Comparison of gaze behaviour of trainee and experienced surgeons during laparoscopic gastric bypass

S. Erridge, H. Ashraf, S. Purkayastha, A. Darzi and M. H. Sodergren

Department of Surgery and Cancer, Imperial College London, London, UK

Correspondence to: Mr M. H. Sodergren, Division of Surgery, Department of Surgery and

Cancer, Imperial College London, Academic Surgical Unit, 10th Floor QEQM, St Mary’s

Hospital, South Wharf Road, London, W2 1NY, UK (e-mail: m.sodergren@imperial.ac.uk;

<TWITTER LOGO> @Simon_Erridge; <TWITTER LOGO> @abdosurgeon)

Presented to the 12th Annual Academic Surgical Congress, Las Vegas, Nevada, USA,

February 2017

Background: Eye tracking presents a novel tool that could be used to profile skill levels in

surgery objectively. The primary aim of this study was to identify differences in gaze

behaviour between expert and junior surgeons performing a laparoscopic Roux-en-Y gastric

bypass (LRYGB) for obesity.

Methods: This prospective observational study used a lightweight eye-tracking apparatus to

determine the difference in gaze behaviours between expert (> 75 procedures) and junior (≤

75 procedures) surgeons at defined stages of LRYGB. Primary endpoints were normalized

dwell time and fixation frequency. Secondary endpoints were blink rate, maximum pupil size

and rate of pupil change.

Results: A total of 20 procedures (12 junior, 8 expert) were analysed. Compared with juniors,

experts showed a prolonged dwell time on the screen during angle of His dissection (median

(range) 91.20 (83.40–94.40) versus 68.95 (59.80–87.60) per cent; P = 0.001), formation of the

retrogastric tunnel (91.50 (85.80–95.50) versus 73.60 (34.60–90.50) per cent; P = 0.001) and

gastric pouch formation (86.95 (83.60–90.20) versus 67.60 (37.10–80.00) per cent P < 0.001).

Juniors had a greater blink frequency throughout all recorded segments (P < 0.010) and had a

larger maximum pupil size during all recorded operative segments (P < 0.010). Rate of pupil

change was greater in juniors in all analysed segments (P < 0.010).

Conclusion: These results suggest that experts display more focused attention on significant

stimuli, alongside experiencing a reduced mental workload and having increased

concentration. This has the potential for future use in validation of surgical skill in high-stakes

assessment.

+A: IntroductionBariatric surgery is an effective method of treating the sequelae of obesity compared with

non-operative approaches1. Laparoscopic Roux-en-Y gastric bypass (LRYGB) has been

shown to be effective in the treatment of type 2 diabetes, hypertension and dyslipidaemia 2–4.

Although there is a clear need for an increased number of surgeons who can perform

LRYGB, it has a significant risk of complications; Li and colleagues3 reported a rate of 19.88

per cent in 3874 procedures. Birkmeyer and co-workers5 have previously demonstrated a

correlation between surgical skill, operative outcomes and case volume. There is a need for

competent surgeons to be trained and subsequently validated in their skill via novel methods.

Eye tracking utilizes infrared lights in combination with cameras to map the user’s

point of focus on to their field of view6. The study of gaze behaviours with eye tracking has

been shown to be of benefit in both training and validating surgical skill7–10.

Gaze training is of use in training surgeons for orientation during laparoscopic

procedures by identifying regions of interest7. It has also been shown to improve the

efficiency of laparoscopic tasks either by verbally informing trainees as to regions of interest

on which they should focus8 or via mapping the region of interest from a trainer on to the

screen to guide a trainee9.

Currently, validation of surgeons in the UK does not require any formal assessment of

surgical skills. Scoring systems have been developed, but there are limitations to their use

during procedures. Objective assessments, such as the Objective Structured Assessment of

Technical Skills (OSATS) scale, although useful for formative assessment of skills during

LRYGB, are time-intensive to undertake11,12. Motion analysis devices have been validated as

bench-side models for assessing laparoscopic skill13, including LRYGB14, but are yet to be

validated for use in determining skill during surgery.

