The order in which the right and left eyes were tested and the choice of charts were randomized.

The VA recorded was the smallest line at which five of five or six of seven letters were identified (HVOT and KM charts) and 70% of the letters of the current line were correctly identified (Monoyer chart). This corresponds to common clinical practice. The assessments were carried out by orthoptists and four experienced school nurses.

Results

  1. Top of page
  2. Abstract.
  3. Introduction
  4. KM chart design
  5. Material
  6. Methods
  7. Results
  8. Further investigation of the data
  9. Discussion
  10. Acknowledgements
  11. References

An initial investigation of the data sets shows that there was only one positive difference, 222 negative differences and 101 ties when subtracting the VA values for each eye obtained with the HVOT chart from the values obtained with the KM chart. This means that in only one case did the HVOT chart give a VA value lower than the corresponding value measured with the KM chart. In 222 cases, the VA value measured with the HVOT chart was higher than the corresponding value measured with the KM chart. In 101 cases, the two charts gave the same VA value. In an equivalent comparison of the results obtained with the Monoyer chart with those obtained with the KM chart, we found 42 negative differences, 27 positive differences and 171 ties. This seems to indicate what we hoped to prove, that the VA values given by the KM and Monoyer charts are more similar than those provided by the KM and HVOT charts.

Figures 2 and 3 show the distribution of the differences of VA measured with the KM chart and the HVOT chart, and VA measured with the Monoyer chart and the KM chart. In Fig. 3, the differences are distributed around zero; in Fig. 2, the differences are skewed to the negative side.

image

Figure 2. Bar-plot of the differences between the KM and HVOT charts.

Download figure to PowerPoint

image

Figure 3. Bar-plot of the differences between the Monoyer and KM charts.

Download figure to PowerPoint

In Figs 2 and 3 we did not separate the left eye from the right eye, but as we cannot exclude the possibility that there is some dependence between the right and left eyes, we chose to separate the eyes in our analysis. Therefore, we obtained four different data sets, comprising the results obtained with the Monoyer chart compared with those obtained with the KM chart, the results obtained with the HVOT chart compared with those obtained with the KM chart, and both comparisons made for the right and left eyes. As we have many observations, a paired t-test was chosen in order to investigate if there were any systematic differences between the VA values obtained with the different charts.

As seen in Tables 1 and 2, the confidence intervals for the systematic differences between the results obtained with the HVOT chart and the Monoyer chart are − 0.005 and 0.017 for right eyes and 0.001 and 0.023 for left eyes.

Table 1.  A paired sample t-test of the differences between the Monoyer and KM charts.
  Paired differencestd.f.Significance (2-tailed)
  MeanSDSEM95% CI of the difference   
     LowerUpper   
  1. SD = standard deviation; SEM = standard error of the mean; 95% CI = 95% confidence interval.

Pair 1Monoyer right eye–KM right eye0.0060.06030.0055−0.0050.0171.1351190.259
Pair 2Monoyer left eye–KM left eye0.0120.06140.00560.0010.0232.0811190.040
Table 2.  A paired sample t-test of the difference between the HVOT and KM charts.
  Paired differencestd.f.Significance (2-tailed)
  MeanSDSEM95% CI of the difference   
     LowerUpper   
  1. SD = standard deviation; SEM = standard error of the mean; 95% CI = 95% confidence interval.

Pair 1HVOT right eye–KM right eye0.1470.11930.00940.1280.16515.6441610.000
Pair 2HVOT left eye–KM left eye0.1480.11360.00890.1310.16616.6361610.000

The corresponding confidence intervals for the differences between the results obtained with the HVOT and KM charts are 0.128 and 0.165 for right eyes and 0.131 and 0.166 for left eyes. The difference between the results obtained with the HVOT and KM charts is highly significant. The Monoyer−KM comparisons generate p-values of 0.04 for the left eye and 0.26 for the right eye. The first p-value might be called significant, but it is much larger than the p-values in Table 2. This further underlines what was visible in Figs 2 and 3, that the Monoyer and KM charts generate results that are more similar than those generated by the HVOT and KM charts. In fact, we can say with a 95% probability that the average difference between VA measurements obtained with the HVOT and KM charts is between 0.13 and 0.17.

The information from the confidence intervals for the Monoyer and KM charts is less clear. It is possible that the KM test slightly overestimates VA compared with the Monoyer chart, but this overestimation seems to be very small compared to the overestimation accounted for by the HVOT chart.

Further investigation of the data

  1. Top of page
  2. Abstract.
  3. Introduction
  4. KM chart design
  5. Material
  6. Methods
  7. Results
  8. Further investigation of the data
  9. Discussion
  10. Acknowledgements
  11. References

Does the difference between the KM chart and the HVOT chart systematically change with the value of VA?

A value of the average VA was calculated. The data sets were divided into quartiles. The quartile with the lowest values of average VA was compared with the quartile with the highest values of average VA. The difference between the KM chart and the HVOT chart is visible in all four quartiles, although we found that the difference between the charts increases when the average VA decreases. This might be explained by the suggestion that the HVOT chart either gives a less accurate result with a lower VA, or that the largest possible value for VA is 1. As a value higher than 1 cannot be obtained, it is natural that the VA values converge as VA measured with either chart approaches 1 ( Tables 3–5).

