Reliability of the Snellen chartBMJ 1995; 310 doi: https://doi.org/10.1136/bmj.310.6993.1481 (Published 10 June 1995) Cite this as: BMJ 1995;310:1481
- Paul McGraw,
- Barry Winn,
- David Whitaker
- Research fellow Department of Vision Sciences, Glasgow Caledonian University, Glasgow G4 0BA
- Professor Senior lecturer Department of Optometry, University of Bradford, Bradford BD7 1DP
Better charts are now available
Historically, visual function has been assessed by determining the finest spatial detail that the visual system can discriminate. A letter acuity chart, such as the Snellen chart, is commonly used. This type of test is simple to perform and is sensitive to the most common sources of visual impairment, such as uncorrected refractive error, cataract, macular disease, and amblyopia. A recent article in the BMJ identified some of the factors reducing the Snellen chart's reliability, such as failure to test visual acuity at the right distance and under recommended levels of illumination.1 But other determinants inherent in the design of the Snellen chart also warrant consideration.
During the measurement of visual acuity only the angular subtense of the letters should change as the subject reads down the chart, which is not the case with the Snellen chart. Variation in the number of letters on each line presents the subject with a task of increasing difficulty rather than providing an equivalent task at all acuity levels. Typically, one letter is presented at the 6/60 level and up to eight letters are presented at the higher levels of acuity. This variation in the number of letters per line creates additional problems. It is now firmly established that the legibility of a letter is impaired if contours (such as other letters) are placed in close proximity.2 This phenomenon has been termed “contour interaction” or “visual crowding,” and many studies have shown that performance is better when letters are presented on their own. Careful consideration should therefore be given to spacing between both letters and rows to control contour interaction at each level of acuity.
Unfortunately, the effects of contour interaction vary throughout the Snellen chart. For example, “uncrowded” acuity is measured at the low end of the acuity scale (6/60) and “crowded” acuity at the higher end of the scale (6/6). The two are clearly not comparable. In addition, the legibility of test letters used in the Snellen chart varies,3 so nominally incremental steps on the chart are not equally capable of being discriminated. This is a particular problem at low levels of acuity, where only one or two letters are presented.
Perhaps the most important problems with the design of the Snellen chart are the irregular progression of the size of the letters on the chart and the lack of an accurate or standardised scoring system. The variation in the ratio of the sizes of the letters between successive lines is somewhat arbitrary. The relatively large gaps between acuity levels at the lower end of the acuity scale on a Snellen chart (6/60-6/24) can result in gross overestimation and underestimation of visual acuity. This makes the assessment of change in visual acuity difficult. An irregular progression in the sizes of the letters also introduces changes in the scaling factor of the chart at reduced test distances that can alter the acuity score. The tradition of scoring Snellen acuity as the smallest line at which a majority of letters is correctly identified has recently been shown to restrict the sensitivity of the test to detecting changes over time.4
These problems5 6 7 have led to the development of alternatives to the Snellen chart.8 9 The most notable innovations are the use of a geometric progression in the size of the letters (that is, the change between lines occurs in uniform steps) and the introduction of an equivalent task for all acuity levels, ensuring that the only variable is the change in the angular size of the letter. This is achieved by using a set of letters that are equally legible and by presenting the same number of letters on each line of the chart.
These charts have been shown to provide accurate and reliable measures of visual acuity10 and have become the gold standard for measuring visual acuity in research.11 Furthermore, their design allows the use of interpolated scoring systems that significantly improve doctors' ability to detect changes in acuity.4 In such systems equal weighting is given to every letter on the chart, and the score for each letter is incorporated in the overall acuity score. An important advantage of this type of chart is that it allows low levels of acuity to be measured with the same precision as higher levels of acuity—a feature that the Snellen chart lacks. Indeed, a reliable measure of visual acuity may be more important in patients with reduced vision as they are likely to require treatment.
Gibson and Sanderson reported that the repeatability of measurements of visual acuity made with a Snellen chart was extremely poor, with up to 13% of subjects displaying discrepancies of two lines or more on repeated testing.6 Taking these results into consideration, doctors must be extremely cautious when assigning clinical importance to changes in acuity of two lines or less as these differences may simply reflect the inherent variability of the Snellen chart.