7

Visual Acuity

Ian l. Bailey

V isual acuity is the spatial resolving capacity of thevisual system. It expresses the angular size of detail

that can just be resolved by the observer. The limits tovisual acuity are imposed by optical and neural factorsor their combination. In the normal eye, the limitationsimposed by optical factors and neural factors are ofsimilar magnitude. 1

OPTICAL LIMITATIONS

When the eye is in ideal focus, a point object is imagedon the retina not as a point but as a small circular patchwith faint surrounding rings; this is the diffractionpattern. The central circular patch is called the Airy disk,and it has an angular size of OJ = 2.44 Ajp (where thediameter OJ is expressed in radians, A. is the wavelengthof light, and p is the pupil diameter). The smaller thepupil is, the larger is the Airy disk. When the quality ofoptical imagery is only limited by diffraction, theRaleigh criterion for resolution says that two Airy diskscan just be resolved when the center of one lies at theedge of the other. In other words, the angle (f3min)between the points is f3min = 1.22 A./p. A useful approx­imation to this equation is f3min = 2.3/p (where f3min is inminutes of arc and p is in millimeters). Applying theRaleigh criterion, a 4.6-mm pupil would be required toachieve a minimum angle of resolution (MAR) of 0.5minutes of arc (minarc); a 2.3-mm pupil is required forMAR equaling 1 minarc, and a 1.1-mm pupil allowsMAR to equal 2 minarc. High resolution cannot beachieved with very small pupils or pinhole apertures.

Obviously, resolution suffers when image quality isdegraded by focusing errors, such as myopia, hyperopia,astigmatism, or the failure to optimize focus by appro­priate accommodation or spectacle lenses. Even withoptimal refractive correction and focusing, there stillmay be image degradation as a result of the chromaticand monochromatic aberrations of the eye. Imagedegradation from aberrations increases with large pupildiameters. With very small pupils, the optical limitation

on resolution is imposed by diffraction; however, withlarge pupils, it is the aberrations that limit optical per­forrnance.':" For maximum visual acuity, the optimalpupil size is about 2.5 mm, and the resolution limit isjust under 1 minarc.

NEURAL LIMITATIONS

The neural limit to resolution is imposed by the packingdensity of the retinal receptors and the neural interac­tions in the retina and subsequent visual pathways. Inthe foveal region, where the retina achieves best resolu­tion, the separation between centers of neighboringcones is about 2 urn. Thus, 4 urn would separate theimages of two points when they fall on the centers oftwo receptors that are separated by one unstimulatedreceptor. Assuming that this situation represents theanatomically imposed limitation to resolution and thatthe nodal point of the eye is 16.67 mm from the retina,it is predicted that the neural limit to resolution shouldbe 0.82 minarc. This is similar in magnitude to theoptical limit.

TESTS OF VISUAL RESOLUTION

A variety of different tests of visual performancemeasure some aspect of the limits of the visual system'sability to discern detail or to recognize detailed targets.

Minimum Detectable Resolution

The minimum detectable resolution is the thresholdsize of a spot or a line required to detect its presenceagainst its background. Consider a light spot displayedagainst a dark background. If the spot is very small, thewidth of the retinal image is determined by diffraction.The width of this image is independent of the width ofthe spot. If the geometrical image of the spot is smallerthan the diameter of one receptor, further reduction of

217

218 BENJAMIN Borishs Clinical Refraction

the spot size simply reduces the total amount of lightfalling on that receptor. The task of the visual systemnow becomes one of contrast discrimination. The visualsystem has to distinguish that the amount of lightfalling on that receptor is greater than that falling on itsneighbors. The functional question becomes, "What isthe size of the smallest spot that can cause a detectableelevation in the total illuminance on the receptor?" Asimilar argument would apply to the detection of a lightline on a dark background or a dark spot or line againsta light background.

Minimum Separable Resolution

The minimum separable resolution is the least separa­tion between two adjacent points or adjacent lines thatallows the two to be seen as separate. The minimum sep­arable value is often used to evaluate the performance orquality of optical systems, and it can be used to measurethe resolution capacity of the human visual system.Popular alternative targets for measuring the minimalseparable resolution are gratings or sets of three lines.For such gratings or three-line targets, the alternatingdark and light lines are of equal width (duty-cycle, 1.0).For the three-line target, the observer's task is to deter­mine the minimum separation of lines that allows themto be distinguished as three different lines. For gratingtargets, the task is to determine the finest grating thatcan just be distinguished from a uniform field of thesame average luminance. Some laboratory tests of visionpresent displays of gratings in which the luminance dis­tribution across the grating has a sinusoidal profile. Forgrating targets, the resolution limit is usually expressedin cycles per degree (cpd). At 30 cpd, there are 30 darkand 30 light lines within each degree so that the averageline width is %0 degree (equal to 1 minarc).

Sometimes, with periodic patterns, "spurious resolu­tion" occurs. If, for example, the angular size of a three­line target is progressively reduced, the three lineseventually become unresolvable. A further reduction inangular size may cause the target to become indistinct,but it might appear that there are two lines rather thanthree. This depends on the luminance profiles of theimages of each line and their combination when theyoverlap. The presence of lines is detectable, but the res­olution is spurious, because three lines appear as two.For grating targets, spurious resolution can cause para­doxical reversals of threshold, because the presence ofthe grating can become more detectable as the spatialfrequency is increased. This effect is more likely to occurwhen the eye is not in clear focus and the limits to res­olution are being determined by optical rather thanneural factors.

Some instruments designed for screening visualacuity use checkerboard targets wherein the patient ispresented with four square areas, three ofwhich contain

a uniform gray or a fine halftone pattern, whereas thefourth square contains a relatively coarse checkerboardor dot pattern. The patient's task is to determine whichof the four areas contains the checkerboard. The meanluminance of the gray halftone squares is matched tothat of the checkerboard pattern so that the four squaresappear to have equal luminance when the checkerboardcannot be resolved.

Recognition Resolution

Most clinical tests of visual acuity are recognition teststhat determine the smallest symbols, letters, or wordsthat can be identified correctly. Test targets used forthese tests are often called optotypes. The Snellen chart(Figure 7-1) uses letters as the optotypes.

Landolt RingsThe Landolt ring target-or "Landolt C"-consists of acircle with a break in it (Figure 7-2). The external diam­eter of the ring is five times the stroke width of the circleso that the internal diameter is three stroke widths. Thebreak or gap is one stroke-width wide. For most Landoltring tests, the gap is presented in four alternative loca­tions: up, down, right, or left. Sometimes, there are eightalternative gap positions (four cardinal and fouroblique). The observer's task is to determine the loca­tion of the gap for each Landolt ring presented. Unlikemost other optotypes, the critical detail in the Landoltring is well defined and unambiguous: it is the gap inthe ring. Thus, the critical detail is one-fifth the heightof the optotype. At threshold or near-threshold levels,the observer does not necessarily see the target as a ringwith a gap in it. Rather, the target appears as a small spotor blob with a region that is marginally asymmetrical orlighter, and it is this irregularity that identifies the gapposition.

Letter OptotypesMost letters designed for visual acuity tests are based ongrid patterns that are five units high. They have usuallybeen five units wide although letter widths of four or sixunits have sometimes been used. The stroke width ofthe letters is usually a fifth of the height and, as muchas is practical, the spacing between adjacent strokes ismade equal to the stroke width. Snellen" introduced theletter chart (see Figure 7-1) for visual acuity measure­ment, and he designed his optotypes so that the majorlimb strokes were one-fifth the letter height. Many of theacuity charts that followed" used a similar approach,and, like the original Snellen design, most used serifs(short lines or blocks added at an angle to the ends oflimbs of the letters) on the letters. More modern lettercharts use sans-serif letters. Today, the most commonlyused sans-serif letters are the Sloan letters." which arebased on a five-by-five grid. In the Sloan letter set, there

Visual Acuity Chapter 7 219

UYACEGL:3

FHKOS3

Figure 7-1Snellen's original chart, shown at about 40% of itsactual size.

the five-by-five grid. Five of these letters are identical tothe Sloan letters (C, H, N, V, Z). There are a new K anda new R with different limb angles, and a new D withdifferent curvatures. There are four letters in addition tothe Sloan set (E, F, P, U), and there are two Sloan lettersthat do not appear in the new British series (0, S). Theselection of limited letter sets and the specification ofthe letter designs are intended to reduce the variabilityof legibility between letters. However, within each letterset, there always remains some variation in the legibil­ity of the individual letters. Chart designers shouldarrange the mixtures of letters at each size so that theaverage legibility is similar at each acuity level. A com­parison of the Sloan and British Standard" letters is con­tained in Table 7-1.

For the recognition of letter targets at or near thresh­old sizes, a variety of clues or combinations may beresponsible for the correct letter identification. Forexample, the letters Nand H are similar in their generalshape, and, when close to threshold size, they might bedistinguished from most other letters with relative ease.The patient might see a squarish letter and narrow downthe choice to H or N. For the final distinction, the crit­ical cue for correctly identifying the N might be thedetection of its diagonal limb, seeing that there is anoffset of the notch in the upper and lower edges of thesquare, or recognizing that there is a concentration ofdarkness in the upper-left and lower-right corners of thesquare.

DESIGNATION OF VISUAL ACUITY

Tumbling EThe tumbling E target, sometimes called the "illiterateE," is based on a five-by-five grid. The E is presentedin different orientations at every acuity level, andthe patient's task is to identify the direction to whichthe limbs of the E point. Most commonly, there arefour alternative directions: up, down, right, and left.Some tests, however, use eight alternatives, with theaddition of the four oblique directions. The letter Eusually has three limbs of equal length. The recentBritish standard" specified an illiterate Ewith the centrallimb one unit shorter than the external limbs. TumblingE targets are most useful when measuring acuity in tod­dlers or other persons who are not familiar with thealphabet.

Numerical and pictorial targets are available, andthey are mainly used with pediatric and illiterate popu­lations." These are further discussed in Chapter 30.

Visual acuity expresses the angular size of the smallesttarget that can just be resolved by the patient, but thereare several different ways in which clinicians specify thisangular quantity (Table 7-2).

4

5

DB

RT

z

p

v

GLN

CE

are 10 letters (C, 0, H, K, N, 0, R S, v, Z), with speci­fied angles and curvatures for each. A previous Britishstandard of optotypes" used a different set of 10 letters(0, E, F, N, H, P, R U, V, Z) based on a five-by-four grid.Figure 7-2 shows examples of a Landolt ring, a five-by­five serif letter, a Sloan (five-by-five) letter, and a 1968British (five-by-four) letter.

The 2003 British standard on optotypes" introduceda new set of 12 sans-serif letters that is also based on

220 BENJAMIN Borishs Clinical Refraction

Serif E Landolt Rin

Non-serif H 5x5 Non-serif E 5x4

Figure 7-2

Examples of optotypes constructed on a grid framework.

Snellen Fraction

The Snellen fraction expressesthe angular size of optotypesby specifyingthe test distance and the height of the letters.In the Snellen notation, the number used to indicate theheight of the letters is the distance at which the letter heightsubtends 5 minarc. In otherwords, a 20-foot (or 6-m) letteris one with a height that subtends 5 minarc at 20 feet (or6 m). The Snellen fraction is written with the test distanceas its numerator and the letter size as its denominator:

Visual acuity = (test distancel/Idistance at which

letters subtend 5 minarc)

A visual acuity score of 20/200 means that the testdistance was 20 feet and the smallest letters that could

be read would subtend 5 minarc when at a distance of200 feet. The angular size of such letters at 20 feet is 50minarc. Provided the retinal image is kept in good focus,the visual acuity should not change with test distance.Thus, 20/200, 40/400, 10/100, 5/50, and 6/60 are allvisual acuity scores that represent the same angle (letterssubtend 50 minarc); the test distances and the thresh­old print sizes are different, but they remain in propor­tion. In the United States, distances are expressed in feet,and clinicians almost invariably use the Snellen fractionwith 20 feet as the numerator. In most other countries,metric units are used, with 6 m being the most commontest distance. Thus, 20/20 is equivalent to 6/6, 20/25 to6/7.5, 20/40 to 6/12, 20/100 to 6/30, 20/200 to 6/60,and so forth (see Table 7-2).

Visual Acuity Chapter 7 221

TABLE 7-1 Comparison of the Sloan letters and British Standard (2003) letters

Sloan British 2003 Letter Letter Stroke Exterior Interior RelativeLetters Letters Height Width Width Radius Radius Angles Legibility

C C 5 5 1 2.5 1.5 0.99D 5 5 1 1.5 0.5 1.01

D 5 5 1 2.5 1.5E 5 5 1F 5 5 1

H H 5 5 1 1.06K 5 5 1 37/128 0.99

K 45/135N N 5 5 1 131 1.050 5 5 1 2.5 1.5 0.90

P 5 5 1 1.5 0.5R 5 5 1 1.5 0.5 116 0.97

R 5 5 1 127S 5 5 1 1.5 0.5 0.93

U 5 5 1 2.5 1.5V V 5 5 1 112/68 1.05Z Z 5 5 1 41 1.10

n =10 n = 12

Five letters are identical (C, H, N, V, Z) in bothfamilies.Three letters (D, K, R) are in bothfamilies, but the shapes are not identical.Two letters (0, S) areonly in the Sloan series.Four letters (E, F, P, U) areonly in the British series.In the British F and E, one horizontal limb is 1 unit shoTter than the other(s).Legibility data are not available for the 2003 British letters.

Decimal Notation

The decimal notation effectively reduces the Snellenfraction to a decimalized quantity. Thus, 20/20 (or 6/6)becomes 1.0,20/200 (6/60) becomes 0.1,20/40 (6/12)becomes 0.5, and so forth. Decimal notation is mostwidely used on the European continent; it gives a singlenumber to quantify an angle, and it does not indicatethe test distance.

Minimum Angle of Resolution

The MAR is typically expressed in minutes of arc, and itindicates the angular size of the critical detail within thejust-resolvable optotype. For letters, the critical detail istaken as one fifth of the letter height. For a visual acuityof 20/20 (or, in metric units, 6/6), the MAR is equal to1 minarc. For 20/40 (or 6/12), the MARis 2 minarc; for20/200 (or 6/60), the MAR is 10 minarc. The MAR inminutes of arc is equal to the reciprocal of the decimalacuity value.

Logarithm of the Minimum Angle ofResolution

The logarithm of the MAR (logMAR)1O is the commonlogarithm of the MAR. When visual acuity is 20/20 (or6/6), the MAR is equal to 1 minarc, so the log MARequals 10glO (1.0) equals 0.0. For 20/40 (or 6/12), theMAR is 2 minarc, so 10gMAR equals loglo (2.0) equals0.30. For 20/200 (or 6/60), the MAR is 10 minarc, so10gMAR equals 10glO (10) equals 1.0.

When the visual acuity score is better than 20/20 (or6/6), the 10gMAR value becomes negative. For example,for 20/16 (or 6/4.8), MAR equals 0.8 minarc and log.,(0.8) equals -0.10. For charts that have a size progressionratio of 0.1 log units and five letters per row, each lettercan be assigned a value of0.02 on the 10gMAR scale.