2

Eye tracking has previously been used as an assessment tool in simulated

environments6,15,16. The potential for eye tracking to identify significant differences in the gaze

behaviour between expert and novice surgeons performing an open inguinal hernia repair has

been reported10. The present study evaluated gaze behaviour for a laparoscopic procedure.

The aim was to identify differences between expert and junior surgeons during laparoscopic

gastric bypass.

+A: MethodsThis prospective observational study was conducted at the Imperial Weight Centre, St Mary’s

Hospital, London, UK. It was performed under ethical approval from the West London and

GTAC Research Ethics Committee.

+B: Patients

All laparoscopic gastric bypass procedures carried out between February and May 2016 were

considered for inclusion in this study, including LRYGB and mini gastric bypass (MGB),

subject to acquisition of consent. All surgeons capable of performing a bypass as the primary

surgeon, who were comfortable and able to wear eye-tracking glasses, were invited to take

part. The patient cohort included those deemed eligible for a gastric bypass procedure after

consultation at the Imperial Weight Centre17. Patients who had undergone previous bariatric

surgery were excluded. Patient demographics including sex, BMI and ASA grade were

recorded. Additionally, the surgeon’s training grade and how many previous LRYGBs they

had done were recorded, alongside the training grade of the surgical assistant.

+B: Consent

Informed consent was obtained from both surgeons and patients on admission, with the option

to withdraw from the study at any point. The consenting process was modified during the

study, to include showing patients the eye-tracking apparatus, in response to feedback

questionnaires to improve patient involvement18.

+B: Data recording

Surgeons wore a pair of lightweight eye-tracking glasses (model 1.4; SensoMotoric

Instruments, Teltow, Germany) while performing the procedures. The iView software

3

(SensoMotoric Instruments) on the glasses collects eye movement and metrics, which are

transmitted via a cable into a portable recording device beneath the surgeon’s gown. The

glasses incorporate a forward-facing scene camera, which records the participant’s field of

view. The glasses additionally utilize dark pupil tracking to record eye movement 19,20. One-

point calibration is carried out before the procedure to allow for integration of pupil

positioning within the field of view. The glasses provide a binocular sampling rate of 30 Hz.

The accuracy of the pupil positioning is within 0.5° in all directions, with a spatial resolution

of 0.1°. The glasses are able to track within a range of 80° and 60° across the horizontal and

vertical axes respectively.

Intraoperative training of the operating surgeon was kept to a minimum, unless this

would impact adversely on patient safety. During data collection, a researcher was present

constantly to assess distractions unrelated to the surgical procedure; these were removed

subsequently at analysis.

+B: Procedure

The Imperial Weight Centre conducts LRYGB in a standard manner, with only minor

deviation in instrumentation used between surgeons. The steps of the procedure analysed in

this study were: preoperative set-up including port placement and liver retraction (segment 1);

dissection of the angle of His (segment 2); formation of the retrogastric tunnel (segment 3);

and construction of the gastric pouch (segment 4). The inclusion of the final segment of the

operation reflects its significance as a critical step in LRYGB21,22.

+B: Assessment

Areas of interest (AOIs) were determined before data analysis (Table 1). The primary study

outcome parameters were fixation frequency and gaze duration. Fixation frequency was

defined as the number of fixations on an AOI per second. Gaze duration is the cumulative

duration of all fixations on an AOI, including short saccades. The secondary parameters

analysed were maximum pupil size, pupil rate of change and blink rate. The maximum pupil

size is the maximum diameter of the pupil (mm) whilst engaging with the task. Pupil rate of

4

change describes the frequency in the change in diameter of the pupil (mm/s). Blink rate

(count/s) is the frequency of blinks recorded during the analysed segment.

