The measurement of visual acuity in children: an evidence-based update



Over the past decade, a number of large clinical trials have provided important information relating to the reliability and repeatability of commonly used paediatric tests of vision and their role in the diagnosis and management of paediatric ocular diseases. The aim of this review is to summarise recent findings on the use of paediatric visual acuity tests in clinical practice and to discuss the validity and accuracy of visual acuity measurements in infants and young children. We provide a broad overview of the benefits and challenges of measuring visual acuity in children and then discuss age-appropriate tests for measuring visual acuity in infants through to school-age children. We also discuss normative values for visual acuity in each age group and, where possible, provide comparisons of results between tests with a particular focus on the importance of optotype design.

Visual acuity provides an estimate of an observer's ability to perceive spatial detail and is the most commonly used measure of visual function in clinical practice. Tests of visual acuity provide information that can be used to determine the presence or absence of refractive error and pathology within the visual pathway and are often considered to be among the most important measures of general visual function. Visual acuity is correlated with quality-of-life measures, such as mobility and the ability of adult patients to live independently.[1, 2] In addition, reduced visual acuity in adults has been associated with a higher risk of falls[3] and higher rates of motor vehicle accidents among older drivers.[4] Changes in visual acuity also aid the detection of many ocular disorders in adults,[5] although visual acuity measurements appear to be poor at distinguishing glaucoma,[6] diabetic retinopathy[7] and posterior capsular opacification after cataract surgery.[8] While visual acuity measurements with standard letter optotypes are made routinely in adult patients and show good reliability and repeatability,[9, 10] the measurement of visual acuity in children can be challenging and results must be interpreted in the context of the specific test used to generate the measurement.

Visual acuity can be measured in a number of ways. Tests may require:

  1. a judgement of whether a target is present or absent (detection acuity)
  2. evidence that the spatial detail contained within a target has been fully resolved (resolution acuity) or
  3. the identification of a target (recognition acuity).

Under normal viewing conditions, detection and resolution acuity at the fovea are limited by the quality of the retinal image;[11] however, the factors limiting recognition acuity are harder to define as not all optotypes are equally recognisable and this measure is more reliant on the observer's cognitive ability and communication skills. This is a particularly important issue when testing children. In the clinical literature, tests that require the discrimination of a grating or optotype from a background with the same mean luminance, such as the Teller acuity cards, are typically referred to as measures of resolution acuity and this terminology will be used for the remainder of the review. We also report visual acuity as the logarithm of the minimum angle of resolution (logMAR) throughout the review. Conversions from logMAR to other acuity metrics are provided in Table 1.

Table 1. Conversion of logarithmic minimum angle of resolution (logMAR) visual acuity notation to the Snellen equivalent fraction in both metric and imperial measurements. For charts with five optotypes per line (Bailey–Lovie, Early Treatment of Diabetic Retinopathy Study chart, Lea symbols, Patti Pics and HOTV) each additional optotype read correctly improves acuity by 0.02 logMAR. For charts with four optotypes per line (crowded Kay pictures and crowded Keeler logMAR) each additional optotype read improves acuity by 0.025 logMAR
Snellen equivalent (metric)6/606/486/386/306/246/196/156/126/9.56/7.56/66/4.86/3.8
Snellen equivalent (imperial)20/20020/16020/12520/10020/8020/6320/5020/4020/3220/2520/2020/1620/12.5

Since the mid-1970s, several standardised charts have been developed for measuring recognition acuity in adults which have greatly improved the accuracy of acuity measurements in both clinical and research settings. The most widely used of these charts, the Bailey–Lovie chart[12] and the Early Treatment of Diabetic Retinopathy Study (ETDRS) chart,[13] have letters of presumed equal legibility, the same number of letters (five) on each row and uniform spacing between each letter and each row of letters. The logarithmic progression of letter sizes used by these charts also means that the task remains constant over the range of visual acuity tested and across different viewing distances. Charts that incorporate these principles are used for almost all clinical trials incorporating visual acuity as an outcome measure and are increasingly being used in clinical practice as well (Table 2). These charts have provided useful information regarding the average recognition acuity values of typical adults without visual impairment (-0.12 ± 0.07 in young adults[14]), the test-retest variability (TVR) of adult observers (approximately 0.1 logMAR or one line of letters)[15, 16] and the minimum difference required to detect a change in recognition acuity (approximately 0.15 logMAR).[15, 17]