Visual Acuity Rating

The visual acuity rating (VAR) 11 is derived from the10gMAR values:

VAR 100 - 50 10gMAR

N N N

TA

BL

E7

-2C

on

ve

rsio

nT

ab

lefo

rV

isu

al

Acu

ity

Sco

res

NE

AR

VIS

ION

DIS

TA

NC

EV

ISIO

NA

T4

0cm

AT

14S

NE

LLE

NF

RA

CT

ION

SS

ne

llen

INC

HE

S

Lo

gM

AR

VA

RM

AR

MA

RD

eci

ma

lG

rati

ng

VE

%B

ased

Bas

edB

ased

no

tati

on

Nx-H

eig

ht

"Re

du

ced

Jae

ge

rS

ne

llen

for

No

tati

on

No

tati

on

Exa

ctN

ota

tio

n'

No

tati

on

'cp

dN

ota

tio

no

n2

0ft

'o

n6

m'

on4

m'

0.4

0m

'M

Un

its

po

ints

'[m

m]

Sn

elle

n'"

(ap

pro

xim

ate

)14

inch

es'

-0.3

011

50.

501

0.50

2.00

6010

9.4%

20/1

06

/34

/20

.40

/0.2

00.

201.

60

.29

101

4/7

-0.2

011

00.

631

0.63

1.60

4810

6.8%

20

/12

.56

/3.8

4/2

.50

.40

/0.2

50.

252.

00

.36

20

/12

.51

4/8

.8-0

.10

105

0.79

40

.80

1.25

3810

3.6%

20

/16

6/4

.84

/3.2

0.4

0/0

.32

0.32

2.5

0.4

72

0/1

614

/11

0.00

100

1.00

01.

001.

0030

100.

0%2

0/2

06

/64

/40

.40

/0.4

00

.40

3.2

0.5

82

0/2

011

14

/14

0.10

951.

259

1.25

0.8

024

95

.6%

20

/25

6/7.

54

/50

.40

/0.5

00

.50

4.0

0.73

20

/25

J1-/

214

/17.

50.

2090

1.58

51.

600.

6319

89

.8%

20

/32

6/9.

54

/6.3

0.4

0/0

.63

0.63

5.0

0.9

22

0/3

2J1

-/4

14

/22

0.30

851.

995

2.0

0.5

015

83

.6%

20

/40

6/1

24

/80

.40

/0.8

00.

806.

31.

162

0/4

0J2

-15

14

/28

0.40

802.

512

2.5

0.4

012

76

.5%

20

/50

6/1

54/

100

.40

/1.0

01.

008

.01.

452

0/5

0J3

-J6

14

/35

0.50

753

.16

23.

20

.32

9.5

67

.5%

20

/63

6/1

94

/12

.50

.40

/1.2

51.

2510

.01.

822

0/6

3J4

-J8

14

/44

0.60

703.

981

4.0

0.2

57.

55

8.5

%2

0/8

06

/24

4/1

60

.40

/1.6

01.

6012

.52.

332

0/8

0J5

-J9

14

/56

0.70

655.

012

5.0

0.2

06.

04

8.9

%2

0/1

00

6/3

04

/20

0.4

0/2

.02.

016

2.91

20

/10

0J8

-112

14

/70

0.8

060

6.31

06.

30

.16

04.

83

8.8

%2

0/1

25

6/3

84

/25

0.4

0/2

.52.

520

3.64

20

/12

5J9

-113

14

/88

0.90

557.

943

8.0

0.1

25

3.8

28

.6%

20

/16

06

/48

4/3

20

.40

/3.2

3.2

254.

652

0/1

60

JlO

-115

14/1

10

1.00

5010

.00

10.0

0.10

03

.020

.0%

20

/20

06

/60

4/4

00

.40

/4.0

4.0

325

.82

20

/20

011

4-11

81

4/1

40

1.10

4512

.59

12.5

0.0

80

2.4

12.8

%2

0/2

50

6/7

54

/50

0.4

0/5

.05.

040

7.27

20

/25

01

4/1

75

1.20

4015

.85

160.

063

1.9

6.8%

20

/32

06

/95

4/6

30

.40

/6.3

6.3

509.

162

0/3

20

14

/22

01.

3035

19.9

520

0.05

01.

53

.3%

20

/40

06

/12

04

/80

0.4

0/8

.08.

063

11.6

20

/40

01

4/2

80

1.40

3025

.12

250

.04

01.

21.

4%2

0/5

00

6/1

50

4/1

00

0.40

/10.

010

.080

14.5

20

/50

01

4/3

50

1.50

2531

.62

320.

032

0.95

0.4

%2

0/6

30

6/1

90

4/1

25

0.4

0/1

2.5

12.5

100

18.2

20

/63

01

4/4

40

1.60

2039

.81

400.

025

0.75

20

/80

06

/24

04

/16

00

.40

/16

1612

523

.32

0/8

00

14

/56

0

1.70

1550

.12

500.

020

0.60

20

/10

00

6/3

00

4/2

00

0.4

0/2

020

16

029

.12

0/1

00

01

4/7

00

1.80

1063

.10

630.

016

0.48

20

/12

50

6/3

80

4/2

50

0.4

0/2

525

20

03

6.4

20

/12

50

14

/88

0

1.90

579

.43

800.

013

0.38

20

/16

00

6/4

80

4/3

20

0.4

0/3

232

25

046

.52

0/1

60

014

/110

0

2.00

010

0.0

100

0.01

00.

302

0/2

00

06

/60

04

/40

00

.40

/40

403

20

58

.22

0/2

00

01

4/1

40

0

•Num

bers

roun

ded

tosi

mpl

ifYse

quen

ces.

Rou

ndin

ger

rors

dono

texc

eed

1.2%

On this scale, a score of 100 corresponds with 20/20(6/6). A VAR that equals 50 corresponds with a Snellenfraction of 20/200 (6/60). The VAR equals 0 when thevisual acuity is at the 20/2000 (6/600) level. The VAR isgreater than 100 when visual acuity is better than 20/20(or 6/6). For example, for 20/16 (or 6/4.8), VAR equals105. On charts that use a O.I-log unit-size progression,the VAR score changes by 5 for each size increment. If, inaddition, there are five letters per size level, each lettercarries a VAR value of 1. The VAR scale can facilitate thescoring of visual acuity. On the VAR scale, a difference of15 points represents a twofold change in the MAR, anda 5-point change represents a change with ratio of 5:4 inthe MAR. The VAR scoring system has been used in theGuides to theEvaluation of Permanent Impairment. 12 A func­tional acuityscore (FAS) is obtained by adding the VAR forthe right eye, the VAR for the left eye, and three times thebinocular VAR and then dividing the sum by 5:

FAS = (VAR oD + YARos + 3 VARolJ)/5

Visual Efficiency

The visual efficiency (VE) scale was introduced in 1925by Snell and Sterling 13, 14 for use when quantifying visualloss for legal and compensation purposes. The scale wasdeveloped on the basis of experiments in which visualresolution was degraded by adding a series of diffusingfilters before the eyes, and it was assumed that visionwas degraded by the same amount as each additionalfilter was introduced. The VE was deemed to be 1.0 (or100%) when visual acuity was 20/20 or 6/6. Arbitrarily,20/200 or 6/60 was said to represent a VEof 0.2 (20%).Given these two chosen benchmarks, a good fit oftheir experimental data was obtained by the followingrelationship:

VE = 0.2(MAR - 1)/9

It is more common for this relationship to beexpressed in the following form:

Log(VE%) = 2.0777 - 0.0777 (MAR)

The American Medical Association (AMA)15 adoptedthe Snell-Sterling scaling of VE. The system wasexpanded by developing VE ratings to quantify losses ofvisual fields and ocular motility. The AMAsystem for theevaluation of permanent visual impairment" allows thecalculation of an overall rating ofVE that is the productof acuity, field, and motility efficiency scores. The AMAsystem combines the monocular VE for the two eyes,giving three times more weight to the VE of the bettereye. This system became obsolete with the AMA's recentpublication of its Guides to the Evaluation of PermanentImpairment, 5th Ed. 12

Visual Acuity Chapter 7 223

VISUAL ACUITY CHART DESIGN

Snellen Chart

Snellen's original chart" had seven different size levels.There was only one letter at the largest size level, andthe number at each size level increased progressively toeight optotypes (seven letters and one number) at thesmallest size (see Figure 7-1). The size sequence in feetwas essentially 200, 100, 70, 50, 40, 30, and 20 (or, inmetric units, 60, 30, 21, 15, 12, 9, 6.) Many modifica­tions were made to Snellen's original chart design, anddetailed descriptions of many of these are provided inthe Bennett's of ophthalmic test types." Despite signifi­cant deviations from Snellen's original design (Le., dif­ferences in letter design and selection, size progressions,spacing relationships, and number of letters at thevarious size levels), it is still common to apply the term"Snellen charts" or even "standard Snellen charts" tocharts that have a single letter at the top and increas­ingly more letters at the smaller sizes.

Bailey-Lovie Design Principles

Bailey and Levie" proposed a set of principles for thedesign of visual acuity charts, and these make the taskessentially the same at each size level (Figure 7-3). Thus,size becomes the only significant variable when chang­ing from one size level to the next. Such standardiza­tion of the visual acuity task requires the following:1. A logarithmic size progression (constant ratio from

one size to the next)2. The same number of letters at each size level3. Spacing between letters and between rows that is

proportional to letter size4. Equal (or similar) average legibility for the

optotypes at each size levelAlong with these chart design principles, they intro­

duced the clinical scoring of visual acuity in 10gMARunits as well as a method for giving equal additionalcredit for each additional letter that is read correctly.

Several charts have since been developed in accor­dance with these principles. Taylor" prepared a tum­bling E chart. Ferris and colleagues" made a chart forthe Early Treatment of Diabetic Retinopathy Study(ETDRS) using Sloan letters rather than the Britishletters that were used in the original version of theBailey-Levie chart. Strong and WOO l 9 arranged Sloanletters with sizes progressing in columns rather thanrows, and they added masking bars to the ends of thecolumns and rows. Iohnstorr" prepared a version usingChinese characters, and Hyvarinen and colleagues" pre­pared charts using abstract "LH symbols" for testingchildren. The Bailey-Lovie chart design using four­position Landolt rings is shown in Figure 7-4. Thesame principles have been used for charts with Arabic,Indian, and Thai characters.F>"

224 BEN.JAMIN Borishs Clinical Refraction

Figure 7-3Visual acuity chart designed according to the principles of Bailey-Lovie (also known as a LogMAR chartdesign), shown at 20% of its actual size. The optotypes used here are Sloan letters, as in the ETDRS charts.LogMAR, Logarithm of the minimum angle of resolution; VAR, visual acuity rating. (Courtesy of the Low VisionResource Centre, Hong Kong Society for the Blind, Kowloon, Hong Kong.)

Design Features for Visual Acuity Charts

Logarithmic Size ProgressionLogarithmic scaling of size on visual acuity charts haslong been advocated by Creen." Sloan." and manyothers," and it is now broadly accepted. Westheimer"provided evidence and argument that logarithmicscaling is more appropriate than other alternatives. Hemeasured peripheral visual acuity at different retinal

eccentncmes, and he found that, across the range ofmeasured visual acuity values, the variance of measure­ment was virtually constant if visual acuity wasexpressed on a logarithmic scale. Thus, just-noticeabledifferences are about equal in size if the scale is loga­rithmic. Although several different logarithmic scalingratios have been suggested, common practice today usesa size progression of 0.1 log unit (10° 1

) . With such a size

Visual Acuity Chapter 7 225

LVRe Near Visual Acuity Test

Meters (Equivalent VA)

8.0 (20/400) cooo oLogMAR (VAR)

1.3 (35)

6.3 (20/320) coooc 1.2 (40)

1.1 (45)

1.0 (50)

0.9 (55)

0.8 (80)

0.7 (65)

0.6 (70)

0.5 (75)0.4 (80)0.3 (85)0.2 (90)

~0.1 (95)0.0 (100)

-4.1 (105)-4.2 (110)

4.0 (201200)-- _

5.0 (20/250)

3.2 (20/180)

oocooCOOOO

OOOOC2.5 (20/125) 0 0 COO2.00 (201100) 0 0 0 C 01.60(20/80) 0 0 0 0 C1.25(20/63) c 0 0 C o1.00 (20/50) 0 COO C

0.80 (20/40) Co ':, ~;0°

0.63 (20132) "....- 00 • e c

0.50 (20/25) ./ ----------------..

0.40 (20/20) -.../'0.32 (20116)0.25 (20/12.5)

Standard test distance=40 cm (16 inches)

Balley-Lovie DesignLandolt Rings #1

LogMAR and VAR values applyto 40 em (16") test distance

For each halving of the viewing distance. LOW VISIONadd 0.3 to LogMAR or -15 to VAR values. REllOURCE c."",.

www.hksb.org.hk

Figure 7-4

A Landolt ring chart following the Bailey-Lovie design, calibrated for a viewing distance of 40 em. This chartis printed on a card that is 228 mm x 176 mm. (Courtesy of the Low Vision Resource Centre, Hong Kong Societyfor the Blind, Kowloon, Hong Kong.)

progression, each successive step represents a change insize by the ratio 1.2589: 1 (approximately 5:4). A changeof 10 increments on this scale represents a change ofexactly 10 times, and a change of three steps representsa change of approximately two times. With a smallamount of rounding to give more convenient numbers,the sequence progresses as follows: 1.0, 1.25, 1.60, 2.0,2.5, 3.2, 4.0, 5.0, 6.3, 8.0, 10, 12.5, 16, and so on. Fora test distance of 6 m, the sequence becomes 6.0, 7.5,9.5, 12, 15, 19,24,30,38,48, 60, 75, 95, and so on. Amore exact sequence for the logarithmic progression isshown in the second column of Table 7-2. This tablealso shows the approximations as they are usuallyapplied when scoring visual acuity in terms of MAR,decimal notation, or Snellen notation based on 20 feetor 6m.

Letter LegibilityThe Landolt ring target has been recommended by theNAS/NRC Committee on Vision" and by the ConciliumOphthalmologicum Universale" as the reference opto­type against which the legibility of all other optotypesshould be calibrated. It is usually assumed that, forLandolt rings, the gap position is equally detectable forall four alternative orientations of the ring, but the gapposition is slightly more difficult to detect when it islocated in oblique positions. However, the EDTRS chartswith the Sloan optotypes have been so commonly usedin research studies around the world that this chart andits optotypes have effectively become the "gold stan­dard" to which alternatives should be compared.

Typically, clinicians prefer letters as the visual acuitytest targets rather than targets such as Landolt rings or

226 BENJAMIN Borishs Clinical Refraction

tumbling Es, which require the patient to identify ori­entations. When patients are naming orientations, it ismore difficult for the clinician to keep track of whichoptotype is being read at any given instant. It is partic­ularly difficult when the patient skips, repeats, or cor­rects their reading of an optotype or of a whole row. Inaddition, some patients make mistakes when calling leftor right (e.g., gesturing right but calling left). Whenthere are only four possible orientations, the probabil­ity of guessing correctly is relatively high (0.25).Although letters of the alphabet show variability in theirindividual legibility, they do offer many advantages thatappeal to clinicians. The probability of guessing cor­rectly is small, being 1 in 26 for random guessing butgreater if the patient realizes that not all 26 letters areused on the chart. It is easier for clinicians to recall lettersequences and to verify that a row has been read cor­rectly, even when the clinician temporarily loses track ofwhich letters are being read by the patient.

The Sloan letters are most widely used today. The 10five-by-five letters show some small variability in theirindividual legibility.5,6,29,30 When there is significant vari­ation in the relative legibility within a set of optotypes,it is desirable to select the group of optotypes at eachsize level so that each group has approximately the sameaverage difficulty. This has been done for the ETDRSchart with Sloan letters and for the Bailey-Levie chartwith 1968 British standard letters.

