Validity and ease of use of a computerized Hess chartPatrick Watts, FRCS, FRCOphth,a Harish Nayak, MRCOphth, FRCS,a May Kim Lim, MRCOphth,a

Alan Ashcroft, BSc,a Hasan Al Madfai, MSc, PhD,b and Hollie Palmer, BScb

PURPOSE To investigate agreement between computerized and conventional methods for obtaining

Author affiliations: aDepartmentUnited Kingdom; and bDivisionWales United KingdomStudy conducted at the UniverPresented at the 33rd Annual

Ophthalmology and StrabismusSubmitted July 4, 2010.Revision accepted December 7,Published online September 29Reprint requests: Patrick Wat

University Hospital of Wales, Hwales.nhs.uk).Copyright � 2011 by the Am

Strabismus.1091-8531/$36.00doi:10.1016/j.jaapos.2010.12

Journal of AAPOS

Hess charts and to compare relative ease of use of both methods.

METHODS Hess charts of 65 patients were obtained by the use of the computerized Assaf Ocular

Motility Analyzer (OMA) and the conventional Lees screen method. The Hess charts pro-duced by each method were compared with a previously described scoring system. Patientscompared the ease of testing by using a 5-point Likert scale.

RESULTS For horizontal deviations of the right eye, the OMA provided a significantly larger (P 5

0.0001) deviation (301� � 267�) than the Lees screen (204� � 306�). The Lees screengave a significantly larger score for vertical deviations of the left eye (117� � 158� vs 96�

� 129�; P 5 0.003). Vertical deviations of the right eye and horizontal deviations of theleft eye did not differ significantly between tests. Patients required the same amount oftime to complete both tests, but the OMA was slightly easier to perform than the Leesscreen (Likert score, 1.2 � 0.5 vs 1.3 � 0.4; P 5 0.046).

CONCLUSIONS The scores measured were larger with OMA in the horizontal and smaller in the vertical

direction than with the Lees screen in some directions. Although the OMA did not savetime, patients found it easier to perform than the Lees screen. The OMA may be consid-ered a useful alternative to the Lees screen. ( J AAPOS 2011;15:451-454)

The Hess chart has been used since 1908 to diag-nose paralytic and mechanical ocular deviationsand to provide a reliable and repeatable record

of the condition.1,2 Although many techniques to obtainHess charts have been described,3 the Lees screen, whichuses a 2-sided mirror to obtain dissociation,4 is one of themost commonly used. Detailed descriptions of the Leesscreen, the equipment required, and the methods usedhave been previously well described.5,6 Recently,a computerized system, the Assaf Ocular MotilityAnalyzer (OMA; Medical Digital, Ltd, Winslow,Buckinghamshire, UK) has been made available fordigitally acquiring and recording Hess charts. Thismethod uses red-green goggles to achieve dissociation sim-ilar to those used with the Lancaster red-green test.7,8

To monitor the progress of an ocular deviation recordedfrom Hess charts, an objective scoring method was intro-duced in 1992.9 This study was designed to prospectively

of Ophthalmology, University Hospital of Wales, Cardiff,of Mathematics and Statistics, University of Glamorgan,

sity Hospital of Wales, Cardiff, Wales, United Kingdom.Meeting of the American Association for Pediatricas a poster, Seattle, Washington, April 11-15, 2007.

2010., 2011.ts, FRCS, FRCOphth, Department of Ophthalmology,eath Park, Cardiff CF14 XW (email: Patrick.Watts@

erican Association for Pediatric Ophthalmology and

.021

compare the agreement between the OMA and the goldstandard conventional Lees screen for analyzing ocularmotility in adult patients. In addition, we compared the re-corded subjective responses of patients in regard to the easeof each method and compared the time taken to completethe test.

Patients and Methods

The study was approved by local research ethics committee and

the University Hospital of Wales Research and Development

Committee and followed the tenets of the Declaration of Hel-

sinki. Adult patients (.16 years) who presented consecutively to

the eye clinic with diplopia with a visual acuity of 20/100 or better

and normal retinal correspondence were included in the study.

Patients who had previously plotted Hess charts (by the use of ei-

ther method) and patients with physical or mental impairments

were excluded from the study.

Patients whomet the study criteria and consented to be a part of

the study were asked to plot Hess charts by using both the Lees

screen (Figure 1A) and the OMA (Figure 1B) in random order.

