- Email: [email protected]
Robotic Assistance Confers Ambidexterity to Laparoscopic Surgeons
Accepted Manuscript Title: Robotic Assistance Confers Ambidexterity to Laparoscopic Surgeons. Author: Souzana Choussein, Serene S. Srouji, Leslie V. Farland, Ashley Wietsma, Stacey A. Missmer, Michael Hollis, Richard N. Yu, Charles N. Pozner, Antonio R. Gargiulo PII: DOI: Reference:
S1553-4650(17)30397-7 http://dx.doi.org/doi: 10.1016/j.jmig.2017.07.010 JMIG 3198
To appear in:
The Journal of Minimally Invasive Gynecology
Received date: Revised date: Accepted date:
17-12-2016 28-6-2017 6-7-2017
Please cite this article as: Souzana Choussein, Serene S. Srouji, Leslie V. Farland, Ashley Wietsma, Stacey A. Missmer, Michael Hollis, Richard N. Yu, Charles N. Pozner, Antonio R. Gargiulo, Robotic Assistance Confers Ambidexterity to Laparoscopic Surgeons., The Journal of Minimally Invasive Gynecology (2017), http://dx.doi.org/doi: 10.1016/j.jmig.2017.07.010. This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
Choussein 1
Robotic assistance confers ambidexterity to laparoscopic surgeons.
2
Souzana Choussein, M.D.1, Serene S. Srouji, M.D.1, Leslie V. Farland ScD2, Ashley
3
Wietsma, M.D.3, Stacey A. Missmer ScD1,4, Michael Hollis , B.S.3, Richard N. Yu,
4
M.D., PhD.3, Charles N. Pozner, M.D.5, Antonio R. Gargiulo, M.D.1
1
5 6
1
7
and Reproductive Biology, Brigham and Women’s Hospital, Harvard Medical School,
8
Boston, Massachusetts
9
2
Center for Infertility and Reproductive Surgery, Department of Obstetrics, Gynecology
Department of Epidemiology, Harvard T.H. Chan School of Public Health, Boston,
10
Massachusetts
11
3
12
Massachusetts
13
4
14
Massachusetts; Channing Division of Network Medicine, Brigham and Women's Hospital
15
and Harvard Medical School, Boston, Massachusetts
16
5
17
School, Boston, Massachusetts; Department of Emergency Medicine, Brigham and
18
Women's Hospital, Harvard Medical School, Boston, Massachusetts.
Department of Urology, Boston Children's Hospital, Harvard Medical School, Boston,
Department of Epidemiology, Harvard T.H. Chan School of Public Health, Boston,
Neil and Elise Wallace STRATUS Center for Medical Simulation, Harvard Medical
19 20
Corresponding author: Antonio R. Gargiulo, MD
21
Center for Infertility and Reproductive Surgery, Brigham and Women’s Hospital,
22
Harvard Medical School, 75 Francis St., ASB1-3, Boston, MA, 02115;
23
Tel: +1-617-732-4285; Fax: +1-617-975-0870; email: [email protected]
Page 1 of 26
Choussein
2
24
Disclosures: Dr. Gargiulo serves as a Consultant for Medicaroid, Inc. and OmniGuide
25
Holdings, Inc. Drs. Choussein, Srouji, Wietsma, Yu, Farland, Missmer, Hollis and Pozner
26
have no conflicts of interest or financial ties to disclose.
27
Funding: Nothing to report.
28 29
Precis: Robot-assisted laparoscopy eliminates the operative handedness observed in
30
conventional laparoscopy, conferring virtual ambidexterity to surgeons in training.
31 32
ABSTRACT
33
Study objective: To examine whether a robotic surgical platform can complement the
34
fine motor skills of the non-dominant hand, compensating for the innate difference in
35
dexterity between the surgeon’s hands, thereby conferring virtual ambidexterity.
36
Design: Crossover intervention study.
37
Design classification: Level II-1 evidence.
38
Setting: Centers for medical simulation of two tertiary care hospitals of Harvard Medical
39
School.
40
Participants: Three different cohorts of subjects were included: 1) surgical novices
41
(medical graduates with no robotic/laparoscopic experience) 2) surgeons in training
42
(PGY3-4 residents and fellows with intermediate robotic and laparoscopic experience)
43
and 3) advanced surgeons (attending surgeons with extensive robotic and laparoscopic
44
experience).
45
Interventions: Each study group completed three dry lab exercises based on exercises
46
included in the Fundamentals of Laparoscopic Surgery (FLS) curriculum. Each exercise
47
was completed four times: using their dominant and non-dominant hand, on a standard
Page 2 of 26
Choussein
3
48
laparoscopic FLS box trainer and in a robotic dry lab set-up. Participants were
49
randomized to the handedness and setting order in which they tackled the tasks.
50
Measurements and Main Results: Performance was primarily measured as time to
51
completion, with adjustments based on errors. Means of performance for the dominant
52
versus non-dominant hand for each task were calculated and compared using a repeated
53
measures ANOVA test.
54
A total of 36 subjects were enrolled study-wide (12 per group). In the laparoscopic
55
setting, the overall time to completion for all three tasks with the dominant hand was
56
significantly different from that with the non-dominant hand (439.4s vs. 568.4s
57
respectively; p=.0008). The between-hand performance difference was nullified with the
58
robotic system (374.4s vs. 399.7s;p=.48). Evaluation of performances for each individual
59
task also revealed a statistically significant disparate performance between hands for all
60
three tasks when the laparoscopic approach was utilized (p=.003, p=.02, p=.01
61
respectively). No between-hand difference was observed when the tasks were performed
62
robotically. When the above analysis was performed within the three surgical experience
63
cohorts, the performance advantage of robotic technology remained significant for the
64
surgical novices and intermediate level experience groups.
65
Conclusion: Robot-assisted laparoscopy may eliminate the operative handedness
66
observed in conventional laparoscopy, allowing for virtual ambidexterity. Such
67
ergonomic advantage is particularly evident on surgical trainees. Virtual ambidexterity
68
may represent an additional aspect of surgical robotics that facilitates mastery of
69
minimally-invasive skills.
Page 3 of 26
Choussein 70
Keywords: robotic surgery, laparoscopic surgery, intraoperative handedness,
71
ambidexterity, daVinci surgical system
4
72 73
Presented at the 45th AAGL Annual Global Congress On Minimally Invasive
74
Gynecology, Orlando, FL. November 14-18, 2016 (oral presentation, Plenary Robotics
75
Session).
76 77
Page 4 of 26
Choussein
5
78
INTRODUCTION
79
Ambidexterity has been identified as a critical factor in the achievement of laparoscopic
80
psychomotor surgical skill proficiency, and constitutes a desirable attribute for all
81
surgeons.[1-3] In terms of efficiency optimization, this is a particularly useful adjunct in
82
minimally invasive surgery, where the capabilities of instrument switching or body
83
position readjustment are limited. This often necessitates the use of the surgeon’s non-
84
dominant hand for delicate and precise tasks.[4] True ambidexterity, however is rare [5],
85
and possibly nonexistent.[6]
86
The robotic surgical platform has provided several ergonomic advancements to the field
87
of minimally invasive surgery, including the physiologic use of the surgeon’s wrist,
88
tremor filtration, motion scaling and three-dimensional vision. These advantages have
89
contributed to its increasing acceptance and growth as an alternative minimally invasive
90
surgical option. These same advantages have been proposed as the main factors that
91
shorten the learning curve of surgical skills: an important benefit, given the relatively
92
extensive learning period required to acquire these skills using laparoscopic platforms.
93
[7, 8] Access to virtual reality simulation, employing commercially available integrated
94
training protocols is another aspect of robotic surgery that has been shown to facilitate
95
negotiation of the surgical learning curve.[9] In summary, several aspects of robotic
96
surgery contribute to technical proficiency being achieved through a shorter and more
97
patient-centered learning curve compared to other types of invasive and minimally
98
invasive surgical techniques. Virtual ambidexterity in the robotic setting may constitute
99
an additional factor explaining the "enabling" role of computer-assisted laparoscopy and
100
its truncated learning curve.
Page 5 of 26
Choussein
6
101 102
There are two studies that previously examined the impact of the robotic platform on
103
surgeon’s handedness. However, none of them used experience-based stratification of the
104
study population, and neither introduced a laparoscopic component, as they both
105
examined open vs. robotic approaches. [10, 11]
106
With the majority of surgical operations now being performed via minimally invasive
107
routes, a study assessing the impact of robotic-assistance on hand dominance as
108
compared to laparoscopy appears more relevant than prior published literature that only
109
used an open surgery arm of comparison.
110
Given the aforementioned gap, the current study aims to examine whether robotic
111
systems, as compared to laparoscopy, can augment fine motor skills of the non-dominant
112
hand, compensating for surgeon’s innate difference in dexterity between each hand, and
113
thus conferring “virtual” ambidexterity.
114 115
MATERIALS AND METHODS
116
Subjects and Design
117
This was a crossover intervention study conducted at the Neil and Elise Wallace
118
STRATUS Center for Medical Simulation of Brigham and Women’s Hospital and at
119
Boston Children’s Hospital; both Harvard affiliated academic medical centers in Boston.
120
Brigham and Women’s Hospital and Boston Children’s Hospital Institutional Review
121
Board approvals were obtained. Volunteers were solicited from the Obstetrics &
122
Gynecology and Urology Departments of the aforementioned hospitals.
