Applied Ergonomics 78 (2019) 251–262

Contents lists available at ScienceDirect

Applied Ergonomics
journal homepage: www.elsevier.com/locate/apergo

Are gestures worth a thousand words? Verbal and nonverbal communication T

during robot-assisted surgery
Judith Tiferesa,b,d,∗, Ahmed A. Husseinb,c, Ann Bisantzd, D. Jeffery Higginbothamd,
Mohamed Sharifd, Justen Kozlowskib, Basel Ahmadd, Ryan O'Harad, Nicole Wawrzyniakd,
Khurshid Gurub
a
Abbott Laboratories, USA
b
ATLAS Program, Department of Urology, Roswell Park Cancer Institute, USA
c
Department of Urology, Cairo University, Egypt
d
University at Buffalo, The State University of New York, USA

A R T I C L E I N F O A B S T R A C T

Keywords: Communication breakdowns in the operating room (OR) have been linked to errors during surgery. Robot-
Robot-assisted assisted surgery (RAS), a new surgical technology, can lead to new challenges in communication owing to the
Surgery remote location of the surgeon away from the patient and bedside assistants. Nevertheless, few studies have
Teams studied communication strategies during RAS. In this study, 11 robot-assisted radical prostatectomies were
Communication
recorded and the interaction events between the surgeon and two bedside surgical team members were cate-
Teamwork
gorized by modality (verbal/nonverbal), topic, and pair (sender and receiver). Both verbal and nonverbal
Patient safety
Nonverbal modalities were used by all pairs. The percentage of nonverbal interactions differed significantly by pair: 66%
for the Surgeon-Physician Assistant, 50% for the Physician Assistant-Scrub Nurse, and 25% for the Surgeon-
Scrub Nurse, indicating different communication strategies across pairs. In addition, there was a significant
dependence between topic and the percentages of verbal and nonverbal events for all pairs. Strategies to improve
team communication during RAS should take into account the use of verbal and nonverbal communication
means and the variation in interaction strategies based on the topic of communication.

1. Introduction Prior research regarding team communication within the OR has

1.1. Study of team communication in the operating room
The operating room (OR) is a high risk dynamic environment where
non-technical surgical skills such as communication are critical to
successful outcomes (Gillespie et al., 2010; Lingard et al., 2002; Roth
et al., 2004). Communication breakdowns have been consistently
linked to human error in surgery and healthcare (ElBardissi et al., 2007;
Lingard et al., 2004; Nagpal and Moorthy, 2010; Nagpal et al., 2010;
Sutcliffe et al., 2004; Wahr et al., 2013). Surgical teams present special
challenges to communication since they deal with (a) inconsistent le-
vels of team familiarity (i.e. the experience team members have
working together), (b) overlapping but different expertise and roles
among team members, (c) time constraints, and (d) hierarchical
structures (Morrow and Fischer, 2013; Morrow et al., 2005; Sutcliffe
et al., 2004).
primarily focused on verbal communication. Nevertheless, a few studies
have challenged the conception that effective communication can only
be achieved verbally. Nonverbal means can support or even replace
verbal exchanges especially for coordinating team actions (Segal,
1995). Theories of communication (e.g., common ground theory) state
that people shape their interactions with others based on assumptions
of their mutual knowledge and beliefs. In particular, multiple commu-
nication modes are important in creating common ground among team
members. Additionally, the process of updating and improving common
ground is affected by both the communication medium and the purpose
of the interaction (Clark and Brennan, 1991).
Nonverbal modes of team communication, including gestures, vi-
sual gaze direction, body positions and movements, facial expressions
and tool manipulations have been identified as critical to successful
communication in other complex work domains (Argyle, 1972;
Hutchins, 2006; Katz et al., 2006; Segal, 1995), but have received less

∗
Corresponding author. A.T.L.A.S (Applied Technology Laboratory for Advanced Surgery) Program, Department of Urology, Roswell Park Cancer Institute, Elm & Carlton St, Buffalo,
NY 14263, USA.
E-mail address: judithti@buffalo.edu (J. Tiferes).