Using the videos created by the glasses, subjects’ fields of view were analysed

manually via semantic gaze mapping. Eye metrics were subsequently determined from this

with respect to the AOIs. All extraneous distractions and segments unrelated to the procedure

were omitted from analysis.

Surgeons were distributed into expert (>75 procedures) and junior (≤75 procedures)

cohorts dependent on the number of LRYGB procedures they had previously performed. This

cut-off was determined based on evidence from a number of studies that define the learning

curve for LRYGB as between 75 and 100 procedures23–25.

After completion of the procedure, surgeons were asked to identify the difficulty of the

recorded segment on a 7-point Likert scale, in order to compare the difficulty of procedure

carried out by each cohort.

+B: Analysis

Demographic data of both junior and expert surgeons was compared using either a t-test or

Mann-Whitney U test. This was dependent on whether the data was parametric or non-

parametric respectively.

The recorded portion was divided into separate sections to account for technical

nuances between sections, as well as variance in anatomy. Segmental procedure analysis was

used to overcome this. The segments analysed included the four stages described above.

A single individual, who had received appropriate training to ensure consistency in

semantic gaze analysis, analysed all videos. The videos were calibrated to a set viewpoint on

which the subject had focused their gaze before analysis. The primary outcomes of dwell time

(per cent) and fixation frequency (count/s) were computed according to the stated AOIs,

whereas the secondary outcomes were calculated across the segment.

Previous studies6,26 have shown the comparison of gaze behaviours between subjects of

varying experience to be non-parametric. This data set was additionally confirmed to be non-

parametric via a Shapiro–Wilk test. Junior–expert analysis was carried out using a one-tailed

5

Mann–Whitney U test. All statistical tests were carried out using SPSS® version 22.0.0.0

(IBM, Armonk, New York, USA). The significance level was set at P < 0.050.

+A: ResultsA total of 23 procedures were recorded for the purpose of analysis. Of these, three (all junior

surgeons) were converted to a sleeve gastrectomy owing to insufficient bowel length to form

a Roux limb. As such they were removed from analysis for all segments apart from segment

1, as the set-up phase is identical. One expert and two junior data sets were excluded from

analysis of segment 1 (failure of equipment, 1; subject refusal to record data on that segment,

2).

Two trainees and two experienced surgeons did the procedure; their experience is

summarized in Table 2.

+B: Eye metrics

Table 3 displays a summary of all AOI metrics across the range of recorded segments, and

Table 4 provides a summary of the physiological eye parameters.

+B: Segment 1: preoperative set-up

During segment 1, experts had a longer normalized dwell time on the operative field

compared with juniors: median 23.20 (range 18.90–37.00) versus 11.40 (2.10–25.30) per cent

respectively (P = 0.004) (Fig. S1, supporting information). Juniors had a greater fixation

frequency on the operating theatre (0.09 (0.02–0.19) per s versus 0.03 (0.02–0.06) per s for

experts; P = 0.023).

Juniors had a greater median (range) blink frequency in segment 1 (1.00 (0.40–1.70)

per s versus 0.30 (0.30–0.40) per s for experts; P = 0.001), alongside a greater rate of pupil

change (0.37 (0.06–0.80) versus 0.24 (0.09–0.28) mm/s respectively; P = 0.009).

+B: Segment 2: dissection of the angle of His

During segment 2, experts had a longer normalized dwell time on the screen compared with

juniors (median 91.20 (range 83.40–94.40) versus 68.95 (59.80–87.60) per cent respectively;

P = 0.001) (Fig. S2, supporting information). Contrastingly, juniors had prolonged dwell

times on the operating theatre (0.15 (0.00–5.30) per cent versus 0.00 (0.00–2.70) per cent for

6

experts; P = 0.045), accompanied by a higher fixation frequency (0.01 (0.00–0.18) versus

0.00 (0.00–0.04) per s respectively; P = 0.036).