Page 6 of 26
Choussein
7
123
An email invitation was extended to the eligible staff members of the aforementioned
124
departments and their satellites, and participation was determined based on matched
125
availability (of the potential participant on the one end and the required facility on the
126
other). The study population was composed of three groups: Group 1: Surgical novices
127
with no robotic or laparoscopic experience i.e medical graduates with no residency
128
training (Surgery-naive subjects); Group 2: Surgeons in training (PGY 3-4 residents and
129
fellows with intermediate robotic and laparoscopic experience); and Group 3: Advanced
130
surgeons (attending surgeons with extensive robotic and laparoscopic experience).
131
Novice group was comprised of surgically untrained medical graduates serving as
132
Research Fellows at Obstetrics & Gynecology and Urology Departments of the
133
aforementioned hospitals. All of the residents and fellows included in this study (as
134
Group 2) had completed structured, stepwise curriculum requirements which included
135
standardized robotic simulation exercises, case observations, and stepwise involvement at
136
the console on live patient surgeries. At the time of their involvement on the study, all
137
had completed the basic requirements and had been fully involved in live console
138
surgery. Similarly, they had had several years of exposure to laparoscopic surgery, as part
139
of the standardized curricula for Accreditation Council for Graduate Medical Education
140
(ACGME)-accredited Obstetrics and Gynecology and Urology residency and fellowship
141
programs. Surgeons in Group 3 held robotic privileges at their institutions and reported
142
performing a minimum of 250 robotic cases and a minimum of 250 laparoscopic cases at
143
the time of enrollment. Individuals who self-reported as ambidextrous were excluded
144
from the study. All enrolled subjects watched a short video tutorial developed by an
145
expert surgeon. The video used a step-by-step process to demonstrate each task –
Page 7 of 26
Choussein
8
146
performed both laparoscopically and robotically-, and oriented participants to the
147
instruments used, ideal body positioning, and skills specific to the robot such as master
148
clutching. Each participant completed a short questionnaire at enrollment that collected
149
information about their demographics, current surgical case loads and prior experience
150
with surgical simulators and computer games.[12]
151
To assess differences in hand dexterity between the laparoscopic and robotic surgical
152
approach, each subject performed two sets (laparoscopic and robotic) of three exercises
153
inspired by skill tests required for the Fundamentals of Laparoscopic Surgery (FLS)
154
certification. Only those FLS skill tests that can be performed both laparoscopically and
155
with the robot were utilized. These are: the peg transfer (PT), the precision cutting (PC),
156
and the intracorporeal suturing (IS). Such skill tests were modified to serve the unilateral
157
performance requirement of the study. Other FLS skill tests not included in our study
158
were the endoloop placement and the extracorporeal suturing/knot tying.[8] Construct
159
validity of the FLS model has been demonstrated for robotic surgery. [5, 13] Exercises
160
were completed on a standard FLS box trainer (Limbs and Things, Savannah, GA)
161
located in a medical simulation center, for the laparoscopic arm. For the robotic arm,
162
exercises were completed in a robotic dry lab set-up. This involved the use of all the
163
components of the robot (surgical console, patient-side cart, EndoWrist instruments, and
164
vision cart) and the lower tray of a standard FLS laparoscopic box trainer, containing the
165
FLS test disposables.
166
We utilized a modified version of the PT task, eliminating the bi-manual subcomponents
167
of the original FLS curriculum, so that one hand at a time could be evaluated.
168
Specifically, each subject was asked to grasp six rubber objects placed on a pegboard,
Page 8 of 26
Choussein
9
169
one at a time, with a Maryland grasper and place them on a peg on the opposite side of
170
the board. (Figure 1.) The PC task required participants to cut out a circle from a 4x4
171
gauze piece suspended between clips, by cutting between two premarked circular
172
tracings. Participants were instructed to use a Maryland grasper in one hand to provide
173
traction and exposure on the gauze and to place it at the best possible angle to the cutting
174
hand being evaluated. No instrument exchange between hands was allowed. (Figure 2.)
175
Being largely bimanual, the IS task was also modified to serve our study aim. Participants
176
were required to drive a 2.0 Silk suture of 15cm length through two marks on a stationary
177
Penrose drain, starting the task with the designated hand being examined. Both hands
178
were occupied with curved needle-drivers. Once the needle was passed, the participants
179
were required to tie two double throws and one single throw by tying with the hand under
180
examination and the other hand holding the suture with the needle. (Figure 3.) These
181
modifications of the original FLS skill tests allowed for each task to be performed using
182
the dominant or the non-dominant hand, one at a time. Each participant completed each
183
task both laparoscopically and robotically on the same day. Robotic procedures were
184
performed using the da Vinci Surgical System (Intuitive Surgical Inc, Sunnyvale, CA).
185
Participants were randomized to which hand (dominant vs. nondominant) and setting
186
(laparoscopic vs. robotic) was completed first; there was no alternation in setting and all
187
three tasks were performed in-series prior to moving to the remaining setting.
188
Performance Assessment
189
To assess the differential in hand performance afforded by the laparoscopic versus the
190
robotic approach, the raw time for task completion was recorded. A time penalty was
191
applied for errors or lack of precision according to Table 1.
Page 9 of 26
Choussein 10 192
Statistical Analysis
193
The sample size of our study with paired subjects was sufficient to detect a 10% mean
194
difference (in seconds) between the subject’s dominant and non-dominant hand. This
195
threshold was chosen a priori to reflect the minimum difference between hands and
196
between laparoscopic and robotic modality that would be clinically meaningful with
197
respect to surgery (here quantified by the skills tests) duration. Given this minimum
198
detectable clinically meaningful difference, we calculated the sample size requiring a
199
minimum of 80% power and an α-level=0.05, using a repeated measures ANOVA
200
statistical approach.
201
Given the crossover study design [14] and the randomization of handedness and setting
202
order in which the participants tackled the tasks, there is little concern about any
203
confounding that remains between groups.[15] Mean values of performance using
204
dominant versus non-dominant hand in the laparoscopic and the robotic setting were
205
compared using a repeated measures ANOVA to take into account correlation between
206
measurements. Analyses were stratified by level of training. Sensitivity analyses were
207
conducted using generalized estimating equations (GEE) with robust “sandwich”
208
standard error estimates for linear regression. Findings from this sensitivity analysis were
209
not different from the original analysis and thus the repeated measures ANOVA results
210
are presented in the manuscript and all tables.
211 212
RESULTS
213
A total of 36 subjects were enrolled in the study with 12 in each group, each with four
214
timed measurements. Among novices, one was unable to complete the laparoscopic IS
Page 10 of 26
Choussein 11 215
task and thus, that individual’s results were excluded from the entire study. Gender and
216
handedness distribution among groups as well as prior experience with computer games
217
and simulators are shown in Table 2. The majority of our sample was male (62.5%) and
218
right-handed (98.2%). Approximately half of our subjects had previous experience with
219
laparoscopic or robotic simulators, not including our novice group which was simulator-
220
naive. Forty three percent reported prior experience with computer games, however
221
experience with computer games was most commonly reported among the novice group.
222
Surgical case load (either laparoscopic or robotic) and time since last surgical case
223
(applicable only to intermediate and advanced experience groups) are also noted. Among
224
the intermediate group, the average number of cases/week was 2.7 and the reported time
225
since last case ranged between 0 and 96 weeks, with a median of 12.5 weeks. Among the
226
expert group, average number of cases/week was 3.7; all of them reported performing a
227
surgery within the preceding two weeks.
228
Table 3 displays an assessment of the overall completion time of all three of the technical
229
tasks combined using laparoscopic and robotic approaches. In the laparoscopic setting,
230
the cumulative execution time with the dominant hand was significantly shorter than with
231
the non-dominant hand (439.4s vs. 568.4s respectively; p=.0008). This difference was
232
nullified when the technical tasks were performed on the robot (dominant: 374.4s vs.
233
non-dominant: 399.7s; p=.48). Of note, the median overall time to completion with either
234
the dominant or non-dominant hand was almost identical when the subjects used the
235
robot (dominant: 322s vs. non-dominant: 324s). Cumulatively, the mean difference in
236
raw time scores between the two hands for all three tasks was also significantly shorter
Page 11 of 26
Choussein 12 237
when the robot was utilized as compared to laparoscopic performance respectively (76.9s
238
vs. 167.8s; p<.0001).
239
Evaluation of each task individually also revealed a statistically significant difference in
240
performance between hands for all PT, PC, IS when the laparoscopic approach was
241
utilized (p=.003, p=.02, p=.01 respectively). No statistically significant between-hand
242
difference was observed when the tasks were performed robotically. (Table 4)
243
When stratifying by experience, the minimization of handedness afforded by the robotic
244
platform remained statistically significant among novice and intermediate operators.
245
In particular, we observed a statistically significant between-hand difference in the
246
laparoscopic setting performance of both novices and intermediate-level operators
247
(dominant: 666.0s vs. non-dominant: 988.6s; p=.01 and dominant: 351.5.0s vs. non-
248
dominant: 437.8s; p=.02, respectively) but not of the expert group (dominant: 319.6s vs.