Table 2. Evaluation of paediatric visual acuity charts (Allen figures, Kay pictures, Lea symbols, Patti Pics, HOTV, Keeler logMAR, Sheridan Gardiner [SG], Snellen and Early Treatment of Diabetic Retinopathy Study [ETDRS] charts) using the International Visual Acuity Chart Guidelines (adapted from the International Council of Ophthalmology [1984] guidelines). Only Lea symbols, Patti Pics and ETDRS charts meet all seven criteria
International Visual Acuity Chart GuidelinesAllenKayLeaPattiHOTVKeeler logMARSGSnellenETDRS
1. Optotypes should be black on white background image image image image image image image image image
2. Crowding elements should be incorporated into the test image image image image image image image image image
3. Optotypes (pictures and letters) used should be of approx. equal legibility   image image      image
4. Horizontal distance between adjacent optotypes should not be less than one optotype  image image image image image    image
5. Vertical distance between optotypes should not be less than height of the larger of the two lines of optotypes  image image image      image
6. At least five optotypes should be displayed on each line   image image image     image
7. Optotype sizes should have a geometrical progression (constant ratio) of step sizes of 0.1 log units per line  image image image image image    image

In children, measurement of visual acuity is used primarily to screen for the presence of amblyopia and/or significant refractive error and other amblyogenic factors. Amblyopia is a prevalent (two to three per cent)[18-23] neurodevelopmental disorder of the visual system that requires early detection and treatment to prevent an enduring impairment of visual function in the affected eye and a loss of binocular vision.[24-26] Measurement of recognition acuity and to a lesser extent resolution acuity, are the primary diagnostic tests for amblyopia and provide the foundation for therapeutic decision-making.[27, 28] Accordingly, measurement of visual acuity is included in almost all public health vision screening programs, most of which rely on a fixed visual acuity threshold as the criterion for triggering a referral for further assessment.

The use of visual acuity measurements to determine the integrity of the visual system in children is complicated by a number of factors. These include variation in the ability of children to understand the requirements of the test (that is, cognitive function), a lack of normative data for many specific age groups and the variety of test designs that are available for use in different circumstances. There is also evidence to suggest that visual acuity does not reach adult levels until after children begin school due to structural and functional changes in the visual system.[29, 30] At birth the macular is still immature,[31-33] with the foveal region only reaching adult morphology at between 15 and 45 months of age.[34, 35] Cortical processing of visual information also continues to develop throughout childhood. For example, the effect of contour interaction (letter crowding) changes with age, with younger children being more susceptible to the effects of crowding.[36] As a result of these factors, there is considerable debate over the value of using only visual acuity measurements in preschool vision screening programs.[37, 38] The large multi-centre Vision in Preschoolers (VIP) study found that measurement of unaided vision was just as effective at detecting significant visual impairment (amblyopia, strabismus and high refractive errors) as direct measurement of ocular refraction with non-cycloplegic retinoscopy and autorefraction.[39, 40] The VIP study group[39, 41] and others[42] have shown that measurement of acuity, using a cut-off point of 0.2 logMAR, is not an accurate method of detecting low to moderate hyperopia and astigmatic refractive error in children, and only performs adequately in detecting myopic children, who are at low risk for amblyopia.

In the remainder of this review, we discuss the techniques available for assessment of visual acuity in preschool children and the considerations that need to be taken into account when interpreting the results of these tests. For very young children (up to 24 months) and those with developmental delay, objective measurements, including electrodiagnostic tests, optokinetic nystagmus and preferential looking tasks, can be employed to measure detection and resolution acuities. Some studies[43] have shown that children as young as two can perform recognition acuity tasks with picture naming or matching; however, recognition acuity testing is generally considered to be more achievable after the age of three.[44, 45] The following sections detail the methodology, expected results and other issues surrounding age-appropriate visual acuity tests for children, beginning with assessment of infants, then toddlers and finally preschool and school-aged children. Our aim is to provide an evidenced-based approach to assessing visual acuity in children with a particular emphasis on testing within an optometric setting. The extent to which different tests meet the international acuity chart guidelines is outlined in Table 2 and normative data[46-54] for a variety for paediatric tests are provided in Table 3.