Number of Optotypes at Each Size LevelThe reliability of visual acuity measures increaseswith increased number of letters at the near-thresholdsizes.31-33 Doubling the number of letters at each sizelevel should reduce the standard deviation of measure­ments (and, correspondingly, the confidence limits fordetecting change) by a factor of 1/-/2 [i.e., 0.71). Simi­larly, a finer size progression would also improve thereliability of measurement. Provided that the size pro-

gression is not excessively coarse, the reliability ofvisualacuity measurements is inversely proportional to thesquare root of the average 10gMAR value per letter (i.e.,the size progression ratio in log units/number of lettersat each size). This represents the sampling frequency

SO = k.JPTr)

where SO is the standard deviation of visual acuitymeasurement in 10gMAR units; p is the size progressionratio in log units; n is the number of letters at each size;and k is a constant that depends on the optotype andchart design. For five-letter rows and a O.l-log unit sizeprogression, standard deviation of letter chart acuity isabout 0.028. To detect or identify change, it is necessaryto establish confidence limits. This is the range of dif­ferences between test and retest values that, if exceeded,is taken as being caused by a real change rather than theresult of noise in the measurement. The standard devi­ation of test-retest discrepancies is equal to -/2 timesthe standard deviation of the measurement. The 95%confidence limits for change may be taken as 1.96 timesthe standard deviation of test-retest discrepancies. Ifvisual acuity is scored on a letter-by-letter basis, to apply95% confidence limits, the criterion for change shouldbe taken as the next scale increment beyond that whichcontains the 95th percentile. Table 7-3 presents a fewexamples to show how size progression ratios andnumber of letters per size level can affect the standarddeviation of measurement, the standard deviation of thetest-retest discrepancies, and the criterion for change.

Spacing Between Letters and Between RowsSpacing between neighboring letters reduces their legi­bility. Flom and colleagues":" coined the term "contourinteraction" to describe the effect that neighboringspatial contours have on the discriminability of smalldetail. They conducted experiments using Landolt rings

TABLE 7-3 Letter Chart Design and Confidence Limits for Change

StandardNo. of Standard Deviation of Confidence

Progression Letters at LogMAR Deviation of Test-Retest Limits Criterion10gMAR Each Size per Letter Measurement Differences Calculated for Change

0.100 5 0.02 0.028 0.040 0.078 5 letters (0.10 log MAR)0.100 10 0.01 0.020 0.028 0.055 7 letters (0.07 log MAR)0.050 5 0.01 0.020 0.028 0.055 7 letters (0.07 log MAR)0.200 10 0.02 0.028 0.040 0.078 5 letters (0.10 log MAR)0.200 5 0.04 0.040 0.057 0.111 4 letters (0.16 log MAR)

Standard deviations and confidence limits in the first row areall based on empirical measurements.The latter four rows give projections madeaccording to the sampling frequency.Thesampling frequency (/ogMAR/letter) is determined by the size progression and the numberof letters at eachsize.

Visual Acuity Chapter 7 227

with "masking bars" located above, below, to the rightof, and to the left of the ring. They found that thediscrimination of the gap position depended on theseparation of the masking bars from the ring. This isconsistent with the clinical observation that letter legi­bility is increased if the letters are isolated or widely sep­arated from their neighbors. Contour interaction shouldbe distinguished from the "crowding" effect, which isrelated to the difficulty of reading letters caused by therequirement of finer eye movements to read letterswhen they are in a tightly packed array." From experi­ments in which the spacing between optotypes wasvaried from 0.5 to 3.0 times the height of the optotype,Bailey and Raasch11,29 found that a twofold change inspacing altered the visual acuity score by 0.03, 0.04, and0.07 log units for British letters, Sloan letters, andLandolt rings, respectively. For low-contrast (10%Michelson) charts, spacing had little effect on visualacuity scores." Although spacing arrangements within achart may influence visual acuity scores, the choice ofthe spacing ratio is arbitrary. The space between adja­cent rows and between adjacent letters is usually madeequal to the letter width. Visual acuity is better when thespacing is wider. It should be recognized that eye-move­ment control and fixation tremor may contribute to thereduction of visual acuity when the letters are tightlyspaced and that the influence of such motor factors isgreater when the threshold print size is smaller.

CLINICAL TESTING OFVISUAL ACUITY

Chart Formats

Visual acuity charts may be prepared as printed panelsor as slides to be projected onto a screen, or they maybe generated for video display. The chart panel, projec­tion screen, or video screen is often viewed directly, but,when the room dimensions do not permit the desiredtest distance, mirrors may be used to lengthen theoptical path from the chart to the patient.

Printed Panel ChartsPrinted panel charts come in a variety of forms. Manyare printed on opaque card or plastic, and these aredirectly illuminated. Others are printed on translucentmaterial and mounted on a light box that provides illu­mination from the rear (back illumination). The differ­ent print sizes on the chart are usually labeled as thedistance in feet or meters at which the letters subtend 5minarc. Most commonly, panel charts are presented ata distance of 20 feet (or 6 m), and the acuity is recordedas the Snellen fraction. Closer test distances are usedwhen the examination room does not permit chart pres­entation at the standard distance or when the patient

has low vision and is unable to read the largest letterson the chart. To ensure that the patient's resolutionthreshold lies within the range of the chart, the patientshould be able to read the letters at the largest size butunable to read the letters at the smallest size. Cliniciansadopting closer test distances typically choose a distancethat is a simple fraction of the standard, because thisfacilitates the comparison of visual acuity scores. Forexample, 10 feet or 5 feet are preferred close distancesfor charts designed for presentation at 20 feet.

When one is using printed panel charts, the distancefrom the patient to the chart and the size of the lettersmust be known to determine the visual acuity. Withsome charts, however, the print size is labeled notaccording to the letter height but rather as the letters'angular size for a specific test distance. Testing with suchcharts at any distance other than the specific "standard"distance requires an adjustment to the score. However,there is then some risk of error when converting scoresto compensate for the use of a nonstandard test dis­tance. For example, the ETDRS charts" are designed fora 4-m presentation distance, and the top row is labeled"20/200," although its letters subtend 5 minarc at 40 m(131 feet) rather than at 200 feet or 60 m. Reading thistop row at 4 m should earn a score of 4/40 (or, in impe­rial units, 13.1/131), which, in angular terms, is equiv­alent to 20/200 (6/60). If the chart is moved to, forexample, 1 m (3.3 feet) and a patient can just read thetop row (40-m letters labeled 20/200), the clinicianmight erroneously assign an acuity score of 3.3/200(1/60), and this would be considered equivalent to20/1200 (6/360). In this example, however, it wouldhave been correct to record the visual acuity score as3.3/131 (1/40), and this can be considered equivalentto 20/800 (6/240).

Although 20 feet or 6 m is the most widely used testdistance, 4 m has been recommended by Hofstetter"and, subsequently, by some authoritative bodies.":" A4-m test distance facilitates making a dioptric allowance(of 0.25 D) to the refractive correction to allow for thechart being closer than optical infinity. Also, using 4 mas the standard for testing distance vision facilitatescomparison with near-vision measurements, in which40 cm is commonly used as a standard test distance.

Projector ChartsIf the projector lens and the patient's eye are equallydistant from the projection screen, the angular size ofthe chart and its component optotypes of the projectorchart image are independent of the observation dis­tance. Consequently, the designation of print size onprojector charts is usually in angular terms. The equiv­alent Snellen fraction is used on most American charts,and decimal acuity notation is used on European pro­jector charts. If the viewing conditions are arranged sothat observation distance is 20 feet (or 6 m) and the

228 BEN.lAMIN Borishs Clinical Refraction

projector is appropriately positioned with respect to thescreen, the expression of visual acuity as a Snellen frac­tion is straightforward. If, however, the optical pathlength from the patient to the screen is some other dis­tance, a proportional change needs to be made to thesize of the projected letters. For example, if 18 feet (5.4m) was the observation distance, the projector systemshould be adjusted so that the row designated as"20/200" has the height of its letters subtending 50minarc at 18 feet (5.4 m); then all other letters' sizeswould be scaled proportionately. If this row is the small­est that the patient can read, the visual acuity as aSnellen fraction should strictly be expressed as 18/180(or 5.4/54); however, it is usual to record such a visualacuity result as 20/200 (6/60), which is the equivalentSnellen fraction. Projector chart systems are usuallyarranged so that the distance from the patient to thescreen is never varied.

The angular width of chart displays for the commonclinical projectors is usually about 2.5 degrees square,and this limits the number of letters that can be dis­played in a single row at the larger sizes. If 11 characterspaces are allowed to display a row of five characters,then the largest presentable row on a 2.5-degree displayhas a visual acuity value of about 20/55 (6/16 or10gMAR = 0.43). The largest angular size available inmost projectors is 20/400 (6/120), and typically onlyone letter of this size can be presented per display. Stan­dard 35-mm slide projectors can present wider fields,and they can readily allow five letters per row up to the20/200 (6/60) level.

Charts on Display ScreensComputer-generated displays are not yet widely used inclinical practice, but they offer distinct advantages. Theyprovide the means to select different optotypes, tochange letter sequences, and to vary stimulus para­meters such as contrast, spacing arrangements, andpresentation time. The computer interface providesopportunities for more detailed recording and analysisof responses. Computer-controlled presentation of testtargets facilitates repeated measurements with randomor semi-random rearrangements of letter sets. Thisprocess avoids some of the memorization problems thatcan occur when using printed or projected charts. Thereare some brightness limitations in that the luminancelevels on cathode-ray tube displays are typically lessthan 150 cd/rrr', but some newer cathode-ray tubemodels and many flat-panel displays provide screenluminance of up to 300 cd/rrr', The sizes of the display'spixels and of the screen itself impose limits on theextreme sizes (small and large) of optotypes and chartsthat can be presented. The pixel structure limits the sizeof the smallest letters, and the screen dimensions limitthe size of the largest letters that can be presented in arow or singly.

At least 20 pixels are required per letter height so thatthe spatial structure or shapes of individual optotypesdo not show significant variation from one size to thenext. Even with this minimal number of pixels, somecompromise must be accepted. Consider Landolt Ringsor tumbling E optotypes. To maintain a 5:1 ratiobetween the height of the optotype and stroke or gapwidth, the number of pixels per letter height must be aninteger multiplied by 5. If the usual logarithmic pro­gression of size is to be preserved, the limb or gapwidth should be incremented in accordance with thefollowing sequence: 1.0, 1.25, 1.6,2.0,2.5,3.2,4.0, 5.0,6.3, 8.0, 10, and so on. The numbers in this sequenceare not all integers and multiples of 5, and so anoptotype will not appear the same when presented indifferent sizes. For example, if the very smallest lettershad 20 pixels per letter height, the limb or gapwidth would be 4 pixels; the next larger letters would be25 pixels high, with a 5-pixel limb width. At the nextlargest size, a letter must be 30 or 35 pixels high, respec­tively, to achieve the correct proportions with a limbwidth of 6 or 7 pixels. However, 32 pixels are requiredto achieve the desired logarithmic size progression ratioindicating the proper level of acuity. At 32 pixels, forinstance, there will be three limbs or spaces of the tum­bling E having a width of 6 pixels and two limbs orspaces composed of 7 pixels. The dilemma is that thechart can supply the proper letter size or the properdetail proportion-but not both at the same time-forthis acuity level.

Today, it is common for display screens to have ascreen resolution of 1600 x 1200 pixels (UXGA),and finer resolutions are available (QXGA = 2048 x1536; QSXGA = 2560 x 2048). Consider the sizerange that could be presented in the Bailey-Lovieformat on a UXGA screen. If the smallest letters were 20pixels high and if they were to subtend 2.5 minarc(20/10, 6/3, or 4/2), the vertical height of the screenwould necessarily subtend an angle of 2.5 degrees (1200pixels, each 0.125 minarc). This would require a screenheight of 17.5 cm at 4.0 meters. If there were to befive letters on each row, with the space aroundeach letter equal to one letter width, then 11 characterspaces would be required for each row of letters.Restricted by the horizontal screen dimension of 1600pixels, the largest characters could be 145 x 145 pixels.If the largest characters were to subtend angles of 50minarc (20/200, 6/60, or 4/40), the screen would needto horizontally subtend an angle of 9.2 degrees. Thus,the screen would need to be 65-cm wide at a viewingdistance of 4 m. Even with two screens (i.e., a largescreen for the large sizes and a small screen for thesmaller sizes), it would be impractical to have a contin­uous chart that maintained a uniform format, and therewould need to be some overlap of the size ranges of thetwo charts.

Visual Acuity Chapter 7 229

Chart Luminance

For most purposes, visual acuity measurements aremade with the visual acuity chart at moderate photopicluminances, and, typically, the general room lighting issubdued. Recommendations for a standardized chartluminance range from 85 to 300 cd/rrr'. Sheedy and col­leagues" showed that, in this luminance range, dou­bling the luminance changes the visual acuity score byabout 0.02 log units (1 VAR unit), which correspondsto one-fifth of a line or a 5% change in MAR. A com­promise chart luminance that is becoming widely usedas a standard is 160 cd/rrr'. The British standard requiresa luminance of at least 120 cd/rrr', It can be difficult toachieve specific luminance levels with different projec­tor, light box, and video display systems, so a clinicaltolerance of 80 to 320 cd/rrr' for test chart luminancemay be reasonable and practical. To better ensure meas­urement consistency within a given clinical setting oramong sites in a clinical study, the chosen luminanceshould be maintained within a 15% tolerance. Whenilluminating charts, one should take care to avoid glaresources within the patient's field of view. The visual per­formance of certain patients-particularly those withretinal pathology-may be considerably influenced byretinal illumination. The clinician may choose to varythe chart luminance to find the patient's specific light­ing dependency.

Contrast is another variable that affects visual acuity.Measurement of visual acuity with low contrast (gray)optotypes is becoming more widely used, mainly forpatients with corneal or lenticular disorders or for thosewho have had refractive surgery. Low-contrast visualacuity and its difference from high-contrast visual acuityare often regarded as measures of contrast sensitivity.The reader is referred to Chapter 8 for a detailed dis­cussion of contrast sensitivity and its assessment.

Refractive Correction

During an eye examination, clinicians are frequentlyconsidering whether to recommend that spectacles orcontact lenses be worn or whether changes should bemade to the corrective lenses that the patient is currentlywearing. In recent times, refractive surgery has becomepart of the range of interventions that may be consi­dered by the clinician and the patient. Visual acuitymeasurements guide the clinician's decisions and rec­ommendations about these various options for treatingrefractive errors. The acuity measurements of most rele­vance are the visual acuity that may be obtained withthe best spectacle or contact lens correction, the visualacuity obtained when no spectacles or contact lenses arebeing worn, and the visual acuity measured with therefractive correction that the patient usually wears whileperforming common distance vision tasks of daily life.The increasing use of surgical treatments of refractive

error and the expected development of methods tocorrect higher-order optical aberrations have created theneed to modify some of the terminology used whenreferring to different kinds of visual acuity measure­ments.

Unaided visual acuity is defined as visual acuity meas­ured without any spectacles or contact lenses (i.e., with"lenses off"). It can apply to eyes that have had refrac­tive surgery and those that have not. The unaided visualacuity becomes a benchmark against which the benefitsof using a refractive correction may be referred. Caremust be taken to ensure that the patient does not squintor narrow the palpebral aperture to reduce the blurcreated by defocus or optical irregularity. Unaided acuityis relevant when predicting how well or how poorlypatients can see if deprived of access to their refractivecorrection. Dimming the ambient illumination causespupil dilation, which is likely to reduce the uncorrectedvisual acuity when there is uncorrected refractive erroror optical irregularity.