249
non-dominant: 313.7; p=.82) No difference in relative hand performance was observed
250
for the novice and intermediate subjects when the tasks were performed robotically;
251
however, a difference was noted for the expert group. (dominant: 251.8s vs. non-
252
dominant: 297.3; p=.004). (Table 5)
253 254
DISCUSSION
255
For surgeons, it has been argued that ambidexterity represents a professional advantage,
256
if not an implicit requirement. [16, 17] Robotic surgical platforms appear to complement
257
motor skills of the non-dominant upper limb, thereby conferring virtual ambidexterity.
258
This is another piece of the puzzle to explain, in a scientific way, the "enabling" nature of
259
robot-assisted surgery.[10, 11] Virtual ambidexterity may represent an additional aspect
Page 12 of 26
Choussein 13 260
of surgical robotics that facilitates mastery of minimally-invasive skills through a
261
truncated, patient-friendly learning curve.
262
Technological innovations, health care cost considerations and patients’ expectations of
263
modern surgery, have all contributed to make minimally invasive surgery, the gold
264
standard for most intra-abdominal operations. Yet, the skill set needed for advanced
265
laparoscopy is hard to master due to the fundamental ergonomic challenges associated
266
with working through a fulcrum (anterior abdominal wall), the lack of three-dimensional
267
vision, and a reduced range of motion as a result of the inability to use wristed action
268
movements. Due to these limitations, even the surgeon’s innate dexterity cannot be fully
269
exhibited, as he or she negotiates the learning curve.[18, 19] Being hard to learn,
270
laparoscopy is harder to teach through the classic “see one, do one, teach one” model.
271
Robotic surgery is poised to address these limitations. To take it a step further,
272
eliminating intraoperative handedness can be expected to positively impact the surgical
273
learning curve, particularly benefitting surgical trainees and surgeons at the early stages
274
of their robotic learning curve, as shown by our data.
275
Studies demonstrating enhanced dexterity with the use of a robotic surgical system over
276
traditional laparoscopy have frequented the literature.[6, 18, 20] In an attempt to quantify
277
the purported advantage of robotic approaches vs. laparoscopy, Moorthy et al. reported a
278
nearly 50% increase in dexterity when employing the robot; attributing this to the tremor
279
cancellation, motion scaling and seven degrees of freedom (i.e specific, defined modes in
280
which a mechanical system can move) afforded by the robot.[18] Also, there is published
281
evidence that bimanual object manipulation is mediated by callosal fibers that
282
interconnect the cerebral hemispheres and that there is a preferential activation of this
Page 13 of 26
Choussein 14 283
human “mirror” system in surgeons operating with the robotic system compared with
284
those performing conventional laparoscopy.[21]
285
As previously mentioned, the impact of a robotic surgical system on a surgeon’s
286
handedness has been examined.[10, 11] Mucksavage et al. conducted manual dexterity
287
assessments of 19 robotic novices using open and robotic approaches to perform the
288
Purdue Pegboard test – in which participants must place as many peg assemblies into the
289
board as possible within a 30-second interval- and a needle targeting test where
290
participants were asked to hit as many bull’s eye paper targets as possible within 10
291
seconds. The performance scores for each hand were statistically different when the open
292
approach was used; however this difference was nullified when the tasks were performed
293
robotically.[11] In a similarly designed study, Badalato et al. concluded that robotic
294
surgery is capable of eliminating handedness among relative novices who performed
295
basic manual dexterity skill tests modified from FLS using the open and the robotic
296
technique.[10]
297
In the era of minimally invasive surgery, comparing the facilitation of ambidexterity
298
between laparoscopic skills and robotic skills is more germane than any comparison with
299
an open approach. Our study is the first to specifically address this question.
300
This study revealed a statistically disparate performance between hands for all three tasks
301
when the laparoscopic approach was utilized. The use of the robot resulted in leveling of
302
performance between the dominant and the non-dominant hand. Equilibration of hand
303
performance has been attributed to enhanced performance of the non-dominant hand as a
304
result of either tremor filtration or motion scaling.[22] Published evidence has established
305
that motion scaling rather than tremor filtration, is primarily responsible for the enhanced
Page 14 of 26
Choussein 15 306
accuracy observed with robotic assistance with more prominent gains in the non-
307
dominant, innately less dexterous, hand.[20, 23]
308
We felt that the multidimensional skill set and versatile technical attributes required
309
during an actual surgical operation are better reflected when combining the results of all
310
three tasks; that was the rationale for calculating and analyzing cumulative execution
311
time. Even in this context, the significant between-hand difference in overall time to
312
completion that was noted when the tasks were performed laparoscopically, was negated
313
when switching to the robotic setting.
314
These results are consistent with the findings of the aforementioned pertinent studies.
315
None of the pre-existing studies, however, has used an experience-based stratification of
316
the study population. At a time where a philosophy of patient-centered medicine is
317
shaping the way we can ethically teach surgery, we strongly believed that this type of
318
study should stratify subjects by surgical experience. Assuming that pure novices and
319
robotic experts occupy the two edges of the robotic surgery experience spectrum, there
320
should be an intermediate one in between them to span it entirely. This is, at its best,
321
represented by senior residents and fellows, and that was the rational for selecting the
322
intermediate study groups as such.
323
In the laparoscopic setting, we found a statistically significant between-hand difference
324
in the performance of both novices and intermediate-level operators. This difference was
325
not identified in the expert group. In the robotic setting, no difference in relative hand
326
performance was observed in the novice and intermediate subjects. However, a difference
327
was found in the expert group.
Page 15 of 26
Choussein 16 328
The inverted pattern observed in the novice-intermediate groups compared to the expert
329
group substantiates previously published work indicating that the performance advantage
330
of robotic technology is more profound for novices than for experts. [6] Re-emergence of
331
handedness among experts in the robotic setting may be reflective of benefit
332
maximization emanating from optimal use of the robotic platform. Once out of the
333
robotic learning curve, both hands still perform better than at conventional laparoscopy,
334
but the dominant hand takes the lead again. Indeed, the dominant hand robotic time of the
335
expert group was the shortest time reported in our study. One could also argue that the
336
intracorporeal suturing task poses a significantly increased level of difficulty to the
337
novice and intermediate-experience groups compared to the expert one, for which may
338
not necessarily comprise an adequately challenging module. It is possible that, had more
339
complex tasks been employed, even experts would have experience a persistent enabling
340
effect of virtual ambidexterity. However, for the purpose of this study, we thought that it
341
was important to limit the comparison to exercises based on a previously validated
342
curriculum, that could be expected to be completed by all three levels of operators.
343
Our study was strengthened by randomizing the order of performance by hand (right or
344
left) and setting (laparoscopic or robotic). This counterbalanced approach controlled for
345
any learning or warming up effects emanating from the completion of the first trial of
346
each session,[15] and likely contributed to the avoidance of a type II error. This was
347
especially important as we found no between-hand difference in the laparoscopic
348
performance of experts. The crossover design of the study,[14] enabled each participant
349
to serve as his or her own control, eliminating the confounding factors of gender, number
350
of cases performed, time since last case, and experience with computer games.[12, 24] Of
Page 16 of 26
Choussein 17 351
note, even within the context of a subgroup analysis of participants with computer game
352
experience, performance patterns remained the same.
353
There are limitations to our study. Although procedure time is considered an acceptable
354
performance metric, it is likely only a proxy indicator of performance, since it does not
355
account for skillful and proper performance. To mitigate this effect, a time penalty was
356
employed where applicable to encourage precise execution. There is published literature
357
where manual dexterity -defined as manual speed and ambidexterity- is measured by
358
utilizing the sums and differences of the right and left hand performance times (only).
359
Based on that, we extrapolated performance time to manual dexterity.[19] Given that the
360
robot-assisted technology is aimed at improving surgical precision and dexterity rather
361
than reducing task time, however, we may have underestimated or failed to reflect the
362
extent of the robotic enabling effect. It is likely that the benefits of robotic assistance will
363
be more profound when the path length and smoothness of task execution are evaluated
364
as well.[25] Also, although the utilized skill drills have been validated to be performed
365
with both hands, they all had be modified to serve the purpose of our study. However, as
366
our study was designed to compare between-hand performance and not laparoscopic and
367
robotic systems nor overall surgical competence per se, implications of these limitations
368
on our conclusions must be insubstantial.
369 370
Conclusion
371
Our study demonstrates that robotic-assistance confers ambidexterity to surgeons in
372
training, compared to conventional laparoscopy. Further investigation is warranted to
Page 17 of 26
Choussein 18 373
define the ultimate clinical value of this particular robotic attribute with respect to actual
374
operative outcomes and patient safety in the surgical learning curve.
375 376
Acknowledgments: The authors would like to thank the staff of the Neil and Elise
377
Wallace STRATUS Center for Medical Simulation at Brigham and Women’s Hospital
378
for their cooperation and overall assistance with study logistics.
379
Page 18 of 26
Choussein 19 380
References
381
1.
382 383
training and evaluation of laparoscopic skills. Am J Surg. 1998;175:482-487. 2.
384 385
3.
Skinner A, Auner G, Meadors M, et al. Ambidexterity in laparoscopic surgical skills training. Stud Health Technol Inform. 2013;184:412-416.
4.
Nutan J. State of the Art Atlas and Textbook of Laparoscopic Suturing. Jaypee Brothers Publishers, 2006.
