A realist process evaluation of robot-assisted surgery: integration into routine practice and impacts on communication, collaboration and decision-making

ROLARR

The current study was conducted alongside an international, multicentre RCT entitled ROLARR, which was funded by the Medical Research Council Efficacy and Mechanism Evaluation programme. 36 The trial sought to establish whether or not RAS improves surgical outcomes in comparison with traditional laparoscopic surgery for the curative treatment of rectal cancer. The assumption underpinning the ROLARR trial was that RAS would facilitate fine tasks, such as dissection and suturing, enabling the surgeon to overcome the challenges experienced in traditional laparoscopic surgery. All of the patients entered into the trial underwent either an anterior resection or an abdominoperineal resection. For patients randomised to the intervention arm of the trial, the surgeon could choose to either undertake the operation totally robot-assisted or undertake a hybrid operation in which the first phase of the operation (mobilisation of the splenic flexure) is undertaken laparoscopically but the rectal mesorectal dissection is undertaken with robot assistance. The primary outcome of interest was conversion to open surgery, which was considered to reflect the ease of surgery; if RAS makes surgery technically easier, there should be fewer conversions to open surgery. The secondary outcomes included the accuracy of the surgery and intraoperative and postoperative complications.

Process evaluations

A process evaluation is ‘a study which aims to understand the functioning of an intervention, by examining implementation, mechanisms of impact, and contextual factors’. 37 Although the RCT design continues to be considered the most reliable method of determining effectiveness,21 process evaluations are now considered to be an essential part of designing and evaluating complex interventions if evaluations are to inform policy and practice. 38 Process evaluations are typically undertaken alongside a trial,39 although there is increasing interest in using process evaluations during the feasibility and piloting phase38 so as to inform the definitive trial, and process evaluations may also be undertaken after a trial. 40 Guidance recommends that process evaluations combine qualitative and quantitative data. 37–39

In examining implementation, process evaluations look at what was delivered, in terms of fidelity (whether or not the intervention was delivered as intended) and dose (how much of the intervention was delivered). 38 This is particularly important in multicentre trials in which the intervention may be implemented in different ways in different sites,39 and so understanding differences in what was delivered can assist in interpreting differences in results. It is also important to look at how the intervention was delivered, in terms of, for example, education and training, support, and communication and management processes. 38 Without this, effective aspects of the intervention may go unmeasured, raising concerns about the validity and reliability of the results of an evaluation41 and preventing replication. 42 For example, an important component of RAS may be the training delivered to the OT team, but if this element of the intervention is not reported and described, health-care organisations may introduce RAS without an equivalent level of training and are unlikely to achieve the same impact.

As well as capturing what and how the intervention is delivered, process evaluations explore the mechanisms through which interventions bring about change,43 which provides important understanding concerning how the impacts of the intervention might be replicated by similar interventions in the future. 38 Alongside this, it is important to capture information about the context in which the intervention is delivered, as an intervention may have different effects in different contexts. 39 Context here is taken to mean anything external to the intervention that may reduce or increase its impact, such as pre-existing circumstances, skills, organisational norms, resources and attitudes. 37

While the objectives of process evaluations have been defined as ‘to assess fidelity and quality of implementation, clarify causal mechanisms and identify contextual factors associated with variation in outcomes’,44 process evaluations do vary in terms of aims. 38 In this study, we wanted to capture what was delivered and how it was delivered within the UK ROLARR sites. All of the centres participating in ROLARR had introduced RAS prior to joining the trial, so there were likely to be variations in how RAS was implemented across the sites. Because ROLARR is an international trial and this study was looking at the UK sites only, it would not be feasible to use the data to understand variations in outcomes across the trial as a whole; in addition, given the relatively small number of operations undertaken by some sites, it is questionable how meaningful it would be to use the data to interpret differences between the UK sites. Nonetheless, given the reported challenges of integrating RAS into routine surgical practice and the fact that this topic has not been explicitly considered by existing studies of RAS, it was considered that an account of the different ways in which RAS was implemented in the UK ROLARR sites, in terms of what was delivered (e.g. totally robot-assisted operations or hybrid operations), how it was delivered (e.g. staff training, organisation of teams) and staff perceptions of the value of those different approaches would provide important information for health-care organisations considering introducing RAS.

In terms of clarifying causal mechanisms, our objective was not primarily to understand how RAS impacts the trial outcome of conversion to open surgery. Complex interventions have the potential to produce unintended consequences, which may be beneficial or harmful, and process evaluations have been identified as providing an opportunity to systematically identify and quantify unexpected unintended outcomes. 45 We wanted to capture impacts outside the scope of the trial, about how RAS, in comparison with laparoscopic surgery, impacts communication, teamwork and decision-making in the OT, and understand how and in what contexts those impacts occurred; thus, to explore this, we carried out a realist evaluation.

Realist evaluation

The evaluation of complex interventions requires a strong theoretical foundation,46 and realist evaluation provides this through a process of eliciting, testing and refining stakeholders’ theories of how an intervention works. Consequently, realist evaluation has been used for studying the implementation of a number of complex interventions in health care. 47 In realist evaluation, interventions in and of themselves are not seen as determining outcomes. Rather, interventions are considered to offer resources to recipients, and outcomes depend on how recipients make use (or not) of those resources, which will vary according to the context. Consequently, patterns in the outcomes of interventions are demi-regularities, the influence of contextual factors making them only semipredictable. 48,49 Realist evaluation seeks to answer not only the question of ‘what works?’ but ‘what works for whom, in what circumstances, and why?’. 50 It seeks to understand not only in what contexts the intended outcomes are achieved, but also unintended outcomes. Further details of how the principles of realist evaluation were applied in this study are provided in Chapter 2.

A note on theory

Given the focus within realist evaluation on the elicitation, testing and refinement of theories, an explanation of how this term is understood within realist evaluation is necessary. Although theories may sometimes be considered to be abstract and irrelevant, separate from the everyday experience of practitioners, the term can also be used simply to refer to practitioners’ ideas and thoughts about how an intervention works. 54 This is how the term is used in realist evaluation. From a realist standpoint, effective theories typically combine both substantive theory and stakeholders’ theories that are derived from experience. 51,54

Realist theories define the mechanism through which a particular outcome is achieved, whereby the mechanism consists of a resource that the intervention provides and the recipients’ reasoning about and response to that resource, and the context in which that mechanism will be activated. Thus, realist theories are often presented in the format Context + Mechanism = Outcome, or C + M = O, referred to as a context–mechanism–outcome (CMO) configuration. The activation of a mechanism should not be thought of as an on/off switch; rather, there is a continuum of activation, which will affect the outcome. 55 For example, there are varying degrees to which an OT team may feel enthusiastic towards RAS, depending on the context.

Literature review

This review constituted the first ‘theory elicitation’ stage of a realist review. Whereas in a full realist review published evidence is used to test and refine stakeholders’ theories,52 our purpose was to catalogue the theories to be refined and tested in subsequent stages of the study. The details of the methods of this review can be found in Appendix 2.

Teacher–learner cycle interviews

Realist evaluation provides a unique approach to undertaking interviews, referred to as ‘teacher–learner cycle’ interviews. 51 In teacher–learner cycle interviews, the researcher’s theory is the subject matter. An iterative approach is taken, whereby the researcher first teaches the interviewee about the theories they want to explore within the interview. The researcher then invites the interviewee to use their experience of the intervention to reflect on these theories, refining and adding to them, so the interviewee is using their experience to teach the researcher. In interviews with members of OT teams with experience of colorectal RAS, we used this approach to refine and test our literature-based theories.

Prioritisation of theories

From a realist perspective, there is an infinite number of potential influences on the interactions between a complex intervention and its intended recipients, and an infinite number of potential impacts resulting from those interactions. 50 It would be impossible to test all of the theories that account for those interactions and, therefore, it is necessary to take some theories on trust, possibly to be tested at some later point, while focusing attention on the testing of other theories. To assist in the process of determining which theories to test, at the SSC meeting and the patient panel meeting, preliminary findings from the interviews were presented and discussion was focused on what outcomes were most important to consider. Once the analysis of the phase 1 interviews was complete, a subset of possible theories to test was identified, based on a consideration of the strength of support offered in the interviews for those theories, the extent to which they concerned outcomes identified by the patient panel and SSC as being important, and the feasibility of testing those theories. This subset was presented to the clinical members of the research team, to capture their perspectives on which theories were most important to test.