All patients were given instructions and supervised by one inves-

tigator. The Lees screen was drawn out by the operator, whereas

the OMA was printed from the computer; both produced similar

Hess charts. Hess charts were completed anonymously and ran-

domized before being scored by other investigators via a previ-

ously described scoring system.4 In brief, each point on the

Hess chart may be considered to have a horizontal and vertical

displacement from normal (each small square represents 5� or

10D). Displacements are taken as positive, and those which are

out of range are assigned an arbitrary 15�. Three separate sums

are calculated for the 16 outer points (Ho or Vo), 8 inner points

451

FIG 1. A, Lees screen used to record a Hess chart. The subject’s eyesare dissociated with a mirror. The examiner uses a wand to presenta target on the illuminated screen and the subject uses a wand onthe opalescent screen to indicate the position of the perceived exam-iners wand. B, OMA used to record a Hess chart. The subject’s eyes aredissociated with red-green goggles. The subject uses the computermouse to indicate the perceived position of the target presented onthe chart by the computer.

Table 1. Conventional and OMA scores in degrees for horizontaland vertical components

Lees screen OMA P value

Right eye horizontal 204 � 306 (SD) 301 � 267 (SD) 0.0001Right eye vertical 105 � 138 (SD) 103 � 145 (SD) 0.79Left eye horizontal 224 � 334 (SD) 268 � 260 (SD) 0.062Left eye vertical 117 � 158 (SD) 96 � 129 (SD) 0.0031

OMA, Ocular Motility Analyzer; SD, standard deviation.

452 Watts et al Volume 15 Number 5 / October 2011

(Hi or Vi), and the single central point (Hc or Vc). The sum of the

outer points is multiplied by 1, the inner points by 4, and the cen-

tral point by 8. The total of the 3 products gives the horizontal or

vertical score for each Hess chart.

Scoring was manually calculated by one observer and cross-

checked by another for accuracy. The time taken to complete

the 2 tests was recorded. Patients were asked to grade the de-

gree of difficulty of the 2 tests on a 5-point Likert scale (1, very

easy; 2, easy; 3, neither easy nor difficult; 4, difficult; 5, very

difficult).

Statistics

On the basis of a pilot study of 20 patients, a sample size of65 subjects was required to detect a mean difference of 10�

with a power of 90%, The pilot data were not included inthe study. The Hess scores obtained by the use of the 2different methods for each eye were then analyzed withmultivariate repeated-measures analysis of variance withOMA/Lees screen, right/left eyes, and horizontal/verticalas main effects in a full model that included the 2- and3-way interactions. This test facilitates the simultaneousevaluation of the effects of individual factors and their in-teractions on the variability of outcomes while takinginto account any influence of left eye–right eye relation-ships on outcomes. The qualitative difficulty scores wererecoded into a 5-point Likert scale from “very easy” to“very difficult.” The Wilcoxon signed-rank test was usedto evaluate the level of significance of degree of difficultybetween the 2 tests. The statistical software usedwas PASW Statistics, version 18.0 (formerly SPSS) onMac OS X.

Results

OMA and conventional Hess charts were obtained from65 patients (45 female patients). The mean age was49.35 � 17.0 years (range, 19-85 years). The scores ofthe right and left eye for the horizontal and vertical com-ponents for the Lees screen and the OMA are presented inTable 1. Hess scores obtained by the OMA method weredifferent from scores obtained via use of the Lees screen(P 5 0.05). For horizontal scores, the OMA system over-estimated the conventional scores for both right and lefteyes. For vertical scores, the OMA underestimated con-ventional scores for left eyes but showed good agreementfor right eyes (Figure 2). The 2-way interaction effect in-dicated that the horizontal and vertical components of theOMA scores were also significantly different from thehorizontal and vertical components of the Lees screen(P 5 0.0001).

Reliability analysis via the use of Bland-Altman plotsdemonstrated that the difference between the OMA hori-zontal and conventional horizontal scores for the left eyewas significant; the intraclass correlation coefficients

Journal of AAPOS

FIG 2. Raw HESS scores from OMA (Comp) and Lees screen (Conv) for right (R) and left (L) eyes for horizontal (Horz) and vertical (Vert).