388 389
Rosser JC, Rosser LE, Savalgi RS. Skill acquisition and assessment for laparoscopic surgery. Arch Surg. 1997;132:200-204.
386 387
Derossis AM, Fried GM, Abrahamowicz M, et al. Development of a model for
5.
Elneel FH, Carter F, Tang B, et al. Extent of innate dexterity and ambidexterity
390
across handedness and gender: Implications for training in laparoscopic
391
surgery. Surg Endosc. 2008;22:31-37.
392
6.
Chandra V, Nehra D, Parent R, et al. A comparison of laparoscopic and robotic
393
assisted suturing performance by experts and novices. Surgery.
394
2010;147:830-839.
395
7.
Mehta S, Dasgupta P, Challacombe BJ. Robotic reconstructive urology:
396
possibilities for the urological surgeon beyond the prostate. BJU Int.
397
2010;106:1247-1248.
398
8.
Moore LJ, Wilson MR, Waine E, et al. Robotic technology results in faster and
399
more robust surgical skill acquisition than traditional laparoscopy. J Robot
400
Surg. 2015;9:67-73.
Page 19 of 26
Choussein 20 401
9.
Culligan P, Gurshumov E, Lewis C, et al. Predictive validity of a training
402
protocol using a robotic surgery simulator. Female Pelvic Med Reconstr Surg.
403
2014;20:48-51.
404
10.
Badalato GM, Shapiro E, Rothberg MB, et al. The da vinci robot system
405
eliminates multispecialty surgical trainees' hand dominance in open and
406
robotic surgical settings. JSLS. 2014;18.
407
11.
innate hand dominance. J Endourol. 2011;25:1385-1388.
408 409
Mucksavage P, Kerbl DC, Lee JY. The da Vinci((R)) Surgical System overcomes
12.
Grantcharov TP, Bardram L, Funch-Jensen P, et al. Impact of hand dominance,
410
gender, and experience with computer games on performance in virtual
411
reality laparoscopy. Surg Endosc. 2003;17:1082-1085.
412
13.
Stefanidis D, Hope WW, Scott DJ. Robotic suturing on the FLS model
413
possesses construct validity, is less physically demanding, and is favored by
414
more surgeons compared with laparoscopy. Surg Endosc. 2011;25:2141-
415
2146.
416
14.
Basic Study Design (Chapter 5). In: Friedman LM, Furberg CT, DeMets DL,
417
Eds. Fundamentals of Clinical Trials. 4th ed. New York: Springer; 2010: 79-
418
80).
419
15.
Repeated Measures Designs. In: Shaughnessy JJ, Zechmeister EB, Zechmeister
420
JS, eds. Research Methods in Psychology. 5th ed. New York: McGraw Hill;
421
2000.
Page 20 of 26
Choussein 21 422
16.
Liebert PS. Surgeons don't have to be ambidextrous. The New York Times:
423
http://www.nytimes.com/1987/02/20/opinion/l-surgeons-don-t-have-to-
424
be-ambidextrous-128587.html (date last accessed: 12/14/15). 1987.
425
17.
Wilder JR. The tyranny of one-handed doctors. The New York Times. :
426
http://www.nytimes.com/1987/02/06/opinion/the-tyranny-of-one-
427
handed-doctors.html?_r=0 (date last accessed: 12/14/15), 1987.
428
18.
surgery. Surg Endosc. 2004;18:790-795.
429 430
19.
431 432
Yang JY, Son YG, Kim TH, et al. Manual Ambidexterity Predicts Robotic Surgical Proficiency. J Laparoendosc Adv Surg Tech A. 2015;25:1009-1018.
20.
Prasad SM, Prasad SM, Maniar HS, et al. Surgical robotics: impact of motion scaling on task performance. J Am Coll Surg. 2004;199:863-868.
433 434
Moorthy K, Munz Y, Dosis A, et al. Dexterity enhancement with robotic
21.
Bocci T, Moretto C, Tognazzi S, et al. How does a surgeon's brain buzz? An
435
EEG coherence study on the interaction between humans and robot. Behav
436
Brain Funct. 2013;9:14.
437
22.
Maniar HS, Council ML, Prasad SM, et al. Comparison of skill training with
438
robotic systems and traditional endoscopy: implications on training and
439
adoption. J Surg Res. 2005;125:23-29.
440
23.
Damiano RJ, Jr., Reichenspurner H, Ducko CT. Robotically assisted endoscopic
441
coronary artery bypass grafting: current state of the art. Adv Card Surg.
442
2000;12:37-57.
443 444
24.
Jenison EL, Gil KM, Lendvay TS, et al. Robotic surgical skills: acquisition, maintenance, and degradation. JSLS. 2012;16:218-228.
Page 21 of 26
Choussein 22 445
25.
Garcia-Ruiz A, Gagner M, Miller JH, et al. Manual vs robotically assisted
446
laparoscopic surgery in the performance of basic manipulation and suturing
447
tasks. Arch Surg. 1998;133:957-961.
448 449 450 451 452
Figure 1. Task 1: Peg transfer (PT) in the (a) robotic and (b) laparoscopic setting.
453
Figure 2. Task 2: Precision cutting (PC) in the (a) robotic and (b) laparoscopic setting.
454
Figure 3. Task 3: Intracorporeal suturing (IS) in the (a) robotic and (b) laparoscopic
455
setting.
456 457
Page 22 of 26
Choussein 23 458
Table 1. Penalty Scoring system TASK Peg transfer
Precision cutting Simple suture with intracorporeal knot
PENALTY Double of average peg transfer time (based on successful transfers only), per each peg dropped 5 sec for every mm of inner or outer circle transected No penalty was applied; task was considered completed when the suture was precisely placed through the two marks and all knots were tied
459 460 461 462
Page 23 of 26
Choussein 24 463
Table 2. Characteristics of study subjects Total Surgical experience, N (%) Gender Female Male Hand dominance Right Left Experience with computer games Yes No Time since last time played <1 yr 1-5 yr >5 yr Among those who play, Mean (SD); hours/week Prior exposure to laparoscopic/robotic technical skills (trainer box/simulatorbased) lab training Yes No Time since last exposure <1 yr >1 yr Cases/week Mean (SD) Weeks since last case Mean (SD)
Novices
Intermediates
Advanced
35
11 (31.4)
12 (34.3)
12 (34.3)
13 (37.1) 22 (62.9)
3 (27.3) 8 (72.6)
8 (66.7) 4 (33.3)
2 (16.7) 10 (83.3)
34 (98.2) 1 (2.9)
11 (100.0) 0 (0.0)
11 (91.7) 1 (8.3)
12 (100.0) 0 (0.0)
15 (42.9) 19 (54.3)
7 (63.6) 4 (36.4)
3 (25.0) 8 (66.8)
5 (41.7) 7 (58.3)
7 (46.7) 2 (13.3) 6 (40.0)
3 (27.3) 2 (18.2) 2 (18.2)
2 (16.7) 0 (0.0) 1 (8.33)
2 (16.7) 0 (0.0) 3 (25.0)
5.3 (9.0)
8.43 (12.15)
3.83 (3.62)
1.00 (0.71)
17(48.6) 17 (48.6)
-11 (100.0)
10 (83.3) 1 (8.3)
7 (58.3) 5 (41.7)
15 (88.2) 2 (11.8)
----
9 (75.0) 1 (8.3)
6 (50.0) 1 (8.3)
2.70 (3.07)
3.73 (2.55)
29.7 (38.5)
0.46 (0.72)
3.3 (2.8) -13.8 (29.3)
464
Page 24 of 26
Choussein 25 465
466 467 468 469 470 471
Table 3. Total performance time using laparoscopic and robotic approaches Non-Dominant Dominant hand, p-value hand, Mean (SD) Mean (SD) Total Lap time, sec
439.4 (244.4)
568.37 (378.2)
0.0008
Total Robotic time, sec
374.43 (229.6)
399.69 (199.3)
0.48
Table 4. Performance time for individual tasks using laparoscopic and robotic approaches Peg transfer
Lap time, sec Robotic time, sec
Dominant hand, Mean (SD)
Non-Dominant hand, Mean (SD)
p-value
46.4 (22.0)
60.2 (33.4)
0.003
48.57 (17.4)
49.03 (22.2)
0.91
Precision cutting
Lap time, sec Robotic time, sec
Dominant hand, Mean (SD)
Non-Dominant hand, Mean (SD)
p-value
190.7 (95.9)
224.43 (135.8)
0.02
145.0 (90.3)
141.57 (57.4)
0.81
Simple suture with intracorporeal knot
Lap time, sec Robotic time, sec
Dominant hand, Mean (SD)
Non-Dominant hand, Mean (SD)
p-value
202.26 (156.7)
283.74 (253.1)
0.01
180.86 (159.2)
209.09 (143.3)
0.33
472 473
Page 25 of 26
Choussein 26 474 475
Table 5. Total performance time using laparoscopic and robotic approaches stratified by level of experience
Group 1 Total Lap time, sec Total Robotic time, sec Group 2 Total Lap time, sec Total Robotic time, sec Group 3 Total Lap time, sec Total Robotic time, sec
Dominant hand, Mean (SD)
NonDominant hand, Mean (SD)
666.0 (196.5)
988.6 (326.1)
609.1 (267.3)
642.6 (164.1)
351.5 (228.7)
437.8 (265.5)
281.9 (110.3)
279.3 (79.8)
319.6 (147.7)
313.7 (93.9)
0.82
251.8 (70.5)
297.3 (75.6)
0.004
p-value
0.01 0.51
0.02 0.92
476 477
Page 26 of 26
S1553-4650(17)30397-7 http://dx.doi.org/doi: 10.1016/j.jmig.2017.07.010 JMIG 3198
To appear in:
The Journal of Minimally Invasive Gynecology
Received date: Revised date: Accepted date:
17-12-2016 28-6-2017 6-7-2017
Please cite this article as: Souzana Choussein, Serene S. Srouji, Leslie V. Farland, Ashley Wietsma, Stacey A. Missmer, Michael Hollis, Richard N. Yu, Charles N. Pozner, Antonio R. Gargiulo, Robotic Assistance Confers Ambidexterity to Laparoscopic Surgeons., The Journal of Minimally Invasive Gynecology (2017), http://dx.doi.org/doi: 10.1016/j.jmig.2017.07.010. This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
Choussein 1
Robotic assistance confers ambidexterity to laparoscopic surgeons.