https://doi.org/10.1016/j.apergo.2018.02.015
Received 1 June 2017; Received in revised form 21 November 2017; Accepted 16 February 2018
Available online 07 March 2018
0003-6870/ © 2018 Elsevier Ltd. All rights reserved.
J. Tiferes et al.
attention in healthcare, particularly surgery (cf. Kolbe et al., 2014;
Moore et al., 2010). In a recent systematic literature review that ana-
lyzed coding schemes for OR communication, none of the studies in-
cluded nonverbal interactions (Tiferes et al., 2015).
1.2. Robot-assisted surgery
During Robot-Assisted Surgery (RAS), the surgeon sits at a console
which is typically located away from the patient, without direct visual
access to the patient and bedside members of the surgical team (Diana
and Marescaux, 2015; Herron and Marohn, 2008). The surgeon ma-
nipulates controls at the console. These movements are reproduced by
the robotic arms holding instruments within the patient, thus elim-
inating surgical tremor and allowing precise microsurgery (Spight et al.,
2014). Robotic instrument changes are performed by other team
members (i.e. the physician assistant or the scrub nurse) situated at the
patient's side at the request of the surgeon. In addition, the physician
assistant aids the surgeon by controlling laparoscopic instruments. A
camera provides the surgeon with a view of the surgical field on a 3D
viewer that is part of the console; the view can be magnified up to 10×.
This video feed is reproduced on multiple (2D) screens throughout the
OR allowing others on the surgical team to see what the surgeon is
viewing. This shared view shows instrument movements (both the ro-
botic instruments manipulated by the surgeon and the laparoscopic
ones manipulated by the physician assistant), the camera view (with
zoom level, horizontal and vertical position controlled by the surgeon;
and changes in focus or camera insertion angle executed by the scrub
nurse), and console display indicators (names of active robotic instru-
ments and the camera insertion angle).
RAS provides advantages to surgeons in terms of improved visua-
lization, precision, access to deep anatomical areas and may also lower
stress and fatigue (Randell et al., 2016). However, it has also brought
new challenges due to the more remote location of the surgeon, a re-
duced ability to maintain vision in the operative field, and the com-
plexity of the robotic equipment (Catchpole et al., 2016; Randell et al.,
2016).
Despite the importance of team communication in surgery, there is a
lack of comprehensive analysis of team communication during RAS.
Some studies have evaluated how RAS differs from laparoscopic sur-
geries in which the surgeon is located next to the OR table and the bed
side assistants. Webster and Cao (2006) compared the steps to perform
instrument changes with both technologies. RAS instrument changes
seem to be more complex than in laparoscopic surgery; changes not
only require more steps, but also that the team to be aware of the ro-
bot's operation mode at every stage. Cao and Taylor (2004) and Nyssen
and Blavier (2010) compared the amount of team communication
during laparoscopic versus RAS and both studies found more commu-
nication during the RAS. However, it is not clear if that disparity was
due to the different technologies per se or the differences in experience
the team had with each technology, which was different in Nyssen and
Blavier (2010) and not reported in Cao and Taylor (2004). Nyssen and
Blavier (2010) also found that the frequency of communications for
some content categories (“orientation”, “manipulation”, “order”, and
“confirmations”) was significantly higher for RAS than for laparoscopic
surgeries. Again, however, we do not know whether these differences
were influenced by the different technological demands, the differences
in experience the team had with each technology, or both. Finally,
Nyssen and Blavier (2010) studied the communication between the
surgeon and one bed side assistant during RAS and suggested that
greater levels of individual experience (measured as number of RAS
performed) allowed for more implicit communication, however they do
not describe what this implicit communication entailed. In summary,
initial studies focused on differentiating the patterns of communication
between robot-assisted and laparoscopic surgery (Cao and Taylor,
2004; Nyssen and Blavier, 2010; Webster and Cao, 2006), or real-time
assessment of verbal communication during RAS (Cunningham et al.,
252
Applied Ergonomics 78 (2019) 251–262
2013; Nyssen and Blavier, 2010), but there are no studies that at-
tempted to analyze verbal and nonverbal communications in RAS.
1.3. Study aims
In this study, we characterized team verbal and nonverbal interac-
tions among the console surgeon, the physician assistant, and the scrub
nurse in order to increase the knowledge of how surgical teams com-
municate during RAS. We hypothesized that team communication
strategies (i.e., use of verbal vs. nonverbal means of communication)
will be associated with the communicating pair and the topic of the
interaction.
2. Materials and methods
2.1. Data collection
This research was conducted as part of an ongoing research in-
itiative at a major cancer research hospital, the “Techno-fields” project.
Techno-fields is intended to study and improve teamwork, commu-
nication, and other non-surgical skills in RAS, and supports audio- and
video-recording of RAS cases using three ceiling-mounted cameras
capturing views of personnel and their movements within the OR, re-
cordings of the console video showing the view of the surgical field
within the patient's body (to provide operative context) and up to eight
audio tracks recorded by lapel microphones (Ahmad et al., 2016; Allers
et al., 2016; Tiferes et al., 2016). As part of the ongoing project, re-
cordings were made for cases in which all members of the surgical team
and the patient had provided consent. Video and audio files were
synchronized via the movie editing software, Adobe Premiere Pro CS6,
resulting in four audiovisual streams per surgery. Noldus Observer XT
12 software was used to code the recordings. In addition to the re-
cordings, time at the console (measured from the moment the surgeon
sat at the console to start the procedure until he or she stood up at the
end of the procedure) was noted.
2.2. Case and participant characteristics
For this study, recordings from 11 robot-assisted radical prostatec-
tomies were analyzed. Surgeries were performed using Intuitive
Surgical's da Vinci Si. Cases were selected to maximize the variability of
team demographics concerning experience and inter-team familiarity
levels for individuals comprising the surgical teams' “main triad” - the
lead surgeon, the physician assistant (PA), and the primary scrub nurse
(SN). These roles were the focus of research because they have a high
degree of interaction during a case. Six surgeons (Ss), two PAs, and
seven SNs participated in the 11 cases that we analyzed. Participants
were surveyed regarding their years of experience in their role and the
number of cases for which they had worked with other members of the
team. Table 1 provides demographic information regarding participant
experience. Table 2 shows familiarity levels among all the pair com-
binations present in each case. Note that for any given case, multiple
individuals may have substituted for another in performing the team
role (e.g., a second scrub nurse may have relieved the first scrub nurse).
Table 1
Number of participants by years of experience in the role.
0-4 years 5-10 years > 10 years
Surgeons 2 3 1
Physician Assistants 1 0 1
Scrub Nurses 4 1 2
J. Tiferes et al. Applied Ergonomics 78 (2019) 251–262

Table 2
Familiarity by pair and case.

Case 1 Case 2 Case 3 Case 4 Case 5 Case 6 Case 7 Case 8 Case 9 Case 10 Case 11

S1&PA High High High High High High High Mid Mid High Mid
S1&SN1 High High High High High High High Low Mid High Low
S1&SN2 – High Low Low Mid Mid – Low Low Mid High
S1&SN3 – – – – – High – – Low – –
S2&PA Mid Mid Mid Mid Mid High High High High High High
S2&SN1 Low Mid Low Low Low Mid Low Mid High High Low
S2&SN2 – Low Low Low Low Low – Low Low Low Low
S2&SN3 – – – – – Low – – Low – –
S3&PA – Mid – – – Mid High Mid – – –
S3&SN1 – Low – – – High Mid Low – – –
S3&SN2 – Low – – – Low – Low – – –
S3&SN3 – – – – – Mid – – – – –
PA&SN1 High High Low High High High High High High High Mid
PA&SN2 – Mid High Low Mid High – High Low Mid High
PA&SN3 – – – – – High – – Mid – –

Low: < 24 cases worked together.

Medium: 25–50 cases worked together.
High: > 50 cases worked together.
Subscript numbers disambiguate between multiple surgeons or scrub nurses who participated during the same case.

Table 3 Table 4
Verbal and nonverbal definitions. Communication topic descriptions.