In segment 2, juniors again had a more frequent blink rate (median 1.15 (range 0.20–

1.70) per s versus 0.10 (0.10–0.20) per s for experts; P = 0.001). Additionally, they had a

larger maximum pupil diameter (4.80 (3.60–5.90) versus 3.50 (2.80–3.90) mm respectively;

P = 0.001] and rate of pupil change (0.27 (0.12–0.60) versus 0.11 (0.09–0.15) mm/s;

P = 0.001).

+B: Segment 3: retrogastric tunnel formation

In segment 3, experts had a longer dwell time on the screen compared with juniors (median

91.50 (range 85.80–95.50) versus 73.60 (34.60–90.50) per cent respectively; P = 0.001)

(Fig. S3, supporting information). Juniors had a greater fixation frequency on the operating

theatre (0.01 (0.00–0.16) per s versus 0.00 (0.00–0.04) per s for experts; P = 0.017).

Additionally, juniors had an increased fixation frequency on the scrub nurse (0.01 (0.00–0.06)

per s versus 0.00 (0.00–0.00) per s for experts; P = 0.011)

Similar to previous segments, juniors had an increased blink rate (median 0.55 (range

0.20–2.00) per s versus 0.10 (0.10–0.20) per s for experts; P < 0.001), a greater maximum

pupil size (5.20 (3.40–7.00) mm versus 3.75 (2.60–4.10) mm for experts; P = 0.001) and

greater rate of pupil change (0.23 (0.15–1.11) versus 0.12 (0.07–0.17) mm/s respectively; P <

0.001).

+B: Segment 4: gastric pouch formation

Throughout segment 4 experts had a longer normalized dwell time on the screen compared

with juniors (median 86.95 (83.60–90.20) versus 67.60 (37.10–80.00) per cent respectively;

P < 0.001) (Fig. S4, supporting information). Juniors had a larger fixation frequency on the

operative field (0.05 (0.02–0.12) per s versus 0.03 (0.02–0.05) per s for experts; P = 0.011).

Juniors again had a more frequent blink rate (median 0.95 (range 0.30–1.90) per s

versus 0.20 (0.10–0.20) per s for experts; P < 0.001). They also had a larger maximum pupil

size (5.30 (3.70–6.80) versus 3.70 (2.80–4.40) mm respectively; P = 0.001) and rate of pupil

change (0.26 (0.12–1.22) versus 0.11 (0.01–0.15) per s; P = 0.001).

7

+A: DiscussionThis study was able to distinguish between surgeons’ skill level in critical segments of live

laparoscopic surgery using eye-tracking technology. It suggests a potential for eye tracking to

be used as a tool to provide objective assessment of surgical performance. It could be used as

a training tool in LRYGB, and indeed the wider field of laparoscopic surgery. The key finding

of this study is the ability to distinguish between junior and expert surgeons with regard to

dwell times and fixation frequency at each operative segment assessed, in addition to blink

rate, maximum pupil diameter and rate of pupil change.

A high dwell time is understood to be a measure of higher cognitive attention, by

ignoring potentially distracting stimuli to focus on the given task27,28. This study shows that

expert surgeons had a significantly higher dwell time on the operative field in the

preoperative segment and on the screen in all segments. This demonstrates the experts’ ability

to concentrate on the most crucial aspect of the operation. This may be a result of an

accumulation of knowledge over a number of procedures, which leads them to have higher

concentration, or this may be a result of a more economical range of eye movements, again as

a result of their experience.

Fixation frequency can be used to demonstrate which AOIs are of most importance to

the surgeon. Throughout all segments the fixation frequency was greatest on the screen,

indicating its importance and validating the gaze strategy of expert surgeons. In preoperative

set-up, juniors had significantly more fixations per second on the operating theatre (median

0.09/s versus 0.03/s for experts; P = 0.023). This suggests that junior surgeons need more

assistance in preoperative set-up via visual and audio cues from the surrounding environment.