2
Souzana Choussein, M.D.1, Serene S. Srouji, M.D.1, Leslie V. Farland ScD2, Ashley
3
Wietsma, M.D.3, Stacey A. Missmer ScD1,4, Michael Hollis , B.S.3, Richard N. Yu,
4
M.D., PhD.3, Charles N. Pozner, M.D.5, Antonio R. Gargiulo, M.D.1
1
5 6
1
7
and Reproductive Biology, Brigham and Women’s Hospital, Harvard Medical School,
8
Boston, Massachusetts
9
2
Center for Infertility and Reproductive Surgery, Department of Obstetrics, Gynecology
Department of Epidemiology, Harvard T.H. Chan School of Public Health, Boston,
10
Massachusetts
11
3
12
Massachusetts
13
4
14
Massachusetts; Channing Division of Network Medicine, Brigham and Women's Hospital
15
and Harvard Medical School, Boston, Massachusetts
16
5
17
School, Boston, Massachusetts; Department of Emergency Medicine, Brigham and
18
Women's Hospital, Harvard Medical School, Boston, Massachusetts.
Department of Urology, Boston Children's Hospital, Harvard Medical School, Boston,
Department of Epidemiology, Harvard T.H. Chan School of Public Health, Boston,
Neil and Elise Wallace STRATUS Center for Medical Simulation, Harvard Medical
19 20
Corresponding author: Antonio R. Gargiulo, MD
21
Center for Infertility and Reproductive Surgery, Brigham and Women’s Hospital,
22
Harvard Medical School, 75 Francis St., ASB1-3, Boston, MA, 02115;
23
Tel: +1-617-732-4285; Fax: +1-617-975-0870; email: [email protected]
Page 1 of 26
Choussein
2
24
Disclosures: Dr. Gargiulo serves as a Consultant for Medicaroid, Inc. and OmniGuide
25
Holdings, Inc. Drs. Choussein, Srouji, Wietsma, Yu, Farland, Missmer, Hollis and Pozner
26
have no conflicts of interest or financial ties to disclose.
27
Funding: Nothing to report.
28 29
Precis: Robot-assisted laparoscopy eliminates the operative handedness observed in
30
conventional laparoscopy, conferring virtual ambidexterity to surgeons in training.
31 32
ABSTRACT
33
Study objective: To examine whether a robotic surgical platform can complement the
34
fine motor skills of the non-dominant hand, compensating for the innate difference in
35
dexterity between the surgeon’s hands, thereby conferring virtual ambidexterity.
36
Design: Crossover intervention study.
37
Design classification: Level II-1 evidence.
38
Setting: Centers for medical simulation of two tertiary care hospitals of Harvard Medical
39
School.
40
Participants: Three different cohorts of subjects were included: 1) surgical novices
41
(medical graduates with no robotic/laparoscopic experience) 2) surgeons in training
42
(PGY3-4 residents and fellows with intermediate robotic and laparoscopic experience)
43
and 3) advanced surgeons (attending surgeons with extensive robotic and laparoscopic
44
experience).
45
Interventions: Each study group completed three dry lab exercises based on exercises
46
included in the Fundamentals of Laparoscopic Surgery (FLS) curriculum. Each exercise
47
was completed four times: using their dominant and non-dominant hand, on a standard
Page 2 of 26
Choussein
3
48
laparoscopic FLS box trainer and in a robotic dry lab set-up. Participants were
49
randomized to the handedness and setting order in which they tackled the tasks.
50
Measurements and Main Results: Performance was primarily measured as time to
51
completion, with adjustments based on errors. Means of performance for the dominant
52
versus non-dominant hand for each task were calculated and compared using a repeated
53
measures ANOVA test.
54
A total of 36 subjects were enrolled study-wide (12 per group). In the laparoscopic
55
setting, the overall time to completion for all three tasks with the dominant hand was
56
significantly different from that with the non-dominant hand (439.4s vs. 568.4s
57
respectively; p=.0008). The between-hand performance difference was nullified with the
58
robotic system (374.4s vs. 399.7s;p=.48). Evaluation of performances for each individual
59
task also revealed a statistically significant disparate performance between hands for all
60
three tasks when the laparoscopic approach was utilized (p=.003, p=.02, p=.01
61
respectively). No between-hand difference was observed when the tasks were performed
62
robotically. When the above analysis was performed within the three surgical experience
63
cohorts, the performance advantage of robotic technology remained significant for the
64
surgical novices and intermediate level experience groups.
65
Conclusion: Robot-assisted laparoscopy may eliminate the operative handedness
66
observed in conventional laparoscopy, allowing for virtual ambidexterity. Such
67
ergonomic advantage is particularly evident on surgical trainees. Virtual ambidexterity
68
may represent an additional aspect of surgical robotics that facilitates mastery of
69
minimally-invasive skills.
Page 3 of 26
Choussein 70
Keywords: robotic surgery, laparoscopic surgery, intraoperative handedness,
71
ambidexterity, daVinci surgical system
4
72 73
Presented at the 45th AAGL Annual Global Congress On Minimally Invasive
74
Gynecology, Orlando, FL. November 14-18, 2016 (oral presentation, Plenary Robotics
75
Session).
76 77
Page 4 of 26
Choussein
5
78
INTRODUCTION
79
Ambidexterity has been identified as a critical factor in the achievement of laparoscopic
80
psychomotor surgical skill proficiency, and constitutes a desirable attribute for all
81
surgeons.[1-3] In terms of efficiency optimization, this is a particularly useful adjunct in
82
minimally invasive surgery, where the capabilities of instrument switching or body
83
position readjustment are limited. This often necessitates the use of the surgeon’s non-
84
dominant hand for delicate and precise tasks.[4] True ambidexterity, however is rare [5],
85
and possibly nonexistent.[6]
86
The robotic surgical platform has provided several ergonomic advancements to the field
87
of minimally invasive surgery, including the physiologic use of the surgeon’s wrist,
88
tremor filtration, motion scaling and three-dimensional vision. These advantages have
89
contributed to its increasing acceptance and growth as an alternative minimally invasive
90
surgical option. These same advantages have been proposed as the main factors that
91
shorten the learning curve of surgical skills: an important benefit, given the relatively
92
extensive learning period required to acquire these skills using laparoscopic platforms.
93
[7, 8] Access to virtual reality simulation, employing commercially available integrated
94
training protocols is another aspect of robotic surgery that has been shown to facilitate
95
negotiation of the surgical learning curve.[9] In summary, several aspects of robotic
96
surgery contribute to technical proficiency being achieved through a shorter and more
97
patient-centered learning curve compared to other types of invasive and minimally
98
invasive surgical techniques. Virtual ambidexterity in the robotic setting may constitute
99
an additional factor explaining the "enabling" role of computer-assisted laparoscopy and
100
its truncated learning curve.
Page 5 of 26
Choussein
6
101 102
There are two studies that previously examined the impact of the robotic platform on
103
surgeon’s handedness. However, none of them used experience-based stratification of the
104
study population, and neither introduced a laparoscopic component, as they both
105
examined open vs. robotic approaches. [10, 11]
106
With the majority of surgical operations now being performed via minimally invasive
107
routes, a study assessing the impact of robotic-assistance on hand dominance as
108
compared to laparoscopy appears more relevant than prior published literature that only
109
used an open surgery arm of comparison.
110
Given the aforementioned gap, the current study aims to examine whether robotic
111
systems, as compared to laparoscopy, can augment fine motor skills of the non-dominant
112
hand, compensating for surgeon’s innate difference in dexterity between each hand, and
113
thus conferring “virtual” ambidexterity.
114 115
MATERIALS AND METHODS
116
Subjects and Design
117
This was a crossover intervention study conducted at the Neil and Elise Wallace
118
STRATUS Center for Medical Simulation of Brigham and Women’s Hospital and at
119
Boston Children’s Hospital; both Harvard affiliated academic medical centers in Boston.
120
Brigham and Women’s Hospital and Boston Children’s Hospital Institutional Review
121
Board approvals were obtained. Volunteers were solicited from the Obstetrics &
122
Gynecology and Urology Departments of the aforementioned hospitals.