In the past, the term uncorrected visual acuityhad beenwidely used to mean the same thing as unaided visualacuity; however, when refractive error has been cor­rected by refractive surgery, the term uncorrected visualacuity literally means the visual acuity without spectacleor contact lenses before the surgical intervention. Afterthe surgery, vision is of course no longer uncorrected.This type of unaided acuity must be measured beforethe surgery; if recorded, it can be a useful referenceagainst which the visual acuity benefits of the surgerymay be quantified.

Habitual visual acuity is defined as the visual acuitymeasured under the refractive conditions that thepatient habitually uses when performing distance visiontasks of daily life. Whether or not the current spectacles,contact lenses, or postsurgical refractive status areoptimal corrections of the refractive error is irrelevant.The question is simply, "What is the visual acuity habit­ually being obtained by the patient?" The habitual visualacuity becomes a benchmark against which the benefitsof changing the refractive correction may be compared.

For patients who do not usually wear eyeglasses orcontact lenses for distance vision, the habitual visualacuity is simply the unaided visual acuity. Often theoptical corrections being worn will be ideal or close toideal, but it is not uncommon for patients to be wearingoptical corrections that are distinctly inappropriate;these may include old corrections prescribed many yearsago or spectacles obtained at a flea market, from arelative, or over the counter. Sometimes patients will behabitually combining two or more different means ofrefractive correction. For example, spectacles may beused over contact lenses to correct residual astigmatism,and spectacles may be used to provide an additionalimprovement in visual acuity after refractive surgery.Many persons who use monovision created by contact

230 BENJAMIN Borishs Clinical Refraction

lenses, refractive surgery, or natural anisometropia willchoose to wear spectacles for tasks such as driving andwatching television.

Corrected visual acuity is defined as the visual acuityobtained with the patient wearing the best availablerefractive correction obtained by conventional spectaclelenses or contact lenses (i.e.. with "lenses on"). The bestavailable refractive correction is usually established bydetermining the best spherocylindrical spectacle correc­tion over and above any other refractive correction thatmay be present, such as contact lenses or refractivesurgery. Thus, the corrected visual acuity could beobtained with a full optical correction in the form of aspectacle lens or with the combination of a spectaclelens over a contact lens. In the case of refractive surgery,a spectacle correction over the surgically modified eye­and perhaps even in combination with a contact lens­could provide the best available correction.

If there are no significant optical irregularities oropacities, the corrected visual acuity indicates the bestresolution achievable by the patient's visual system. Inthe presence of corneal surface irregularities (e.g., kera­toconus, traumatically induced corneal distortion, somecases of distortion resulting from refractive surgery), aspectacle correction over a rigid contact lens might benecessary to provide the best measure of corrected visualacuity (see Chapter 34).

The corrected visual acuity provides a benchmark orreference for determining whether visual acuity haschanged as a result of disorders affecting the optical orneural components of the visual system. Changes in cor­rected visual acuity can be critically important whenmaking diagnoses, when determining whether additionalvision loss has occurred, and when deciding whetherchanges should be made to eye disease treatments.

During the past decade, there have been significantadvances in the technology required to measure andcorrect optical aberrations of the eye (see Chapter 19).There is some promise that optimizing the quality of theretinal image may lead to measurable and significantimprovements in the visual acuity as compared with thatwhich can be obtained through the use of conventionalspectacle lenses and contact lenses. Optimal control ofaberration might be achievable by surgical shaping of thecorneal surface or through having individualized aspheric­ity built into the surface configurations of the patient'sspectacle lenses, contact lenses, or intraocular lenses.

Optimal visual acuity can be defined as the visualacuity that will be obtained when the optical quality ofthe retinal image is optimized. It is a long establishedand common clinical practice to use the term "best-cor­rected visual acuity" to mean the acuity that is obtainedwith best correction in the form of conventional sphe­rocylindrical lenses. Thanks to technological advancesin the control of higher-order aberrations, the conceptofwhat is "best correction" is changing. Although "best"

and "optimal" may be synonyms, it seems appropriateto use "optimal visual acuity" to refer to the very bestpossible visual acuity and to possibly discourage thefuture use of the term "best-corrected visual acuity" toavoid confusion between the old and new meanings of"best correction." The term "best-corrected spectacleacuity" is now sometimes used to indicate the bestacuity derived from the wear of the spectacle refraction.

Pinhole acuity refers to visual acuity measured usingpinhole apertures (usually having a diameter of 1.0-1.5mm) placed before the patient's eye to determinewhether a reduced visual acuity is a result of opticaldefects. The pinhole increases the depth of focus so thatthe blur created by optical irregularities or refractiveerror becomes reduced; consequently, visual acuityimproves. Pinhole tests are used when the best correctedvisual acuity is poorer than expected or when there isreason to suspect optical irregularities. A pinhole can beexpected to improve visual acuity in patients withkeratoconus or with cortical or posterior subcapsularcataracts, because it can channel light through a betterregion of the eye's optics. Defocus and optical irregu­larities become less important as depth of focus isincreased. However, a pinhole should not have any sig­nificant impact on visual acuity that is reduced becauseof amblyopia or some retinal disorders. The pinholedoes reduce the illuminance of the retinal image;through this mechanism, the pinhole may sometimesreduce visual acuity, especially in patients with retinaldiseases that make visual performance particularly sen­sitive to changes in retinal illumination.

The Potential Acuity Meter (PAM) by Marco Oph­thalmic Instruments (Jacksonville, Fla) is an instrumentthat presents an image of a visual acuity chart to the eyeusing a Maxwellian view optical system. Maxwellianview, explained in Chapter 1, confines the beam enter­ing the eye to an area that is smaller than 1 mm at theplane of the pupil. In theory, using the PAM is similarto using a pinhole, except there is better control of theretinal illumination."

Visual acuity measurement under special illuminationconditions may be indicated to evaluate the potentialfunctional difficulties that depend on illumination con­ditions. Bright conditions cause pupil constriction, andthis may have adverse effects on visual acuity in casesof centrally located optical opacities or irregularities.On the other hand, more peripherally located opticaldefects-as often occur after refractive surgery-mightcause a visual acuity reduction when illumination isreduced and the pupils dilate to expose the regions ofoptical irregularity. In some patients (especially thosewith retinal disease), visual performance may bestrongly affected by retinal illuminance, and the clini­cian may choose to vary the chart luminance over a widerange to identify and quantify the patient's specific light­ing dependencies.

Visual Acuity Chapter 7 231

Testing Distance

In a given clinical setting, a standard testing distance isestablished. It may be a specific distance such as 20 feet,6 m, or 4 m to facilitate scoring in Snellen notation.With projector displays, the test distance is usuallychosen according to spatial constraints of the examina­tion room. Most examination rooms are too short toallow a direct observation path of 20 feet (6 m), somirrors are used in both the projection and observationpaths to achieve longer testing distances. In mostcircumstances, the test distance is close to 20 feet (6 m).Although variations from the "standard" are notuncommon, they rarely fall outside of a 10- to 30-footrange.

If patients cannot read all letters at the largest angularsize available at the standard distance, a shorter test dis­tance should be used. Short test distances are mosteasily achieved using printed panel charts. For patientswith very low visual acuity, close distances such as 5 feet,1 m, 1 foot, and 40 cm might be considered. Whenpatients cannot read the largest letters available on aprojected chart, a printed panel chart should be used.

When close test distances are used, it may be neces­sary to modify the refractive correction by adding theappropriate plus lens power to ensure optimal focus onthe retina. If a plus lens is used to cause the image ofthe chart to be at optical infinity, some prepresbyopicpatients might not achieve their best possible acuity,because proximal accommodation may create somedefocus. Proximal accommodation is only of potentialsignificance when testing patients with good visualacuity at distances closer than 10 feet. If short viewingdistances are adopted to enable the testing of patientswith very low vision, it is not usual to make any refrac­tive compensation; modest levels of defocus are notlikely to affect the legibility of their threshold-sizeletters, because they are so large in angular size.

Testing Procedure

Monocular visual acuities are tested with one eyeviewing the test chart while an occluder is placed beforethe other eye. If the hand of the patient or the clinicianis being used to occlude the other eye, care should betaken to use the palm, because otherwise the patientmight look through a narrow gap between the fingers.Usual practice is to measure the right eye first, but theleft eye might occasionally be measured first if it isknown that the patient has poorer vision in that eye. Atalmost every eye examination, the visual acuity of theright eye (00) and the left eye (OS) are measured sep­arately. Typically, clinicians measure the binocular (OU)visual acuity as well. This is measured with both eyesopen, and it is usually expected that the binocular visualacuity will be marginally better than-or at least equalto-the visual acuity of the better eye. Rarely is the

binocular visual acuity poorer than the better of the twomonocular acuities. This may happen in some cases ofbinocular vision disorders, nystagmus, or metamor­phopsia, and it can occur in monovision when thepatient is unable to alternate central suppression fromone eye to the other (see Chapter 28).

Some clinicians ask patients to read from the largestletters at the top of the chart through to the smallest thatcan be read. More commonly, the patient is asked tobegin reading at a size level that is expected to be a littlelarger than patient's resolution limit. For example, apatient expected to have an acuity of 20/20 or bettermight be asked to begin at the 20/40 level. The patientis instructed to read down the chart as far as possible.There is often a hesitancy that indicates that the patientis earnestly struggling to read as many letters as possi­ble. For clinical testing, it is common practice to ignorean occasional error if all letters at the next smallest sizeare read correctly. When reading letters at sizes close tothreshold, the patient should be encouraged to guess.One widely used rule is that, if patients correctly iden­tify 50% or more of the letters correct at a given size(e.g., three of five letters), they should be obliged toguess the remaining letters at that size level and then toguess at all letters at the next smallest size. Carkeet"modeled visual acuity responses and supported guess­ing when 40% or more of the letters were read correctlyat the previous size level. When there was a high prob­ability of guessing correctly (e.g., as with four-positionLandolt rings), a more stringent 20% criterion (i.e., oneout of five) was deemed appropriate. To ensure that thepatient has been tested at sizes both larger and smallerthan the threshold size, all optotypes at a larger sizeshould be read correctly, and no optotypes at thesmallest size should be read at all.

Special problems sometimes arise in patients withdisorders affecting macular function. Patients withmacular scotomas may miss letters at many differentsize levels, and patients with amblyopia may behavesimilarly. There may be a tendency to completely missletters at the start or end of rows. Some patients appearto have to search for individual letters, and they mayname the letters out ofsequence. The clinician may helpsuch patients keep their bearings by pointing to indi­vidual letters. Eccentric viewing helps some patientswith macular scotomas achieve better visual acuityscores. The clinician may encourage eccentric viewing byhaving the patient look above, below, to the right, andto the left of the letters being read; this may improvevisual acuity performance. Patients with amblyopia ormacular disorders are likely to achieve better resolutionif presented with isolated single letters rather than aseries of letters in a row or chart.

Flip charts are sometimes used to isolate differentsize levels and to facilitate the isolation of individualletters. With a view to expediting testing, Rosser"

232 BENJAMIN Borishs Clinical Refraction

designed charts with an abbreviated size range andfewer letters per row. Camparini and colleagues" rec­ommended having the patient read only the first letterin each row until difficulties or errors were encountered.These quick testing procedures will generally reduce thereliability and validity of the test results.

Assigning Visual Acuity Scores

Row-by-Row ScoringUnfortunately, it is a common practice to assign a visualacuity score on a row-by-row basis. The visual acuityscore records the smallest size at which at least a spe­cific proportion (typically 50%, but up to 80%) of allof the letters of that size are correctly identified. The pos­sible scores correspond with the size levels on the chart.Scoring row by row is too coarse to reliably detect smallchanges in visual acuity. For example, when using achart with five letters per row, a one-row change invisual acuity score could be caused by as little as a one­letter difference or as much as one letter short of twofull rows. With row-by-row scoring, the visual acuityscore must change by at least two size levels for clini­cians to be confident that there has been a significantchange." Despite its relative insensitivity, this is themethod that remains the most widely used by eye-carepractitioners.

Many clinicians do give partial credit, qualifying avisual acuity score by adding plus or minus signs toindicate that the patient actually did a little better or alittle worse than the performance indicated by thenumerical value recorded. A patient reading all letters inthe 20/25 (6/7.5) row and correctly identifying twoletters on the 20/20 row could be given a score of20/25+2 (6/7.5+2).

Letter-by-Letter ScoringGiving credit for every letter read provides more sensi­tivity for the detection of changes in acuity. Cliniciansmay record a visual acuity score followed by a plus signwith a number to indicate the number of letters read atthe next smallest size or a minus sign with a number toindicate the number of letters missed at that size level:for example, 20/25+2,20/25- 1

, 20/30-1,+2. If the chart hasthe same number of letters on each row, the qualifiers(e.g.,-2,+1,-1,+2) carry the same value at all levels of thechart. By giving credit for every letter, 20/25+1 can beconsidered equivalent to 20/25-1,+2. If the number ofletters at the different size level varies throughout thechart, the weight given to the qualifying numberdepends on the specific number of letters in the rowsconcerned.

If visual acuity is being recorded in logarithmic units(logMAR or VAR), each letter can be assigned a valuethat is added to the score when that letter is read cor­rectly. On charts with five letters per row and a size pro-

gression of 0.10 log units, each letter can be assigned avalue of 0.02 10gMAR units. For each additional letterread, 0.02 is deducted from the 10gMAR score. Similarly,scoring in VAR units gives a value of one point per letterso that each extra letter read adds one extra point to thescore. Table 7-4 provides three examples of how letter­by-letter scoring can be used to give scores in terms of10gMAR, VAR, or Snellen fractions with qualifiers. Forthese three examples, it has been assumed that the chartcomplies with the Bailey-Lovie design principles so thateach letter carries equal value. Table 7-5 also shows howcredit can be assigned for each letter, even when chartsdo not have a regular size progression and the numberof letters varies per row, Then, for each size level, theper-letter value in 10gMAR or VAR units is determinedby subtracting the 10gMAR or VAR values for that rowfrom that of the proceeding row and dividing this dif­ference by the number of letters,

Visual Acuity Measurement in Research

In many research projects involving visual acuity meas­urement, visual acuity tests are likely to be administeredfrequently, and more sampling is likely to be required.When only one or a small number of charts is available,there can be problems caused by patients memorizingletter sequences, particularly in the threshold region.There are several ways to reduce or eliminate thisproblem; these include having more charts available orusing modest variations in the test distance so that thepatient's resolution threshold moves to a new region ofthe chart. For charts that use British or Sloan letters withfive letters by row, chart pairs can be designed so thatthere is no replication of any of the 10 letters at each ofthe size levels. This allows for the presentation of 10letters at each size without there being any repeats ofletters that might be more difficult or easier for thepatient. Computer generation of new letter sequencesprovides a good solution to problems that result frompatients memorizing letters or sequences.

When visual acuity is being measured for researchpurposes, it is important to have the testing conditionsand procedures rigidly defined. Standard refraction pro­cedures may be required. Testing conditions such aschart luminance and contrast, viewing distance, and cri­teria for changing to alternate viewing distances shouldbe specified. There should be standard instructions foradvising the patient that the chart contains letters only,that all letters should be attempted, reading should beat a steady pace, and that guessing is permitted. Theexaminer should not point to individual letters or rows,all errors should be recorded, and patients may notmake correct a response once the next letter has beenread. Procedures to encourage guessing and rules forstopping should be applied when the visual acuitythreshold is approached.