Table 3. Visual acuity norms for paediatric populations by test type. Mean (± standard deviation where data were available) values for visual acuity are shown in the units published, as well as logMAR and Snellen equivalents, where appropriate. For crowded charts, * denotes the use of crowding bars with a single optotype and denotes the use of crowding bars with a line of optotypes. Note that the spacing of crowding bars/lines of optotypes is not standardised across charts (see Table 2)
TestAuthorsAge (years)Mean visual acuitylogMAR equivalentApproximate Snellen fraction
Teller acuity cardsMayer and colleagues46(1 month)1 cycle per degree1.56/180
  (6 months)6 cycles per degree0.706/30
  425 cycles per degree0.076/7
 Hargadon and colleagues475–624.5 cycles per degree0.106/7.5
Single Lea symbolBecker and colleagues482–60.88 Snellen decimal0.076/7
Crowded Lea symbolsBecker and colleagues482–60.74 Snellen decimal0.126/8
 Chen and colleagues 494.5–8.5 0.08 ± 0.096/7.5
Crowded Patti PicsMercer and colleagues503–5 0.266/10
Single Kay picturesNorgett and colleagues364–6 -0.15 ± 0.116/4
Crowded Kay picturesJones and colleagues512.5–16 -0.046/6
 Norgett and colleagues364–6 -0.10 ± 0.096/4.8
Sheridan Gardiner uncrowdedNorgett and colleagues364–6 -0.18 ± 0.086/4
Sheridan Gardiner crowded*Simmers and colleagues275–61.13 ± 0.09 modified logMAR-0.08 ± 0.096/5
HOTV crowded*Drover and colleagues523–4 0.086/7.5
 Drover and colleagues525 0.036/6
 Drover and colleagues526 -0.036/6
 Pan and colleagues533–4 0.17 ± 0.136/9
  4–6 0.08 ± 0.116/7.5
 Birch and colleagues545.5–12 -0.10 to 0.306/4.8 to 6/12
Crowded Keeler logMARJones and colleagues512.5–16 0.046/6
 Simmer and colleagues275–60.9 ± 0.08 modified logMAR0.1 ± 0.086/7.5
 Norgett and colleagues364–6 0.00 ± 0.086/6
Early Treatment of Diabetic Retinopathy StudyBirch and colleagues545.5–12 -0.10 to 0.406/4.8 to 6/15

Objective Measurement of Visual Acuity in Infants

Several studies have used visual evoked potentials (VEPs) to assess detection/resolution acuity in infants, as this technique does not require behavioural responses.[55] The typical approach is to present the infant with a pattern reversal stimulus that evokes a steady state VEP. The spatial frequency of this stimulus is then varied and the mean amplitude of the VEP is estimated for each spatial frequency. Because the VEP amplitude decreases as the stimulus becomes less visible, regression techniques can be applied to the measured VEP amplitudes to estimate the acuity threshold.[55] Initial studies measured VEP amplitude and latency individually for each spatial frequency. The subsequent introduction of the ‘sweep’ VEP technique, during which multiple spatial frequencies are presented in rapid succession, allowed for visual acuity estimates to be acquired rapidly.[56-60] While VEPs can provide useful measures of visual acuity in young pre-verbal infants, this is primarily a research technique that is not often employed in clinical or screening settings.

Assessment of the presence or absence of optokinetic nystagmus (OKN) in response to a patterned rotating drum or a drifting stimulus presented on a computer screen can also be used to objectively assess visual function without the need for behavioural responses.[61-63] While OKN is most often used to provide a general assessment of the presence or absence of visual function, the spatial frequency of the inducing pattern can be varied to provide an objective measure of resolution acuity.[63] The logic is similar to that applied to VEPs whereby the visual acuity threshold is the spatial frequency at which an OKN is no longer present. Recent studies in adults have demonstrated that computerised OKN tests incorporating precise infrared eye-tracking techniques can provide estimates of resolution acuity.[62, 64] These techniques have not been tested in children, presumably due to the difficulty of using eye tracking equipment that requires a fixed head position and good compliance.