Page 6 of 26
Choussein
7
123
An email invitation was extended to the eligible staff members of the aforementioned
124
departments and their satellites, and participation was determined based on matched
125
availability (of the potential participant on the one end and the required facility on the
126
other). The study population was composed of three groups: Group 1: Surgical novices
127
with no robotic or laparoscopic experience i.e medical graduates with no residency
128
training (Surgery-naive subjects); Group 2: Surgeons in training (PGY 3-4 residents and
129
fellows with intermediate robotic and laparoscopic experience); and Group 3: Advanced
130
surgeons (attending surgeons with extensive robotic and laparoscopic experience).
131
Novice group was comprised of surgically untrained medical graduates serving as
132
Research Fellows at Obstetrics & Gynecology and Urology Departments of the
133
aforementioned hospitals. All of the residents and fellows included in this study (as
134
Group 2) had completed structured, stepwise curriculum requirements which included
135
standardized robotic simulation exercises, case observations, and stepwise involvement at
136
the console on live patient surgeries. At the time of their involvement on the study, all
137
had completed the basic requirements and had been fully involved in live console
138
surgery. Similarly, they had had several years of exposure to laparoscopic surgery, as part
139
of the standardized curricula for Accreditation Council for Graduate Medical Education
140
(ACGME)-accredited Obstetrics and Gynecology and Urology residency and fellowship
141
programs. Surgeons in Group 3 held robotic privileges at their institutions and reported
142
performing a minimum of 250 robotic cases and a minimum of 250 laparoscopic cases at
143
the time of enrollment. Individuals who self-reported as ambidextrous were excluded
144
from the study. All enrolled subjects watched a short video tutorial developed by an
145
expert surgeon. The video used a step-by-step process to demonstrate each task –
Page 7 of 26
Choussein
8
146
performed both laparoscopically and robotically-, and oriented participants to the
147
instruments used, ideal body positioning, and skills specific to the robot such as master
148
clutching. Each participant completed a short questionnaire at enrollment that collected
149
information about their demographics, current surgical case loads and prior experience
150
with surgical simulators and computer games.[12]
151
To assess differences in hand dexterity between the laparoscopic and robotic surgical
152
approach, each subject performed two sets (laparoscopic and robotic) of three exercises
153
inspired by skill tests required for the Fundamentals of Laparoscopic Surgery (FLS)
154
certification. Only those FLS skill tests that can be performed both laparoscopically and
155
with the robot were utilized. These are: the peg transfer (PT), the precision cutting (PC),
156
and the intracorporeal suturing (IS). Such skill tests were modified to serve the unilateral
157
performance requirement of the study. Other FLS skill tests not included in our study
158
were the endoloop placement and the extracorporeal suturing/knot tying.[8] Construct
159
validity of the FLS model has been demonstrated for robotic surgery. [5, 13] Exercises
160
were completed on a standard FLS box trainer (Limbs and Things, Savannah, GA)
161
located in a medical simulation center, for the laparoscopic arm. For the robotic arm,
162
exercises were completed in a robotic dry lab set-up. This involved the use of all the
163
components of the robot (surgical console, patient-side cart, EndoWrist instruments, and
164
vision cart) and the lower tray of a standard FLS laparoscopic box trainer, containing the
165
FLS test disposables.
166
We utilized a modified version of the PT task, eliminating the bi-manual subcomponents
167
of the original FLS curriculum, so that one hand at a time could be evaluated.
168
Specifically, each subject was asked to grasp six rubber objects placed on a pegboard,
Page 8 of 26
Choussein
9
169
one at a time, with a Maryland grasper and place them on a peg on the opposite side of
170
the board. (Figure 1.) The PC task required participants to cut out a circle from a 4x4
171
gauze piece suspended between clips, by cutting between two premarked circular
172
tracings. Participants were instructed to use a Maryland grasper in one hand to provide
173
traction and exposure on the gauze and to place it at the best possible angle to the cutting
174
hand being evaluated. No instrument exchange between hands was allowed. (Figure 2.)
175
Being largely bimanual, the IS task was also modified to serve our study aim. Participants
176
were required to drive a 2.0 Silk suture of 15cm length through two marks on a stationary
177
Penrose drain, starting the task with the designated hand being examined. Both hands
178
were occupied with curved needle-drivers. Once the needle was passed, the participants
179
were required to tie two double throws and one single throw by tying with the hand under
180
examination and the other hand holding the suture with the needle. (Figure 3.) These
181
modifications of the original FLS skill tests allowed for each task to be performed using
182
the dominant or the non-dominant hand, one at a time. Each participant completed each
183
task both laparoscopically and robotically on the same day. Robotic procedures were
184
performed using the da Vinci Surgical System (Intuitive Surgical Inc, Sunnyvale, CA).
185
Participants were randomized to which hand (dominant vs. nondominant) and setting
186
(laparoscopic vs. robotic) was completed first; there was no alternation in setting and all
187
three tasks were performed in-series prior to moving to the remaining setting.
188
Performance Assessment
189
To assess the differential in hand performance afforded by the laparoscopic versus the
190
robotic approach, the raw time for task completion was recorded. A time penalty was
191
applied for errors or lack of precision according to Table 1.
Page 9 of 26
Choussein 10 192
Statistical Analysis
193
The sample size of our study with paired subjects was sufficient to detect a 10% mean
194
difference (in seconds) between the subject’s dominant and non-dominant hand. This
195
threshold was chosen a priori to reflect the minimum difference between hands and
196
between laparoscopic and robotic modality that would be clinically meaningful with
197
respect to surgery (here quantified by the skills tests) duration. Given this minimum
198
detectable clinically meaningful difference, we calculated the sample size requiring a
199
minimum of 80% power and an α-level=0.05, using a repeated measures ANOVA
200
statistical approach.
201
Given the crossover study design [14] and the randomization of handedness and setting
202
order in which the participants tackled the tasks, there is little concern about any
203
confounding that remains between groups.[15] Mean values of performance using
204
dominant versus non-dominant hand in the laparoscopic and the robotic setting were
205
compared using a repeated measures ANOVA to take into account correlation between
206
measurements. Analyses were stratified by level of training. Sensitivity analyses were
207
conducted using generalized estimating equations (GEE) with robust “sandwich”
208
standard error estimates for linear regression. Findings from this sensitivity analysis were
209
not different from the original analysis and thus the repeated measures ANOVA results
210
are presented in the manuscript and all tables.
211 212
RESULTS
213
A total of 36 subjects were enrolled in the study with 12 in each group, each with four
214
timed measurements. Among novices, one was unable to complete the laparoscopic IS
Page 10 of 26
Choussein 11 215
task and thus, that individual’s results were excluded from the entire study. Gender and
216
handedness distribution among groups as well as prior experience with computer games
217
and simulators are shown in Table 2. The majority of our sample was male (62.5%) and
218
right-handed (98.2%). Approximately half of our subjects had previous experience with
219
laparoscopic or robotic simulators, not including our novice group which was simulator-
220
naive. Forty three percent reported prior experience with computer games, however
221
experience with computer games was most commonly reported among the novice group.
222
Surgical case load (either laparoscopic or robotic) and time since last surgical case
223
(applicable only to intermediate and advanced experience groups) are also noted. Among
224
the intermediate group, the average number of cases/week was 2.7 and the reported time
225
since last case ranged between 0 and 96 weeks, with a median of 12.5 weeks. Among the
226
expert group, average number of cases/week was 3.7; all of them reported performing a
227
surgery within the preceding two weeks.
228
Table 3 displays an assessment of the overall completion time of all three of the technical
229
tasks combined using laparoscopic and robotic approaches. In the laparoscopic setting,
230
the cumulative execution time with the dominant hand was significantly shorter than with
231
the non-dominant hand (439.4s vs. 568.4s respectively; p=.0008). This difference was
232
nullified when the technical tasks were performed on the robot (dominant: 374.4s vs.
233
non-dominant: 399.7s; p=.48). Of note, the median overall time to completion with either
234
the dominant or non-dominant hand was almost identical when the subjects used the
235
robot (dominant: 322s vs. non-dominant: 324s). Cumulatively, the mean difference in
236
raw time scores between the two hands for all three tasks was also significantly shorter
Page 11 of 26
Choussein 12 237
when the robot was utilized as compared to laparoscopic performance respectively (76.9s
238
vs. 167.8s; p<.0001).
239
Evaluation of each task individually also revealed a statistically significant difference in
240
performance between hands for all PT, PC, IS when the laparoscopic approach was
241
utilized (p=.003, p=.02, p=.01 respectively). No statistically significant between-hand
242
difference was observed when the tasks were performed robotically. (Table 4)
243
When stratifying by experience, the minimization of handedness afforded by the robotic
244
platform remained statistically significant among novice and intermediate operators.
245
In particular, we observed a statistically significant between-hand difference in the
246
laparoscopic setting performance of both novices and intermediate-level operators
247
(dominant: 666.0s vs. non-dominant: 988.6s; p=.01 and dominant: 351.5.0s vs. non-
248
dominant: 437.8s; p=.02, respectively) but not of the expert group (dominant: 319.6s vs.
249
non-dominant: 313.7; p=.82) No difference in relative hand performance was observed
250
for the novice and intermediate subjects when the tasks were performed robotically;
251
however, a difference was noted for the expert group. (dominant: 251.8s vs. non-
252
dominant: 297.3; p=.004). (Table 5)
253 254
DISCUSSION
255
For surgeons, it has been argued that ambidexterity represents a professional advantage,
256
if not an implicit requirement. [16, 17] Robotic surgical platforms appear to complement
257
motor skills of the non-dominant upper limb, thereby conferring virtual ambidexterity.