Similarly, during formation of the retrogastric tunnel the juniors had a higher fixation rate on

the operating theatre (0.01/s versus 0.00/s for experts; P = 0.017), and also on the scrub nurse

(0.01 versus 0.00/s respectively; P = 0.011). This indicates that juniors have a higher reliance

on the scrub nurse for instrument choice, potentially leading to reduced concentration on

important AOIs. During gastric pouch creation junior surgeons had an increased fixation

frequency on the operative field (0.05/s versus 0.03/s for experts; P = 0.011). This is most

8

likely the result of the higher number of instrument changes during this segment, and the

difference results from a reduced familiarity with inserting instruments within ports compared

with the expert cohort. Additionally, experts do not withdraw their concentration from the

screen as much when requesting and inserting another instrument.

These results are supported by previous work10 using eye tracking during open inguinal

hernia repair, in which experts also had a significantly increased dwell time on the operative

site. The statistically significant increase in dwell time for experts on critical AOIs also

conforms to the information reduction hypothesis29, which states that experts selectively

neglect redundant information and focus on pertinent information, such as the screen.

Previous work15,16 from a group in British Columbia, Canada, demonstrated that expert

surgeons had a reduced dwell time on a laparoscopic screen and increased saccades to patient

vital signs in a simulated environment. Although this contradicts the data presented in the

present study, it should be noted that the Canadian study was in a simulated environment

without the assistance of allied health professionals. The reason for this difference could be

that in an operating theatre experts are able to use other members of the team and verbal

communication skills to monitor the vital signs whilst maintaining concentration on the

operative screen.

With regard to secondary parameters, it was shown that experts had a significantly

lower blink rate throughout all recorded segments. A reduced blink rate has been shown

previously to correlate with improved concentration in allied fields such as aviation30 and

video game tasks31. This supports the previous deductions from dwell times and fixation

frequency regarding improved concentration and efficiency in expert surgeons.

Juniors had significantly larger maximum pupil diameters in all segments apart from

segment 1. Hess and Polt32 originally demonstrated a positive correlation between pupil

diameter and high cognitive workload. This has been consequently verified through multiple

examples across different fields33–35. The rate of change in pupil diameter has been shown to

be a better indicator of sustained cognitive workload throughout a recorded segment35,36. Here,

experts were shown to have a lower rate of change in all segments. This, along with

9

maximum pupil size, suggests that experts had a lower cognitive workload throughout the

whole procedure compared with juniors.

It is important to consider the limitations of this study. LRYGB is a technically

challenging procedure and consequently performed only by senior trainees and consultant

surgeons. This resulted in a limited number of surgeons being able to take part in this study.

Ideally the study should be replicated in additional centres to validate the findings.

Additionally, this study contained no sample size calculation, as no other data for live

laparoscopic surgery exist. It is possible that surgeons could succumb to the Hawthorne effect

and manipulate their eye movements in order to produce a falsely high reading on specific

AOIs. However, this is an issue that should affect both cohorts; in the future these data could

be used to provide a cut-off score to exclude samples that are deemed to have been

manipulated. Another limitation was that there was no standard for potential distractions in

the operating theatre. To minimize any potential impact, a researcher was present throughout

data collection to make note of, and exclude, any extraneous distractions from the analysis.

The experience of the operating theatre staff may have influenced the gaze behaviour of the

surgeons. The junior surgeons had a greater number of consultant grade assistants, which

should produce the opposite bias to the results presented. The analysis of maximum pupil size

has the potential to be affected by ambient lighting conditions; standard lighting was used

throughout, with ambient white light during preoperative set-up and ambient green lighting

throughout the other segments. The use of dark pupil-tracking technology reduced any

potential effects of ambient light on mapping fixations37. Finally, the analysis did not focus on

the other two critical steps of the bypass: the gastrojejunal and jejunojejunal anastomoses.