Resolution Acuity in Infants and Toddlers

The preferential looking technique is the most commonly used clinical method for assessing visual acuity in infants and toddlers. This technique typically employs square wave grating stimuli to provide psychophysical estimates of resolution acuity and can be used with newborn babies through to five-year-old children.[65] Preferential looking measurements tend to become more challenging as children become more easily distracted with age, which makes the reasonably lengthy[66] behavioural testing required more difficult. The principle behind this method of visual acuity assessment is that when infants and children are simultaneously presented with a patterned target and a blank target of equal luminance, they will preferentially look toward the patterned target.[67] By varying the spatial frequency of the stimulus shown on the patterned target, it is possible to identify the point at which the infant can no longer resolve the stimulus and therefore no longer shows a preference for either the patterned or blank target. The most commonly used preferential looking test in clinical practice is the well-known Teller acuity cards test (TAC).[65, 68] Estimates of visual acuity using the TAC grating targets show a rapid increase in acuity during the first six months of life from 1.0 cycle per degree at one month of age to five cycles per degree by six months of age, then a more gradual increase until adult-like levels (40 cycles per degree) are reached at five years of age.[46, 69]

The usefulness of grating acuity targets to detect amblyopia and visual impairment remains equivocal. Drover and colleagues[70] have proposed a criterion based on normative data established by Mayer and colleagues,[46] whereby any eye that fell below the lower limit of age-appropriate 95 per cent confidence intervals was classified as having amblyopia. Using this criterion, Teller acuity cards had a sensitivity of 80 per cent and a specificity of 74 per cent for detecting children with amblyopia; however, other authors argue that the absence of crowding and contour interaction effects, phenomena that are not generally present in tests of grating acuity, may make resolution acuity tests less sensitive to the detection of amblyopia.[71] Consistent with this idea, acuity measurements using grating stimuli provide significantly higher (better) estimates of acuity in strabismic and anisometropic amblyopia, when compared with recognition tasks using either picture or letter optotypes.[72, 73] Acuity measured with grating stimuli also appears to be relatively insensitive to dioptric blur,[74] possibly due to spurious resolution. Interestingly, Teller acuity cards have proved accurate in the detection of meridional amblyopia in children with high astigmatism, when the difference in grating acuity with the stripes presented both vertically and horizontally is assessed.[75] The reason for this improved detection of astigmatism and meridional amblyopia has not been fully explained but may be due to better detection of acuity deficits when both vertical and horizontal presentations of the grating stimuli are employed.[76]

The Cardiff Acuity Test (CAT) is a preferential looking test that uses pictures as vanishing optotypes,[77, 78] as toddlers often quickly become bored with grating targets.[79, 80] The pictures (fish, car, boat, train, house and duck), are designed so that the picture's outline is a white band surrounded by two black bands. The mean luminance of the outline approximately matches the card's grey background. All pictures are the same size but as higher levels of visual acuity are tested, the black and white bands become narrower. As this task is also a resolution-based test, it has been argued to provide an alternative to grating measurements; however, the pictures contain complex spatial components and as such may not be equivalent to grating targets at all spatial frequencies.[81] Despite this, close agreement between CAT and grating acuity measurements have been reported in children with and without neurological impairment.[79, 82] In literate children aged five to six years with no significant refractive error or ocular pathology, comparable monocular visual acuity results were found between TAC II cards (mean visual acuity of 24.5 cycles per degree; approximately 0.10 logMAR equivalent) and ETDRS letters (mean visual acuity of 0.040 logMAR).[47] In addition, a statistically significant correlation between the CAT and Snellen acuity measurements was found within a cohort of older adults (47 to 99 years), although the CAT gave a higher estimate of visual acuity than Snellen letters.[83] This may be due to the fact that, like grating acuity, Cardiff acuity cards are poor at detecting uncorrected refractive errors.[84]