TA

BL

E7

-4le

tte

r-b

y-l

ett

er

Sco

rin

go

fV

isu

al

Acu

ity

inU

nit

so

flo

gM

AR

,V

AR

,a

nd

Sn

elle

nF

racti

on

sfo

ra

nE

TD

RS

or

Ba

ile

y-l

ovie

Ch

art

SIZ

ELA

BE

LSO

NC

HA

RT

PA

TIE

NT

AP

AT

IEN

TB

PA

TIE

NT

C

Lo

gM

AR

VA

RS

ne

llen

Sn

elle

nN

o.

corr

ect

Lo

gM

AR

VA

RN

o.

corr

ect

Lo

gM

AR

VA

RN

o.

corr

ect

Lo

gM

AR

VA

Rr-

r-

0.4

080

20

/50

6/1

55

of

50

.40

805

0f5

0.4

080

5o

f5

0040

80

0.30

852

0/4

06

/12

5o

f5

0.30

855

0f5

0.3

085

5o

f5

0.3

085

0.20

902

0/3

26

/9.5

50

f50.

2090

4o

f5

0.2

289

5o

f5

0.2

090

0.10

9520

/25

6/7

.51

of

50

.18

912

of

50

.18

915

of

50.

1095

0.0

010

02

0/2

06

/60

0.1

891

00

.18

914

of

50.

0299

-0.1

010

52

0/1

66

/4.8

00

.18

910

0.1

891

3o

f5

-0.0

410

2

log

MA

R=

VA

R=

log

MA

R=

VA

R=

.lo

gM

AR

=V

AR

=

Eac

had

diti

onal

lett

erad

ds-0

.02

_~1~__

____~_~l

____

0.1

891

-0.0

410

2L

__

__

__

__

__

_L

__

__

__

__

L_

__

__

__

__

__

__

__

__

__

toth

eL

ogM

AR

scor

ean

d1

toS

nell

en=

20

/3rl

Sne

llen

=2

0/3

2-1,

+2

Sne

llen

=2

0/2

0-1 ,+

3

the

VA

Rsc

ore.

or

20/3

2+1

or

20/2

0+2

Thr

eeex

ampl

esof

scor

in}(

visu

alac

uity

lette

r-by

-let

ter

ona

char

twit

hjil

1ele

tters

per

row

and

a0.

1lo

gun

itsi

zepr

ogre

ssio

n.

TA

BL

E7

-5le

tte

r-b

y-l

ett

er

Sco

rin

go

fV

isu

al

Acu

ity

inU

nit

so

flo

gM

AR

,V

AR

,a

nd

Sn

elle

nF

racti

on

sfo

ra

"Sta

nd

ard

"S

ne

llen

Ch

art

SIZ

ELA

BE

LSO

NC

HA

RT

No

.o

fle

tte

rsL

og

MA

RV

AR

PA

TIE

NT

XP

AT

IEN

TY

Sn

elle

nS

ne

llen

log

MA

RV

AR

pe

rR

ow

pe

rle

tte

rp

er

Le

tte

rN

o.

Co

rre

ctL

og

MA

RV

AR

No

.C

orr

ect

log

MA

RV

AR

20

/20

06

/60

1.00

501

I1

of

11.

0050

1o

f1

1.00

I-

-50

I

20/1

006

/30

0.70

652

I-0

.15

7.5

1o

f2

0.85

582

of2

0.70

65I

20/7

06/

210.

5473

3I

-0.0

52.

6o

of3

1o

f3

0.65

68I

--

20

/50

6/1

5I

oo

f4

0.40

804

I-0

.04

1.8

--

oo

f4

--

log

MA

R=

VA

R=

log

MA

R=

VA

R=

Eac

had

diti

onal

lett

erca

rrie

sa

valu

eth

atde

pend

son

0.85

580.

6568

L_

__

__

__

__

_~

__

__

__

_L

__

__

__

__

__

_~

__

__

__

_

the

num

ber

ofle

tters

inth

atro

wan

don

the

incr

emen

tofs

ize,

Sne

llen

=20

/100

-1S

nell

en=

20/1

00+

1

or

20/2

00+

1

1lvo

exam

ples

ofsc

orin

gvi

sual

acui

tyle

tter-

by-l

ette

rus

ing

atr

adit

iona

lSne

llen

char

t.

234 BEN.lAMIN" Borishs Clinical Refraction

S-ChartsThe S-chart test is a Landolt ring test designed by Flomand colleagues":" for clinical research. The S-chart testconsists of a series of 35-mm slides. Each slide containsa panel of 25 symbols arranged in a five-by-five squaregrid. The 16 external symbols and the central symbolare all tumbling Es in randomized orientations.Forming a square around the central E is a series of eightLandolt rings, with their gap orientations (right, left, up,down) arranged randomly. The observer's task is tobegin with the top left Landolt ring and progress in aclockwise direction, naming the location of the gap foreach of the eight Landolt rings. Each 35-mm slide pres­ents a new size. Originally the size progression followedthe Snell-Sterling visual efficiency scale, but later alogarithmic size progression was used. Visual acuityscore was determined by fitting a sigmoid or S-shapedpsychometric function to a plot of percent letters correctto letter size. Probit analysis methods achieve the sameresult. With four alternative gap orientations, chanceperformance is represented by 25% correct. Conse­quently, the print size that allows 62.5% of letters to beread correctly represents the 50% probability of seeingpoint. The 50% probability of seeing point is commonlytaken as the threshold for the visual acuity value.

PEDIATRIC TESTS OF VISUAL ACUITY

A wide variety of visual acuity tests are available formeasuring visual acuity in infants, toddlers, and otherswho have a limited ability to respond to standard teststimuli." The clinician selects a visual acuity test that isappropriate given the response capability of the patient.There is a hierarchy of tests." ranging from visuallyevoked potentials for the least responsive patients andprogressing to techniques of preferential looking; obser­vation of optokinetic nystagmus; and responses topicture flash card tests, picture charts, symbol and letterflash cards, symbol and letter charts, and reading charts.

Grating Acuity Tests

Striped or checkered grating targets are used in some ofthe more elementary tests ofvisual resolution in infants.The size of the detail is varied to determine the finestpattern that can elicit a response from the infant. Thepatient's responses may be determined by objective orsubjective means. The angular size of the detail withinthis pattern is expressed as cycles per degree, and it istaken as the visual acuity. Within a grating pattern, onecycle embraces a dark and a light stripe. An equationused to covert grating acuity score (cpd) to MAR is asfollows:

MAR 30/cpd

For a 30-cpd grating, each cycle is one-thirtieth of adegree, so one cycle is 2 minarc. Each dark and eachlight stripe is 1 minarc. Thus, a 30-cpd grating has anominal equivalence to 20/20 or 6/6 (MAR= 1 minarc),a 3-cpd grating is equivalent to 20/200 or 6/60 (MAR=10 minarc), and so forth.

Visually Evoked Potential TestsVisually evoked potential tests involve measuring elec­trical potentials from the back of the head as the patientlooks toward a screen on which flickering striped orcheckered patterns are presented. The magnitude of theelectrical response declines as the detail within thepattern is made finer. The size of the detail (expressedas spatial frequency in cycles per degree) in the finesttarget that evokes a measurable response is taken as thevisual acuity."

Preferential Looking TestsThese tests evolved from research procedures developedby Dobson and Teller46 and McDonald and coworkers"for studying the development of vision in infants. Theinfant is presented with two target areas: one contain­ing a black-and-white spatial pattern, the other con­taining uniform gray. The clinical adaptation of thelaboratory technique uses Teller Acuity Cards." whichare large, rectangular, gray cards with a black-and-whitestriped grating pattern off to one side. When the card ispresented to the patient, the clinician observes whetherthe patient (usually an infant) moves the eyes to fixateto the side with the grating. Within a test card series,there is a progression of spatial frequency (fineness ofgrating), and the task of the clinician is to determine thespatial frequency of the finest grating that still attractsthe patient's attention.

Cardiff cards" are somewhat similar in concept tothe Teller cards. Each Cardiff card presents a linedrawing of an object. The picture is formed by a linethat consists of a central white line with finer blackflanking lines on either side. The luminance averagedacross the black-white-black line matches the lumi­nance of the gray background. Consequently, when thelines are too fine to be individually resolved, theybecome indistinguishable from the gray of the back­ground. The clinician determines the finest line drawingthat still attracts the child's attention.

Optokinetic NystagmusStriped patterns are presented on a video screen, on arotating drum, or by other methods, and they are movedin one direction in front of the patient. If the stripedpattern is visible, the patient's eyes will make "railroadnystagmus" eye movements as they follow the move­ment of the stripes. The clinician determines the size ofthe finest grating having motion that elicits the nystag­mus response when it is moving.

Flash Card Tests

Flash cards with pictures or symbols as targets can beused for patients who have some ability to respond toinstructions. Patients may be asked to point to or toname pictures or symbols on flash cards, to find a match­ing target, or to play other matching games. At the sim­plest end of the range of such tests is the Bailey-Hallcereal test." in which pairs of flash cards are presented.One card of the pair shows a picture of a cereal ring(Cheerio), whereas the comparison card simply has asquare. The patient is asked to identify which card hasthe picture of the cereal ring. On successful identifica­tion, the patient is rewarded with a piece of cereal to eat.The size of the picture of the cereal ring and the furthestdistance at which it can be recognized provide the basisfor the visual acuity estimate. More advanced testsinvolve the use ofalternative pictures or symbols that areto be identified. The LH symbols" are an example ofsuch optotypes. The four alternative LH symbols, towhich patients can apply their own name, are most com­monly called square, apple, circle, and house. Otherflash card series similarly call for identifications ormatching responses; these include Lighthouse flashcards (umbrella, apple, house); the broken wheel test,"which requires distinguishing a car with wheels shownas Landolt rings; and the Allen picture cards, whichcontain a series of simple line drawings of easily namedobjects.

Letter Flash CardsThere are several alternative series of letter flash cardsfor which the patient is required to name the target letteror letters or to make a match':" Some use selectedletters, including the mirror-reversible H, 0, T, and V;others use tumbling E or Landolt ring optotypes forwhich the patient is required to identify the orientationof the symbol. The HOTVset of optotypes is used in thecomputerized display of the Baylor-Visual Acuity Tester(B-VAT) by Mentor 0&0 (Norwell, Mass}."

Picture or Symbol Charts

Picture or symbol charts present an array of simpledrawings of objects or simple symbols of progressivelydecreasing sizes, and the targets are to be named bythe patient. Such charts are available with LH symbolsthat include different contrasts to assess contrast sensi­tivity; others have different spacing arrangements,which are used to identify problems with symbolcrowding.

Letter Charts for ChildrenIf the child is capable of making the appropriateresponses, it is preferable to use adult charts withletters or equivalent optotypes arranged in the usualchart format. Similarly, if the child is capable of reading,

Visual Acuity 0 Chapter 7 235

there may be value in performing reading acuity testsusing typeset materials as the test target. Some specialseries of reading tests using simpler words have beendeveloped for children (e.g., the Sloan reading cardseries).53

NEAR VISUAL ACUITY

Near visual acuity is measured at distances within arm'slength. A testing distance of 40 cm is usually consideredto be the standard. If the test chart design and the lumi­nance levels are comparable, the near visual acuity scoreshould be equal to the score of distance visual acuity,provided that the eye is accommodated or optically cor­rected to provide good focus for the retinal image.However, there are rare exceptions, such as patients withposterior subcapsular cataract whose pupil constrictionat near-vision tasks causes the pupil area to becomemore completely filled with the cataract so that thevisual acuity becomes degraded. Most of the tests ofnear visual acuity do not use letter chart formats that arecomparable with the charts used for testing at distance.Usually, the near vision tests use typeset material that issimilar in style to the print of newspapers and books.The material may be arranged in sentences or para­graphs or in series of unrelated words.

Designation of Near Visual Acuity

Specification of near visual acuity usually includes therecording of both the observation distance and the sizeof the smallest print that can be read. Several differentmethods are used to specify the size of print in near­vision tests; the relationships among them can be seenin Table 7-2.

M UnitsM units are a measure of print size introduced by Sloanand Habel. 54 They are used to specify the size of printby indicating the distance in meters at which the heightof the smaller letters (the lowercase x-height of typesetprint) of the printed material sub tends S minarc. Printthat is I.O-M units subtends S minarc at I m; accord­ingly, it is IAS-mm high. Regular newsprint is usuallyabout 1.0 M in size. Visual acuity may easily be recordedas a Snellen fraction in which the clinician records thetest distance in meters in the numerator, and thedenominator indicates the M-unit size of the smallestprint that can be read at that distance. A patient whocan just read I.O-M print at 40 cm would have his or hervisual acuity recorded as 0040/1.0 M. Jose and Atcher­son" pointed out that the M-unit rating of a sample ofprint can easily be estimated by measuring the height ofthe smallest letters in millimeters and multiplying thisnumber by 0.7.

236 BENJAMIN Borishs Clinical Refraction

PointsPoints are units used to specify the size of typeset printand are used in the printing industry; one point is equalto Y72 of an inch. The point size of a specimen of printessentially indicates the size of the print extending fromthe bottom of the descenders (as in letters g, j, p. q. y)to the top of ascenders (b. d, f, i. j, k, 1, t). For print stylesthat are most commonly used for newspaper text, theheight of the smaller lowercase letters (a, c, e, m, n, 0,

r, s, u, v, w, x, z) is about half of the total height.Newsprint is often 8 points in size, so the x-height isabout 4 points. Because ~72 inches is equal to 1.41 mm.8-point print in a newsprint style font can be given anM-unit rating of about 1.0 M. Thus, for print in fontstyles similar to common newsprint, the M-unit ratingfor the lower case letters can be estimated by dividingthe point size by 8. Capital letters and numbers are tallerthan the lowercase letters (often by 1.5 times), and, forthese larger characters, 8-point print is equal to 1.5 Mrather than to 1.0 M. Common sans-serif fonts such asHelvetica have ascenders and descenders that aresmaller as a proportion of the overall size. Conse­quently, for samples of print at the same point size, thex-height will be larger for Helvetica than for TimesRoman. Fonts presented on computer screen displaysusually have their sizes expressed in points, which referto the size when the document is printed as hard copy.The size on the display screen varies with screen size andpixel density. It is handy to remember the following:

1.0 M units = 1.45 mm '" 8 points (lowercase,

newspaper style) '" typical newsprint

N NotationTo standardize the testing of near vision, the Faculty ofOphthalmologists of the United Kingdornv-" adoptedthe Times New Roman font as the standard font fortesting near vision, and they recommended that theprint size be indicated in points. The size label "N8"indicates that the standard near test font is being usedand that the size is 8 points. The near visual acuity per­formance is recorded as the smallest print that can beread (recorded in N notation), and the distance is spec­ified (e.g., N8 at 40 cm). A print size recorded in N­notation can be converted to M-units by dividing thenumber by 8 (e.g., N20 = 2.5 M-units).