Verbal Event Categories
Request action: Requesting/suggesting someone to do something
Request acknowledgment/repetition: Asking for a repetition or a confirmation if a
previously given message was heard or a previously requested action was
completed
Request information (other than acknowledgment/repetition): Asking to for
new/extra information about something, other than "confirmation/repetition"
Give acknowledgment/repetition: Giving acknowledgment that a previously
received message was heard or that a previously requested action was completed
or Repeating/restating previously given information/request
Give information (other than acknowledgment/repetition): Giving new/extra
information about something, other than "confirmation/repetition"
Give teaching: Providing general information that would apply to other cases
Request teaching: Asking for general information that would apply to other cases
Nonverbal Event Categories
Instrument movement on screen: Interaction via instrument movements visible on
the shared screen (e.g., pointing with an instrument)
Camera view change: Interaction via changes or adjustments on camera position,
zoom, focus, or angle
Display indicators on screen: Interaction via changes of display indicators (e.g., the
name of the robotic instrument appears on the screen as it is being installed)
"Face-to-face" interaction: Interaction via body, head, or hand position and
movements & visual gaze
2.3. Communication analysis
Interactions between the console surgeon, PA, and SN were identi-
fied and coded by a Human Factors Ph.D. candidate (JT) and two
biomedical engineering undergraduate students (RO and NW). The
coding scheme was based on a prior study which synthesized codes
from previous research of surgical communications (Tiferes et al.,
2015). Since there were a great volume of interaction events categor-
ized as “performed action” under the “statement type” dimension sug-
gested by (Tiferes et al., 2015), this dimension—now renamed “mod-
ality”—was broken down into verbal and nonverbal events as described
below. Each interaction event was characterized in terms of the sender,
the recipient, modality (verbal or nonverbal), and topic. Verbal events
were coded as request action, acknowledgment/repetition, other in-
formation (apart from acknowledgment/repetition), teaching, give ac-
knowledgment/repetition, give information (other than acknowl-
edgment/repetition), or give teaching. Nonverbal events were
categorized by the communication medium through which they
1- Camera angle/position: Changing or adjusting the camera angle or position
2- Camera clean: Cleaning camera lens
3- Camera focus: Adjusting the camera focus or light intensity
4- Specimen bag: Manipulations of specimen's bag
5- Catheter: Manipulation of the urethral Foley catheter
6- Staple: Manipulation of staple instrument
7- Clip: Placing a surgical clip to control/seal a blood vessel
8- Cut: Interactions about cutting tissue without including cutting sutures
9- Dissecting needle: Manipulation of dissecting needle
10- Hold/retract: Assisting the surgeon by retracting tissue with a laparoscopic
instrument
11- Patient condition/information: Current case's patient general condition or
information
12- Remove: Removing tissue without including bag manipulations
13- Stitching/needle: Manipulation of the needle and assistance during stitching
14- Suction: Removal of excess fluid (saline, urine, blood) present in the surgeon's
operative field with the laparoscopic suction instrument
15- Instrument change: Changing robotic instrument or installing them for the first
time
16- Instrument preparation/organization: Getting instruments ready and organized
17- Wash: Cleaning the surgeon's working area by irrigating saline solution and
suctioning it once clean
18- Workflow/time management: Interactions about what's going to happen in the
surgery or how long is it going to take
occurred (i.e. tool movement on shared screen, camera view change,
display indicators, or ‘face-to-face’). These interactions primarily in-
clude events such as instrument movements seen on the shared screen,
camera adjustments such as zooming, focusing, or panning, and body
gestures (Tiferes et al., 2016). Note that nonverbal communication in-
teractions included actions whose intended purpose may not have been
explicitly to communicate but which served as a means of implicit
communication. For example, the primary purpose of handing an in-
strument is the physical exchange of the instrument; however, it also
serves implicitly to acknowledge the request for the instrument. Defi-
nitions of verbal and non-verbal codes are provided in Table 3. “Topic”
identified the theme of the interaction and included eighteen categories
(Table 4).
A secondary, more in-depth analysis was performed by grouping
single interaction events into sequences comprised of a series of se-
quential events that (1) shared the same pair and topic, and (2) were
related to the same task instance. For example, a clip sequence between
surgeon and PA would include all the interaction events that were
performed to apply that clip. Any additional clip applications would be
included in another sequence. Note that the specific order of actions

253
J. Tiferes et al. Applied Ergonomics 78 (2019) 251–262

Table 5 Table 6
Percentages of communication interactions by team familiarity and individual experience Frequency of verbal and nonverbal event categories by pair.
(n refers to the number of communication interactions for each pair, or individual).
Nonverbal event categories S & PA PA & SN S & SN
Team Familiarity S-PA PA-SN S-SN
(Number of cases worked (n = 9300) (n = 3639) (n = 934) "face-to-face" interaction 0% 100% 3%
together) camera view change 3% 0% 45%
display indicators on screen 2% 0% 47%
Low: < 24 3% 7% 43% tool movement on screen 95% 0% 4%
Medium: 25-50 42% 25% 15%
High: > 50 55% 68% 42% Verbal event categories S & PA PA & SN S & SN

Individual Experience S PA SN Give acknowledgment/repetition 20% 18% 22%

(Number of years of (n = 8721) (n = 12903) (n = 6032) Give information 31% 56% 37%
experience) (other than acknowledgment/repetition)
Give teaching 0% 1% 0%
Low: < 5 37% 51% 33% Request acknowledgment/repetition 4% 6% 4%
Medium: 5–10 41% 0% 36% Request action 37% 9% 27%
High: > 10 22% 49% 31% Request information 7% 10% 10%
(other than acknowledgment/repetition)
Request teaching 0% 0% 0%
and associated communicative interactions related to a topic varied
from instance to instance (because patient and surgical conditions
vary). Identifying sequences allowed the analysis of sets of interactions 2.4. Statistical analysis
related to a topic. Two measures were used to characterize sequences:
the percentage of verbal events (V%) and the verbal grounding criterion Data were summarized using descriptive statistics. Dependencies
level (Verbal-GCL) per sequence. The grounding criterion refers to the between the number of verbal and nonverbal interaction events and the
level of understanding that is enough for achieving communication independent variables pair and topic were evaluated with Chi-Square
goals. The use of verbal acknowledgments and repetitions may show analysis. Sequences were grouped into ranges to evaluate how the
evidence of a higher criterion level (Clark and Brennan, 1991). In this distribution of V% and Verbal-GCL was influenced by combination of
study, we defined Verbal-GCL as the number of acknowledgments and role and topics. A nonparametric Median Mood Test was conducted to
repetitions divided by the number of interaction events in a sequence. evaluate if there was an association between familiarity levels and
The necessary grounding criterion may vary according to the amount of median %V and median Verbal-GCL.
information known at the time (higher criterion the less is known), the IRR was determined by calculating the percent agreement. P-values
complexity of the information (higher criterion with increased com- less than or equal 0.05 were considered significant. All statistical ana-
plexity), and the importance of the information (higher criterion when lyses were performed using Minitab 16.
information directly related to the task goal)(Traum and Dillenbourg,
1996). A higher or lower value of Verbal-GCL is not necessarily good or 3. Results
bad, rather it reflects an adaptation in communication strategies as the
information needs and context changes. Nevertheless, we expect that 3.1. Data characteristics
dyads with lower familiarity to will have less “common ground” which
will be reflected in a higher %V and a higher Verbal-GCL. The 11 surgeries comprised 13,873 unique interaction events (41%
A randomly selected portion of each surgery (10% of overall re- verbal and 59% nonverbal) over 37.1 console hours, with an average of
cording time) was selected and coded by a second observer to calculate
inter-rater reliability (IRR) for each coding dimension.

Fig. 1. Percentage (and number) of verbal and nonverbal interaction events by pair of Fig. 2. Most frequent topics that accounted for 70% of more of interaction events for each
participants and across surgeries. χ2 (2, n = 13873) = 787.8, p < 0.001. pair.