This was due to limitations in the length of recording available from the eye-tracking

apparatus.

Future aims for eye-tracking research should focus on both training and assessment.

For training, specifically identifying anatomical structures that experts use should improve

efficiency and orientation throughout the procedure. Studies27,28 have identified that gaze

strategies are a result of top-down processes to focus on useful stimuli and thus can be taught.

10

Additionally, work should continue to develop eye tracking as a tool for validating surgeon

technical skill level. The next step is to analyse any correlative or predictive relationships

between gaze behaviours and either clinically relevant patient outcomes or other assessment

metrics, such as the OSATS.

+A: Acknowledgements

The authors thank N. Fakih, C. Tsironis, S. Hakky and K. Moorthy of Imperial Weight

Centre, Imperial College Healthcare Trust, London, United Kingdom

This work was supported by a grant from the Royal College of Surgeons of England.

Disclosure: The authors declare no conflict of interest.

+A: References

1 Gloy VL, Briel M, Bhatt DL, Kashyap SR, Schauer PR, Mingrone G et al. Bariatric

surgery versus non-surgical treatment for obesity: a systematic review and meta-

analysis of randomised controlled trials. BMJ 2013; 347: f5934.

2 Ikramuddin S, Korner J, Lee W, Connett JE, Inabnet WB, Billington CJ et al. Roux-en-

Y gastric bypass vs intensive medical management for the control of type 2 diabetes,

hypertension, and hyperlipidemia: the Diabetes Surgery Study randomized clinical

trial. JAMA 2013; 309: 2240–2249.

3 Li JF, Lai DD, Lin ZH, Jiang TY, Zhang AM, Dai JF. Comparison of the long-term

results of Roux-en-Y gastric bypass and sleeve gastrectomy for morbid obesity: a

systematic review and meta-analysis of randomized and nonrandomized trials. Surg

Laparosc Endosc Percutan Tech 2014; 24: 1–11.

4 Carlin AM, Zeni TM, English WJ, Hawasli AA, Genaw JA, Krause KR et al. The

comparative effectiveness of sleeve gastrectomy, gastric bypass, and adjustable gastric

banding procedures for the treatment of morbid obesity. Ann Surg 2013; 257: 791–797.

5 Birkmeyer JD, Finks JF, O’Reilly A, Oerline M, Carlin AM, Nunn AR et al; Michigan

Bariatric Surgery Collaborative. Surgical skill and complication rates after bariatric

surgery. N Engl J Med 2013; 369: 1434–1442.

6 Tien T, Pucher PH, Sodergren MH, Sriskandarajah K, Yang G, Darzi A. Eye tracking

for skills assessment and training: a systematic review. J Surg Res 2014; 191: 169–178.

11

7 Sodergren M, Orihuela-Espina F, Froghi F, Clark J, Teare J, Yang G et al. Value of

orientation training in laparoscopic cholecystectomy. Br J Surg 2011; 98: 1437–1445.

8 Vine SJ, Masters RSW, McGrath JS, Bright E, Wilson MR. Cheating experience:

Guiding novices to adopt the gaze strategies of experts expedites the learning of

technical laparoscopic skills. Surgery 2012; 152: 32–40.

9 Chetwood AS, Kwok K, Sun L, Mylonas GP, Clark J, Darzi A et al. Collaborative eye

tracking: a potential training tool in laparoscopic surgery. Surg Endosc 2012; 26: 2003–

2009.

10 Tien T, Pucher PH, Sodergren MH, Sriskandarajah K, Yang G, Darzi A. Differences in

gaze behaviour of expert and junior surgeons performing open inguinal hernia repair.

Surg Endosc 2015; 29: 405–413.

11 Zevin B, Bonrath EM, Aggarwal R, Dedy NJ, Ahmed N, Grantcharov TP; ATLAS

Group. Development, feasibility, validity, and reliability of a scale for objective

assessment of operative performance in laparoscopic gastric bypass surgery. J Am Coll

Surg 2013; 216: 955–965.e8.