Equivafent Snellen NotationEquivalent Snellen notation (also known as "reducedSnellen") is widely used to indicate the size of print usedin testing vision at near. Equivalent or reduced Snellennotation ostensibly expresses the distance visual acuityvalue that is mathematically equivalent to the nearvisual acuity. Usually-but not always-a standard testdistance of 40 cm is assumed. Thus, a specimen of print

of size 1.0 M presented at 40 cm might be labeled as20/50 equivalent, because the Snellen fraction 20/50 isequal to 0.40/1.00. At any other test distance, thissame 1.0-M print no longer has the same equivalence.Unfortunately, it is a relatively common practice touse the equivalent or reduced Snellen notation as ameasure of the height of the test print. 58 Clinicians whouse the equivalent Snellen system might record, forexample, "20/50 at 20 em" to indicate that 1.0-M print(equivalent to 20/50 when the chart is at 40 cm) isbeing read at 20 em. This visual acuity performance isactually equivalent to 20/100, and it could be moreappropriately recorded as 0.20/l.00 M. Despite its wide­spread use, "equivalent" or "reduced" Snellen notationis inappropriate for specifying the size of print used fortesting near vision. First, it is inappropriate to use anangular measure (Snellen fraction) to specify the heightof letters. Second, it is inappropriate to use a term thatsuggests a test at 20 feet when this distance is not rele­vant to the near-vision test distance or to the size of theprint.

Jaeger NotationJaeger notation indicates the size of print by using theletter J followed by a number, and it is widely used,mainly by ophthalmologists. The near visual acuity isindicated by recording the print size and the test dis­tance (e.g., J3 at 40 cm). Unfortunately, there is no stan­dardization of the Jaeger sizes; hence, there is nointrinsic meaning to the number that indicates the printsize. Sizes J1 to J3 usually indicate that the print is small,and J5 to J8 indicate fairly small print. Jose and Atcher­son" found that, among charts from different manu­facturers, there can be as much as a twofold differencein the sizes of print samples that carry the same Jaegersize label. The Jaeger notation should not be used formeasurements of visual acuity, but it remains popularin the ophthalmological community.

Acceptable Size Notations forTesting Near VisionM units and points specify the height of letters, and bothare satisfactory. M units have the advantage that they canbe used in the traditional Snellen fraction format, whichspecifies the testing distance and the print size; pointshave the advantage that they are well known and widelyused outside of the ophthalmic professions. Visualacuity is an angular measure, and its specification usingM units or points requires that the testing distance begiven as well (e.g., 6 points at 40 em, or 1.5 Mat 40 cm).When using point sizes, test material in capital lettersshould be specified. Otherwise, it should be assumedthat the height of the lowercase letters is the critical sizedimension.

Currently there is no international standardization ofthe font style that should be used for testing reading

Visual Acuity Chapter 7 237

acuity, but the Times font has long been the establishedstandard in the United Kingdom. Most regular newsprintis in Times or in fonts of similar styles, and it seems rea­sonable to use such fonts for clinical testing, becausethey are representative of everyday reading tasks.

Reading Acuity and Letter Chart Acuity

Letter chart acuity at near is directly comparable withletter chart acuity at distance. Reading acuity is a morecomplex function. Reading acuity tests use typeset printas the test target and the resolution of more congestedand complex components arranged in sequences thatmust be recognized. Patients with disorders affectingmacular function (e.g., age-related macular degenera­tion, amblyopia) are likely to have a reading acuity thatis significantly worse than the letter-chart acuity. Taskcomplexity can have a substantial effect on the visualacuity scores. Kitchin and Bailey" studied a group ofsubjects with age-related macular degeneration andfound that visual acuity scores showed substantial dif­ferences depending on the task complexity. Their resultsare summarized in Table 7-6; it can be seen that, forthis group of subjects, there was a fivefold differencebetween their averaged grating acuities and their readingacuities.

Reading charts come in a wide variety of formats.Most use a series of passages of text or simple sentenceswith print size diminishing with each successivepassage. Some charts use a series of unrelated words, 60.61

because this avoids context being used to help thepatient guess at the word. Charts that require thereading of sentences or passages'":" are obviously more

TABLE 7-6 Visual Acuity andTask Complexity inAge-Related MacularDegeneration Subjects

AVERAGE VISUALACUITY'

Task for Visual SnellenAcuity Test LogMAR (SOl Notation

Grating display 0.61 (+/-0.11) 20/81Single letters 0.76 (+/-0.31) 20/115Landolt rings 0.89 (+/-0.36) 20/155

with maskingbars

Letter chart 0.97 (+/-0.32) 20/189Word reading 1.31 (+/-0.25) 20/408

chart

'16 age-related macular degeneration subjects.From Kitchin fE, Bailey /1. 1981. Task complexity and visual acuityin senile macular degeneration. Aust J Optom 63:235-242.

representative of real reading tasks, in which contextand syntax contribute to reading accuracy and efficiency.Word reading charts can be said to test the ability to seeto read rather than the ability to read.

Near Visual Acuity VersusNear Vision Adequacy

Many tests said to be near visual acuity tests are notreally used to measure visual acuity but rather near­vision adequacy. A visual acuity test determines theangular size of the smallest print that a patient can read.A test of near visual acuity requires that some of the pre­sented material be beyond the patient's limit. On mostreading charts, the smallest letters are 0.4 M or larger(0.4/0.4 M is equivalent to 20/20 or 6/6). When beingtested with such charts at the usual 40-cm test distance,patients with normal visual acuity can be expected toread the entire near-vision chart at 40 ern. and, conse­quently, their resolution limit is not established. Onreading charts, the progression of print sizeshould extend to 0.25 M (2 points) or smaller if visualacuity is to be measured at a test distance of40 cm (0.25M at 40 cm is equivalent to 20/12.5 or 6/3.8 perform­ance). Although tests of near vision adequacy shouldnot be confused with tests of reading acuity, they arenevertheless useful, because they do demonstrate thatthe patient is capable of reading print that is quite small.

Visual Acuity and ResolutionLimit at Near

Sometimes it is important for the clinician to distin­guish between near visual acuity and the patient's reso­lution limit. Near visual acuity is determined from thesmallest print that can be read when the retinal imageis in good focus. The resolution limit simply determinesthe size of the smallest print that can be read withoutthe requirement of a sharp retinal image. A low-visionpatient with a 2.50-D addition might read 4.0-M printat 40 cm and 3.2-M print at 32 ern, with the retinalimage being in sharp focus at both distances. However,this patient might have a resolution limit of 2.5 M whenthe print is at 16 cm. Moving the print closer to 16 cmhas enlarged the retinal image, but now defocus is sig­nificant, so best visual acuity is not achieved. Reading2.5-M print at 16 cm represents better resolution(smaller print can be read), but visual acuity is poorerthan 0.40/4.0 M or 0.32/3.2 M. To predict near-visionmagnification needs for low-vision patients, cliniciansshould rely on measurements of reading acuity takenwhen the retinal image is in good focus.

Reading Efficiency

As patients read the reading acuity test charts, readingefficiency decreases as the print size approaches the

238 BENJAMIN" Borishs Clinical Refraction

patient's resolution limit. For normally sighted subjects,maximum reading efficiency is usually obtained whenthe print is about three times larger than the smallestresolvable print.":" When testing patients with lowvision, clinicians commonly note the smallest print forwhich good reading efficiency can be obtained in addi­tion to noting the smallest print that can just be read,The size limit for good efficiency and the resolutionlimit should both be accounted for when prescribingmagnification for near vision, For normally sighted sub­jects, maximum reading efficiency is usually obtainedwhen the print is about three times larger than thesmallest resolvable print. 37

,6 5 ,66 This relationship holdsover wide ranges of luminance and contrast.

Some near-vision test charts have been designed tofacilitate the clinical measurement of reading effi­ciency,(,0-62 These charts allow for the assessment ofchanges in reading speed as a function of the print size,Reading speed slows as the threshold size is approached.Bailey and colleagues" pointed out that cognitive andmotor demands of the reading task can affect readingspeeds and the degree to which speed changes withprint size, Bailey-Lovie word reading charts'? have 17size levels that range from 10 M to 0,25 M, For thesmallest 11 rows, there are 42 letters (two words eachwith lengths of 4, 7, and 10 letters). At the six largestsizes, there are fewer words, The MNREAD charts ofLegge and colleagues" have 19 different sizes rangingfrom 8,0 M to 0,12 M; at each size level, there is a simplesentence that has 60 character spaces distributed aboutevenly over three rows,

Logarithmic Scaling of Reading Charts

If the progression of print size on a reading chart followsa logarithmic or constant ratio of progression, there arespecial advantages that facilitate the prescribing ofnear-vision additions or magnifiers for the purpose ofallowing the patient to resolve-or read with effi­ciency-print of a particular size. Several such chartsare available. 60

.62 ,63, 68 The labels for the progression forprint size on the chart can be used as a sequence ofalternative viewing distances or as a sequence ofalternative dioptric powers (Figure 7-5). To change theresolution limit by a certain number of size levels onthe chart, the viewing distance (or dioptric distance)should be changed by an equivalent number of steps onthe scale. Consider the following progression of theM-unit ratings of print sizes on a logarithmically scaledchart:

10, 8,0, 6.3, 5.0, 4,0, 3,2, 2.5, 2,0, 1.6, 1.25, 1.00,

0,80, 0,63, 0.50, 0040, 0.32, 0.25, 0.20, and so on.

Assume that a patient can just resolve 1,6-M printand can efficiently read 2.5-M print when the chart is at

40 cm. The clinician's goal might be to enable thispatient to read 1.00-M print with efficiency, and thus0.63 M would be the resolution limit. Here, two stepsof size separate the efficiency and resolution limits: firstwith 2.5 M and 1,6 M and second with 1,00 M and0.63 M, Four steps of improvement are required, Theefficiency limit must shift by four steps from 2,5 M to1,00 M, and the resolution limit must also shift by foursteps from 1.6 M to 0,63 M. To achieve this, four stepsof change must be made to the viewing distance (andto the dioptric demand), The clinician may use thesequence of size labels on the charts as a guide to thesequence of changes in viewing distance, Since there isa four-step change in size from 4,0 to 1.6, a change inviewing distance from 40 cm to 16 cm will be a changeof equal proportion, Thus, the four-step change in theviewing distance required in this example goes from astarting point of 40 cm to 16 ern. Similarly, the dioptricdemand needs to be increased by four steps from a start­ing point of 2,50 OS to 6,25 OS, Provided that theretinal image remains in satisfactory focus, the readingresolution and reading efficiency goals for this patientcan be achieved by spectacles or magnifying systemsthat provide an equivalent viewing distance of 16 cm oran equivalent viewing power of 6,25 OS,

PURPOSES OF VISUALACUITY MEASUREMENT

Refraction and Prescribing Decisions

Refractionists use visual acuity charts as test objects forrefraction procedures to determine the lens power thatprovides the sharpest retinal image-and thus the bestvisual acuity-for the individual patient. Comparisonbetween the patient's habitual visual acuity and thevisual acuity with the newly determined refractive cor­rection often influences the clinician's or the patient'sjudgment on the advisability of obtaining newspectacles or contact lenses. Some insurance programsrequire documentation of an improvement in visualacuity before they will pay for a change in opticalcorrection, Some worthwhile changes in optical correc­tion do not produce measurable increases in visualacuity, Depending on the patient's accommodationabilities, a correction for hyperopia might serve torelieve an excessive accommodative response while pro­viding little or no improvement in measured visualacuity. Often, for patients with low vision, a change inrefractive correction will provide a substantial improve­ment in perceived clarity without changing the visualacuity score.

The relationship between visual acuity and uncor­rected refractive error is complex, and there is consider­able variation between individuals, even when optical

card

(8.0

)63

{6.3

J50

{4.0

J32

{3.2

J25

Fig

ure

7-5

Bai

ley-

Lov

iew

ord

read

ing

char

tw

ith

loga

rith

mic

size

prog

ress

ion,

show

nat

actu

alsi

ze.

M,

Mun

its;

N,

Nno

tati

onpo

ints

;Lo

gMA

R,

loga

rith

mo

fth

em

inim

um

angl

eo

fre

solu

tion

;VA

R,vi

sual

acui

tyra

ting.

(Cou

rtes

yof

the

Nat

iona

lV

isio

nRe

sear

chIn

stit

ute

ofA

ustr

alia

.)

12

---0

.7

{6

5J­

8 ~

---0

.4

(7

0)--

. .--0

.0(1

00

1-

BA

ILE

Y-L

OV

IEW

OR

DR

EA

DIN

GC

HA

RT

Nat

iona

lV

isio

nR

esea

rch

Inst

itute

ofA

ustr

alia

eC

op

yrig

ht

19

79

.

s III c ~ ~ c ~ '":r "~ ; .. -. N W -0

240 BENJAMIN Borishs Clinical Refraction

factors are similar. Acuity depends on optical factorssuch as pupil size, the presence of astigmatism, the axisof astigmatism, and optical aberrations. Atchison andcolleagues" studied visual acuity for wide ranges ofsimulated myopia and pupil size, and they tested at twodifferent luminance levels. There was a significant reduc­tion in uncorrected visual acuity as a function of themagnitude of the myopia and the diameter of the pupil,but both of these relationships are nonlinear. Peters"studied clinic records of patients in the age groups of 5to 15 years, 25 to 35 years, and 45 to 55 years, with over2000 subjects in each group. He analyzed uncorrectedvisual acuity as a function of spherical and astigmaticrefractive error and their combinations. His resultsshowed high similarity for all three age groups withregard to myopia. In hyperopia, the two younger agegroups obtained better acuities; this was clearly as aresult of their ability to use accommodation to fully orpartially overcome spherical hyperopic error. There is anapproximation or rule of thumb that is compatible withthe results of Peters" and Atchison and colleagues." aswell as the results of others: up to about 2 D of refrac­tive error, visual acuity reduces by about 0.1 log unit(one row on the chart) per 0.25 D. This means that thereis a visual acuity reduction of 2.5 times (0.4 log units)for 1.00 D of spherical defocus and 6.25 times (0.8 logunits) for 2.00 D. In hyperopia, young patients willlikely use their accommodation to reduce the amountof apparent refractive error such that the expected visualreduction does not occur in the magnitude expected. Forsimple astigmatism, the visual acuity reduces at abouthalf the rate. Thus, visual acuity will be reduced byabout a factor of 2.5 times (0.4 log units) for 2.00 D ofastigmatism and 1.6 times (0.2 log units) for 1.00 D ofcylindrical defocus.

Monitoring Ocular Health

Many disorders affecting the optical or neural compo­nents of the visual system cause a change in visualacuity. When it is known that a vision-reducing disor­der is present, monitoring visual acuity can provide ameans of detecting deterioration or improvement in thecondition. In many eye diseases, a change in acuity isoften a major determinant of whether treatments areimplemented, altered, or continued.

When it is clinically important to detect smallchanges in visual acuity, extra care should be taken withthe measurement. First of all, the clinical test conditionsand the procedures should be carefully standardized.For better sensitivity to change, the visual acuity scoresshould be as reliable as is practical; this usually meansthat acuity should be scored in a manner thatgives credit for each letter read. Bailey and colleagues"showed that, for normally sighted subjects, scoring letter

by letter gave 95% confidence limits for changes at ±5letters. They compared test and retest scores and foundthat the discrepancy between the two exceeded ±4 lettersfor fewer than 5% of the comparisons. If a five-letter dis­crepancy between test and retest scores occurs less fre­quently than 5% of the time by the vagaries ofsampling,it is reasonable to consider a five-letter differencesufficient evidence that a real change has occurred. Inthe results of these researchers, the 95th percentile dis­crepancy is found at ±4 letters, so the next largest dis­crepancy (±5 letters) becomes the 95% confidencelimit-or criterion-for change. If visual acuity is scoredrow by row, the 95% confidence limit for change isdetermined to be ±2 rows, because the 95th percentilediscrepancy lies in the next smallest category, which is±1 row (Figure 7-6). This means that, when visual acuityis scored by row, at least two rows of change must occurbefore the clinician can determine that there has beena real change in acuity. Thus, there is a twofold differ­ence in the confidence limits (five letters versus two rows= 10 letters) when scoring letter-by-letter rather thanscoring row-by-row (without adding qualifiers to indi­cate how many letters were read).