254
J. Tiferes et al. Applied Ergonomics 78 (2019) 251–262

a)
Surgeon-PA CommunicaƟon Events by Topic
100% Verbal Verbal
90% 223 195 Verbal Verbal
80% Verbal 329 254 Verbal

Percentage of Events
587 348
70%
60%
50% Nonverbal Nonverbal
40% 1638 1134 Nonverbal Nonverbal
30% Nonverbal 744 546 Nonverbal
777 361
20%
10%
0%
SucƟon Clip Wash Hold SƟtching/Needle Catheter
Topics

b)
PA-SN CommunicaƟon Events by Topic
100%
Verbal
90%
155
80% Verbal
Verbal 115 Verbal
Percentage of Events

70% 125
776
60%
50%
Nonverbal
40%
695
30% Nonverbal
Nonverbal 164 Nonverbal
20% 115
561
10%
0%
Tool Prep. & Clip SƟtching/Needle Remove
OrganizaƟon
Topics

c) Surgeon-SN CommunicaƟon Events by Topic

100%
90%
80%
Percentage of Events

70% Verbal
Verbal Verbal
191
60% 132 122
50%
40%
30%
20% Nonverbal
Nonverbal Nonverbal
114
10% 46 48
0%
Tool Change Camera Clean Camera Angle/PosiƟon
Topics
Fig. 3. Percentage and number of verbal and nonverbal interaction events by Topic between: 3.a) Surgeons and PAs. χ2 (5, n = 7136) = 683.1, p < 0.001. 3.b) PAs and SNs. χ2 (3,
n = 2706) = 343.9, p < 0.001. 3.c) Surgeons and SNs. χ2 (2, n = 653) = 8.28, p = 0.016.

6.8 interaction events per minute of console time, 95% CI [6.07, 7.46]. 3.2. Communication differences by role and topic
The mean console time was 3.4 h per surgery. IRR with respect to
identifying events themselves was 82% (i.e. assessing that an event with Communication patterns were affected by the role of the partici-
consistent start and end times, with the same sender, recipient, topic, pants involved and the topic of their communications. The S-PA was the
and type, within a specified time tolerance, was identified by both most active pair and participated in 9300 interaction events, followed
raters). Event categorization agreement across each dimension was high by 3639 events for PA-SN, and 934 for S-SN. Table 5 shows the per-
(modality, 85%, topic, 82%, sender, 98%, and receiver, 96%), and centage of communication interactions by familiarity level by pair, and
higher than chance (chance outcomes would be 50%, 6%, 11%, and experience by role.
11% for modality, topic, sender, and receiver respectively). We also considered the degree to which the triad pair (i.e. S-PA, PA-
SN, or S-SN) and the topic affected the percentages of verbal vs. non-
verbal events. First, as seen in Fig. 1, both verbal and nonverbal

255
J. Tiferes et al. Applied Ergonomics 78 (2019) 251–262

Table 7 interaction events are shown (Fig. 2). We have only included the
Number of events per sequence. communication events associated with these (most frequent) topics in
the following analyses. The remaining topics were excluded from the
Pair Topic Number of Number of events per sequence
sequences Chi-square analysis because they represented infrequent events which
Min Q1 Median Q3 Max were less representative of the interactions (5% or less of S&PA events,
6% or less for PA&SN events, and 9% or less for S&SN events) and to
S & PA Catheter 331 1 1 1 2 79
better satisfy Chi-Square sample size assumptions.
Clip 189 1 5 6 9 46
Hold 573 1 1 1 2 21 The percentage of verbal and nonverbal events by pair/topic is
Stitching/ 50 2 11 14 19 43 shown in Fig. 3. Chi-square tests showed a significant dependence be-
Needle tween topic and the percentage of verbal and nonverbal events for all
Suction 1512 1 1 1 1 14 three pairs (Surgeons and PAs, χ2 (5, n = 7136) = 683.1, p < 0.001;
Wash 585 1 2 2 2 13
Total 3240 1 1 1 2 79
PAs and SNs, χ2 (3, n = 2706) = 343.9, p < 0.001; Surgeons and SNs,
PA & SN Clip 185 2 4 4 5 19 χ2 (2, n = 653) = 8.28, p = 0.016).
Remove 50 1 2.25 3.5 6 17
Stitching/ 51 2 4 6 9 25
Needle 3.3. Detailed communication results by sequence
Instrument 263 1 3 4 6 23
Preparation/ The interaction events associated with the most frequent topics by
Organization pair were grouped into individual sequences. Table 7 shows the ob-
Total 549 1 3 4 6 25
served number of sequences and number of interaction events per se-
S & SN Camera Angle/ 48 1 2 3 4 10
Position quence (minimum, first quartile, median, third quartile, and maximum)
Camera Clean 49 1 2 3 4 10 for each combination of pair and topic.
Instrument 81 1 2 2 4 21 Tables 8 and 9 show the observed counts of sequences by amount of
Change
verbal communication (%V) and Verbal-GCL per sequence for the most
Total 178 1 2 3 4 21
Grand Total 3967 1 1 1 3 79 frequent topics by pair. Range categories for %V and Verbal-GCL were
established to satisfy Chi-Square sample size assumptions while trying
to keep range sizes as similar as possible (All cells need to have ex-
communications were present for all pairs. A Chi-square test showed pected counts at least (a) two, or (b) one and 50% or fewer of the cells
that the percentage of verbal vs. nonverbal interactions differed sig- have expected counts of less than 5). Chi-Square tests showed a sig-
nificantly by pair χ2 (2, n = 13873) = 787.8, p < 0.001: 34% of nificant dependence between topic and the percentage of verbal events
events were verbal and 66% nonverbal for the S-PA pair; 50% verbal per sequence (%V) for the three interaction pairs. Additionally, there
and 50% nonverbal for the PA- SN pair, and 75% verbal and 25% was a significant dependence between topic and the percentage of ac-
nonverbal for the S-SN (Fig. 1). In addition, Table 6 details the fre- knowledgements and repetitions per sequence (Verbal-GCL) for S&PA
quency of verbal and nonverbal categories by pair. and PA&SN, but not for S&SN.
Communication topics also differed across pairs. The most frequent Figs. 4 and 5 shows the individual value plots for familiarity level
communication topics by pair that accounted for 70% or more of all the versus %V and Verbal-GCL respectively. Lines connect median values.
The Median Mood test was significant across pairs, meaning that there

Table 8
Number (%) of sequences by %V range and topic, by pair.

S & PA %V Ranges Catheter Clip Hold Stitching/Needle Suction Wash

0–15 221 (67%) 13 (7%) 422 (74%) 9 (18%) 1336 (88%) 444 (76%)
15–30 0 (0%) 34 (18%) 12 (2%) 18 (36%) 9 (1%) 49 (8%)
30–45 7 (2%) 81 (43%) 32 (6%) 17 (34%) 61 (4%) 52 (9%)
45–60 46 (14%) 43 (23%) 59 (10%) 5 (10%) 73 (5%) 23 (4%)
60–75 24 (7%) 15 (8%) 29 (5%) 1 (2%) 18 (1%) 10 (2%)
75–100 33 (10%) 3 (2%) 19 (3%) 0 (0%) 15 (1%) 7 (1%)

χ2(25, n = 3240) = 1166, p < 0.001

PA & SN %V Ranges Clip Remove Stitching/Needle Instrument Preparation/Organization

0–15 109 (59%) 13 (26%) 10 (20%) 47 (18%)

15–30 37 (20%) 1 (2%) 9 (18%) 7 (3%)
30–45 22 (12%) 14 (28%) 15 (29%) 58 (22%)
45–60 9 (5%) 7 (14%) 8 (16%) 41 (16%)
60–75 7 (4%) 10 (20%) 6 (12%) 45 (17%)
75–100 1 (1%) 5 (10%) 3 (6%) 65 (25%)

χ2(15, n = 549) = 190, p < 0.001

S & SN %V Ranges Camera Angle/Position Camera Clean Instrument Change

0–20 2 (4%) 1 (2%) 18 (22%)

20–40 0 (0%) 2 (4%) 2 (2%)
40–60 12 (25%) 16 (33%) 31 (38%)
60–80 23 (48%) 17 (35%) 18 (22%)
80–100 11 (23%) 13 (27%) 12 (15%)

χ2 (8, n = 178) = 26, p = 0.0012

256
J. Tiferes et al. Applied Ergonomics 78 (2019) 251–262

Table 9
Number (%) of sequences by Verbal-GCL range and topic, by pair.