Settings and participants

A multisite case study was used63 to generate findings with relevance beyond a single setting. 64 There is no consensus regarding how many case sites to include in a multisite case study. 63 The number of sites depends on the number of aspects of the context anticipated to impact on the phenomenon of interest,63 while also involving a trade-off between breadth and depth of investigation. 65 Four sites were used to enable the identification of organisational-level factors that impacted the deployment of the robot, while providing confidence in the generalisability of findings across sites. This approach has been successfully deployed by members of the research team in a previous multisite case study of the introduction of new technologies. 66 Therefore, it was decided that four case sites would be selected from the nine trusts included in phase 1. Case sites were purposively sampled to ensure variation in the experience of the surgeon and the team with RAS, as this was identified as an important contextual factor in the prioritised theories. In addition, we made sure that the case sites included both large teaching hospitals and district general hospitals, and that three of the case sites were participating in the ROLARR trial and one was not.

Once agreement to participate in phase 2 of the study was obtained from the appropriate colorectal surgeon at each site, visits were made to the sites to explain the study to other surgeons and members of the OT team and to answer any questions or address any concerns (see Appendix 5 for the staff information sheet and Appendix 6 for the staff consent form). Research governance approval was then obtained from each trust for phases 2 and 3.

Once data collection began, it was typically the colorectal surgeon or the research nurses at each site who obtained written consent from patients for us to observe and video record their operations (see Appendix 7 for the patient information sheet and Appendix 8 for the patient consent form). In each of the ROLARR case sites, while recruitment to ROLARR continued, we observed operations involving patients in the trial. Once recruitment to ROLARR ended, and in the case site not participating in ROLARR, we initially used the same inclusion and exclusion criteria as the ROLARR trial, to ensure that the operations observed were comparable (see Appendix 9). Following challenges in recruitment, the inclusion criteria were extended slightly, although these changes were discussed with the clinical members of the research team to ensure that they would not change the nature of the operation observed. Although we observed both laparoscopic and robot-assisted operations in each site, the number of laparoscopic and robot-assisted operations observed in each site depended on the extent to which these two techniques are used within the site and, for case sites in the ROLARR trial, the randomisation of patients to the different arms of the trial.

There is no consensus regarding how many periods of observation are necessary to provide an adequate overview of current practice in a particular setting. 65 We initially proposed to observe 10 rectal cancer resections in each site (n = 40), on the basis that this would be a feasible number to observe within the time frame of the project and would provide over 200 hours of data, which would constitute a substantial corpus. However, the number of suitable operations varied substantially between sites, and data collection at one site ended early due to RAS no longer being used at that trust. In total, between June 2014 and November 2015, we observed 22 rectal cancer resections, 16 robot assisted and six laparoscopic, and we were able to secure 202 hours of data collection, as many of the operations were longer than we had initially anticipated. In addition, the observation of a further 10 colorectal operations was undertaken across three of the sites at different points during data collection, five of which were robot assisted. These provided the opportunity to learn more about colorectal surgery and become familiar with the setting, to trial the structured observation tool OTAS (Observational Teamwork Assessment for Surgery) we would be using and to get to know theatre staff. Details of the number of operations observed in each site are provided in Table 2. In consultation with our SSC, we agreed that we had collected a substantial corpus of data that achieved the anticipated number of hours of observation. Therefore, rather than continue data collection beyond the proposed period, we focused on data analysis.

Data collection

Data analysis

An iterative approach to data collection and analysis was taken, to enable the ongoing testing and refinement of the theories and the gathering of further data in the light of such revisions. Field notes were entered into NVivo 10 following transcription. As a first step in analysing the data, we produced a series of matrix displays based on the case dynamics matrices described by Miles and Huberman. 62 Although these are normally produced once the data have been indexed, we found it helpful to go straight to producing the matrices, as a way of getting an overview of the data and keeping our analysis focused on the testing of our theories. A matrix display was created for each theory and, with one row for each robot-assisted operation, we summarised the anticipated contextual factors from the theories (whether or not they were present), other contextual factors that appeared to exert influence, anticipated mechanisms from the theories (whether or not they appeared to be at play), other mechanisms that appeared to be at play, and anticipated and unanticipated outcomes. We used data from the field notes, interviews undertaken after the operations, responses to the SURG-TLX questionnaire and the OTAS ratings.

As we began to write up the analysis based on the matrix displays, further questions became apparent, so we returned to the field notes for additional information. This involved indexing the data, using codes relevant to the questions and inductive codes to capture other aspects of the contexts, mechanisms and outcomes relevant to our theories.

Interviews with other surgical specialties

A 1-day workshop on designing for robot-assisted surgery

In July 2016 we held a workshop at the University of Leeds entitled ‘Designing for robotic surgery: challenges and opportunities’. The workshop had 25 attendees, made up of a mixture of engineers, computer scientists, psychologists, social scientists, surgeons, theatre nurses, ODPs and members of our patient panel. The findings from the study were presented and then participants were asked to consider the implications of the findings for the design of RAS. When we first envisioned the workshop, before beginning the study, we had anticipated that it would be purely about designing technology, whether designing surgical robots or designing other technologies for the OT that support RAS. However, given the issues that came up in the study around training and the influence of the physical configuration of technology and people in the OT on teamwork, participants were invited to think beyond the design of technology if they so wished.

Issues of implementation

The robot-assisted surgery implementation chain

The term ‘implementation chain’ is used to refer to the series of interconnected processes through which an intervention is introduced and delivered to produce immediate and intermediate outcomes and ultimate impacts (both short and long term). The introduction of complex interventions into practice typically involves long implementation chains, influenced by stakeholders at different levels within and beyond the organisation. 50 The way in which an intervention is implemented across different settings is likely to vary, affecting the ultimate impacts of that intervention and often leading to unintended consequences. 50,143 Thus, it is recommended that, both within realist reviews and prior to entering the field as part of a realist evaluation, the implementation chain of an intervention is mapped, identifying which intermediate outcomes need to occur in order to create the context for positive final outcomes, as well as examining the blockages to effective implementation and the contexts that support the flow of implementation. 50,52 Our intention was not to look at the evolution of RAS as a technology over time. Rather, our interest was in the implementation chain from the point at which the purchasing of a surgical robot by a hospital is considered as an option to the point at which RAS is used on a routine basis in that hospital.

The review revealed a number of decision points in the implementation chain of RAS, which are summarised in Figure 3. The first decision is the decision of whether or not to purchase a da Vinci robot, and there appeared to be a number of stakeholders involved in this decision. In most studies that described the introduction of RAS, this had been led by surgeons. However, also involved was whoever would be paying for the initial purchase of the robot (in many NHS trusts, some or all of the funds came from the hospital’s local charity16) and whoever would be paying the ongoing costs associated with RAS.

FIGURE 3.

Decision points on the RAS implementation chain.

For the surgeons, a number of motivations for the introduction of RAS were described. The theory, promoted by Intuitive Surgical, that the increased precision and control RAS offers the surgeon will result in improved patient outcomes was reiterated in many papers. In robot-assisted colorectal surgery, rates of conversion to open surgery are generally found to be low, providing support for this theory144–146 (although, with such studies being non-randomised and authored by RAS enthusiasts, this could be due to case selection). Another motivation was the ergonomic benefits that the robot is anticipated to offer to the surgeon. With the surgeon using natural hand movements147 and sitting at the console,148 the awkward and unnatural positions often required during laparoscopy are removed. This suggests motivation to undertake RAS would be greater among those surgeons who were familiar with Intuitive Surgical’s marketing of the da Vinci robot and/or the published literature on the clinical and ergonomic benefits of RAS. RAS also was seen as a way of providing more patients with the benefits of minimally invasive surgery, being easier to learn than laparoscopy,149,150 suggesting that the motivation to undertake RAS would be greater in those surgical specialties for which laparoscopic surgery was found to be harder and/or among those surgeons who had struggled to adapt to laparoscopic surgery. Surgeons’ natural interest in the development of new tools was also seen as a contextual factor supporting the growth of RAS. 151

At the organisational level, the drivers behind and contextual factors influencing the decision to support the purchase of a robot were less clear, although a couple of benefits for the organisation were identified. For example, RAS was perceived as attractive to surgical trainees, so a hospital may support the decision if ownership of a da Vinci robot was seen as a way to attract the best trainees,15,152 or as a way of attracting highly trained surgeons who already have experience of RAS. 16 The importance of surgeons obtaining the support of the hospital administration and nursing management was emphasised. This support was necessary not only for the initial purchase of the robot but also to ensure the provision of adequate resources while staff are on the learning curve, such as additional OT time. 152,153 How to obtain this support was not explicated, although one report described the need to create a ‘shared vision’ of what the introduction of RAS would enable, starting with the administrators. 154 The underlying theory seems to be that, by being engaged in this process of imagining potential future benefits of RAS, the hospital administration and nursing management will perceive RAS as an innovation that can assist in achieving the organisation’s goals and so will be willing to invest the necessary resources to assist its integration into routine practice.