Visual Acuity for Normalcy

Although 20/20 (6/6) is commonly held to representnormal vision, most normally sighted persons haveacuity that is measurably better than 20/20; the tradi­tional 20/20 is more a limit at the poorer end of thenormal range. Brown and Levie-Kitchin" emphasizedthat, when evaluating a given patient, it is important toestablish the patient's individual baseline visual acuityand reliability against which subsequent measurementscan be compared. For patients with normal or near­normal vision, five letters on a Bailey-Lovie or ETDRSchart is a reasonable criterion for identifying change.

Elliott and colleagues" presented a meta-analysisof data that they had collected in different experiments,and they showed a systematic decrease in visualacuity with age (Figure 7-7). For inclusion in theiranalysis, they selected only subjects who had no signif­icant ocular disorders and no substantial reductionin acuity. For their groups younger than the age of50 years, the average visual acuities were better than20/16 (6/4.8); along with the reported standard devia­tions, it was seen that, at least up to age 50, visual acuitywas expected to be better than 20/20 (or 6/6). Even forsubjects who were more than 75 years old, the averagebest-corrected visual acuity was slightly better than20/20 (6/6).

Visual acuity scores in patients with significant dis­orders affecting vision are likely to be less reliable.Depending on the cause of the condition and on indi­vidual factors, it can become more difficult to set theconfidence limits for change. Taking visual acuity meas-

>­ozW::Jowa:u,

Visual Acuity Chapter 7 241

A TEST/RETEST DISCREPANCY (# LETTERS) B TEST/RETEST DISCREPANCY (# LETTERS)

Figure 7-6A, Distribution of test-retest discrepancies when scoring visual acuity letter by letter. B, Distribution of test­retest discrepancies when scoring visual acuity row by row.

0.1

VISUAL ACUITY WITH AGEFor subjects with healthy eyes

Comparison between the visual acuity scores for the twoeyes can also be useful for identifying deviations fromnormalcy."

Figure 7-7

Visual acuity with age for healthy eyes. (Adapted fromElliott DB, Yang KCH, Whitaker D. 1995. Visual acuitychanges throughout adulthood in normal, healthy eyes:seeing beyond 6/6. Optom Vis Sci 72:188.)

urements using letter-by-Ietter scoring at each clinicalvisit allows the clinician to accumulate data to identifyand document reliability characteristics for the individ­ual patient; real change is then identified when a visualacuity score is significantly outside the range of theusual "noise" in visual acuity scores for that patient.

12-c::2: 0.0Cl

d.~'5;;j -0.1

tii::::lIns~ -0.2~Q)

~

-0.3o 10 20 30 40 50 60 70 80

Age (years)

20/20

20/16

20/12.5

Visual Acuity MeasurementApplied to Vision Standards

Clinicians are often required to provide visual acuityscores that will be used by others to determine whetherthe patient meets eligibility standards specified forcertain occupational tasks, for licenses, or for certainbenefits. Visual acuity measurements have been used asthe basis for determining the amount of financial com­pensation in insurance claims or legal suits involving aloss of vision.

Chart design, testing distance, scoring method, andtest procedures can significantly affect acuity scores andthe consequent decisions. As an example, consider thevisual acuity standard for legal blindness, which is"visual acuity should be 20/200 or less." It is usuallyeasier for a patient to meet this criterion if a "standardSnellen chart" is used at a test distance of 20 feet. Suchcharts have a single letter at the top 20/200 level, thereare two letters at the next 20/100 level, and then thereare progressively more letters at the smaller size levelson the chart. The patient is given a score of 20/200 forreading the largest letter but failing to read the pair ofletters that constitute the 20/100 row. On such a chart,the "20/200 or less" criterion effectively becomes "failsto achieve a visual acuity of 20/100." Decisions mightchange if this chart were presented at 10 feet. A patient

242 BENJAMIN Borishs Clinical Refraction

who did not achieve the 20/100 acuity level mightobtain a visual acuity score of 10/70. This performanceis equivalent to 20/140, so the patient would no longermeet the legal blindness requirement. An apparent butartifactual improvement in visual acuity has occurred asa result of the chart design. On the other hand, a patientwho might have been able to read the two letters at the20/100 level might be unable to achieve equivalentsuccess when the chart is presented at a lO-foot distance,because the ostensibly equivalent task (10/50) is likelyto involve more letters that are more closely spaced and,as a consequence, less legible. If testing for legal blind­ness is performed with a chart that uses a 0.1 log unitsize progression, the next size smaller than the 200-feetletters is 160 feet. To meet the definition of legal blind­ness, the practical criterion would now become "fails toachieve an acuity of 20/160." A patient who achieves20/200+1 or 20/200+2 technically fails to meet therequirement for legal blindness; however, should theclinician use row-by-row scoring without adding quali­fiers, the acuity would be recorded as 20/200, and thepatient would be considered legally blind. The test chartdesign, the testing distance, and the scoring method canall significantly affect the visual acuity score and thedecision as to whether the patient meets specified visualacuity standards. Until chart formats, scoring methods,and test conditions and procedures are specified or stan­dardized.":" opportunities for substantial inconsisten­cies remain.

TOWARD STANDARDIZATION OFVISUAL ACUITY MEASUREMENT

There is still no single accepted international standardfor the clinical measurement of visual acuity. Threemajor authoritative bodies have produced sets ofsimilar-but not identical-principles for chart design.The authorities are (1) the National Research Council'sCommittee on Vision":": (2) the Concilium Ophthal­mologicum Universale, Vision Functions Committee":and (3) the International Standards Organization.":"They agreed that the standard test distance should be4 m, that the standard optotype against which othersshould be calibrated is the four-choice Landolt ring, andthat the size progression ratio should be 0.1 log unit(1.259x). There were some disagreements about thenumber of optotypes at each size, the spacing betweenadjacent optotypes. and the spacing between adjacentrows, The recent British Standard 4274-1 8 unfortunatelyintroduced new "standard" optotypes and allowed con­siderable variations in the size progression, the numberof letters at different sizes, the spacing between rows,and the chart luminance. It did, however, require thatthe space between letters be equal to the letter width.The International Standards Organization standards

were those accepted by the American National Stan­dards Institute.

The ETORS chart with its Sloan Letters andBailey-Lovie layout has become the de facto standardfor research in Western countries." Versions are avail­able with Landolt ring or tumbling E optotypes andcharacters from other languages. It seems probable thatthe ETORS chart will become more consolidated as thestandard for clinical research, with some variationsallowed to accommodate populations that lack famil­iarity with the 10 alphabetical letters used in the ETORScharts. The clinical community may slowly come toincorporate this chart design into routine clinical prac­tice. When computerized display screens eventuallybecome widely used for the clinical measurement ofvisual acuity, the flexibility of manipulating the displaywill create a need for decisions about chart-designparameters. Then, perhaps, the clinical community willembrace the Bailey-Levie design principles.

OTHER APPLICATIONS OFVISUAL ACUITY TESTING

Contrast Sensitivity

Low-contrast visual acuity charts (usually light-grayletters on a white background) are sometimes used as ameasure to identify changes that affect contrast sensi­tivity. Visual acuity is poorer when the contrast is lower.The extent to which acuity is degraded by the contrastreduction can identify patients whose general contrastsensitivity has been affected by their visual disorder. TheRegan letter charts" are available as a series of visualacuity charts at several different contrasts, and alO%-contrast Bailey-Lovie chart is available.Y"Haegerstrom-Portnoy and colleagues" produced a low­contrast chart by printing black letters in a Bailey-Lovieformat on a dark gray background, This chart, which isknown as the Smith-Kettlewell Institute Low-Luminancechart (SKILL chart), was found to be more sensitive fordetecting changes in visual function after retinal andoptic nerve disease than other tests of visual acuity andcontrast sensitivity.

The Small Letter Contrast Sensitivity (SLCS) test hasa series of Sloan letters that are all of the same size(5.5 M), and the contrast reduces progressively in stepsof 0.1 log units.":" There are 10 letters at each of the 14levels of contrast. As compared with standard visualacuity charts of the Bailey-Lovie design, the results ofthe SLCS test are more sensitive to small changes inrefractive error, and they show enhanced performanceunder binocular viewing. Part or all of this sensitivityadvantage may the result of increased sampling. Forcontrast levels greater than 10%, the relationshipbetween log contrast sensitivity and log visual acuity is

Visual Acuity Chapter 7 243

almost a linear function, with a slope of approximately7 to 10. In other words, the contrast should be changedby 0.7 to 1.0 log units to produce a 0.1 log unit changein visual acuity. In addition, there are twice as manyletters per row, so the SLCS test effectively offers a 14­to 20-fold sampling advantage. Consequently, sensitiv­ity to differences should be increased by about fourtimes (.Ji4 to .J2O). To obtain equivalent sampling fre­quency with a visual acuity chart, it would be necessaryto have more rows with much finer increments of size(0.01 to 0.014 log units instead of the usual 0.1 logunits) and to double (from 5 to 10) the number ofletters per row. Alternatively, an equivalent samplingfrequency could be obtained by averaging the results of14 to 20 independent visual acuity measurements withETDRS charts. The subject of contrast sensitivity iscovered in detail in Chapter 8.

Tests of Disability Glare

Visual acuity may be affected by glare when light is scat­tered by optical elements of the eye. Light scatter reducesthe contrast of the retinal image, which in turn reducesthe visual acuity. For a given glare situation, the reduc­tion in visual acuity can serve as an indicator of theseverity of light scatter. More subtle levels of light scatterare detected more easily if low-contrast charts are used,because low-contrast visual acuity is more substantiallyreduced by the additional reduction in contrast thatresults from the scatter.85

,86 Disability glare is covered indetail in Chapter 8.

Measurement of Potential Acuity

With cataracts or other optical degradations of vision, itcan be useful to know the visual capabilities of theretina, because this may influence decisions regardingsurgical intervention. The effect of the optical opacitiesor irregularities must be eliminated or reduced to assessthe retina's potential to achieve good visual acuity. Thesimplest and best-known procedure is the measurementof visual acuity with a pinhole aperture before the eye,as mentioned earlier in this chapter. Similar in princi­ple was the PAM, which is an instrument that projectsa Maxwellian view image of a visual acuity chart intothe eye via a narrow beam that is of pinhole size as itenters the pupil." After surgical treatment, the visualacuity should be at least as good as that obtained withthe pinhole of the PAM.

When coherent light is used to form two-pointimages in the plane of the pupil, interference occurs,and a high-contrast grating pattern is formed on theretina." The spatial frequency of the grating depends onthe separation of the two spots, with wider separationsgiving higher spatial frequencies. The optical quality ofthe optics of the eye has relatively little effect on thequality (i.e., contrast) of the interference pattern on the

retina. A similar effect can be achieved by presentinggrating patterns of variable frequency to the eye in aMaxwellian view." Hence, the projection of coherenthigh-contrast gratings of different special frequenciesonto the retina through distorted or clouded optics hasbeen a method for the assessment of potential acuity.Elliott and colleagues" used tests of reading speed andVianya-Estopa and colleagues" used flicker fusion testsas alternative methods for the assessment of potentialvision behind cataracts or other optical obstructionsto vision. Yet another method for the evaluation ofpotential vision is by measurement of vernier acuity.Vernier acuity measures the accuracy with whichthe patient can judge whether targets are aligned,such as judging whether two spots are placed one under­neath the other. The accuracy of alignment is relativelyimpervious to optical blur. Enoch and colleagues":"advocated using vernier acuity to evaluate the integrityof the retina behind dense cataracts as a means of esti­mating the visual improvement that might result fromcataract extraction. Vernier acuity measurements neces­sarily require that patients make multiple settings,because it is the variance of alignment errors that is themeasure of vernier acuity. The average alignment erroris expected to be zero, unless there is an optical distor­tion of the retinal image or some disruption of retinalstructure.

SUMMARY

Visual acuity remains the most widely used and mostuseful single clinical measurement for determiningwhether a significant abnormality or change is affectingthe visual system. It is sensitive to refractive error and tomany abnormalities that affect the optical media, theretina, the optic nerve, and the visual pathways. It isused routinely by eye-care practitioners during refractiveprocedures and during decision making when diagnos­ing or monitoring ocular disorders that affect vision.Letter charts are likely to remain the test of choice forthe clinical measurement of visual acuity.

Since the 1970s, there has been slow progress towardstandardizing many of the factors that can affect visualacuity measurements. The clinical research communityhas generally accepted and now almost exclusively usescharts that standardize the test task, and visual acuityscores are assigned in accordance with the principlesproposed by Bailey and Lovie.'" Practitioners have beenslower to change their methods. The chart design prin­ciples have not become widely used in clinical practice,except perhaps for low-vision care. Popular projectorsystems have restricted display areas that are not wideenough to allow five letters per row when the acuityrating is 20/63 or greater; these small display areas aremore compatible with the restricted angular size of theviewing apertures of phoropters. These constraints makeit unlikely that the clinical-practitioner community will

244 BENJAMIN Borishs Clinical Refraction

change to charts that have the same number of lettersin each row across all size levels. Alternatively, logarith­mic size progressions, standardized spacing ratios (atleast between adjacent letters), and using letter sets ofapproximately equal legibility can readily be incorpo­rated into projector chart formats. Pressures from rec­ommendations published by authoritative bodies andfrom the practices of clinical researchers may influencethe adoption of new chart designs for projector chartsfor clinicians.

Computer-driven flat-panel displays will graduallybecome the standard method for presenting visualacuity tests in clinical practice. Having the chart displaysgenerated by computer allows for the randomizationof letter sequences and controlled variation of opto­type, contrast, luminance, and spacing. Similarly, fortesting near vision, computer displays will allow for theeasy variation of the test targets and display parameters,and it may become easy to monitor eye movementrather than listen to the patient reading aloud to assessreading efficiency.The use of computers will also enablebetter scoring of visual acuity, because it will be easy torecord exactly which letters were read correctly. Appro­priate algorithms can then be applied to generate moreprecise scores, to provide reliability information, and toenable the analysis of response speeds as a function ofangular size.

The use of finer scaling to record acuity scores is evenmore important than the standardization of chartdesign. Among the community of clinical practitioners,there is not yet a broad appreciation of the extent towhich clinical decision making can be enhanced by theuse of finer scaling.Y" Again, the methods of scoringletter by letter are widely used by researchers, and thismight eventually influence general eye-care practition­ers. One can expect resistance to adopting new unitssuch as 10gMAR or the more user-friendly VAR. Withoutchanging units, practitioners would gain better sensitiv­ity for detecting changes in visual acuity if they makemore frequent and more disciplined use of pluses andminuses to qualify visual acuity scores; alternatively,they could use approximate interpolations (e.g., 20/20+might be called 20/19).