S & PA GCL Ranges Catheter Clip Hold Stitching/Needle Suction Wash

0–0.1 305 (92%) 117 (62%) 537 (94%) 42 (84%) 1498 (99%) 566 (97%)
0.1–0.2 4 (1%) 42 (22%) 8 (1%) 6 (12%) 0 (0%) 2 (0%)
0.2–0.3 6 (2%) 21 (11%) 15 (3%) 2 (4%) 4 (0%) 11 (2%)
0.3–1 16 (5%) 9 (5%) 13 (2%) 0 (0%) 10 (1%) 6 (1%)

χ2 (15, n = 3240) = 654, p < 0.001

PA & SN GCL Ranges Clip Remove Stitching/Needle Instrument Preparation/Organization

0–0.1 163 (88%) 33 (66%) 27 (53%) 138 (52%)

0.1–0.2 11 (6%) 3 (6%) 7 (14%) 20 (8%)
0.2–0.3 9 (5%) 6 (12%) 10 (20%) 42 (16%)
0.3–0.4 1 (1%) 2 (4%) 4 (8%) 23 (9%)
0.4–0.5 0 (0%) 2 (4%) 2 (4%) 16 (6%)
0.5–1 1 (1%) 4 (8%) 1 (2%) 24 (9%)

χ2 (15, n = 549) = 81, p < 0.001

S & SN GCL Ranges Camera Angle/Position Camera Clean Instrument Change

0–0.1 22 (46%) 26 (53%) 58 (72%)

0.1–0.2 3 (6%) 1 (2%) 4 (5%)
0.2–0.3 10 (21%) 7 (14%) 7 (9%)
0.3–0.4 8 (17%) 9 (18%) 7 (9%)
0.4–1 5 (10%) 6 (12%) 5 (6%)

χ2 (8, n = 178) = 12, p = 0.66

is an association between the familiarity level and the number of ob-
servations below and above the medians of %V and Verbal-GCL. While
significant associations between familiarity levels and the medians of %
V and Verbal-GCL were found across all pairs, an examination of the
median values reveals that our hypothesis — which predicted that %V
and Verbal-GCL would be lower for dyads with higher familiarity— was
only supported for S&SN pairs, but not for S&PA and PA&SN.
4. Discussion
Communication problems have been suggested as a root cause of
adverse effects in surgery (Wahr et al., 2013). However, there have
been few studies that investigated the specific characteristics of the
team communication process in RAS or that have attempted coding
nonverbal communication events. The current study analyzed com-
munication strategies among robot-assisted surgical teams including
the use of both verbal and nonverbal communication methods. Better
understanding of communication during RAS can guide targeted in-
terventions to reduce medical errors related to poor team interactions.
Our observational study showed that the RAS team combined verbal
and nonverbal modes to communicate. Additionally, different roles
(i.e., surgeon, PA, or SN) used different communication strategies: the
rate of communication activity, percentage of verbal and nonverbal
events, and most frequent communication topics were different for each
pair. The differences in communication activity between pairs appear to
reflect the needs of each pair to collaborate throughout the surgery.
Since the role of the PA is to continuously assist the surgeon, this pair
communicated more frequently and their communications accounted
for 67% of all events. The interactions between the PA and SN ac-
counted for 26% of all events, reflecting the degree to which the SN
provides the PA with laparoscopic instruments for PA-specific work and
robotic instruments for the PA's side. Also, the PA has an instrument
table which is limited in size; this forces the PA to transfer and keep
instruments on the SN's table resulting in recurring communications
about when and what instruments to transfer. Interactions between the
surgeon and SN are the least frequent (7%), and primarily concern
robotic instrument changes on the SN's side of the patient and camera
adjustments. These tasks are less frequent than other RAS tasks.
4.1. Role of available communication media
Another reason for difference in communication patterns between
the pairs may be that each of the interaction pairs of the “key triad” has
different communication media available. Fig. 6 summarizes the
available communication media between the surgeon at the console and
the bedside assistants (PAs and SNs). The shaded area indicates inter-
actions that occur via the shared view. Bidirectional arrows indicate
that both members of the pair are able to emit and receive information
via that communication medium. Similarly, directional arrows indicate
that information may only go from source to receiver. The surgeon and
PA are not in direct visual proximity with one another; however, they
can both see and coordinate each other's actions via the shared console's
view. That is, the surgeon and PA can both see the robotic instruments
(operated by the surgeon) and the laparoscopic instruments (operated
by the PA). This virtual shared space allows the surgeon and the PA to
minimize the need for verbal communication. For example, sometimes
the PA can proactively perform the suction task when he or she eval-
uates this is needed according to what is shown in the shared view,
without the surgeon always requesting it. The surgeon can also request
a specific task and complement verbal commands with nonverbal ac-
tions. For instance, the surgeon may point at a location with the robotic
instrument during a request for a clip application.
Most collaborative tasks between the surgeon and PA are visually
available in the shared view which may explain why 66% of their in-
teraction events were found to be nonverbal. In contrast, the PA and SN
are the only triad pair that interacts “face-to-face”. They can directly
see what the other is doing and interact by combining speech, visual
gaze, body positions and instrument handling —which resulted in a
50%–50% proportion between verbal and nonverbal events. Finally,
interactions between Surgeon and SN are mostly verbal (75%) perhaps
because the surgeon has little visual evidence of SN activities: the
surgeon cannot directly see the SN and the shared view only shows the
beginning and end actions when the SN is removing or installing the left
robotic instrument or adjusting the camera. Thus, any other

257
J. Tiferes et al. Applied Ergonomics 78 (2019) 251–262

Fig. 4. Individual value plots of %V by familiarity between: 3.a) Surgeons and PAs. χ2 (2, n = 3240) = 21, p < 0.001. 3.b) PAs and SNs. χ2 (2, n = 549) = 10.26, p = 0.006. 3.c)
Surgeons and SNs. χ2 (2, n = 178) = 8.03, p = 0.018.