Resolution acuity tasks, such as the CAT are not only used with infants and toddlers but are often useful in other populations, such as children with developmental delay,[85] patients with neurological conditions[86] and deaf-blind populations.[87] Visual impairment can be prevalent in these populations and appropriate management may be delayed due to difficulties in obtaining accurate measurements of visual acuity using recognition-based tests that place demands on cognitive function and communication. Resolution acuity tasks can provide a rapid and objective assessment in these populations. For example, success rates with both TAC and Cardiff acuity cards are very high with many studies reporting co-operation for over 90 per cent of patients.[85, 87]

Grating acuity measures can also be valuable in older patients with cognitive dysfunction;[88-90] however, this is not always the case. Chriqui and colleagues[91] recently reported results from three different groups of older participants: a group of young adults (mean age 25 years), a group of older adults without neurological disease (mean age 70 years) and a group of older adults with cognitive impairment. In young and older adults without neurological disease, there were no significant differences in visual acuity measures obtained with charts that measured resolution acuity (TAC) and those that measured recognition acuity (Snellen, ETDRS, Tumbling E and Patti Pics). In patients with moderate to severe dementia, TAC and Patti Pics gave both statistically and clinically worse measures of visual acuity than ETDRS, Snellen and Tumbling Es. The authors suggested that may this be due to the disease process itself, as letters are familiar symbols that are learnt early in life and my be easily retrieved from long-term memory in patients with Alzheimer's disease.

Measurement of Recognition Acuity in Preschool Children

Picture optotype tests

By the age of four years, most children can complete a recognition acuity task that requires either naming or matching of optotypes. A number of picture tests have been designed for use in paediatric populations; however, many of these tests, such as the Allen cards[92] and the Wright figures,[93] lack standardisation, especially in the construction and detail of the pictures and the acceptable names that children may use to identify the pictures.[94] The Allen cards in particular have significant design problems, including variable inter-line gap widths and shape cues, which allow some symbols to be more readily identified than others. The Wright figures contain five uniform optotypes that subtend equal angles of resolution; however, due to the complexity of the pictures, the symbols are 2.7 times larger than their Snellen letter equivalents. Both Allen cards and Wright figures over-estimate visual acuity in children with amblyopia by one to two lines of letters.[93]

To provide greater standardisation, Lea symbols were developed using the same principles as the Bailey–Lovie logMAR chart,[9, 12] with equal numbers of optotypes on each line and a standardised progression of optotype sizes.[95] Four symbols (square, circle, apple/heart and house) were chosen, as they all blur equally with plus lenses or diffusing filters and appear as a circle when presented below threshold acuity. The Lea symbols had to be 1.5 times larger than the equivalent Snellen E optotype to provide the same level of visual acuity in adult participants. In general, children co-operate well with the Lea symbols test, particularly children three years of age and older;[45, 48, 96] however, the Lea symbols may not be suitable for all children, as in one study only 31 per cent of 21 to 48 month old infants were able to complete the test monocularly.[48] Estimates of measurements of visual acuity using Lea symbols vary but most studies[49, 97, 98] find a mean visual acuity of around 0.10 logMAR in children, when measured with crowded Lea symbol charts (Table 3). In a group of Native American children aged five to seven, who wore spectacle corrections for astigmatism or other significant refractive error, visual acuity was 0.29 ± 0.18 logMAR.[75] This result is significantly worse than other populations that have been examined and may reflect mild amblyopia in this group of children.

In older children and adults, there appears to be a high correlation between visual acuity measurements made using the Lea symbols chart and those made using a Bailey–Lovie chart; however, overall visual acuity measurements obtained with the Lea symbol chart were approximately one line (0.09 ± 0.11 logMAR) better than those obtained with the Bailey–Lovie chart.[99] Similar results are seen when comparing Lea symbols and the ETDRS chart, with Lea symbols visual acuity being approximately half a line of letters better in children aged five to seven years old.[100] Lea symbols appear to be sensitive to amblyopia, particularly when using the crowded forms of the test.[94] The VIP study has shown that in a screening setting, with specificity set at 0.90, Lea symbols have a sensitivity of 61 per cent for detecting amblyopia, strabismus, significant refractive error or other causes of reduced visual acuity, which was similar to non-cycloplegic retinoscopy and autorefraction.[39]