258
This is another piece of the puzzle to explain, in a scientific way, the "enabling" nature of
259
robot-assisted surgery.[10, 11] Virtual ambidexterity may represent an additional aspect
Page 12 of 26
Choussein 13 260
of surgical robotics that facilitates mastery of minimally-invasive skills through a
261
truncated, patient-friendly learning curve.
262
Technological innovations, health care cost considerations and patients’ expectations of
263
modern surgery, have all contributed to make minimally invasive surgery, the gold
264
standard for most intra-abdominal operations. Yet, the skill set needed for advanced
265
laparoscopy is hard to master due to the fundamental ergonomic challenges associated
266
with working through a fulcrum (anterior abdominal wall), the lack of three-dimensional
267
vision, and a reduced range of motion as a result of the inability to use wristed action
268
movements. Due to these limitations, even the surgeon’s innate dexterity cannot be fully
269
exhibited, as he or she negotiates the learning curve.[18, 19] Being hard to learn,
270
laparoscopy is harder to teach through the classic “see one, do one, teach one” model.
271
Robotic surgery is poised to address these limitations. To take it a step further,
272
eliminating intraoperative handedness can be expected to positively impact the surgical
273
learning curve, particularly benefitting surgical trainees and surgeons at the early stages
274
of their robotic learning curve, as shown by our data.
275
Studies demonstrating enhanced dexterity with the use of a robotic surgical system over
276
traditional laparoscopy have frequented the literature.[6, 18, 20] In an attempt to quantify
277
the purported advantage of robotic approaches vs. laparoscopy, Moorthy et al. reported a
278
nearly 50% increase in dexterity when employing the robot; attributing this to the tremor
279
cancellation, motion scaling and seven degrees of freedom (i.e specific, defined modes in
280
which a mechanical system can move) afforded by the robot.[18] Also, there is published
281
evidence that bimanual object manipulation is mediated by callosal fibers that
282
interconnect the cerebral hemispheres and that there is a preferential activation of this
Page 13 of 26
Choussein 14 283
human “mirror” system in surgeons operating with the robotic system compared with
284
those performing conventional laparoscopy.[21]
285
As previously mentioned, the impact of a robotic surgical system on a surgeon’s
286
handedness has been examined.[10, 11] Mucksavage et al. conducted manual dexterity
287
assessments of 19 robotic novices using open and robotic approaches to perform the
288
Purdue Pegboard test – in which participants must place as many peg assemblies into the
289
board as possible within a 30-second interval- and a needle targeting test where
290
participants were asked to hit as many bull’s eye paper targets as possible within 10
291
seconds. The performance scores for each hand were statistically different when the open
292
approach was used; however this difference was nullified when the tasks were performed
293
robotically.[11] In a similarly designed study, Badalato et al. concluded that robotic
294
surgery is capable of eliminating handedness among relative novices who performed
295
basic manual dexterity skill tests modified from FLS using the open and the robotic
296
technique.[10]
297
In the era of minimally invasive surgery, comparing the facilitation of ambidexterity
298
between laparoscopic skills and robotic skills is more germane than any comparison with
299
an open approach. Our study is the first to specifically address this question.
300
This study revealed a statistically disparate performance between hands for all three tasks
301
when the laparoscopic approach was utilized. The use of the robot resulted in leveling of
302
performance between the dominant and the non-dominant hand. Equilibration of hand
303
performance has been attributed to enhanced performance of the non-dominant hand as a
304
result of either tremor filtration or motion scaling.[22] Published evidence has established
305
that motion scaling rather than tremor filtration, is primarily responsible for the enhanced
Page 14 of 26
Choussein 15 306
accuracy observed with robotic assistance with more prominent gains in the non-
307
dominant, innately less dexterous, hand.[20, 23]
308
We felt that the multidimensional skill set and versatile technical attributes required
309
during an actual surgical operation are better reflected when combining the results of all
310
three tasks; that was the rationale for calculating and analyzing cumulative execution
311
time. Even in this context, the significant between-hand difference in overall time to
312
completion that was noted when the tasks were performed laparoscopically, was negated
313
when switching to the robotic setting.
314
These results are consistent with the findings of the aforementioned pertinent studies.
315
None of the pre-existing studies, however, has used an experience-based stratification of
316
the study population. At a time where a philosophy of patient-centered medicine is
317
shaping the way we can ethically teach surgery, we strongly believed that this type of
318
study should stratify subjects by surgical experience. Assuming that pure novices and
319
robotic experts occupy the two edges of the robotic surgery experience spectrum, there
320
should be an intermediate one in between them to span it entirely. This is, at its best,
321
represented by senior residents and fellows, and that was the rational for selecting the
322
intermediate study groups as such.
323
In the laparoscopic setting, we found a statistically significant between-hand difference
324
in the performance of both novices and intermediate-level operators. This difference was
325
not identified in the expert group. In the robotic setting, no difference in relative hand
326
performance was observed in the novice and intermediate subjects. However, a difference
327
was found in the expert group.
Page 15 of 26
Choussein 16 328
The inverted pattern observed in the novice-intermediate groups compared to the expert
329
group substantiates previously published work indicating that the performance advantage
330
of robotic technology is more profound for novices than for experts. [6] Re-emergence of
331
handedness among experts in the robotic setting may be reflective of benefit
332
maximization emanating from optimal use of the robotic platform. Once out of the
333
robotic learning curve, both hands still perform better than at conventional laparoscopy,
334
but the dominant hand takes the lead again. Indeed, the dominant hand robotic time of the
335
expert group was the shortest time reported in our study. One could also argue that the
336
intracorporeal suturing task poses a significantly increased level of difficulty to the
337
novice and intermediate-experience groups compared to the expert one, for which may
338
not necessarily comprise an adequately challenging module. It is possible that, had more
339
complex tasks been employed, even experts would have experience a persistent enabling
340
effect of virtual ambidexterity. However, for the purpose of this study, we thought that it
341
was important to limit the comparison to exercises based on a previously validated
342
curriculum, that could be expected to be completed by all three levels of operators.
343
Our study was strengthened by randomizing the order of performance by hand (right or
344
left) and setting (laparoscopic or robotic). This counterbalanced approach controlled for
345
any learning or warming up effects emanating from the completion of the first trial of
346
each session,[15] and likely contributed to the avoidance of a type II error. This was
347
especially important as we found no between-hand difference in the laparoscopic
348
performance of experts. The crossover design of the study,[14] enabled each participant
349
to serve as his or her own control, eliminating the confounding factors of gender, number
350
of cases performed, time since last case, and experience with computer games.[12, 24] Of
Page 16 of 26
Choussein 17 351
note, even within the context of a subgroup analysis of participants with computer game
352
experience, performance patterns remained the same.
353
There are limitations to our study. Although procedure time is considered an acceptable
354
performance metric, it is likely only a proxy indicator of performance, since it does not
355
account for skillful and proper performance. To mitigate this effect, a time penalty was
356
employed where applicable to encourage precise execution. There is published literature
357
where manual dexterity -defined as manual speed and ambidexterity- is measured by
358
utilizing the sums and differences of the right and left hand performance times (only).
359
Based on that, we extrapolated performance time to manual dexterity.[19] Given that the
360
robot-assisted technology is aimed at improving surgical precision and dexterity rather
361
than reducing task time, however, we may have underestimated or failed to reflect the
362
extent of the robotic enabling effect. It is likely that the benefits of robotic assistance will
363
be more profound when the path length and smoothness of task execution are evaluated
364
as well.[25] Also, although the utilized skill drills have been validated to be performed
365
with both hands, they all had be modified to serve the purpose of our study. However, as
366
our study was designed to compare between-hand performance and not laparoscopic and
367
robotic systems nor overall surgical competence per se, implications of these limitations
368
on our conclusions must be insubstantial.
369 370
Conclusion
371
Our study demonstrates that robotic-assistance confers ambidexterity to surgeons in
372
training, compared to conventional laparoscopy. Further investigation is warranted to
Page 17 of 26
Choussein 18 373
define the ultimate clinical value of this particular robotic attribute with respect to actual
374
operative outcomes and patient safety in the surgical learning curve.
375 376
Acknowledgments: The authors would like to thank the staff of the Neil and Elise
377
Wallace STRATUS Center for Medical Simulation at Brigham and Women’s Hospital
378
for their cooperation and overall assistance with study logistics.
379
Page 18 of 26
Choussein 19 380
References
381
1.
382 383
training and evaluation of laparoscopic skills. Am J Surg. 1998;175:482-487. 2.
384 385
3.
Skinner A, Auner G, Meadors M, et al. Ambidexterity in laparoscopic surgical skills training. Stud Health Technol Inform. 2013;184:412-416.
4.
Nutan J. State of the Art Atlas and Textbook of Laparoscopic Suturing. Jaypee Brothers Publishers, 2006.
388 389
Rosser JC, Rosser LE, Savalgi RS. Skill acquisition and assessment for laparoscopic surgery. Arch Surg. 1997;132:200-204.
386 387
Derossis AM, Fried GM, Abrahamowicz M, et al. Development of a model for
5.