Letters should remain the target of choice for theclinical measurement of visual acuity for distancevision. Landolt rings, tumbling Es, numbers, and gratingpatterns will continue to be important, especiallyfor population groups who do not use or who cannotread the English alphabet. The standard letters arelikely to remain the five-by-five Sloan letters thatfollow the traditional framework used for earlier serifedletters and Landolt rings. The additional five-by-fiveletters and the redesigned variants of the Sloan lettersintroduced by the recent British Standard 4274-1 8 areunlikely to challenge the broad acceptance of the Sloanletters. Although sans-serif fonts may be more easily dis-

played on pixilated screens, it seems probable thatserifed fonts (Times Roman or Times New Roman) willbecome an official or de facto standard for readingacuity tests that are presented on printed charts ordisplay screens.

References1. Bennett AG, Rabbetts RB. 1989. Clinical Visual Optics, 2nd ed.,

pp 23-72. Boston: Butterworths.

2. Campbell I-W, Green DG. 1965. Optical and retinal factorsaffecting visual resolution. J Physiol 181:576-593.

3. Campbell loW, Gubisch RW. 1966. Optical quality of the humaneye. J Physiol 86:558-578.

4. Snellen H, 1862. Letterproeven tot Bepaling derGezigtsscherpte. Utrect: PW van der Weijer. Cited by BennettAG. 1965. Ophthalmic test types. A review of previous workand discussions on some controversial questions. Br J PhysiolOpt 22:238-271.

5, Bennett AG. 1965. Ophthalmic test types. A review of previouswork and discussions on some controversial questions. Br JPhysiol Opt 22:238-271.

6. Sloan LL. 1959. New test charts for the measurement ofvisual acuity at far and near distances. Am J Ophthalmol48:807-813.

7. British Standard 4274. 1968. Test Charts for Determining DistanceVisual Acuity. London: British Standards Institute.

8. British Standard 4274-1. 2003. Visual Acuity Test 'lyPes. London:British Standards Institute.

9. Fern K, Manny RE. 1986. Visual acuity of the preschool child: areview. Am J Optom Physiol Opt 63:319-345.

10. Bailey IL, Lovie IE. 1976. New design principles for visual acuityletter charts. Am J Optom Physiol Opt 53:740-745.

11. Bailey II.. 1988. Measurement of visual acuity-towardsstandardization. In Vision Science Symposium, A Tribute toGordon G. Heath, pp 217-230. Bloomington: IndianaUniversity.

12. American Medical Association. 2001. Evaluation ofpermanent visual impairment. In Cocchiarrella L,Andersson GBI (Eds), Guides to the Evaluation of PermanentImpairment, 5th ed, pp 277-304. Chicago: American MedicalAssociation.

13. Snell AC, Sterling S. 1925. The percentage evaluation ofmacular vision. Arch Ophthalmol 54:443-461.

14. Snell AC, Sterling S. 1926. An experimental investigation todetermine the percentage relation between macular acuity ofvision and macular perception. In Contributions to OphthalmicScience. Menasha, Wise: Banta Publishing.

15. American Medical Association. 1955. Council on IndustrialHealth. Special report. Estimation of loss of visual efficiency,Arch Ophthalmol 54:462-468.

16. American Medical Association. 1984. Evaluation of permanentvisual impairment. In Guides to the Evaluation of PermanentImpairment, AMA. Reprinted in Physicians' Desk Reference forOphthalmology, 1996, Medical Economics.

17. Taylor fiR. 1978. Applying new design principles to theconstruction of an illiterate E chart. Am J Optom Physiol Opt55:348-351.

18. Ferris FL, Kassoff A, Bresnick GH, Bailey II.. 1982. Newvisual acuity charts for clinical research. Am J Ophthalmol94:91-96.

19. Strong G, Woo Gc, 1985, A distance visual acuity chartincorporating some new design principles. Arch Ophthalmol103:44-46.

Visual ACUity Chapter 7 245

20. Johnston AW. 1985. Near visual acuity tests using Chinesecharacters and the 10gMAR principle. Singapore Med J26:484-455.

21. Hyvarinen L. 1984. LH Visual Acuity Tests. Espoo, Finland:Vistest.

22. Al-Mufarrej MM, Abo-Hiemed FA, Oduntan AO. 1996. Anew Arabic distance visual acuity chart. Optom Vis Sci 73:59-61.

23. Khamar BM, Vyas UH, Desai TM. 1996. New standardizedvisual acuity charts in Hindi and Gujarati. Indian J Ophthalmol44:161-164.

24. Ruamviboonsuk P,Tiensuwan M. 2002. The Thai logarithmicvisual acuity chart. J Med Assoc Thai 85:673-681.

25. Green I. 1905. Notes on the clinical determination ofthe acuteness of vision including the construction andgraduation of optotypes. Trans Am Ophthalmol Soc 10:644­654.

26. Westheimer G. 1979. Scaling of visual acuity measurements.Arch Ophthalmol 97:327-330.

27. National Academy of Sciences/National Research Council,Committee on Vision. 1979. Recommended standardprocedures for the measurement and specification of visualacuity. Adv OphthalmoI41:103-148.

28. Concilium Ophthalmologicum Universale. 1984. Visual acuitymeasurement standards. COU Vision Functions Committee.Reprinted in ltal J OphthalmoI1988;2:15-22.

29. Raasch TW, Bailey IL. 1984. Choice of optotype and spacingaffect visual acuity scores. Invest Ophthalmol Vis Sci25(Suppl): 145.

30. Raasch TW, Bailey IL, Bullimore MA. 1998. Repeatability ofvisual acuity measurement. Optom Vis Sci 75:342-348.

31. Arditi A, Cagenello B. 1993. On the statistical reliability of letterchart visual acuity measurement. Invest Ophthalmol Vis Sci34:120-129.

32. Bailey IL, Bullimore MA. Raasch TW, Taylor HR. 1991. Clinicalgrading and the effect of scaling. Invest Ophthalmol Vis Sci32:422-432.

33. Lovie-Kitchin IE. 1988. Validity and reliability of visual acuitymeasurements. Ophthalmol Physiol Opt 8:363-370.

34. F10m Me. Heath G, Takahashi E. 1963. Contour interactionand visual resolution, contralateral effects. Science 142:979-980.

35. F10m Me. Weymouth FW, Kahneman K. 1963. Visual resolutionand contour interaction. J Opt Soc Am 53:1026-1032.

36. Flom MC 1991. Contour interaction and the crowding effect.Problems in Optometl}' 2:237-257.

37. Bailey IL, Clear R, Berman SM. 1993. Size as a determinant ofreading speed. J IlIum Eng Soc 22: 102-117.

38. Hofstetter HW. 1973. From 20/20 to 6/6 or 4/4. Am J Optom50:212-222.

39. Sheedy IE, Bailey [L, Raasch 1W. 1984. Visual acuity and chartluminance. Am J Optom Physiol Opt 61:595-600.

40. Minkowski IS, Guyton DL. 1984. New methods for predictingvisual acuity after cataract surgery. Ann Ophthalmol 16:511,513-516.

41. Carkeet A. 2001. Modeling 10gMAR visual acuity scores: effect oftermination rules and alternative forced-choice options. OptomVis Sci 78:529-538.

42. Rosser DA, Laidlaw DAH, Murdoch IE. 2001. The developmentof a "reduced 10gMAR" visual acuity chart for use in routineclinical practice. Br J Ophthalmol 85:432-436.

43. Camparini M, Cassinari P, Ferrigno L, et al. 2001. ETDRS-Fast:Implementing psychophysical adaptive methods to standardizedvisual acuity measurement with ETDRS charts. Invest OphthalmolVis Sci 42: 1226-1231.

44. Bailey IL. 1994. Optometric care for the multi-handicappedchild. Ptact Optom 5:158-166.

45. Norcia AM, Tyler Cw. 1985. Spatial frequency sweep VEP, visualacuity during the first year of life. Vision Res 25:1399-1408.

46. Dobson V, Teller DY. 1978. Visual acuity in human infants.Vision Res 18:1469-1483.

47. McDonald MA, Dobson V, Sebris SL, et al. 1985. The acuitycard procedure, a rapid test of infant acuity. Invest OphthalmolVis Sci 26: 1158-1162.

48. Adoh TO, Woodhouse M, Odwaiye KA. 1992. The Cardiff test: anew visual acuity test for toddlers and children with intellectualimpairment. Optom Vis Sci 69:427-432.

49. Hyvarinen L, Nasanen R, Laurinen P. 1980. New visualacuity test for pre-school children. Acta Ophthalmol 58:507­511.

50. Richman IE, Petito GT, Cron MT. 1984. Broken wheel acuitytest, a new and valid test for preschool and exceptionalchildren. JAm Optom Assoc 55:561-565.

51. McGraw PV, Winn B. 1985. Glasgow Acuity Cards, a new testfor the measurement of letter acuity in children. OphthalmolPhysiol Opt 13:400-404.

52. Moke PS, Turpin AH, Beck RW, et al. 2001. Computerizedmethod of visual acuity testing: adaptation of the amblyopiatreatment study visual acuity testing protocol. Am J Ophthalmol132:903-909.

53. Sloan LL. 1977. Reading Aids for the Partially Sighted, pp 37-51.Baltimore: Williams & Wilkins.

54. Sloan LL, Habel A. 1956. Reading aids for the partially blind.Am J Ophthalmol 42:863-872.

55. lose RT, Atcherson RM. 1977. Type size variability for nearpoint acuity tests. Am J Optom Physiol Opt 54:634-638.

56. Law FW. 1951. Standardization of reading types. Br J Ophthalmol35:765-775.

57. Law FW. 1952. Reading types. Br J Ophthalmol 36:689-690.

58. Bailey IL. 1982. A call for the elimination of the Iaeger andreduced Snellen notations. Optom Monthly 72:676-679.

59. Kitchin IE, Bailey IL. 1981. Task complexity and visual acuity insenile macular degeneration. Aust J Optom 63:235-242.

60. Bailey IL, Lovie IE. 1980. The design and use of a new near­vision chart. Am J Optom Physiol Opt 57:378-387.

61. Baldasare l. Watson GR, Whittaker SG, Miller-Schaffer H.1986. The development and evaluation of a reading test forlow vision macular loss patients. J Vis Impair Blindness80:785-789.

62. Legge GE, Ross IA. Luebke A, LaMay 1M. 1989. Psychophysics ofreading. VIII. The Minnesota Low-Vision Reading Test. OptomVis Sci 66:843-853.

63. Lewis HT. 1985. University of North Carolina near vision chart.J Vis Rehabil 3:8-9.

64. Sloan LL, Brown DI. 1963. Reading cards for the selection ofoptical aids for the partially sighted. Am J Ophthalmol55:1187-1199.

65. Bailey IL. 1987. Mobility and visual performance under dimillumination. In Night Vision-Current Research and FutureDirections. Symposium Proceeding, Committee on Vision of theNational Research Council and American Academy of Science,pp 220-230. Washington, DC; National Academy Press.

66. Bailey IL. 1993. Detecting early visual loss in the elderly. OptomVision Sci 70:299-305.

67. Bailey IL, Lueck AH, Greer R. et al. 2003. Understanding therelationships between print size and reading in low vision. JVisual Impairment & Blindness 97:325-334.

68. Keeler CH. 1956. On visual aids for the partially sighted. TransOphthalmol Soc UK 76:605-614.

246 BENJAMIN Borishs Clinical Refraction

69. Atchison DA, Smith G, Efron N. 1979. Effect of pupil size onvisual acuity in uncorrected and corrected myopia. Am J OptomPhysiol Optics 56:315-323.

70. Peters HB. 1961. The relationship between refractive error andvisual acuity at three age levels. Am J Optom Physiol Opt Arch AmAcad Optom 38:194-198.

71. Brown B, Levie-Kitchin IE. 1993. Repeated visual acuitymeasurement: establishing the patient's own criterion forchange. Optom Vis Sci 70:45-53.

72. Elliott DB, Yang KCH, Whitaker D. 1995. Visual acuity changesthroughout adulthood in normal healthy eyes: seeing beyond6/6. Optom Vis Sci 72:186-191.

73. Brown B, Yap MKH. 1995. Differences in visual acuity betweeneyes: determination of normal limits in a clinical population.Ophthalmol Physiol Opt 15:163-168.

74. National Academy of Sciences/National Research Council,Committee on Disability Determination for Individuals withVisual Impairments. 2002. In: Lennie P, Van Hernel SB (Eds),Visual Impairments: Determining Eligibility for Social SecurityBenefits. Washington, DC: National Academy Press.

75. National Academy of Sciences/National Research Council,Committee on Vision. 1994. Measurement of Visual Field andVisual Acuity for Disability Determination. Washington, DC:National Academy Press.

76. International Standards Organization. 1994. ISO 8596, Opticsand Optical Instruments-Visual acuity testing. ISO CentralSecretariat, Geneva, Switzerland.

77. International Standards Organization. 1994. ISO 8597, Opticsand Optical Instruments-Standard optotype and its presentation.ISO Central Secretariat, Geneva, Switzerland.

78. Ferris FL, Bailey IL. 1996. Standardizing the measurement ofvisual acuity for clinical research studies. Ophthalmology103:181-182.

79. Regan 0, Neima D. 1983. Low-contrast letter charts as a test ofvisual function. Ophthalmology 90: 1192-1200.

80. Bailey ll., 1982. Simplifying contrast sensitivity testing. Am JOptom Physiol Opt 59:12.

81. Bailey IL, Raasch 'IW, Koh P, et al. 1993. Contour interactionwith high and low contrast charts. Tech Dig of Topical Meeting

on Non-Invasive Assessment of the Visual System, OpticalSociety of America, Washington, DC, 3:228-231.

82. Haegerstrorn-Portnoy G, Schneck M, Brabyn L lampolskyA. 1997. The SKILLcard: a test of acuity under reducedcontrast and luminance. Invest Ophthalmol Vis Sci 38:207­218.

83. Rabin I. 1995. Luminance's effects on visual acuity and smallletter contrast sensitivity. Optom Vis Sci 71:685-688.

84. Rabin L Wicks I. 1996. Measuring resolution in the contrastdomain: the small letter contrast test. Optom Vis Sci73:398-403.

85. Bailey IL, Bullimore MA. 1991. A new test of disability glare.Optom Vision Sci 68:911-917.

86. Regan OM, Giashi E, Fresco BB. 1993. Measurement of glaresusceptibility using low contrast letter charts. Optom Vis Sci70:969-975.

87. Minkowski IS, Palese M, Guyton DL. 1983. Potential acuitymeter using a minute aerial pinhole aperture. Ophthalmology90: 1360-1368.

88. Enoch 1M, Bedell HE, Kaufman HE. 1979. Interferometric visualacuity testing in anterior segment disease. Arch Opht1llllmol79: 1916.

89. Lotmar W. 1980. Apparatus for measurement of retinalvisual acuity by Moire fringes. Invest Ophthalmol Vis Sci19:393-401.

90. Elliott DB, Patel B, Whitaker D. 2001. Development of areading speed test for potential-vision measurements. InvestOphthalmol Vis Sci 42:1945-1949.

91. Vianya-Estopa M, Douthwaite WA, Pesudovs K, et al. 2004.Development of a critical flicker/fusion frequency test forpotential vision testing in media opacities. Optom Vis Sci81:905-910.

92. Enoch 1M, Knowles R. 1995. Hyperacuity test to evaluate visionthrough dense cataracts. Optom Vison Sci 72:630-642.

93. Enoch 1M, Williams RA, Essock EA, Fendick MG. 1985.Hyperacuity: a promising means of evaluating vision throughcataract. In Osborne NN, Chader GI (Eds]. Progress in RetinalResearch, vol 4, pp 67-68. Oxford: Pergamon.