A contextual factor that appeared to influence the decision-making of both surgeons and administrators was the attitude of patients and the public to RAS. Intuitive Surgical’s marketing of the da Vinci robot has been characterised as aggressive151 and in the USA it has included direct-to-patient advertising on billboards and the internet,155 with the result that RAS has become a symbol of providing enhanced care. 151 In the context of competition among surgeons and hospitals, this perception of RAS among patients and the public is seen as a significant factor in the rapid growth in the purchase of da Vinci robots. 151 This creates a cycle in which, having purchased a robot and wanting to maximise its use, surgeons and hospitals promote RAS to patients, further increasing public enthusiasm for RAS. 150 This promotion of RAS by surgeons and hospitals is not unique to the USA; a number of NHS trust websites were found to promote the benefits of RAS, particularly in relation to prostate cancer, emphasising the advanced, cutting-edge technology and the high cost of the system, with some trust websites describing hospital open days during which members of the public could see and try the robot. Furthermore, the presence of a robot was seen as a means of ‘raising the profile’ of a trust and enabling it to attract patients from ‘elsewhere’ (presumably outside the local catchment area), thus increasing patient volume within the trust. 156

The decision to buy a robot includes a decision about what model of da Vinci robot to purchase, and this appeared to be a decision with implications for the subsequent integration of RAS into practice, including the feasibility of using the robot for particular operations. To date, there have been five different models of the da Vinci robot in use: standard, S, S HD, Si HD and Xi. The standard da Vinci, introduced in 1999, has been discontinued, and technical support in terms of the availability of parts and services stopped in 2014. 157 The larger size of the standard da Vinci compared with later models potentially increases the likelihood of external collision of the robotic arms and also makes it more difficult to manoeuvre. The increased likelihood of collision of the robotic arms and the shorter arms have been cited as reasons why rectal cancer resections must be performed as hybrid operations when using the standard system. 157

Following the decision to purchase a robot, presumably led by a particular surgical specialty, there is then the decision about which (other) surgical specialties the robot should be used for. In many NHS trusts, RAS was initially introduced in urology, for undertaking radical prostatectomy for prostate cancer. 16 The advantages of RAS seem to apply especially to procedures carried out in the pelvis or rectum because surgeons are operating in limited space, where laparoscopic surgery may be difficult. 144,158

Alongside decisions about who will be using RAS are decisions about what training the surgeons and OT teams should receive, although who was involved in such decisions and what influenced the decisions was not reported in the studies. The studies we reviewed described a number of features of RAS surgeons have to become familiar with: the procedure of docking the robot, including the set-up of the robotic arms; how to control the robot and troubleshoot; how to deal with the absence of tactile information; and how to mentally visualise the spatial relationships of the robotic arms, to minimise external clashing of robotic arms and to optimise manoeuvrability and range of motion. 159–162 Training provided for surgeons by Intuitive Surgical was reported to consist of a two-day course with practice in an animal laboratory, but without performance-based end points or verification of skill, leading some to argue that it is an ‘antiquated educational method’ that should be replaced by carefully structured standardised simulator curricula. 163 Intuitive Surgical has since introduced online training modules and there is a simulator curriculum that can be undertaken on the robot itself.

In relation to training for the rest of the team, a well-trained team was described as necessary to ensure the smooth running of a robot-assisted operation. 147 The scrub practitioner and circulating practitioner need to acquire new skills and need to learn their part in draping and docking the robot, how to change instruments, and troubleshooting. 164,165 The first assistant will need to learn all of this, as well as have an understanding of basic laparoscopic surgery. Specific training for first assistants, involving dry-laboratory training, personal instruction sessions, videos and surgery observation, was described. 166 Although studies occasionally talked of theatre nurses participating in training,167 generally there was little mention of training for the wider OT team.

Following on from this is the decision of which operations to use the robot for and on which patients, a decision typically made by the consultant surgeon. This is not a one-off decision but one that is likely to evolve over time as a surgeon’s experience with RAS increases. Some authors recommended starting with more straightforward procedures and then gradually moving to increasingly difficult procedures,167 the underlying theory seeming to be that, by gaining experience with simpler procedures, the surgeon builds up his or her skills, providing him or her with both the skills and the confidence to be able to take on more challenging procedures. A contextual factor that influences the decision of whether or not to undertake an operation with robot assistance is the patient condition, as the longer operation duration of RAS (discussed below) may be prohibitive in certain patients, although it can be difficult to assess who those patients are. 168 Another contextual factor that influences the decision is the anticipated benefit, which is based on the details of the operation to be performed and characteristics of the patient. For example, in the context of rectal cancer surgery, maximum benefit was reported to be expected in cases involving mid to lower rectal cancer, male patients (due to the narrower pelvises), obese patients and patients who have had preoperative chemoradiation therapy. 169,170

Alongside this are decisions about how to undertake a particular operation with the robot, again a decision typically undertaken by the consultant surgeon and again not a one-off decision but one likely to evolve over time. For example, in robot-assisted rectal cancer surgery, there is no standardised technique and the decision has to be made whether to perform it as a hybrid operation or totally robot assisted. 171 As noted above, this decision will be influenced by the model of robot the hospital has acquired. However, there are also other factors that influence the decision. Performing a hybrid operation removes the need to redock the robot, which can add to the duration of the operation,171 and thus the decision may be influenced by concerns about time. This ties in with concerns about the patient condition; if patient factors mean that the operation is likely to be difficult and require considerable time, a hybrid operation is more likely. Experience seems to be an important contextual factor that influences surgeons’ decisions about how to perform an operation. For example, it has been suggested that, for surgeons experienced in laparoscopic surgery, a hybrid approach is preferable, as their increased experience will mean that they feel more comfortable carrying out the first phase of the operation that way,171,172 while a totally robot-assisted procedure may be preferable for surgeons who have less experience of laparoscopic surgery, despite the need to redock. 173 Another decision relates to the positioning of the ports through which the instruments are inserted, and some authors described not wanting to follow the layouts suggested by other surgeons experienced in RAS because they wanted to adapt the positioning they used for laparoscopic operations, again presumably wanting to build on their existing experience. 167

Impact of robot-assisted surgery on communication and teamwork

Reporting on their experience of undertaking RAS, surgeons described how the physical separation of the surgeon from the rest of the team and the lack of visual contact make it harder for the team to hear the surgeon’s oral instructions,152 particularly if the surgeon becomes immersed in the console. 154 Consequently, it has been suggested the team needs to listen more carefully,152 again implying the need for a motivated team. There is a perception that, if the team does not respond in this way, communication is compromised. 152,158 Although the consequences of this were not explicated in the literature we reviewed, one consequence may be a further increase in operation duration, as communication failure is known to be a significant predictor of deviation in expected length of operation. 186 The use of directional cues is considered to be problematic in RAS, potentially resulting in confusion, time-wasting and patient injury. 164 This problem occurs, presumably, because the separation means that team members do not have the same physical context as the surgeon to understand such deictic instructions and, again, the surgeon is unable to support those instructions with gestures.

