ARTICLE

Safety of nondominant-

hand ophthalmic surgeryJonathan Park, FRCOphth, BSc, Olayinka Williams, MBChb, Salman Waqar, MRCS(Ed), BSc,

Neil Modi, MBChb, Thomas Kersey, FRCOphth, Tamsin Sleep, FRCOphth, MA

SubmittFinal revAccepte

From thNHS Tru

Steve Shhelped w

CorrespmologyDevon,

Q

P

2112

ed: Junision sd: July

e Ophst, Tor

aw, Phith the

ondingDepartTQ2 7A

2012 A

ublished

PURPOSE: To establish the safety of nondominant-hand simulated intraocular surgery.

SETTING: Horizon Suite Simulation Centre, South Devon Foundation NHS Trust, Devon, UnitedKingdom.

DESIGN: Cohort study.

METHODS: Physicians with no previous ophthalmic surgical experience completed an introductoryprogram on the EyeSi ophthalmic surgical simulator to eliminate the learning curve. They then com-pleted the validated level-4 forceps module 4 times with their dominant hand and then 4 times withtheir nondominant hand. Simulator total score, odometer movement, corneal injury, lens injury, andtotal time were recorded. Acuity (Snellen near) and stereoacuity (Frisby) were also recorded.

RESULTS: All 30 physicians showed good acuity (6/6 and N6 or better) and stereopsis (mean35 seconds of arc). The total score was lower (mean 60.8 versus 65.6; PZ.019), operating timeswere longer (mean 71.6 versus 70.0; PZ.026), and lens injury was greater (mean 0.93 versus 0.79,PZ.021) when operating with the nondominant hand than with operating with the dominant hand.Those with higher scores with the dominant hand had higher scores with their nondominant hand.

CONCLUSIONS: Simulated nondominant-hand ophthalmic surgery resulted in less efficient, lesssafe, and slower surgery. This observation was more marked in those with less skill with theirdominant hand. This has practical implications for trainee and trainer if 1 surgeon is left handedand 1 right handed. It also suggests that a higher degree of competence with the dominant handis required before performing nondominant-hand surgery.

Financial Disclosure: No author has a financial or proprietary interest in any material or methodmentioned.

J Cataract Refract Surg 2012; 38:2112–2116 Q 2012 ASCRS and ESCRS

Patient expectations after ophthalmic surgery con-tinue to rise. Historically, cataract surgery was consid-ered necessary to clear the media and to allow, withfresh spectacle correction, a reasonable level of func-tional vision. Now, there are increasing demands toroutinely combine clear media cataract surgery with

e 13, 2012.ubmitted: July 13, 2012.15, 2012.

thalmology Department, South Devon Foundationbay, Devon, United Kingdom.

D, University of Plymouth, Devon, United Kingdom,statistical analysis.

author: Jonathan Park, FRCOphth, BSc, Ophthal-ment, South Devon Foundation NHS Trust, Torbay,A, United Kingdom. E-mail: jonathanpark@nhs.net.

SCRS and ESCRS

by Elsevier Inc.

additional steps to minimize surgically induced astig-matism. Steps that can minimize surgically inducedastigmatism are therefore becoming increasingly com-mon, and some require major tasks to be performedwith the nondominant hand. These include bimanualphacoemulsification, bimanual irrigation/aspiration(I/A), and the facilitation of on-meridian surgery.

At a relatively earlier stage of learning, modern oph-thalmic trainees are expected to operate in a positionthat is more challenging than would be the caseif they were just aiming for clear media surgery.When the trainee's dominant hand is not the same ashis or her supervisor's, it poses an additional challengefor the trainee or the supervisor when surgical stepsare shared between surgeons. Other bimanual surger-ies, such as pars plana vitrectomy (PPV), also requirea high level of skill with both hands. For example, inphakic patients, when the vitrectomy cutter cannotpass the lens, it is necessary to use the nondominanthand to perform some of the main surgical steps.

0886-3350/$ - see front matter

http://dx.doi.org/10.1016/j.jcrs.2012.07.030

2113SAFETY OF NONDOMINANT-HAND SURGERY

It is not ethical to observe the effect of nondominant-hand surgery in a live operating room environmentbecause it could cause harm to patients. It is, however,possible to do so with the use of an ophthalmic simu-lator, which we previously used to show other vari-ables that cannot be observed on real patients, suchas the effect of distraction, fatigue, and loss of stereop-sis on surgical performance.1–3

The aim of this study was to establish the safety ofnondominant-hand simulated intraocular surgery.