communication events between the surgeon and scrub nurse (i.e., events embedded in a certain situation and environment. For instance,
which instrument to change, type camera adjustment needed, style of all members of the triad used spatial deixes such as “here” or “there”
needle and length of thread) need to be verbal. that would not have had any meaning unless used in combination with
According to common ground theory, people shape their interac- other communication events such as nonverbal actions (like pointing or
tions according to the available communication media and the purpose handing an object) within a certain context.
of their interaction (Clark and Brennan, 1991). It is therefore not sur-
prising that the percentages of verbal and nonverbal events was sig-
4.2. Role of topic
nificantly influenced by the participant pair, since pairs had access to
different communication modes and needed to communicate about
The most frequent topics among each pair reflect the type of ac-
different things.
tivities each pair collaborates on throughout the surgery. Surgeons and
It is important to emphasize that the “modality” dimension of our
PAs interact about tasks that occur in the shared view such as: suction,
coding scheme assigned the interaction events to one of two categories:
clip applications, wash, retractions, and needle and catheter manip-
verbal or nonverbal (plus their subcategories). However, considering
ulations. The most frequent interactions between PA and SNs include
communication acts as purely verbal or nonverbal is a false dichotomy
topics that require transferring items between the left and right side
and an artificial by-product of this (and many) coding schemes, as
(instrument preparation, clip charging, needle transfer, and tissue re-
communication is actually multimodal. Therefore, the use of “verbal”
moval). Finally, most interactions between surgeon and SN involve
and “nonverbal” to classify interaction events here should not be in-
robotic camera adjustments and robotic instrument changes on the left
terpreted as isolated verbal or nonverbal acts, since a single commu-
side.
nication event cannot be understood in isolation: meaning emerges
There was a significant dependence between the percentage of
from the combination of multiple verbal or nonverbal interaction
verbal vs. nonverbal events and communication topic for each pair.

258
J. Tiferes et al.
Fig. 5. Individual value plots of Verbal-GCL by familiarity between: 3.a) Surgeons and PAs. χ2 (2, n = 3240) = 17, p < 0.001. 3.b) PAs and SNs. χ2 (2, n = 549) = 6.7, p = 0.035. 3.c)
Surgeons and SNs. χ2 (2, n = 178) = 12.2, p = 0.002.
Some pair-topic combinations resulted in the lowest percentage of
verbal interactions, such as suction and wash between S and PA or clip
between PA and SN, while others presented the highest, such as in-
strument changes, camera clean, and camera focus between S and SN.
Accordingly, a lower the percentage of verbal interactions was present
when most of collaborative actions needed to complete the task were
visually accessible for both team members in a shared workspace (ei-
ther in virtual form on the console's shared view or “face-to-face”) and
allowed some response actions to be inferred. Table 10 provides a
summary of why topics may have impacted communication mode for
each pair.
4.3. Communication analysis by sequence
The communication analysis by sequence further confirmed that dif-
ferent topic pair combinations exhibit different communication strategies,
as topic was associated with the number of (i) verbal events and (ii) ac-
knowledgments and repetitions. Additionally, the communication strategies
259
Applied Ergonomics 78 (2019) 251–262
within the same topic and pair varied from instance to instance as shown in
the wide range of number of events, percentage of verbal events, percentage of
acknowledgments and repetitions observed for each pair-topic combination.
These variations may indicate that there are other contextual factors, be-
yond pair and topic, that further influence the communication strategies of
each specific sequence. For instance, Traum and Dillenbourg (1996) state
that the necessary grounding criterion varies according to the amount of
information known at the time (higher criterion the less is known), the
complexity of the information (higher criterion with increased complexity),
and the importance of the information (higher criterion when information
directly related to the task goal).
We expected that pairs that had performed fewer cases together (i.e.
had a lower familiarity level) would have less common ground, re-
sulting in an increasing need for more explicit communication (higher
%V and Verbal-GCL). Our results overall did not support our hypothesis
(except for the S-SN pair). One possible explanation for these results is
that familiarity levels do not, in fact, affect the communication mea-
sures tested here. Or, these communication measures were not sensitive
J. Tiferes et al. Applied Ergonomics 78 (2019) 251–262

enough to capture differences due to familiarity. Alternatively, the

measure of familiarity, or the distribution of cases with different fa-
miliarity levels, used in this study may have not been sensitive enough
to capture the difference (if it was indeed present). Unfortunately, while
an effort was made to select cases with varied demographics, it was not
possible to break familiarity down into more levels due to the fact that
the data was collected during actual surgeries (e.g., we could not assign
participants to teams or to surgeries). In addition, there might have
been unknown confounding factors affecting the possible association
between team familiarity and communication. Perhaps, some con-
textual factors override differences in familiarity.

4.4. Limitations and future research

This study has several limitations. First, the characteristics of the

communication process reported here are inherently limited to the
cases that were sampled and to the specific surgeons, PAs and SNs who
participated in this study. For example, PAs in this study seemed to act
proactively and did not always wait for the surgeons to request an ac-
tion before assisting. However, we acknowledge that other surgeons
may prefer for the PAs not to act unless explicitly asked which could
change the use of verbal and nonverbal strategies among the identified
topics. Also, we did not sample from cases in which team members had
Fig. 6. Available communication media between the surgeon, physician assistant, and not provided prior consent which may have produced biased results as
scrub nurse. the communications being analyzed were among individuals who
agreed to be recorded. Our analysis only considered communication

Table 10
Interpretation of the percentage of verbal interactions by pair and topic.

Surgeon-PA pair

Suction (12% of verbal interactions) and Wash (15% of verbal Minimal need for verbalization could be due to two reasons: 1) the PAs did not always wait for the surgeon to
interactions) verbally request a suction or wash action and they completed it when they saw that it was needed, or 2) even when
the action was requested the PAs needed almost no indications on how to perform the task as they could see what
area needed to be suctioned or washed directly in the shared view.
Hold (31% of verbal interactions) During these interactions, the PA utilized a laparoscopic instrument to improve the field of view of the surgeon which
was sometimes done proactively, without any prior verbal command from the surgeon, while on other occasions
(e.g., when the surgeon needed the PA to retract in a particular way that was not completely obvious) the pair would
engage in verbal negotiations.
Stitching/Needle (32% of verbal interactions) Some verbal negotiations are needed during stitching/needle activities. The surgeon needs to ask for the needle while
specifying the type needle. Also, the associated risk involved in dropping the needle into the patient cavity stimulates
verbal exchanges between PA and surgeon when introducing and removing the needle. However, some other
stitching/needle interactions do not need much negotiations; for instance, when the surgeon ties the knot and tighten
the thread, this is an unspoken cue for the PA to cut the suture.
Clip (43% of verbal interactions). The surgeon typically requests the PA a clip verbally, and then indicates its location through verbal and/or
nonverbal means (such as pointing at the location, holding the section to be clipped, or/and adjusting the camera
position and zoom). Then the PA usually places the clip at the required location without closing it until the surgeon
confirms that the location is correct. Finally, the PA closes the clip and removes the laparoscopic clip applicator.
Catheter (49% of verbal interactions). The surgeon needs the catheter positioned in many different locations throughout the surgery which encourages
verbal exchanges. Also, since sometimes the catheter in not visually available and the PA needs to describe
verbally its location to the surgeon.