The Patti Pics test is a variation on the Lea symbols and contains an additional ‘star’ optotype. There is a paucity of published data on this test, despite it being used widely in clinical practice, particularly in North America. It has been reported that measurements of visual acuity in adults are constant across the Landolt C, Patti Pics and Lea symbols tests;[101] however, Mercer and colleagues[50] recently found that while the Patti Pics gave more consistent results in adults than Sloan letters, the visual acuity measurements using Patti Pics were poorer than those made using the Lea symbols by one line of optotypes. This may indicate that Patti Pics provide estimates of acuity that are similar to those obtained using the Bailey–Lovie chart, which also provides acuity estimates that are one line of letters worse than the Lea symbols in adults.[99]

The Kay pictures test was created to provide a recognition-based visual acuity measurement for use in children aged two to three years. The Kay pictures test is based on the presentation of eight familiar pictures that a child can either match or name. It comes in both uncrowded and crowded formats and is widely used, especially in the UK and Europe. This test also incorporates Bailey–Lovie chart construction principles, using a logarithmic progression of acuity levels, an equal number of optotypes (four) on all but the first two rows and a constant level of crowding between lines. The symbols used in the Kay pictures chart are based on Snellen sizing, with a stroke width that is the same as the equivalent Snellen letter.[43] Due to the overall complexity of constructing these optotypes, the size of the pictures was increased to twice the equivalent Snellen letter size even though the stroke width remained the same.

The manufacturer's website suggests that the normative values for crowded Kay pictures are 0.10 logMAR for children under four years and 0.05 logMAR for children aged between four and five years ( In children with normal visual acuity, comparisons between letter matching and the crowded Kay pictures test have found less than one line difference in acuity, although a statistically significant (but clinically small) difference was seen in children with amblyopia, whose visual acuity was 0.074 ± 0.036 logMAR better when measured with the Kay pictures test.[102] A comparison between the crowded Kay pictures and the gold standard ETDRS chart in 30 adult subjects with pathology and 40 amblyopic children (four to 15 years) has shown that Kay pictures consistently produce better acuity measurements than those obtained with a standard adult letter chart.[103] This difference was approximately one line of letters better in adults with stable ocular pathology and nearly two lines better in amblyopic children. Test-retest variability was similar between both charts (± 0.14 and ± 0.16 logMAR for adults and children, respectively), which suggests good repeatability of measurements as long as the same chart is used at each visit. The bias toward better acuity measurements achieved with the crowded Kay pictures test implies that caution should be used when interpreting measurements made with crowded Kay pictures and ETDRS charts in the same patients, particularly for children.

Letter matching tests

Tests that require naming of letters are often too difficult for preschool children; however, letter matching tests have a high degree of testability[45] and similar test-retest variability,[104] when compared to picture optotype tests, which require naming and/or matching. These tests all use a limited range of letters with a key card for matching and a reduced testing distance (usually three metres) to improve testability with younger children.[105, 106] Letter matching can be performed by children as young as three years (approximately one-third of children co-operating) and becomes almost universally testable by the age of four years.[107, 108]

One of the most popular tests employed for the assessment of visual acuity in children 36 months of age and older in the United Kingdom (as well as New Zealand and some parts of Australia) is the Sheridan–Gardiner test, which was adapted from the original STYCAR (Screening Tests for Young Children and Retardates) in the 1970s.[109] This single optotype letter matching test has gained much popularity due to its ease of use, speed and simplicity and is often employed in both vision screening programs and in private practice for the measurement of visual acuity in young children. The test has several disadvantages, including irregular progression size of the letters, the absence of a 6/7.5 (0.10 logMAR) line, truncation of the test at Snellen 6/6 acuity in the standard booklet, with measures of better visual acuity requiring the use of an additional booklet and a lack of contour interaction in the original uncrowded version.[27] The Sheridan–Gardiner test has been used widely in vision screening programs; however, preschool vision screening programs that rely solely on measurements of visual acuity have low specificity[110] and often produce high false positive referral rates[111] compared to more comprehensive testing that is carried out by trained eye-care providers.[39, 112] Sheridan suggests that normal monocular visual acuity in children aged five to seven is 6/4.5 (-0.125 logMAR) on the single optotype version of the test[113] and other authors[27, 114] have also reported better than 6/6 visual acuity obtained with Sheridan–Gardiner letters in young children.