SUBJECTS AND METHODS

Physicians with no previous ophthalmic microsurgical expe-rience were recruited through local announcement of thestudy on a first-come, first-served basis.

This study used the EyeSi virtual reality cataract surgerysimulator (VRmagic Holding AG) to measure intraocularsurgical performance (Figure 1). Such simulators are becom-ing increasingly common and are useful in surgical trainingbecause they allow trainees to obtain complex motor skillsin a safe environment; this not only aids their training butalso improves patient care. This simulator comprises a man-nequin head with a model eye that rotates and pivots whenmanipulated by the surgeon. Probes inserted into the eyesimulate different intraocular instruments with appropriatevisual feedback. The microscope provides stereoscopicimages that can be manipulated like a real operating roommicroscope using a footpedal.

The simulator allows measurements of a standardizedsurgical task and provides feedback to the user in the

Figure 1. Intraocular surgical simulator. Note the phacoemulsifica-tion and microscope pedals, instructor screen, head prop withelectronic eye, and viewing microscope on an adjustable platform.

J CATARACT REFRACT SURG - V

following categories: surgeon efficiency, achievement of sur-gical goal, and tissue injury.A The forceps module level4 (Figure 2) was found to have construct validity4; thus,this module was selected for the study. The surgeon mustgrasp 6 triangular objects from the periphery of the anteriorchamber and place them in a net in the center of the anteriorchamber. This module was designed to teach surgeons howto accurately grasp the edge of a capsulorrhexis flap whilekeeping the eye in the primary position to maintain the redreflex and avoid injury to the cornea or lens.

For each attempt, the simulator calculates a total score thatcan vary from 0 to 100. Positive points are awarded for com-pleting the task. Negative points are subtracted from thisscore if the surgery is inefficient (excessive time, excessivemovement of instruments or the eye, operating withouta red reflex, operating while out of focus) or unsafe (injuryto the lens or cornea).

For each participant, monocular and binocular distancevisual acuity using a Snellen chart at 3 m was recordedand then converted to 6 m for analysis. Near vision wasrecorded with a standard Times New Roman Faculty ofOphthalmologists reading chart. Stereopsis was measuredusing the Frisby test.5 These recordings were made to ensureparticipants had an adequate level of acuity and stereopsis toperform simulated intraocular surgery. Arbitrary poorvalues of acuity or stereopsis for exclusion criteria were con-sidered but were not required because all participantsshowed good acuity and stereopsis.

Participants then received standardized introduction onthe use of the simulator. This introduction included basicinformation relating to the anatomy of the eye that wassufficient to allow subsequent simulator task completion.The introduction also involved 8 practice sessions on the

Figure 2. Anterior segment forceps module, level 4. Note the trian-gular objects in the peripheral anterior chamber, which have to begrasped with the forceps and placed in a central net.

OL 38, DECEMBER 2012

2114 SAFETY OF NONDOMINANT-HAND SURGERY

forcepsmodule (1 attempt level 1, 1 attempt level 2, then 6 at-tempts level 4) to help eliminate the learning curve. The prac-tice data were not recorded. Main data collection thencommenced and consisted of 4 attempts of the level 4 forcepsmodule with the dominant hand followed by 4 attempts ofthe level 4 forceps module with the nondominant hand.The dominant hand was defined as the hand that the partic-ipant writes with, and the other hand was defined as thenondominant hand. Simulator total score, odometer move-ment, corneal injury, lens injury, and total time wererecorded. The odometer movement score reflects efficiency;more-skilled surgeons will complete the same task withfewer movements.

It was decided that the dominant-hand trials should beperformed before the nondominant-hand trials rather thanvice versa for the following rationale: If the dominant handgoes first, any possible learning effect would potentially im-prove the subsequent nondominant hand scores and reducethe likelihood of the dominant hand obtaining a better score.If the nondominant hand goes first, any possible learningeffect would potentially improve the subsequent dominanthand scores and increase the likelihood of the dominanthand obtaining a better score. It was therefore decided to al-low the dominant hand to go first so if there were anylearned effect, it would slightly reduce (rather than increase)the dominant-hand score. With this sequence, if the domi-nant hand were found to have a higher score, it would bedue to the hand being more skillful and such skill differencebetween the hands would be a more substantial factor thanany possible opposing learning effect.

Data tests of normality (histograms and Q-Q plots) wereapplied to determine data distribution. Parametric testing(paired t test) was used for normally distributed data andnonparametric testing (Wilcoxon signed-rank test) forskewed data. Multiple comparison corrections were not re-quired due to the small sample. The Spearman correlationwas calculated to establish whether high scores fordominant-hand surgery correlated with high scores fornondominant-hand surgery. A P value less than 0.05 wasconsidered statistically significant. All data were analyzedin SPSS software (SPSS, Inc.).