PA-SN pair

Clip (18% of verbal interactions) During clip application, the SN follows the actions on the shared screen and when the PA removes the laparoscopic
clip applicator the SN has a new applicator with a clip ready to be exchanged with the empty one. These actions
usually need no words unless there is a problem with the applicator or another clip size is needed.
Stitching/Needle (41% of verbal interactions) Verbal exchanges are needed to specify the type of needles that would be needed, plus extra exchanges, to ensure
that needle transfer from side to side happens quickly and safely.
Remove (52% of verbal interactions) Remove interactions involve the transfer of patient tissue from the PA's side to the SN side. Verbal interactions are
needed to ensure a secure and fast transfer and that the removed tissue is labeled correctly.
Instrument preparation & organization (58% of verbal Instrument preparation & organization interactions include the transfer of various laparoscopic and robotic
interactions) instruments, verbal exchanges are needed to specify which instruments are needed and when.

S-SN pair

Instrument Changes (63% of verbal interactions) The surgeon typically requests a specific instrument verbally and the SN installs it which needs a high percentage of
verbal events.
Camera clean (74% of verbal interactions) and camera angle/ These interactions are typically requested verbally by the surgeon, and since these two actions take longer than the
position (72% of verbal interactions) instrument change, the SN may update the surgeon verbally as soon as the task is completed to minimize stop times.