The impact of RAS on communication between the surgeon and the first assistant was another area in which we found conflicting theories. Some authors argued that communication between the surgeon and first assistant is particularly important in RAS, especially during instrument exchanges, when failure in communication could lead to ‘inadvertent adjustment, movement and complete removal of an instrument that is in use’. 164 However, it has also been argued that less co-ordination and communication may be required between the surgeon and the first assistant because the surgeon controls the camera and, if four robotic arms are used, can do more of the retraction (instead of this being done by the first assistant). 157,187

Strategies for overcoming the communication challenges focused on the use of standardised communication. The use of ‘readback’, in which team members repeat back instructions in a precise, clear, standardised manner, has been advocated, particularly for instrument exchanges and other key transition points. 154 This allows the surgeon to check that his or her instructions have been heard correctly (remembering that, being in the console, they are unable to draw on visual cues to determine this) and, if not, to correct any misunderstandings before they result in actions that could have negative consequences for the patient. The use of agreed terms has been recommended for RAS. 164 The use of anatomic or OT references by the surgeon, rather than directional cues, has been recommended while moving the patient or robot during docking, again to reduce the risk of misunderstandings. 164

The impact of robot-assisted surgery on decision-making

There were competing theories in the literature concerning the impact of RAS on decision-making. Some surgeons reported a ‘tendency for surgeons to bury themselves in the console’, with surgeons ‘block[ing] out the operating room’. 154 This suggests a reduction in the surgeon’s situational awareness. Some surgeons recommended positioning the console so that the surgeon has a clear view of the patient and can immediately see the patient when looking up from the console. 164

There may be beneficial impacts of RAS on surgeon decision-making. It has been argued that, immersed in the console, the surgeon’s ‘distractibility’ is reduced, which could potentially have a positive impact on patient outcomes. 188 This is supported by recent research that reveals that the number of intraoperative interruptions is significantly associated with surgeons’ experienced distraction, and interruptions in the form of case-irrelevant communication in particular are linked to increased surgeon distraction. 189 Others suggest that the 3D image creates a sense of immersion, which presumably contributes to reduced distractibility. 190

Robot-assisted surgery may also impact on surgeon decision-making by reducing the surgeon’s level of stress. Some surgeons have argued that stress arising from a difficult operation may lead a surgeon to decide to convert from laparoscopic surgery to open surgery. 171 By removing the awkward and unnatural movements required during laparoscopy191 and enabling the surgeon to sit comfortably at the console,148 RAS reduces physical discomfort. 192 This leads to the theory that, owing to the ergonomic benefits of RAS, surgeon stress is reduced, which may influence the decision of whether to convert, potentially resulting in a lower rate of conversion. 171

Robot-assisted surgery also impacts on surgeon decision-making by changing the ability of the surgeon to use tactile perception to determine anatomic information. This is considered to be a major limitation of RAS. 193 A significant contextual factor here is the surgeon’s experience, with surgeons finding visual information sufficient for informing their decision-making as their experience of RAS increases. 172

Organisational support

One of the theories identified from the review was that, by being engaged in the process of imagining potential future benefits of RAS, the hospital administration and nursing management will perceive RAS as an innovation that can assist in achieving the organisation’s goals and so will be willing to invest the necessary resources to assist its integration into routine practice. Our participants emphasised the vital role of support of the hospital administration in enabling the acquisition of a robot, perceiving the introduction of RAS as impossible without it:

It was paramount that we had the administrators behind the decision and the way things are happening with the NHS at the moment it’s very difficult to introduce anything through a bottom-up scenario and going through your divisional manager or clinical director and then pushing from that point of view. It really has to come from the top, from the board-level administration down.

Site 4, surgeon

Emphasis was placed on financial support, because of the initial cost of the robot and the ongoing costs of maintenance and disposable items:

[Support of the hospital administration] is important because . . . you can’t function in isolation. Obviously it’s expensive . . . so it needs the full input of everybody, and . . . you need the hospital on board with that, and there’s ongoing contracts and warranties.

Site 1, surgeon

It needs a lot of support as in financial support, that is the support that the robot needs. Because everything we use on the robot is expensive and each item that we use is . . . well I say each item, the majority of items that we use are disposable, they only have 10 uses and it won’t let you use it more than 10 times.

Site 1, ODP

It was also felt that the support of the hospital administration was important because of the possible negative consequences of RAS, in terms of the longer operation duration and the impact that this could have, something surgeons in particular were aware of. Consequently, surgeons would not accept responsibility for the implementation of RAS without support from the hospital administration.

Our participants also provided insight into some of the ways in which this support was achieved. Creating a shared vision in some cases literally meant giving the hospital administration the opportunity to see the robot in action:

We had two very important people come and watch an operation and one of the directors of finance and one of the, I can’t remember quite what their titles were, but they came in and one of our surgeons invited them in to come and have a look at the robot. They came and watched a full case and I talked to them afterwards and they said it was very, very informative to actually see what goes on compared to what they hear. And to actually see it they realised how impressive it was and also the benefits to the patient [. . .] I got a lovely e-mail off both of them saying it was very informative and [. . .] when they can go to the board of management [. . .] they can then have a better idea of what they’re talking about to promote robotic surgery.

Site 4, ODP

Whereas this quotation emphasises the perception of patient benefits, other participants emphasised the hospital administration’s awareness of competition, which could outweigh concerns about cost:

Even the chief executive I think has a very sensible approach to it and says, quite frankly, it is expensive, it doesn’t pay for itself, but he sees it as cutting edge, I think as a kudos thing, probably, although he hasn’t quite said that, but I think reading between the lines it’s a kudos thing. And I think he bought robot number 2, I think on the basis that a neighbouring trust was talking buying a robot and he felt that some of our specialities weren’t able to progress because of access to the robot. So he wants to facilitate us being there, you know, in the [region] leading as it were.

Site 4, surgeon

I think the fact that we were the first in this part of the country to have it. I sort of got the impression that there was a little bit of a race on between whether [nearby trust] would have it or [nearby trust] would have it, or we would have it, and just the little snippets that I gleaned from the consultants, you know, that it was clear that it was going to be us. It was considered a very prestigious move, so yes it was considered, you know, to be such a futuristic addition to our theatres that it was very exciting. Everybody was curious.

Site 5, nurse

Functioning in an increasingly competitive situation, RAS was perceived as a mark of prestige or kudos for the trust and enabled it to be viewed as providing cutting-edge services, which in turn enhanced the likelihood that such services would be retained, particularly prostatectomies.

Rather than nursing management, our participants talked about the importance of team leaders in supporting the introduction of RAS, a role taken on by experienced theatre nurses and ODPs. A supportive team leader could facilitate integration of the robot by:

• gaining access to training for the whole team, which contributed to confidence in using the equipment and to safety

• co-ordinating the staff rota to ensure that the right skill mix was available to carry out robot-assisted operations

• ensuring the availability of the robot across specialties in order to co-ordinate use of the robot across the trust

• managing OT schedules to allow, at least initially, for longer set-up times and for the availability of an OT suitable to accommodate the equipment and personnel safely, without risk of desterilisation or compromising access to the patient.

Although in most sites participants felt that there had been good team leader support, the importance of team leader support was most apparent in those accounts of participants who felt that it had been lacking:

It would have lightened the workload off me and the burden on me to fight all the time to get cases done. etc. . . . Theatre time, theatre space, flexibility when you’re introducing a new technology.

Site 2, surgeon

Finally, support from surgical colleagues was perceived as important in accepting the impact of RAS on operation duration and subsequent impacts on waiting lists. That RAS was more likely to become embedded in practice if surgical colleagues were supportive was not one of our original ideas. Rather, in one of the early interviews, a surgeon stated that he felt it was difficult for RAS to become embedded in the NHS because of what he perceived as a focus on surgeon-specific outcomes. An individual surgeon’s outcomes may worsen during the time they are learning RAS, putting them under increased scrutiny, which may deter some surgeons from adopting RAS:

Everyone’s looking to knock you down, they want the first robotic death, you know, to pooh-pooh the robot. No one looks at it with a positive thing and there’s a real lack of support. So you’re always looking over your shoulder [. . .] the scrutiny as a surgeon is just ridiculous. So you put yourself under more scrutiny and you think, well why would I possibly want to do this, you’ve got to be mad.

Site 2, surgeon

When we asked other surgeons about this in subsequent interviews, they indicated that support from colleagues could mitigate many of the potential threats of RAS on performance indicators (e.g. on extending waiting lists). However, lack of support from colleagues and jealousy from others could limit the use of RAS. As one participant explained:

You need the absolute support of your [surgical] colleagues . . . First of all if you’re going to start spending 10, all day lists on your first 10 cancers then your waiting list increases or the pressure on others increases. If there’s any murmuring from the background . . . you will start to avoid doing this [RAS]. Secondly if colleagues hate the idea of others learning a skill or getting a reputation which they don’t have yet, they could scupper this happening. I’ve been lucky that those things don’t count here and that’s one of the reasons why I can progress. When I speak to colleagues they cite one or all of those, say they not actually allowed to progress.