RESULTS

Thirty physicians were recruited. Seventeen were menand 13 were women. Twenty-eight were right handedand 2 were left handed. All 30 participants showed

Table 1. Comparison of surgical scores between dominant handand nondominant hand.

Mean for Group

ParameterDominantHand

NondominantHand P Value*

Total score 65.6 60.8 .019Odometer 196.5 196.1 .29Lens injury 0.79 0.93 .021Corneal injury 0.28 0.47 .87Time (seconds) 70.0 71.6 .026

*For difference (Wilcoxon signed-rank test; P!.05 statistically significant)

J CATARACT REFRACT SURG - V

good acuity (6/6 and N6 or better) and stereopsis(mean 35 seconds of arc G 18 [SD]).

The total score was lower (PZ.019), operating timeswere longer (PZ.026), and lens injury was greater(PZ.021) when participants operated with thenondominant hand than when they operated withthe dominant hand. Differences for other parameterswere not statistically significant (Table 1).

Positive Spearman correlation analysis showed thatparticipantswithhigher scores (more efficient and safersurgery) for their dominant hand were more likely toscore highly with their nondominant hand (Table 2).

DISCUSSION

In this study, simulated microsurgery with the non-dominant hand was less efficient, less safe, and slowerthan simulatedmicrosurgerywith the dominant hand.This is perhaps not unexpected; however, to ourknowledge, this is the first study to test this variablein a formal manner. The relatively recent introductionof surgical simulators allows the testing of variablesthat were not previously examined because testingin a real-life operating room would be unethical anddangerous to patients.

The forceps module task was chosen because at thetime of this study, it was the only module to showconstruct validity.4 Recently, the capsulorrhexis mod-ule was validated,6 suggesting that this simulator,which is becoming increasingly popular in ophthalmictraining departments, is quite realistic.

If this observed effect for a validated simulator taskcould be transferred to a real-life operating session,there would be 2 key practical implications. First,when a trainee's dominant hand is not the same asthe supervisor's, the proposed surgical steps thateach surgeon is to perform should be first plannedand agreed on. The trainee should be encouraged touse his or her dominant hand to perform main tasks.Although this would mean that the supervisor mustperform nondominant-hand surgery at times, overallit would be safer for the patient because this study

Table 2. Surgical scores and Spearman correlation (r value) andP values for comparison between hands.

Mean for Group

ParameterDominantHand

NondominantHand r Value P Value

Total score 65.6 60.8 0.65 !.001Odometer 196.5 196.1 0.77 !.001Lens injury 0.79 0.93 0.59 .001Corneal injury 0.28 0.47 0.50 .005Time (seconds) 70.0 71.6 0.76 !.001

OL 38, DECEMBER 2012

2115SAFETY OF NONDOMINANT-HAND SURGERY

found that surgeons with greater skill in their domi-nant hand are more likely to have a greater degree ofskill with their nondominant hand. The supervisorcould alternatively create another port for his or herdominant hand; however, such an additional incisionmay not be required if the supervisor were able toproceed with the nondominant hand safely. Second,for tasks that require a high level of skill with thenondominant hand, it is sensible to first recommendthat the dominant hand is well adept before attempt-ing such tasks with the weaker hand. This is perhapsmost relevant for the steps in cataract surgery thatare becoming increasingly common tominimize surgi-cally induced astigmatism, such as bimanual phaco-emulsification, bimanual I/A, and facilitation ofon-meridian surgery. It is also relevant for PPV be-cause in phakic patients the vitrectomy cutter cannotsafely reach to the other side without inducing lenstouch (an unfortunate complication), which requirespart of the vitrectomy to be done with the nondomi-nant hand. Trainees new to PPV could therefore startlearning vitrectomy on pseudophakic eyes, in whichthe majority of the vitrectomy can be performed withthe dominant hand. This extrapolation to PPV surgeryremains a speculation, however, because simulatormodules for PPV are yet to be validated.

There are some limitations of this study. First, thenumber of physicians who could be recruited limitedour sample size. Acquisition of similar simulators bymore hospitals around the United Kingdom couldprovide a larger sample size through multicentercollaboration, although at present there are a limitednumber of these simulators available. However, oursample is larger than in previously published studiesof ophthalmic surgical simulators relating to valida-tion4 (15 subjects) or the effect of other variables onsurgical performance such as distraction1 (21 subjects),fatigue2 (7 subjects), and stereopsis3 (30 subjects).Furthermore, this sample size managed to yield statis-tically significant differences between dominant-handand nondominant-hand surgical performance.