The HOTV test also represents a modification of the STYCAR test and includes selected letters with vertical symmetry that are present on the standard Snellen chart. The Amblyopia Treatment Study (ATS) groups have reported good testability with a single optotype HOTV test surrounded by crowding bars with 87 per cent of four-year-old children and 96 per cent of children five to seven years able to complete this test.[104] The protocol employed by the ATS group, also showed excellent test-retest reliability with 93 per cent of eyes within 0.1 logMAR of the initial test score, which is similar to adult values on the ETDRS chart.[17] Using the ATS protocol implemented as an automated, computer-based staircase algorithm to find the threshold visual acuity of children aged three to 10 years old, Drover and colleagues[52] found that visual acuity improves from 0.08 logMAR in three- to four-year-old children to -0.06 logMAR in eight- to 10-year-old children. There was a statistically significant improvement in visual acuity from three years to adulthood and Snellen equivalent 6/6 visual acuity was reached between the ages of five to six years. The VIP study group found that using a single crowded HOTV optotype displayed on the Baylor Video Acuity Tester (BVAT) system gave better visual acuity than a multiple line, crowded Lea symbols chart.[45] The authors suggested that the use of a single optotype may be less confusing for children and the use of an internally illuminated television monitor may provide a more interesting stimulus, which is less prone to interference from external distraction than a printed letter chart.

Glasgow acuity cards (marketed as Crowded Keeler logMAR™) are comprised of six letters (X, V, O, H, U, Y) of approximately equal legibility with a 0.1 log unit progression between lines. The letters were chosen to be symmetrical around the vertical midline to avoid right-left confusion in younger children. Four letters make up a line and each line is surrounded by a contour box. The box is positioned one half an optotype size from the letters to improve sensitivity for detecting amblyopia. In adult observers this chart configuration produces measurements that are in close agreement with measurements obtained on the Bailey–Lovie chart.[115]

Use of adult letter charts in paediatric practice

Few studies have compared visual acuity measurements obtained using paediatric tests with the gold standard ETDRS chart, due to obvious difficulties with co-operation and cognitive ability in younger children. However, by the time children reach school age, the use of standard adult charts is achievable, with only 1.8 per cent of six-year-olds in a sample of 1,738 Australian children being unable to complete visual acuity measurements on the ETDRS chart.[116] Similarly, it has been reported that 52 per cent of five-year-olds, 87 per cent of six-year-olds and 100 per cent of seven-year-olds can complete visual acuity measurements using an electronic version of the ETDRS chart.[117] Furthermore, adult levels of test-retest variability on the ETDRS chart were found for children aged six to 12, whose 95 per cent confidence intervals were between 0.10 and 0.20 logMAR;[118-120] however, despite evidence that the ETDRS chart can be used to test visual acuity in these younger age groups, adult levels of acuity for this test may not be reached until as late as 13 years of age.[121, 122]

In older children, there appears to be a good correlation between measurements obtained with Lea symbols and the ETDRS chart;[100] however, it was noted that in children aged five to seven years, visual acuity measurements tended to be approximately two to three lines better when measured with Lea symbols relative to the ETDRS chart[100] and 0.06 logMAR (three letters different) relative to the HOTV chart.[117] These discrepancies may be due to a number of factors, including fundamental differences in the spatial properties of picture and letter optotypes and the cognitive demands of each test. There is also recent evidence to suggest that many commonly used paediatric and adult visual acuity charts have some optotypes that are more easily recognisable than others.[101] This is important as visual acuity measurements are interpreted under the assumption that all optotypes are equally recognisable/discriminable and differences in recognisability could systematically bias the results of the test.