12 Moorthy K, Munz Y, Sarker SK, Darzi A. Objective assessment of technical skills in

surgery. BMJ 2003; 327: 1032–1037.

13 Grantcharov TP, Bardram L, Funch-Jensen P, Rosenberg J. Assessment of technical

surgical skills. Eur J Surg 2002; 168: 139–144.

14 Boza C, Varas J, Buckel E, Achurra P, Devaud N, Lewis T et al. A cadaveric porcine

model for assessment in laparoscopic bariatric surgery – a validation study. Obes Surg

2013; 23: 589–593.

15 <OTHER>Tien G, Atkins MS, Zheng B, Swindells C. Measuring situation awareness

of surgeons in laparoscopic training. Proceedings of the 2010 ACM Symposium on Eye-

Tracking Research and Applications, 2010; 149–152.

16 Zheng B, Tien G, Atkins SM, Swindells C, Tanin H, Meneghetti A et al. Surgeon’s

vigilance in the operating room. Am J Surg 2011; 201: 673–677.

12

17 <EPATH>National Institute for Health and Care Excellence. Obesity: Identification,

Assessment and Management. NICE Guideline (CG189); 2014.

https://www.nice.org.uk/guidance/cg189 [accessed 17 April 2016].

18 Erridge S, Ashraf H, Dilley J, Darzi A, Sodergren M. Eye tracking research: seen

through the patient’s eyes. BMJ Simul Technol Enhanced Learn 2016; 2: 101–102.

19 Wilson M, McGrath J, Vine S, Brewer J, Defriend D, Masters R. Psychomotor control

in a virtual laparoscopic surgery training environment: gaze control parameters

differentiate novices from experts. Surg Endosc 2010; 24: 2458–2464.

20 Wilson MR, McGrath JS, Vine SJ, Brewer J, Defriend D, Masters RSW. Perceptual

impairment and psychomotor control in virtual laparoscopic surgery. Surg Endosc

2011; 25: 2268–2274.

21 Schauer PR, Ikramuddin S, Hamad G, Eid GM, Mattar S, Cottam D et al. Laparoscopic

gastric bypass surgery: current technique. J Laparoendosc Adv Surg Tech 2003; 13:

229–239.

22 Scally CP, Varban OA, Carlin AM, Birkmeyer JD, Dimick JB; Michigan Bariatric

Surgery Collaborative. Video ratings of surgical skill and late outcomes of bariatric

surgery. JAMA Surg 2016; 151, e160428.

23 Oliak D, Ballantyne GH, Weber P, Wasielewski A, Davies RJ, Schmidt HJ.

Laparoscopic Roux-en-Y gastric bypass: defining the learning curve. Surg Endosc

2003; 17: 405–408.

24 Schauer P, Ikramuddin S, Hamad G, Gourash W. The learning curve for laparoscopic

Roux-en-Y gastric bypass is 100 cases. Surg Endosc 2003; 17: 212–215.

25 Søvik TT, Aasheim ET, Kristinsson J, Schou CF, Diep LM, Nesbakken A et al.

Establishing laparoscopic Roux-en-Y gastric bypass: perioperative outcome and

characteristics of the learning curve. Obes Surg 2009; 19: 158–165.

26 Edmondson MJ, Pucher PH, Sriskandarajah K, Hoare J, Teare J, Yang GZ et al.

Looking towards objective quality evaluation in colonoscopy: analysis of visual gaze

patterns. J Gastroenterol Hepatol 2016; 31: 604–609.

13

27 Gaschler R, Frensch PA. Is information reduction an item-specific or an item-general

process? Int J Psychol 2007; 42: 218–228.

28 Gaschler R, Marewski JN, Frensch PA. Once and for all – how people change strategy

to ignore irrelevant information in visual tasks. Q J Exp Psychol (Hove) 2015; 68: 543–

567.