260
J. Tiferes et al.
among the surgeons, physician assistants and scrub nurses. The OR
team also includes additional personnel such as circulating nurses, and
anesthesiologists, and future research in RAS team communication
could include these additional communicative partners. Additionally,
as all cases were recorded in the same hospital and OR, with unique
layout, while performing the same type of surgery with the same type of
technology, our results do not take into account the possible implica-
tions that variations in hospital culture and norms, OR layout design,
type of surgery, and technology may have on communication.
Therefore, future research should evaluate the use of nonverbal com-
munication for different RAS settings, surgery types, and team com-
position. In addition, we should investigate how to incorporate non-
verbal aspects into the design of interventions (such as communication
protocols), and if doing so could improve team communication without
increasing overall communication time for surgical teams or any other
team working in time sensitive environments. Finally, our study did not
include a detailed analysis of miscommunications in terms of the si-
tuations in which those miscommunications occurred or how they im-
pact communication strategies or surgical tasks and outcomes. Future
work could focus on miscommunications specifically to better under-
stand how verbal and nonverbal communication strategies could be
affected by, or used to mitigate the effects of, communication problems.
While some few studies have reported on the existence of nonverbal
communication in surgery (Kolbe et al., 2014; Moore et al., 2010), we
are unaware of other research which explicitly coded verbal and non-
verbal communication events in any type surgery. Therefore, it is un-
clear if our findings, in terms of the relative proportions of verbal to
nonverbal events, are unique to RAS, or may be similar to other forms
of surgery (e.g., laparoscopic or open surgery). However, role- and task-
based differences in the number of verbal and nonverbal events have
been found in other domains. For example, Segal (1994) found that the
ratio of nonverbal events to the number of verbal events between pilots
was significantly affected by role (pilot flying vs pilot not flying) and
checklist interface design (paper, electronic, automatic). Also, this ratio
was also influenced by the type of task (normal checklist vs emergency
procedure). These results validate findings presented here: the pro-
portion of verbal to nonverbal communication seems to be affected by
role and the topic of communication.
4.5. Design implications
In this study, we found that surgical team members combine verbal
and nonverbal interactions to communicate, and that 59% of these in-
teractions were nonverbal. Thus, team communication could poten-
tially be enhanced if team members were aware of how nonverbal
actions can repeat, contradict, substitute, or complement verbal in-
formation. Also, since these nonverbal behaviors occurred either “face-
to-face” or mediated by technology, OR design and its layout need to
support both verbal and nonverbal communication. This means that
surgical staff members need to be able to hear each other speak, and
that screens reproducing the console's view should be accessible for the
entire surgical team to allow nonverbal, mediated communication. In
addition, these findings suggest that usability testing of surgical
equipment should go beyond testing surgeons performing tasks in-
dividually as surgical technology mediates team communication. That
is, it is important to include tests that accurately replicate how the new
technology may change, assist, or hinder the way the team commu-
nicates (in which both verbal and nonverbal communication must be
included). Finally, solutions designed to improve team communication
cannot be “one-size fits all” as the way team communicates is influ-
enced by the people involved, how their communication is mediated,
what the communication is about, and the context.
5. Conclusions
Although team communication issues have been studied in the OR,
261
Applied Ergonomics 78 (2019) 251–262
the study of nonverbal interactions during surgeries (RAS and non-RAS)
has been limited. Nonverbal behaviors are a crucial aspect of the team
communication process and should not be left out of OR communica-
tion research. In addition, the way the team combines verbal and
nonverbal behaviors changes according to the available communication
media, the roles involved in an interaction, and the specific activity
being performed.
Acknowledgements
This work has been supported by the Roswell Park Alliance
Foundation.
References
Ahmad, N., Hussein, A.A., Cavuoto, L., Sharif, M., Allers, J.C., Hinata, N., Ahmad, B.,
Kozlowski, J.D., Hashmi, Z., Bisantz, A., 2016. Ambulatory movements, team dy-
namics and interactions during robot-assisted surgery. BJU Int. 118, 132–139.
Allers, J.C., Hussein, A.A., Ahmad, N., Cavuoto, L., Wing, J.F., Hayes, R.M., Hinata, N.,
Bisantz, A.M., Guru, K.A., 2016. Evaluation and impact of workflow interruptions
during robot-assisted surgery. Urology 92, 33–37.
Argyle, M., 1972. Non-verbal communication in human social interaction. In: Hinde, R.A.
(Ed.), Non-verbal Communication. Cambridge University Press, pp. 243–269.
Cao, C.G.L., Taylor, H., 2004. Effects of New Technology on the Operating Room Team.
DTIC Document.
Catchpole, K., Perkins, C., Bresee, C., Solnik, M.J., Sherman, B., Fritch, J., Gross, B.,
Jagannathan, S., Hakami-Majd, N., Avenido, R., 2016. Safety, efficiency and learning
curves in robotic surgery: a human factors analysis. Surg. Endosc. 30, 3749–3761.
Clark, H.H., Brennan, S.E., 1991. Grounding in communication. Perspect. Soc. Shared
Cognit. 13, 127–149.
Cunningham, S., Chellali, A., Jaffre, I., Classe, J., Cao, C.G.L., 2013. Effects of experience
and workplace culture in human-robot team interaction in robotic surgery: a case
study. Int. J. Soc. Robot. 5, 75–88.
Diana, M., Marescaux, J., 2015. Robotic surgery. Br. J. Surg. 102, e15–28.
ElBardissi, A.W., Wiegmann, D.A., Dearani, J.A., Daly, R.C., Sundt, T.M., 2007.
Application of the human factors analysis and classification system methodology to
the cardiovascular surgery operating room. Ann. Thorac. Surg. 83, 1412–1419.
Gillespie, B.M., Chaboyer, W., Longbottom, P., Wallis, M., 2010. The impact of organi-
sational and individual factors on team communication in surgery: a qualitative
study. Int. J. Nurs. Stud. 47, 732–741.
Herron, D., Marohn, M., 2008. A consensus document on robotic surgery. Surg. Endosc.
22, 313–325.
Hutchins, E., 2006. The distributed cognition perspective on human interaction. In:
Enfield, N.J., Stephen, C.L. (Eds.), Roots of Human Sociality: Culture, Cognition and
Interaction, pp. 375–398 Berg Oxford and New York.
Katz, L.C., Kambe, G., Kline, K.F., Grubb, G.N., 2006. Nonverbal Communication and
Aircrew Coordination in Army Aviation: Annotated Bibliography. DTIC Document.
Kolbe, M., Grote, G., Waller, M.J., Wacker, J., Grande, B., Burtscher, M.J., Spahn, D.R.,
2014. Monitoring and talking to the room: autochthonous coordination patterns in
team interaction and performance. J. Appl. Psychol. 99, 1254–1267.
Lingard, L., Espin, S., Whyte, S., Regehr, G., Baker, G., Reznick, R., Bohnen, J., Orser, B.,
Doran, D., Grober, E., 2004. Communication failures in the operating room: an ob-
servational classification of recurrent types and effects. Qual. Saf. Health Care 13,
330–334.
Lingard, L., Reznick, R., Espin, S., Regehr, G., DeVito, I., 2002. Team communications in
the operating room: talk patterns, sites of tension, and implications for novices. Acad.
Med. 77, 232–237.
Moore, A., Butt, D., Ellis-Clarke, J., Cartmill, J., 2010. Linguistic analysis of verbal and
non-verbal communication in the operating room. ANZ J. Surg. 80, 925–929.
Morrow, D., Fischer, U., 2013. Communication is socio-technical systems. In: Lee, J.D.,
Kirlik, A. (Eds.), The Oxford Handbook of Cognitive Engineering. Oxford University
Press, USA, pp. 178–199.
Morrow, D., North, R., Wickens, C.D., 2005. Reducing and mitigating human error in
medicine. Rev. Hum. Factors Ergon. 1, 254–296.
Nagpal, K., Moorthy, K., 2010. Information Transfer and Communication in Surgery: a
Need for Improvement, Key Topics in Surgical Research and Methodology. Springer,
pp. 771–780.
Nagpal, K., Vats, A., Lamb, B., Ashrafian, H., Sevdalis, N., Vincent, C., Moorthy, K., 2010.
Information transfer and communication in surgery: a systematic review. Ann. Surg.
252, 225–239.
Nyssen, A.-S., Blavier, A.l.d, 2010. Integrating Collective Work Aspects in the Design
Process: an Analysis Case Study of the Robotic Surgery Using Communication as a
Sign of Fundamental Change, Human Error, Safety and Systems Development.
Springer, pp. 18–27.
Randell, R., Honey, S., Alvarado, N., Pearman, A., Greenhalgh, J., Long, A., Gardner, P.,
Gill, A., Jayne, D., Dowding, D., 2016. Embedding robotic surgery into routine
practice and impacts on communication and decision making: a review of the ex-
perience of surgical teams. Cognit. Technol. Work 18, 423–437.
Roth, E., Christian, C., Gustafson, M., Sheridan, T.B., Dwyer, K., Gandhi, T., Zinner, M.,
Dierks, M., 2004. Using field observations as a tool for discovery: analys ing cognitive
and collaborative demands in the operating room. Cognit. Technol. Work 6, 148–157.
J. Tiferes et al.
Segal, L., 1995. Designing team workstations: the choreography of teamwork. In: Local
Applications of the Ecological Approach to Human-machine Systems.
Segal, L.D., 1994. Effects of Checklist Interface on Non-verbal Crew Communications.
Ames Research Center. NASA, Moffett Field, California, pp. 97.
Spight, D.H., Hunter, J.G., Jobe, B.A., 2014. Minimally invasive surgery, robotics, natural
orifice transluminal endoscopic surgery, and single-incision laparoscopic surgery. In:
In: Brunicardi, F.C., Andersen, D.K., Billiar, T.R., Dunn, D.L., Hunter, J.G., Matthews,
J.B., Pollock, R.E. (Eds.), Schwartz's Principles of Surgery, vol. 10e McGraw-Hill
Education, New York, NY.
Sutcliffe, K.M., Lewton, E., Rosenthal, M.M., 2004. In: Communication Failures: an
Insidious Contributor to Medical Mishaps, vol. 79. Academic Medicine, pp. 186–194.
Tiferes, J., Bisantz, A.M., Guru, K.A., 2015. Team interaction during surgery: a systematic
review of communication coding schemes. J. Surg. Res. 195, 422–432.
Tiferes, J., Hussein, A.A., Bisantz, A., Kozlowski, J.D., Sharif, M.A., Winder, N.M., Ahmad,
N., Allers, J., Cavuoto, L., Guru, K.A., 2016. The loud surgeon behind the console:
262
Applied Ergonomics 78 (2019) 251–262
understanding team activities during robot-assisted surgery. J. Surg. Educ. 73,
504–512.
Traum, D., Dillenbourg, P., 1996. Miscommunication in multi-modal collaboration. In:
AAAI Workshop on Detecting, Repairing, and Preventing Human–machine
Miscommunication, pp. 37–46.
Wahr, J.A., Prager, R.L., Abernathy 3rd, J.H., Martinez, E.A., Salas, E., Seifert, P.C.,
Groom, R.C., Spiess, B.D., Searles, B.E., Sundt 3rd, T.M., Sanchez, J.A., Shappell, S.A.,
Culig, M.H., Lazzara, E.H., Fitzgerald, D.C., Thourani, V.H., Eghtesady, P.,
Ikonomidis, J.S., England, M.R., Sellke, F.W., Nussmeier, N.A., American Heart
Association Council on Cardiovascular, S., Anesthesia, C.o.C., Stroke, N., Council on
Quality of, C., Outcomes, R, 2013. Patient safety in the cardiac operating room:
human factors and teamwork: a scientific statement from the American Heart
Association. Circulation 128, 1139–1169.
Webster, J.L., Cao, C.G.L., 2006. Lowering communication barriers in operating room
technology. Hum. Factors: J. Human Factors Ergon. Soc. 48, 747–758.