Second, it remains difficult to remove the effect ofthe learning curve that surgeons experience as theyrepeat an identical task on the simulator. The currentstudy protocol is a reasonable way of reducing thiswithout compromising performance with fatigue, aswe established in other studies examining simulatedsurgical performance and distraction,1 fatigue,2 andloss of stereopsis.3 It is likely that the learning curveis not entirely removed and that there would besome residual small improvement in scores whilesubjects progressed through the research protocol.Given that the nondominant hand was tested afterthe dominant hand, any residual learning curvewouldbe likely to improve the nondominant hand scores and

J CATARACT REFRACT SURG - V

hence reduce the differences between the nondomi-nant hand and the more skillful dominant hand. Thedifference in scores between the hands is thereforepotentially larger than the raw data would imply.

Third, it is uncertain to what extent observations ina simulated environment can be transferred to a realoperating room. However, from a practical perspec-tive, variables such as nondominant surgery can onlybe assessed by simulation because to do so in a realoperating environment would endanger patients andbe unethical.

Also, simulators are very useful for surgical trainingin addition to completing research projects such asthis. The simulator allows the trainee to practice withhis or her nondominant hand ad nauseum. Then,when it is necessary for the trainee to perform the pro-cedure on his or her patients, it will be at a relativelysafer stage of progression.

To conclude, we have found that simulated ophthal-mic surgery with the nondominant hand was less effi-cient, less safe, and slower than simulated ophthalmicsurgery with the dominant hand. This observationwas more marked in those with less skill with theirdominant hand. Until now, surgeons have assumedtheir nondominant hand is perhaps less surgicallyskillful, and this study is the first to confirm this as-sumption and to quantify the difference in simulatedsurgical ability between the dominant hand and thenondominant hand.

This has practical implications for trainee andtrainer if 1 surgeon is left handed and 1 right handed.It also suggests that a higher degree of competencewith the dominant hand is required before the physi-cian performs nondominant hand surgery.

This study yields results that before the advent ofsimulation could not be obtained from real-life sur-gery. The project also highlights another training ad-vantage that simulators providedthe opportunity topractice with the nondominant hand until the surgeonis, potentially, surgically ambidextrous.

O

WHAT WAS KNOWN

� Surgeons perhaps assume that their nondominant hand isless skillful than their dominant hand; however, no re-search exists to confirm or quantify this.

WHAT THIS PAPER ADDS

� This paper is the first to confirm and also quantify the de-gree to which nondominant-hand surgery is less skillfulthan dominant-hand surgery and also highlights theimportance of recognizing this in the context of trainingophthalmic surgeons.

L 38, DECEMBER 2012

2116 SAFETY OF NONDOMINANT-HAND SURGERY

REFERENCES1. Park J, Waqar S, Kersey T, Modi N, Ong C, Sleep T. Effect of dis-

traction on simulated anterior segment surgical performance.

J Cataract Refract Surg 2011; 37:1517–1522

2. Waqar S, Park J, Kersey TL, Modi N, Ong C, Sleep TJ. Assess-

ment of fatigue in intraocular surgery: analysis using a virtual

reality simulator. Graefes Arch Clin Exp Ophthalmol 2011; 249:

77–81

3. Waqar S, Williams O, Park J, Modi N, Kersey T, Sleep T. Can vir-

tual reality simulation help to determine the importance of stere-

opsis in intraocular surgery. Br J Ophthalmol 2012; 96:742–746

4. Mahr MA, Hodge DO. Construct validity of anterior segment anti-

tremor and forceps surgical simulator trainingmodules; attending

versus resident surgeon performance. J Cataract Refract Surg

2008; 34:980–985

5. FrisbyJP.TheFrisbystereotest:amended instructions.BrOrthopt

J 1980; 37:108–112

J CATARACT REFRACT SURG - V

6. Privett B, Greenlee E, Rogers G, Oetting TA. Construct validity of

a surgical simulator as a valid model for capsulorhexis training.

J Cataract Refract Surg 2010; 36:1835–1838

OTHER CITED MATERIALA. EYESi Ophthalmic Surgery Simulator User Guide. Mannheim,

Germany, VRmagic Holding AG, 2006; 58�59

OL

38, DECEMBER 2012

First author:Jonathan Park, FRCOphth, BSc

Ophthalmology Department,South Devon Foundation NHS Trust,Torbay, Devon, United Kingdom