Table 3.  Median and percentile values for average VA obtained with the HVOT and KM charts.
nValid162
 Missing0
Median 0.8500
Percentiles250.7250
 500.8500
 750.9000
Table 4.  The result of the paired t-test for the quartile with the lowest average VA.
  Paired differences td.f.Significance (2-tailed)
  MeanSDSEM95% CI of the difference   
     LowerUpper   
  1. SD = standard deviation; SEM = standard error of the mean; 95% CI = 95% confidence interval.

Pair 1HVOT right eye–KM right eye0.1680.11130.01760.1320.2039.522390.000
Pair 2HVOT left eye–KM left eye0.1610.11180.01770.1250.1979.122390.000
Table 5.  The result of the paired t-test for the quartile with the highest average VA.
  Paired differences td.f.Significance (2-tailed)
  MeanSDSEM95% CI of the difference   
     LowerUpper   
  1. SD = standard deviation; SEM = standard error of the mean; 95% CI = 95% confidence interval.

Pair 1HVOT right eye–KM right eye0.0690.10870.01680.0350.1034.115410.000
Pair 2HVOT left eye–KM left eye0.0560.09770.01510.0260.0863.713410.001

Here, functional amblyopia is defined as a difference in VA of two lines or more between the two eyes. Comparisons between the different charts' tendencies to diagnose functional amblyopia are shown in Tables 6 and 7. A McNemar test was performed. For the KM−Monoyer comparison, the p-value was 0.45; for the HVOT− KM comparison, the p-value was 0.04. These results show that the KM and Monoyer charts give more similar results when diagnosing functional amblyopia than do the KM and HVOT charts.

Table 6.  Functional amblyopia diagnosed with the Monoyer and KM charts, respectively.
  MonoyerTotal
  Not functional amblyopiaFunctional amblyopia 
KMNot functional amblyopia1025107
 Functional amblyopia21113
 Total10416120
Table 7.  Functional amblyopia diagnosed with the HVOT and KM charts, respectively.
  MonoyerTotal
  Not functional amblyopiaFunctional amblyopia 
HVOTNot functional amblyopia13015145
 Functional amblyopia51217
 Total13527162

Discussion

  1. Top of page
  2. Abstract.
  3. Introduction
  4. KM chart design
  5. Material
  6. Methods
  7. Results
  8. Further investigation of the data
  9. Discussion
  10. Acknowledgements
  11. References

For many years the HVOT letter matching chart has been accepted as one of the most accurate methods for evaluating VA in preschool children from the age of 3.5 years. Contrary to the previously used Sheridan Gardiner letters in rows, the HVOT chart was designed to elicit a crowding effect, with smaller interletter separations of one letter width only. Nevertheless, it has been presumed that the HVOT method overestimates VA in some children when compared to the Monoyer method, once the children are mature enough to have their VA tested with the Monoyer chart.

This study clearly confirms the clinical impression that the VAs for children with amblyopia obtained with the HVOT method are somewhat higher than those obtained with the new KM method. The same testing principles are used in both methods, except for the letter shape and the separation distance between the letters: the HVOT chart has a letter size of 5 × 5 units with a interletter separation of 5 units, whereas the corresponding sizes in both the KM and Monoyer charts are 4 × 5 units and 4 units, respectively. This 20% reduction in interletter separation space implies that separation difficulties or crowding should increase, a fact that is the most probable explanation for the lower acuity levels found with the KM method.

Crowding phenomena can be elicited among subjects with normal visual functions but is strongly exaggerated in amblyopia (Stuart & Burian 1962). The KM test was found to be about twice as effective as the HVOT test at detecting mild and moderate amblyopia when amblyopia was defined as a difference in VA of two lines or more between the eyes.

The results obtained with Monoyer chart showed good consistency with those obtained with the KM chart. This indicates that KM chart is a good substitute for the Monoyer chart. As the letter separations are the same on these two charts, this finding concurs with the opinion that constant separation distances are important when comparing between different set-ups for VA testing.

The viewing distance chosen for the KM chart is 3 metres in order to make it easier to maintain the interest of children. In terms of accommodation power, the difference of only 0.13 dioptres between the 3-metre distance used for the KM chart and the 5-metre distance used for the Monoyer chart has no practical implications for the assessment of VA in children in this age group.

In our series the KM chart was mainly used by naming the different letters, but for some children it was also used as a matching test. It turned out to be a great advantage that the same test can be used in the two ways, not least for psychological reasons when testing younger schoolchildren, who have varying levels of skill at identifying letters by name. In practice, the KM chart turned out to be as practical and attractive for use with children as the HVOT chart. However, the KM chart, with its greater number of letters, seemed to be somewhat more demanding than the HVOT chart was in our previous experience with younger preschool children. Therefore, we have arrived at the conclusion that the HVOT method should be maintained for the assessment of VA in the 3.5–4.5 years age group, but should be replaced by the KM chart for older children. The KM chart was found to be an accurate and practical method of testing VA when screening a mixed population of illiterate and literate children in the range of approximately 5–8 years of age.

Acknowledgements

  1. Top of page
  2. Abstract.
  3. Introduction
  4. KM chart design
  5. Material
  6. Methods
  7. Results
  8. Further investigation of the data
  9. Discussion
  10. Acknowledgements
  11. References

This work was supported by the Sigvard and Marianne Bernadotte Research Foundation for Children's Eye Care awarded to Konstantin Moutakis. Konstantin Moutakis has an economic interest in the KM chart.

References

  1. Top of page
  2. Abstract.
  3. Introduction
  4. KM chart design
  5. Material
  6. Methods
  7. Results
  8. Further investigation of the data
  9. Discussion
  10. Acknowledgements
  11. References