Site 7, surgeon

This suggests that the introduction of RAS could potentially increase surgical waiting times, because a longer operation duration leads to fewer operations being scheduled or to operations being cancelled. Surgical waiting times for cancer treatment are the focus of government targets; increases in waiting times could attract fines for the trust and possibly lead to patients choosing to have surgery in other trusts and thus to a loss of income for the trust. In this context, a lack of support from surgical colleagues could tip the balance and add to existing constraints to the use of RAS, which in turn could mean that RAS is avoided in order to reduce these risks.

Team involvement

A candidate theory we began the study with was that if the whole OT team can feel that the advantages of RAS outweigh its disadvantages and are involved in the decision to introduce it, they will be more motivated to work together to develop solutions to problems that may arise when using it. The OT team had not been involved in the decision to introduce RAS in any of the nine sites. However, it did not seem that involvement in the decision was essential for the OT team to have the motivation to persist with RAS. In most sites, there was a positive attitude among the OT team towards RAS. For example, one nurse noted that, for them, there was a sense of pride as the robot added ‘another string to their bow’. This view was echoed by a nurse at another site, who described the robot as a ‘good opportunity’ with regard to their CV and professional development, particularly because not all trusts have a robot.

However, attitudes at one site, site 1, were notably different. Whereas the OT team members at the other sites appeared accepting of the fact that they were not involved in the decision, one nurse at site 1 expressed disappointment about this:

I think it’s a nice piece of equipment and I would love to have been asked to be involved in making that decision, not just it being given to me, or handed to me. Because for anyone, it would be nice to have somebody to say, yes I would like to have involvement in that, it seems to be interesting to me, because that would mean they’re curious and they will have that . . . they will be driven to learn more than if they had just been told. They can learn it more intimately than someone who has been given the job. It’s something that the person made the decision to actually get involved with the robot procedures.

Site 1, nurse

What this quotation seems to highlight is a perceived lack of control over aspects of their work; the decision they wanted to be involved in was not the one about whether or not to purchase a robot, but the one about extending the use of that robot to colorectal surgery. An ODP at the same site expressed similar sentiments and felt that having greater staff involvement in the decision would have positively impacted staff engagement:

That element of communication and knowing and agreeing that this is what we’re going to do from the start and this is how we’re going to implement certain areas, and this is what you need, and these are the dangers and these are the benefits, and things like that. I think it’s really important that the team know. And it will make them work better together, you know, you feel more comfortable if you know the bigger picture as opposed to little bits thrown in.

Site 1, ODP

Although theatre nurses and ODPs at this site expressed an appreciation of the potential benefits of RAS for the patient, attitudes to the use of it within their trust were generally negative, and this appeared to be related to a lack of confidence in undertaking the tasks associated with RAS. It was suggested the robot was not very popular because the team were not provided with an opportunity to learn how to use it:

We were actually kind of upset when we were told we were doing it because where was the training. We were all questioning, well I’m not trained, I wasn’t particularly happy with that because I wasn’t trained. I don’t know I’ll be safe, or my patient won’t be safe when I started to do it.

Site 1, nurse

Thus, it seems that motivation to persist with RAS can be achieved without involvement in the initial decision to purchase a robot, but what is essential is training that enables team members to feel confident as they take on the new tasks.

Dedicated team

Participants reported that, in many cases, the people who trained together became a dedicated robotic team, at least initially. Our theory from the literature review was that, by working through all robot-assisted cases, a dedicated team more quickly becomes familiar and confident with the equipment and tasks associated with setting up the robot, allowing members to complete the necessary tasks more quickly, reducing set-up time. Certainly, participants perceived that having a dedicated team could reduce operation duration:

When we had a dedicated team of people who could manoeuvre the robot and position patients . . . to start with you do need a core knowledge . . . it definitely did reduce the time having the same skill set.

Site 4, surgeon

Within most teams, however, it was not always possible to maintain a dedicated team owing to staff changes, holidays and sickness. The tendency for theatre nurses and ODPs to work only within one or two specialties also made a dedicated team hard to achieve, especially where there was a low volume of robot-assisted cases. When a dedicated team was not feasible, a larger pool of people, trained by the experienced staff, was established:

Within our team we have four out of seven people who know how to use the robot . . . So ideally if three or four of those people were at work on any given day and could be put in the same theatre where the robot was, then that would be the ideal. But at worst we should have at least one or two people at work on that day who could be moved in order to do the robot case. But the aim is that over a period of time that everybody within our team should be able to use the robot more efficiently.

Site 1, ODP

At some sites as many as 50% of the staff had been trained, and, at one site, which carried out a large volume of robot-assisted cases, all of the staff could manage the cases.

Dedicated operating theatre

A theory from the review was that having a dedicated OT helps RAS to become embedded in practice because the team does not have to move the robot, and this reduces set-up time and speeds turnover to the next case. Participants agreed with this and, although only three sites had a dedicated OT, participants felt that a dedicated OT would be the ideal situation, to avoid having to move the robot:

I think the biggest issue with it is it’s, from my point of view, it needs to have a specific designated theatre because it’s the time it takes to move it from storage to the theatre it’s going to be used in, what speciality it’s going to be used in and the complexities associated with that.

Site 9, nurse

When there was no dedicated OT, team leaders were perceived to play a vital role in ensuring that a suitably sized OT was available. In fact, an adequately large OT appeared to be a greater concern than necessarily having a dedicated OT. Participants felt that a suitably sized OT would make RAS more efficient because a cramped working environment meant that staff struggled to move around quickly and safely. It could also cause them to break the sterile field:

We always worry about doors opening and bashing the robot . . . or we get desterilised because we’re prepping near the door that leads to the corridor.

Site 1, nurse

The desterilisation of equipment has implications for patient safety and it is then necessary to replace or redrape the equipment, potentially extending the operation duration, with the concomitant risks of having to cancel or reschedule operations planned for later that day. However, OT teams at some sites had developed strategies for reducing the risk of desterilisation:

What we do is we cover the robot and the robot’s arms with sterile drapes, so if anybody pushes against them we don’t desterilise. So we take measures to minimise the chance of contamination of the sterility of the robotic arms anyway.

Site 4, ODP

Communication

Some participants suggested that RAS reduced some challenges to communication experienced during laparoscopic surgery. For example, ‘mumbling’ from the surgeon is reduced because they do not wear a surgical mask when seated at the console. However, most participants agreed that RAS resulted in loss of both oral and non-verbal communication, compared with laparoscopic surgery. A theatre nurse explained that during a laparoscopic case:

We’re so close to each other that when you say something, most of the time I would hear it or I would even know what he [the surgeon] intends to do next because I can see it from his movements, his mannerisms.

Site 1, nurse

During laparoscopic surgery, the proximity of the surgeon enables the scrub practitioner and first assistant to hear the surgeon’s requests. However, as suggested, the surgeon’s ‘movements’ and ‘mannerisms’ also convey the message.

Constraints on the use of non-verbal cues were also noted by surgeons, who described that ‘no one notices if they are struggling’ when they are positioned at the console. Non-verbal cues appeared to enable team members to gauge the progress of the operation. These appeared to be particularly important when the operation was at a difficult stage or was not progressing to plan, signalling the need for support from other team members who could step in without the need to be told to do so. However, in RAS, surgeons had to verbalise their concerns to team members, as they could not rely on team members to pick up on their non-verbal cues of stress.

The constraints on the use of non-verbal cues emphasised the importance of oral communication during RAS, but participants discussed how this could be compromised. Although during RAS a microphone is fitted to the console to amplify the surgeon’s voice, team members reported that it was still difficult to hear the surgeon because this microphone did not always work well:

In a robot case, because they [the surgeons] are somewhere in the corner and we hear them through a speaker and sometimes that doesn’t even work very well, and we always have to ask or repeat what he said just to be absolutely sure that whatever we’re going to do is the right thing.

Site 1, nurse

This results in the repetition of instructions from the surgeon and more frequent requests for the surgeon to repeat their instruction. This idea was used to inform a tentative theory that was presented to participants, as follows:

The physical separation [the surgeon from the team] means that it can be difficult for the team to hear what the surgeon says, resulting in increased repetition of instructions, requests to repeat information, reduced co-ordination and longer operation duration.