Elneel FH, Carter F, Tang B, et al. Extent of innate dexterity and ambidexterity
390
across handedness and gender: Implications for training in laparoscopic
391
surgery. Surg Endosc. 2008;22:31-37.
392
6.
Chandra V, Nehra D, Parent R, et al. A comparison of laparoscopic and robotic
393
assisted suturing performance by experts and novices. Surgery.
394
2010;147:830-839.
395
7.
Mehta S, Dasgupta P, Challacombe BJ. Robotic reconstructive urology:
396
possibilities for the urological surgeon beyond the prostate. BJU Int.
397
2010;106:1247-1248.
398
8.
Moore LJ, Wilson MR, Waine E, et al. Robotic technology results in faster and
399
more robust surgical skill acquisition than traditional laparoscopy. J Robot
400
Surg. 2015;9:67-73.
Page 19 of 26
Choussein 20 401
9.
Culligan P, Gurshumov E, Lewis C, et al. Predictive validity of a training
402
protocol using a robotic surgery simulator. Female Pelvic Med Reconstr Surg.
403
2014;20:48-51.
404
10.
Badalato GM, Shapiro E, Rothberg MB, et al. The da vinci robot system
405
eliminates multispecialty surgical trainees' hand dominance in open and
406
robotic surgical settings. JSLS. 2014;18.
407
11.
innate hand dominance. J Endourol. 2011;25:1385-1388.
408 409
Mucksavage P, Kerbl DC, Lee JY. The da Vinci((R)) Surgical System overcomes
12.
Grantcharov TP, Bardram L, Funch-Jensen P, et al. Impact of hand dominance,
410
gender, and experience with computer games on performance in virtual
411
reality laparoscopy. Surg Endosc. 2003;17:1082-1085.
412
13.
Stefanidis D, Hope WW, Scott DJ. Robotic suturing on the FLS model
413
possesses construct validity, is less physically demanding, and is favored by
414
more surgeons compared with laparoscopy. Surg Endosc. 2011;25:2141-
415
2146.
416
14.
Basic Study Design (Chapter 5). In: Friedman LM, Furberg CT, DeMets DL,
417
Eds. Fundamentals of Clinical Trials. 4th ed. New York: Springer; 2010: 79-
418
80).
419
15.
Repeated Measures Designs. In: Shaughnessy JJ, Zechmeister EB, Zechmeister
420
JS, eds. Research Methods in Psychology. 5th ed. New York: McGraw Hill;
421
2000.
Page 20 of 26
Choussein 21 422
16.
Liebert PS. Surgeons don't have to be ambidextrous. The New York Times:
423
http://www.nytimes.com/1987/02/20/opinion/l-surgeons-don-t-have-to-
424
be-ambidextrous-128587.html (date last accessed: 12/14/15). 1987.
425
17.
Wilder JR. The tyranny of one-handed doctors. The New York Times. :
426
http://www.nytimes.com/1987/02/06/opinion/the-tyranny-of-one-
427
handed-doctors.html?_r=0 (date last accessed: 12/14/15), 1987.
428
18.
surgery. Surg Endosc. 2004;18:790-795.
429 430
19.
431 432
Yang JY, Son YG, Kim TH, et al. Manual Ambidexterity Predicts Robotic Surgical Proficiency. J Laparoendosc Adv Surg Tech A. 2015;25:1009-1018.
20.
Prasad SM, Prasad SM, Maniar HS, et al. Surgical robotics: impact of motion scaling on task performance. J Am Coll Surg. 2004;199:863-868.
433 434
Moorthy K, Munz Y, Dosis A, et al. Dexterity enhancement with robotic
21.
Bocci T, Moretto C, Tognazzi S, et al. How does a surgeon's brain buzz? An
435
EEG coherence study on the interaction between humans and robot. Behav
436
Brain Funct. 2013;9:14.
437
22.
Maniar HS, Council ML, Prasad SM, et al. Comparison of skill training with
438
robotic systems and traditional endoscopy: implications on training and
439
adoption. J Surg Res. 2005;125:23-29.
440
23.
Damiano RJ, Jr., Reichenspurner H, Ducko CT. Robotically assisted endoscopic
441
coronary artery bypass grafting: current state of the art. Adv Card Surg.
442
2000;12:37-57.
443 444
24.
Jenison EL, Gil KM, Lendvay TS, et al. Robotic surgical skills: acquisition, maintenance, and degradation. JSLS. 2012;16:218-228.
Page 21 of 26
Choussein 22 445
25.
Garcia-Ruiz A, Gagner M, Miller JH, et al. Manual vs robotically assisted
446
laparoscopic surgery in the performance of basic manipulation and suturing
447
tasks. Arch Surg. 1998;133:957-961.
448 449 450 451 452
Figure 1. Task 1: Peg transfer (PT) in the (a) robotic and (b) laparoscopic setting.
453
Figure 2. Task 2: Precision cutting (PC) in the (a) robotic and (b) laparoscopic setting.
454
Figure 3. Task 3: Intracorporeal suturing (IS) in the (a) robotic and (b) laparoscopic
455
setting.
456 457
Page 22 of 26
Choussein 23 458
Table 1. Penalty Scoring system TASK Peg transfer
Precision cutting Simple suture with intracorporeal knot
PENALTY Double of average peg transfer time (based on successful transfers only), per each peg dropped 5 sec for every mm of inner or outer circle transected No penalty was applied; task was considered completed when the suture was precisely placed through the two marks and all knots were tied
459 460 461 462
Page 23 of 26
Choussein 24 463
Table 2. Characteristics of study subjects Total Surgical experience, N (%) Gender Female Male Hand dominance Right Left Experience with computer games Yes No Time since last time played <1 yr 1-5 yr >5 yr Among those who play, Mean (SD); hours/week Prior exposure to laparoscopic/robotic technical skills (trainer box/simulatorbased) lab training Yes No Time since last exposure <1 yr >1 yr Cases/week Mean (SD) Weeks since last case Mean (SD)
Novices
Intermediates
Advanced
35
11 (31.4)
12 (34.3)
12 (34.3)
13 (37.1) 22 (62.9)
3 (27.3) 8 (72.6)
8 (66.7) 4 (33.3)
2 (16.7) 10 (83.3)
34 (98.2) 1 (2.9)
11 (100.0) 0 (0.0)
11 (91.7) 1 (8.3)
12 (100.0) 0 (0.0)
15 (42.9) 19 (54.3)
7 (63.6) 4 (36.4)
3 (25.0) 8 (66.8)
5 (41.7) 7 (58.3)
7 (46.7) 2 (13.3) 6 (40.0)
3 (27.3) 2 (18.2) 2 (18.2)
2 (16.7) 0 (0.0) 1 (8.33)
2 (16.7) 0 (0.0) 3 (25.0)
5.3 (9.0)
8.43 (12.15)
3.83 (3.62)
1.00 (0.71)
17(48.6) 17 (48.6)
-11 (100.0)
10 (83.3) 1 (8.3)
7 (58.3) 5 (41.7)
15 (88.2) 2 (11.8)
----
9 (75.0) 1 (8.3)
6 (50.0) 1 (8.3)
2.70 (3.07)
3.73 (2.55)
29.7 (38.5)
0.46 (0.72)
3.3 (2.8) -13.8 (29.3)
464
Page 24 of 26
Choussein 25 465
466 467 468 469 470 471
Table 3. Total performance time using laparoscopic and robotic approaches Non-Dominant Dominant hand, p-value hand, Mean (SD) Mean (SD) Total Lap time, sec
439.4 (244.4)
568.37 (378.2)
0.0008
Total Robotic time, sec
374.43 (229.6)
399.69 (199.3)
0.48
Table 4. Performance time for individual tasks using laparoscopic and robotic approaches Peg transfer
Lap time, sec Robotic time, sec
Dominant hand, Mean (SD)
Non-Dominant hand, Mean (SD)
p-value
46.4 (22.0)
60.2 (33.4)
0.003
48.57 (17.4)
49.03 (22.2)
0.91
Precision cutting
Lap time, sec Robotic time, sec
Dominant hand, Mean (SD)
Non-Dominant hand, Mean (SD)
p-value
190.7 (95.9)
224.43 (135.8)
0.02
145.0 (90.3)
141.57 (57.4)
0.81
Simple suture with intracorporeal knot
Lap time, sec Robotic time, sec
Dominant hand, Mean (SD)
Non-Dominant hand, Mean (SD)
p-value
202.26 (156.7)
283.74 (253.1)
0.01
180.86 (159.2)
209.09 (143.3)
0.33
472 473
Page 25 of 26
Choussein 26 474 475
Table 5. Total performance time using laparoscopic and robotic approaches stratified by level of experience
Group 1 Total Lap time, sec Total Robotic time, sec Group 2 Total Lap time, sec Total Robotic time, sec Group 3 Total Lap time, sec Total Robotic time, sec
Dominant hand, Mean (SD)
NonDominant hand, Mean (SD)
666.0 (196.5)
988.6 (326.1)
609.1 (267.3)
642.6 (164.1)
351.5 (228.7)
437.8 (265.5)
281.9 (110.3)
279.3 (79.8)
319.6 (147.7)
313.7 (93.9)
0.82
251.8 (70.5)
297.3 (75.6)
0.004
p-value
0.01 0.51
0.02 0.92
476 477
Page 26 of 26