Other Factors That Influence The Measurement of Visual Acuity

The inclusion of contour interaction in visual acuity tests, commonly referred to as the crowding phenomenon, is important to avoid the over-estimation of visual acuity and to improve amblyopia detection rates.[123, 124] The deleterious effects of crowding on visual acuity are well documented, particularly in amblyopic children and manifest as better acuity with single optotypes than with a linear or multi-line test.[125, 126] In the crowded version of the Sheridan–Gardiner test, flanking bars surround a single optotype; however, this appears to produce less pronounced crowding effects than a linear test.[36] For example, Hilton and Stanley[127] compared visual acuity in amblyopic children using both a standard Snellen letter chart and the Sheridan–Gardiner test. They found that all children showed a reduction in visual acuity of at least one line of letters on the Snellen chart compared with measurements previously recorded on the Sheridan–Gardiner test, with more than a quarter of the children showing a reduction of four or more lines of letters. The authors suggest that the Sheridan–Gardiner test is useful in obtaining measurements on children who are unable to complete adult letter tests but that the level of visual acuity measured may not be comparable to standard Snellen charts. Recent evidence also suggests that the effect of contour interaction is more pronounced in children than in adults, with children as old as 11 years being more effected by contour interaction than adults,[128] possibly due to immaturities in brain areas beyond the primary visual cortex.

It is well known that not all letters have equal legibility[12, 129] and even though the 10 Sloan letters chosen for the ETDRS chart were originally thought to have equal legibility, when this chart was designed,[130] recent studies have shown that the letter ‘C’ is more likely to be misidentified than other letters on the ETDRS chart.[13] In a sample of 60 amblyopic children aged five to 13 years, Mathew, Shah and Simon[28] found that certain Snellen letters (B, C, F, S) are more difficult to determine than other letters on the chart and that this is the case for both amblyopic and non-amblyopic eyes. These results suggest caution should be used in interpreting very small changes in visual acuity, as both the degree of crowding (contour interaction) and the legibility of the optotype can influence the measurement of visual acuity.

One additional factor that can influence the outcome of visual acuity measurements in children is the testing distance. While the use of a six metre testing distance is considered standard for the measurement of adult visual acuity, shorter working distances such as 3.0 or 1.5 metres allow for excellent repeatability and reliability when measuring visual acuity in children.[40, 115] The use of a shorter testing distance brings the vision tester closer to the child, which significantly improves co-operation and concentration.[131]


Children seen in clinical practice commonly have their visual acuity measured with a variety of tests, as they develop from infancy to adolescence. Caution must be used when comparing measurements between tests, particularly if the task has changed, for example from a resolution to a recognition test. If visual acuity changes are observed, this may reflect a change in chart design rather than an actual change in visual function. This is also true for changing from one optotype chart to another, as picture and letter charts may not always give comparable results.

Design features such as contour interaction, the level of crowding, optotype design and the number of optotypes, are all factors which influence acuity measurement. Even in adult charts, which strive for equal legibility of letters used, some letters appear more difficult to recognise than others. This issue is more pronounced in paediatric charts, as the choice of letters is often restricted to those that are symmetrical around the vertical midline to avoid laterality confusions.

Many tests have little published and validated data and therefore, it is difficult to provide normative acuity for children, as results vary depending on the child's age, the chart used and the methodology employed for establishing visual threshold. It is also apparent that not every paediatric chart meets the International Visual Acuity Chart Design Guidelines and therefore, it is difficult to directly compare and interpret results using different charts. Recent efforts to standardise testing protocols, such as those used by the Amblyopia Treatment Study group, are helping to provide a better understanding of recognition acuity norms in children. Both in clinical practice and in research, uniform principles need to be employed to ensure accurate, reliable and repeatable measurements in paediatric populations. Further research is required to establish appropriate age-specific limits of normal visual acuity for many of the paediatric charts currently used in clinical settings as well as to develop cut-off values for normal and abnormal results for use when screening. Appropriate values of specificity and sensitivity are also required for each test.[96] Further large cohort, multicentre studies, similar to the VIP, ATS and Pediatric Eye Disease Investigator Group (PEDIG) studies, are required to provide appropriate validation of many of the clinical tests commonly used with preschool populations to ensure that appropriate assessment techniques are employed, which will provide the clinician with diagnostically valid results.