Subsequent findings reinforced the relevance of this theory. In discussing this theory, participants described strategies (implicit or explicit) that they have used to overcome the separation. Although the literature review suggested that the team might use a number of strategies to overcome the communication challenges of RAS, including the use of standardised communication strategies such as ‘readback’, we did not initially ask participants about the use of such strategies, and so it was interesting to see the similarities in their accounts:

I would always say, scissors coming out, Maryland [dissection instrument] going in, whatever, so they [the surgeon] always knew loud and clear that we’ve [the nurses] not only heard them but we’re actually doing it.

Site 5, nurse

This strategy was perceived to improve information transfer and avoid complications that could extend operation duration.

When presented with the tentative theory, one surgeon commented:

I think that’s possibly true [that it is more difficult to hear the surgeon during RAS], but I think working with two surgeons together makes it a lot easier in that one surgeon can be at the table and the other at the console, and they can communicate very well with each other and then to the theatre staff.

Site 4, surgeon

This quotation reveals another strategy: the use of two surgeons to conduct the operation. At the majority of sites, consultant surgeons initially worked in pairs as they became familiar with the robot. Thus, this strategy seemed to be used more frequently when teams were at the earlier stages of gaining experience in the use of RAS. The participant describes how this strategy enabled effective communication, with one surgeon making ‘many trips’ between the table and the console. The use of two surgeons increases the cost of the intervention, but ensured that the scrub practitioner and the first assistant understood the information communicated by the operating surgeon, again avoiding complications that might extend operation duration.

In an extension of our tentative theory, interviewees suggested that the lack of non-verbal communication and visual contact between the team and the surgeon also resulted in confusion about to whom surgeons’ requests were directed. As one surgeon explained:

If you say, suction and suck the smoke [smoke is sometimes generated when the diathermy is used to cut and cauterise], the nurse might look at the assistant say, who is he talking to? So I think you need clear instructions, who should be doing what. I think that’s probably a skill on its own, sort of to say things a bit more clearly and then in a sort of crisp concise manner I think.

Site 6, surgeon

To counteract this problem, the surgeon explicitly announces to whom he or she is talking and provides ‘clear’ instruction. We refer to this strategy as ‘explicit communication’.

Engagement

As part of the discussion of how and why RAS impacted on communication and teamwork, participants described changes to their duties and responsibilities:

The robot does a lot of the work that the scrub nurse used to do of handing instruments . . . the exchange of instruments is massively reduced because of the dexterity of the instruments we use and the type of instruments we use. And the surgical assistant would . . . you know, if he needs a suture then they just hand it to the surgical assistant and he will put it in and take it out. So it’s basically just the odd handing of instruments.

Site 4, ODP

Consequently, some participants described RAS as ‘monotonous’ and said that ‘there is a lot of standing around’. One surgeon provided some insight into the repercussions of this change:

I’m the only one working and because I’m the only one working, everyone else’s attention gets distracted, including the assistant and they start chitting chatting away and then occasionally I’ll be sitting there with a tied suture waiting for them to cut. I look over and they’re chatting away.

Site 7, surgeon

A preliminary analysis of these accounts was used to develop a new proposition that encapsulated the impact of the reconfigured roles and responsibilities produced by the use of RAS. Owing to this reconfiguration, some team members experienced a reduction in their task load, which might reduce their engagement in the procedure and mean that they stop monitoring the screen and may engage in more case-irrelevant communication than they would during a laparoscopic operation. This affects co-ordination, either because the team do not anticipate what the surgeon wants or because they do not hear him when he asks for something, potentially increasing operation duration. Participants supported this idea.

Teaching and learning were associated with maintaining engagement:

Sometimes I’ve noticed the scrub staff are less engaged because they’re kind of sat back a little bit, [. . .] I suppose a lot of the surgical registrars are still learning, usually they seem totally engaged from my perspective.

Site 4, anaesthetist

A theatre nurse supported this view, reporting that learning had maintained their engagement when RAS was introduced at their trust, and that teaching others prevents boredom now that this is established in practice. However, without these aspects (learning or teaching) they are ‘falling asleep’ during robot-assisted procedures. One surgeon reported using a strategy to maintain engagement that incorporated a teaching element:

The way I deal with my assistant, I say, oh this is the plane you would cut so that they’re not looking away somewhere else chatting or something or someone, and then you say, well this is the normal sort of thing you see, this is the plane you would cut, so I explain the technique and then give them sort of tips on how they can do the operation when it comes to their turn.

Site 6, surgeon

The surgeon provided an educational commentary so that their first assistant would listen to the information and remain focused on the task at hand. This idea was discussed as a potential strategy with remaining participants, who explained that its success was dependent on the recipient:

If you’re interested in the operation yes, that’s [commentary] good, you would take attention and you would concentrate in an operation, if you had not, if the surgeon is still talking to me you wouldn’t be more interested.

Site 7, trainee

Some surgeons felt that they would not be able to provide a commentary during more challenging points of the procedure:

Personally I’m likely to clam up more probably at just the time when you need someone to help you because you’re concentrating more and you don’t want to distract yourself by describing what you’re doing because you’re too busy in yourself doing it.

Site 3, surgeon

Co-ordination

One of our candidate theories, which remained unrevised following the literature review, related to co-ordination between the surgeon and the team:

When the surgeon is separated from the rest of the OT team, the team is less aware of the surgeon’s actions, making it more difficult to co-ordinate their actions during the operation and so the operation takes longer.

Participants refuted this idea, stating that co-ordination is the same in RAS and laparoscopic surgery because in both types of operations the first assistant and the scrub practitioner watch the screen, so that they know what the surgeon is doing and can anticipate what he or she needs:

I don’t really know whether I go along with that [the proposition] because I think if all the theatre team are watching the [. . .] screens, they’re watching exactly what the surgeon does in the same way as they would if it was ordinary laparoscopic surgery.

Site 1, ODP

However, as we explored this idea further, it became clear that this depends on having a team who are familiar with the operation and who do watch the screen, which ties in with the issue of engagement described in the previous section. When this is present, there should be reduced oral communication and improved co-ordination.

Particularly important seemed to be the experience of the first assistant, as this surgeon describes:

If it’s a consultant assisting or very experienced trainee then they will do something, they’ll notice that there’s something they can do to help exposure and they will make the operation easier. And it’s quite difficult to communicate that to a junior or an inexperienced assistant.

Site 3, surgeon

Although in the review there were conflicting ideas about co-ordination between the first assistant and the surgeon, the surgeons we interviewed felt that they were more dependent on the first assistant in a robotic operation because of the physical separation. As one surgeon said, ‘You can’t just grab it [the instrument] yourself’. Surgeons also reported finding it harder to guide the first assistant, because of the communication challenges but also, again, because of the physical separation – ‘you can tell but not show’ – and therefore they felt that it was more important to have an experienced first assistant in a robot-assisted operation. For this reason, some sites were using experienced theatre nurses and ODPs who had done the necessary training as assistants, whereas previously surgical trainees would have assisted.

Training

Among our candidate theories, one was that it is harder for the surgeon to explain what is happening and monitor the trainee’s understanding because of the different views of the surgical site, resulting in the trainee not learning as much as they would in other forms of surgery. However, the findings presented above suggest that the reduced opportunities for learning are due not to the different views of the surgical site but to the reduced role of the first assistant in robot-assisted operations and the greater use of theatre nurses and ODPs as first assistants in place of surgical trainees.

The idea that RAS reduced opportunities for learning was described by the surgical trainees we interviewed:

You really get less opportunity; unless the surgeon lets you into the console you don’t get very much learning experience.

Site 1, trainee

When we discussed this with surgeons, some questioned the appropriateness of involving surgical trainees in robot-assisted operations, on the basis that use of RAS was not widespread enough, an attitude also expressed by some trainees; they were unlikely to need skills in RAS in their first consultant post. However, surgeons did try to find ways to increase the learning opportunities available for surgical trainees within robot-assisted operations. Some surgeons did their robot-assisted operations as hybrid operations, allowing the surgical trainee more involvement in the laparoscopic phase of the operation. As described above, some surgeons would provide an educational commentary, although both surgeons and trainees felt that the surgeon would be less likely to provide a commentary while undertaking a robot-assisted operation:

Because the surgeon is also really by himself concentrating on the operation so it is not as interactive as having each of you side to side or opposite each other.