Volume 15 Number 5 / October 2011 Watts et al 453

between the horizontal scores were 0.78 and 0.77 and thevertical scores were 0.92 and 0.94 for the left and righteyes, respectively. This finding indicated that the OMAand Lees screen scores were closer in the vertical scorescompared to the horizontal scores.The average time taken to complete the Lees screen was

232 � 89 seconds, and for the OMA, 226 � 163 seconds(P 5 0.74; 95% confidence interval for difference,�39.36 to 28.16 seconds). The correlation coefficient be-tween age and the time taken to complete the Lees screen(r5 0.45, P5 0.0001) and the OMA (r5 0.35, P5 0.004)indicated that there was a weak direct relationship betweenage and testing times.When we used the Mann-Whitney test, we found that

sex had no significant effect on the time it took to completethe test for the Lees screen (U 5 371, P 5 0.26) and theOMA (U5 362.5, P5 0.214). For the female patient sam-ple, when we used theWilcoxon signed-rank test, we foundthere was insufficient evidence to conclude that the mediantimes for Lees screen and the OMA were different(P5 0.13); for the male sample (n5 20), there was likewiseinsufficient evidence to conclude a difference (P 5 0.61).The difficulty score of the Lees screen (1.3 � 0.5) was

significantly greater (P 5 0.046 by Wilcoxon signed-ranktest) than that for the OMA (1.2� 0.4); however, both testshad a median difficulty score of 1, with a range of 1-3.Conventional Hess charts were graded as “very easy” to

Journal of AAPOS

perform by 76% of patients, compared with 85% ofpatients for the OMA.

Discussion

The high standard deviations recorded for both themethods reflect the different ocular motility disorders ineach patient; however, when an objective scoring systemis used, it appears that the OMA and the conventionalLees screen generated Hess charts that are not inter-changeable in terms of the magnitude of misalignment.The difference was largest in the horizontal deviation ofthe right eye.

An explanation for the right eye/left eye difference is notobvious, and further work on the OMA will be necessary todetermine whether our current findings are reproducible.The discrepancy may be attributable to a systematic errorwithin the OMA software package or to the difference inworking distance used by the 2 methods. The working dis-tance in the Lees screen should be 50 cm5; in this study,the working distance was 55.5 cm. For the OMA we used27.3 cm. It is possible that a shorter working distance hasan effect on accommodation and thus horizontal displace-ments. Assaf10 has suggested that amodification of thework-ing distance to 35 cm gives a more accurate measurement.

The OMA uses a large, flat screen monitor, which maybe wall mounted. This can thus occupy a small space

454 Watts et al Volume 15 Number 5 / October 2011

compared to the Lees screen, freeing up valuable clinicspace. The OMA’s digital output facilitates records man-agement and assessment.

In conclusion, although the OMA registered largerscores than the Lees screen for right eye horizontal devia-tions and smaller scores for left eye vertical deviations.Thanks to ergonomic advantages and the equivalentperformance, the OMA has replaced the Lees screen inour practice.

References

1. HessW. Eine neue Untersuchungsmethode bei Doppelbildern. ArchAugenheilk 1909;62:233-8.

2. Roper-Hall G. The Hess screen test. Am Orthopt J 2006;56:166-74.

3. Roodhoft JM. Screen tests to map out ocular deviations. Bull SocBelge Ophtalmol 2007;305:57-67.

4. Lees VT. A new method of applying the screen test for inter-ocularmuscle balance. Br J Ophthalmol 1949;33:54-9.

5. Mein J, Harcourt B. Ocular deviation. In: Diagnosis and managementof ocular motility disorders. Oxford: Blackwell Scientific Publica-tions; 1986:70-74.

6. Timms C. The Lees screen test. Am Orthopt J 2006;56:180-3.7. Von Noorden GK, Campos EC. Examination of the patient—II. In:

Binocular vision and ocular motility. 6th ed. St. Louis, MO: Mosby;2002:190-94.

8. Christoff A, Guyton DL. The lancaster red-green test. Am Orthopt J2006;56:157-65.

9. Aylward GW, McCarry B, Kousoulides L, Lee JP, Fells P. A scoringmethod for Hess charts. Eye 1992;6:659-61.

10. Assaf A. Validity of a computerized Hess chart. J AAPOS 2008;12:107.

Journal of AAPOS