29 Haider H, Frensch PA. Eye movement during skill acquisition: more evidence for the

information-reduction hypothesis. J Exp Psychol Learn Mem Cogn 1999; 25: 172–190.

30 Veltman J, Gaillard A. Physiological workload reactions to increasing levels of task

difficulty. Ergonomics 1998; 41: 656–669.

31 Yamada F. Frontal midline theta rhythm and eyeblinking activity during a VDT task

and a video game: useful tools for psychophysiology in ergonomics. Ergonomics 1998;

41: 678–688.

32 Hess EH, Polt JM. Pupil size in relation to mental activity during simple problem-

solving. Science 1964; 143: 1190–1192.

33 <OTHER>Marshall SP. The index of cognitive activity: measuring cognitive workload.

Proceedings of the 2002 IEEE 7th conference on Human factors and power plants,

2002.

34 <OTHER>Pomplun M, Sunkara S. Pupil dilation as an indicator of cognitive workload

in human–computer interaction. Proceedings of the International Conference on

Human–Computer Interaction, 2003; 542–546.

35 <OTHER>Palinko O, Kun AL, Shyrokov A, Heeman P. Estimating cognitive load

using remote eye tracking in a driving simulator. Proceedings of the 2010 ACM

Symposium on Eye-Tracking Research and Applications, 2010; 141–144.

36 Ahlstrom U, Friedman-Berg FJ. Using eye movement activity as a correlate of

cognitive workload. Int J Ind Ergonomics 2006; 36: 623–636.

37 Green P. Review of eye fixation recording methods and equipment. IVHS Technical

Report 1992.

14

Supporting information

Additional supporting information may be found online in the supporting information tab for this article:

Fig. S1 Normalized dwell times for areas of interest in operative segment 1 (Word document)

Fig. S2 Normalized dwell times for areas of interest in operative segment 2 (Word document)

Fig. S3 Normalized dwell times for areas of interest in operative segment 3 (Word document)

Fig. S4 Normalized dwell times for areas of interest in operative segment 4 (Word document)

15

Table 1 Areas of interest and their descriptions

Area of interest DescriptionScreen Any display containing the intra-abdominal imageAssistant The surgical assistantOperative field The area of the abdomen bordered by sterile drapesSterile field The area within the sterile drapes, excluding operative siteInstruments Any surgical instrumentation and hands whilst using instrumentAnaesthetist The anaesthetist or operating department practitionerScrub nurse The scrub nurseScrub table The table from which the scrub nurse provides sterile instrumentsOperating theatre An area within the operating theatre not defined by the previous terms

16

Table 2 Experience of the four surgeons involved in the study

Junior (n = 2)

Expert (n = 2) P‡

Sex ratio (M : F) 2 : 0 2 : 0 1.000No. of procedures 12 8

LRYGB 12 6Mini gastric bypass 0 2

Difficulty of procedure† 5 5 0.609Previous procedures performed* 20.5 (15–55) 300 (300–

1000)< 0.00

1Training grade

Bariatric fellow 1 0Consultant 1 2

Assistant gradesSpecialist registrar in surgery 1 2Bariatric fellow 1 1Consultant 3 0

Procedures performed per assistant grade

Specialist registrar in surgery 3 3Bariatric fellow 2 5Consultant 7 –

*Values are median (range). †Assessed by surgeon on a 7-point Likert scale. LRYGB, laparoscopic Roux-en-Y gastric bypass. ‡Mann–Whitney U test.

17

Table 3 Summary of area of interest in each of the four operative segments

Values are median (range). Op field – Operative field, Op Theatre – Operating Theatre

*P < 0.050, **P < 0.010, ***P < 0.001 (Mann–Whitney U test).

18

Table 4 Physiological eye parameters during each operative segment

*

P < 0.050, **P < 0.010, ***P < 0.001 (Mann–Whitney U test).

19