Site 7, trainee

Situational awareness

Our candidate theory suggested that, when the team is more experienced in RAS, they understand that the surgeon’s situational awareness is dependent on them orally communicating information, and so they respond by using more oral communication about the patient’s state, which in turn improves the surgeon’s situation awareness. The majority of surgeons perceived that their situational awareness is potentially reduced during RAS, stating that they are focused on a small area and therefore are less aware of their environment: they have ‘tunnel vision’. One surgeon provided the example of ‘sucking fluid’; the surgeon can request that their assistant provide suction, but they are unaware if the assistant experiences difficulties fulfilling their request or the reasons why:

Sometimes the assistants struggle to do something and when you’re with them you can see why they’re struggling because something is in the way or whatever. Whereas when you’re away from them in the box you say, can you just suck that fluid out there, and they can’t get the sucker to where you want them to go. And it’s like what’s the matter, what’s the matter? And they’re like, oh I’m . . . and you’re like, come on suck it through there. And it’s like, I can’t, I can’t get to it. And you know, you don’t know why because you’re not there and you can’t see the problem that they’re having.

Site 1, surgeon

Views about the seriousness of this reduced situational awareness varied. One surgeon described being ‘vastly less aware’ of what is going on in the OT but he had not been ‘hindered’ by this and did not think it made any difference to the operation. The strategy described above of two consultant surgeons working together was seen as a way to counteract the problem of reduced situational awareness, with one surgeon stating that he would be very concerned about reduced situational awareness if he operated without a second surgeon present:

Well that’s why with another consultant there who you trust, it’s like a wingman, like a co-driver. I wouldn’t do it without one. That’s why it’s more labour intensive. So in answer to your question, arguably if I was on my own I’d be very concerned about it.

Site 2, surgeon

Only two surgeons felt that their situational awareness was not reduced. One described that he continued to listen to the ‘banter’ among the team in the OT, while the other surgeon made a conscious effort to intermittently ask the team about the patient’s status.

Team members also perceived the surgeon’s situational awareness to be reduced due to their position in the console; for example, the surgeons do not have lateral vision and their sensory feedback, which can indicate problems, is reduced. The consequences of the surgeon’s reduced situational awareness, as described by the team, included the robotic arms impinging on each other, which could damage the robot or prevent the surgeon achieving their aim. Respondents explained the surgeon only realises the robot arms are clashing when he or she is unable to manoeuvre the instruments as desired. On one occasion the robot arms nearly collided with a patient’s head; this problem was averted by the theatre nurse, who intervened.

The overarching strategy described by the surgeons to increase situational awareness was to establish good communication links between the surgeon and the team:

We’re pretty keen that they tell us if they’re concerned. Sometimes when the arm . . . because we dock between the legs, and often one of the joints of the arm is quite close to the thigh, and so often what I will do is say to the nurses, if they think it’s going to close then to let me know and make sure they’re aware that we want to know if there’s an issue like that. And they’re all pretty good at saying that, because we make sure that they’re not nervous about talking or, you know, we don’t have a silent theatre, they can chat and we let them come and talk to us and tell us if they think there’s an issue.

Site 3, surgeon

Good communication was seen as an essential part of RAS. Trust between the surgeon and the team was also emphasised, as the surgeon has to rely on the rest of the team to communicate information outside their field of vision to avoid complications. If the surgeon trusts their team to communicate problems to them, their concern over their reduced situational awareness is lessened. One surgeon commented that a more experienced team might be better able to communicate the necessary information to them.

Communication was also described by the OT teams as the main strategy to increase situational awareness; they saw it as their responsibility to act as the ‘surgeon’s eyes and ears’. In contrast to our tentative theory, the information they described communicating to the surgeon was less about the patient state and more often about the robot, as in the examples described above. Some noted that they just ‘tell the surgeon’ when there are problems and that everyone in their team knows to do this whether it is them or the robot that is struggling. Others described that, because of the physical separation of the surgeon from the team, it is important that team members have voices ‘strong enough for the surgeon to hear’. Good communication was seen as dependent on the relationship between the surgeon and the team. Although this is dependent on individual personalities and approaches, the strategies of training together as a team and having a dedicated team described above were considered by interviewees to increase team members’ confidence to speak up and to transcend existing professional hierarchical boundaries.

Lack of tactile information

One of the theories emerging from the review was that surgeons progress more slowly through a robot-assisted operation because they do not have tactile information to inform their assessment of the situation and to determine whether to persist with or revise their course of action, although this effect becomes less pronounced as their experience with RAS increases. Certainly in the interviews, several surgeons described initial experiences with the robot where, owing to the absence of tactile information, they had not realised how much force they were applying and consequently had, for example, snapped a suture. However, none of the surgeons considered the lack of tactile information a significant problem. Although a couple of the surgeons described being ‘a bit more careful’, ‘a bit more hesitant’, the surgeons we interviewed did not consider that the lack of tactile information led to an increased operation duration. They felt that they had adapted quickly to relying on visual cues, learning to look, instead of feel, for tension. Several surgeons related this to their experience of laparoscopic surgery; with laparoscopic surgery, they had already learnt to work with reduced tactile information. As one surgeon described:

I don’t find that a great problem because you know what . . . from previous experience you know what you’re looking for, so you know the tension that you’re putting on the tissues from what you can actually see. So actually not having the haptic feedback I don’t find that a great problem providing the instruments are in view.

Site 3, surgeon

Interviewees contrasted this with the experience of urology surgeons, who had moved straight from doing open prostatectomies to doing them with robot assistance.

Immersion

Another theory related to decision-making that emerged from the review was that the sense of immersion the robot provides means that the surgeon is more focused, resulting in improved decision-making and patient outcomes. The majority of surgeons we interviewed agreed that the robot produces a sense of immersion. One surgeon described how they can ‘lose themselves’ during the operation, and, referring to the level of concentration, he described this feeling as ‘quite intense’. Other surgeons commented that it is not that the robot creates a sense of immersion but just that they have to concentrate more because they have less experience with RAS than with laparoscopic surgery:

I think the intensity is, your level of concentration because it’s less familiar than say laparoscopic surgery, your level of concentration has to be at a higher level for a longer time, whereas you could probably relax on occasion doing laparoscopic surgery.

Site 2, surgeon

Two surgeons refuted the idea that the robot produces a sense of immersion, commenting that they are immersed in the procedure regardless of whether it is laparoscopic or robot assisted and that technology should not determine whether or not the surgeon is immersed.

A number of theories were suggested by the participants about the contexts in which a sense of immersion occurs. One surgeon anticipated that, although he already experiences a sense of immersion when using the robot, the feeling will probably increase when the ‘mundane’ and routine tasks related to using the robot (e.g. port positioning) have been mastered. In contrast, another surgeon commented that he feels immersed using the console, particularly during complex cases, but that this feeling would probably lessen over time (i.e. it was a feature of his limited experience with the robot). One surgeon described immersion as being dependent on who he has assisting; for example, if he trusts the assistant he can be immersed, as the assistant fulfils requests with ‘silver service’, whereas otherwise he is ‘constantly looking’ as there is anxiety about where the assistant is ‘pointing the instrument’.

Some surgeons described the OT as quiet during RAS, enhancing their concentration, and that there are no distractions. By comparison, in open and laparoscopic cases, the surgeon can chat with the team.

Perceptions of the impact of the sense of immersion varied. Some respondents commented that heightened concentration might lead to better decision-making, but how or why was not articulated. One surgeon described the sense of immersion as making him more focused, which should enable a more precise dissection. Others felt unable to comment on whether or not immersion would be reflected in patient outcomes. One surgeon said that he felt the sense of immersion would not impact his decision-making, except that he may persevere longer with an operation because he is less aware of time. However, this could cause concern if the patient is operated on for an ‘excess amount of time’.

Impact of ergonomics

The final theory that came from the review was that the ergonomic benefits of the console mean that the surgeon is less stressed and tired, resulting in better decision-making and reduced conversion to open surgery. Some surgeons discussed that, for them, performing operations using the robot is more stressful than laparoscopic surgery because they are in the early stages of implementation (i.e. they have not used the robot on many occasions). In this context, the surgeons stated that it was sharing the operation with a colleague that reduced their levels of stress, as opposed to the ergonomics of the robot. Other surgeons felt that the robot was an improvement on laparoscopic surgery (ergonomically); how and why it was an improvement were not fully explored, although one surgeon described that they were in a ‘less awkward position’. The surgeons also discussed that using the robot might be physically less tiring than laparoscopic surgery, but it is mentally more so because they have less experience of RAS than laparoscopic surgery. For this reason, they have a higher level of concentration for a longer time using the robot; in comparison, they could relax on occasion during laparoscopic cases.

Two surgeons described how the level of stress is affected by the how the team acts; as one surgeon described it:

I think it probably makes you physically less tired. I think you’re probably mentally more tired [. . .] We’ve all done less robotics than we have laparoscopic, so you’re carrying more of a burden I think robotically. And sometimes you feel like you’re only the person in theatre that knows what’s going on. [. . .] Because you’re there, and you’re the only one there looking.

Site 1, surgeon

The extent to which the robot reduces surgeons’ stress levels was also described as dependent on the stage of the procedure. For example, talking about suturing, one surgeon said:

If I was doing that laparoscopically, it would be a nightmare. It’s just a joy to do it robotically because of the ergonomics.

Site 2, surgeon

In contrast, this surgeon described how with dissection his ‘fear’ of bleeding is increased, because the magnified image means that he notices tiny blood vessels he would not notice otherwise. It was also noted that stress can be dependent on the type of operation performed; for example, a low anterior resection is stressful using both approaches, whereas operations that do not go down to the pelvic floor are less stressful and demanding. Being able to take breaks when using the robot was noted as a benefit.

The extent to which the ergonomics of the robot impacted on decision-making, particularly the decision to convert to open surgery, was difficult to ascertain. It was suggested by some respondents that if the surgeon is more comfortable during surgery he or she might persevere with a difficult operation rather than convert to open surgery. It was also noted that the surgeon can ‘take five minutes’ to consider their decisions during RAS, whereas they might feel more pressure in decision-making during laparoscopic surgery. However, it was also acknowledged that the decision to convert to open surgery is often due to circumstances outside the surgeon’s control; for example, conversion was described as a ‘technical’ matter not linked to ergonomics or how stressed the surgeon was. One surgeon stated that, if anything, he would persevere longer with laparoscopic surgery because that is the technique with which he has more experience and so feels more confident.

How and why the ergonomics of the robot reduced surgeons’ stress levels was also postulated by the wider team. Team members discussed a number of ergonomic benefits of the robot, for example because the surgeon is sitting down and that must mean they are more relaxed, the surgeon can adjust the console’s head piece, the console is padded, and it is easier to have coffee breaks as no scrubbing or descrubbing is required to step away from the console. However, it was also suggested that stress might be dependent on the surgeon’s experience (i.e. those learning how to use the robot do not seem as relaxed using the console). Participants also noted that if a surgeon found a stage of the procedure difficult, this would cause stress regardless of the ergonomics of the robot. The difference between mental and physical tiredness was also highlighted by the team; some described that the surgeon gets tired looking at the 3D image, or that RAS is stressful for their eyes and requires more mental concentration. It was also noted that the surgeon can be hunched in the same position for hours.

The role of the first assistant in robot-assisted and laparoscopic surgery

Actions first assistants were observed performing in the robot-assisted phases of operations included assisting with retraction, providing suction, inserting and removing swabs (to absorb blood), applying clips to a vessel, changing the robotic instruments and temporarily removing the robotic camera for cleaning. However, the extent of what the first assistant did varied according to their level of experience. For example, in one robot-assisted operation a FY2 took on the role of first assistant and with the guidance of an experienced scrub practitioner, she provided suction, inserted and removed swabs, and applied clips to ligate a vessel, but there were also periods during which she was not holding any instruments, and she did not assist with retraction.

In laparoscopic surgery, generally the first assistant would be holding the laparoscopic camera and a grasper, to assist with retraction. There would be no changes of instruments and the surgeon’s requests would relate to movement of the camera or grasper. The surgeon would be holding in their right hand an instrument for cutting, changing it for another instrument when needing suction or to apply a clip for example, and holding in their left hand a grasper for retraction. However, although this comparison initially suggests that the first assistant does less in a laparoscopic operation, in the laparoscopic operations and the laparoscopic phases of the robot-assisted operations, there was more variation in the extent of the first assistant’s involvement in the operation, depending on his or her level of experience. With more experienced first assistants who were surgical trainees, the operation took on a teaching element and the surgeon and first assistant would frequently swap roles; the first assistant would assist the surgeon for the more difficult parts and then for other parts the first assistant would operate, guided by the surgeon, while the surgeon controlled the laparoscopic camera and assisted with retraction. We observed a laparoscopic case in which two surgical trainees operated without a consultant surgeon present and the laparoscopic phase of a robot-assisted operation in which the two surgical trainees carried out the first part of the operation without the consultant surgeon present. At the other extreme, in several operations across two sites when an inexperienced trainee took on the role of first assistant, the first assistant held the camera but was directed by the surgeon, who would sometimes physically move the first assistant’s hand or move the camera himself and then get the first assistant to hold the camera again.

Does experience of colorectal surgery support co-ordination?

The actions that a first assistant is required or expected to undertake in the robot-assisted phase of an operation are largely tasks that, if they are experienced in laparoscopic surgery in terms of having had the opportunity to operate under the consultant surgeon’s guidance for parts of the operation, they will already be familiar with. The exception to this is changing the robotic instruments. On this basis, we may anticipate that it is experience of the particular procedure that is important, providing the first assistant with knowledge of the steps in the procedure and how to correctly complete them (e.g. when and how to apply clips to a vessel, when suction is likely to be required and how to provide it), enabling them to anticipate the surgeon’s requests and react to events without prompting.

In the laparoscopic operations and the laparoscopic phases of the robot-assisted operations that we observed, the difference in the ability of an experienced first assistant and an inexperienced first assistant to work without prompting was clear, particularly in relation to movement of the camera; whereas an inexperienced first assistant would need to be more explicitly guided by the surgeon, an experienced first assistant would need much less guidance, typically moving the camera without prompting. Similarly, in assisting with retraction, an experienced first assistant would adjust the position of their grasper without prompting. However, while actions were largely performed correctly and in a timely manner in the robot-assisted phases of operations, analysis of the field notes revealed few examples of unprompted action from the first assistant, except for the occasional unprompted provision of suction. It was only when we carefully reviewed the videos looking for such examples that the types of unprompted action performed by the more experienced first assistants became more apparent. These were similar to the unprompted action in the laparoscopic operations, involving small movements, such as using the suction instrument to assist with retraction, pushing back the tissue following the surgeon’s initial request for suction. Additionally, although we saw few examples of unprompted action, with the experienced first assistants there was a sense that they were doing more than just following instructions. For example, although the surgeon may ask the first assistant to provide suction, with an experienced first assistant what then follows is a series of movements by the first assistant as they co-ordinate their movement of the instrument with the movement of the surgeon’s instruments, so as to provide suction without obstructing the movement of the surgeon’s instruments. In contrast, with an inexperienced first assistant, the surgeon would direct the first assistant where to put the suction instrument and then guide them in the movement of that instrument.

Interestingly, in five of the eight video-recorded robot-assisted operations, we identified instances of first assistants offering assistance, for example:

Do you want me to let some of this gas out?

Yes please.

Site 6

Do you want some suction on that left side?

Yeah.

Site 6

Shall I suck that tonsil swab?

Site 7

The surgeon cuts and a pool of blood appears on the screen.

Do you want suction?

Yeah.

Site 7

Most of the offers of assistance related to providing suction or helping with retraction, the two tasks most commonly undertaken by the first assistant in the robot-assisted phase of an operation. We did not identify any instances of offers of assistance in the laparoscopic operations and the laparoscopic phases of robot-assisted operations but, when acting as a first assistant (as opposed to operating under the guidance of the surgeon), the first assistant’s responsibilities are consistent and clearly defined: to move the camera and to assist with retraction. It seems that the first assistant’s experience of colorectal surgery provides them with knowledge of what actions are likely to be required, and they make use of that by making an oral offer of assistance. However, the ability to make these offers of assistance did not appear to require significant experience; the first assistants in these instances varied substantially in their experience of colorectal surgery, with one first assistant being a FY2.

Making an offer of assistance also avoids the situation, observed on a couple of occasions in one operation, of providing assistance unprompted and having that assistance rejected:

As the surgeon cuts, a pool of blood becomes visible on the screen.

Oh lovely.

Suction. Sucker coming in.

I should be able to get it with this, I think [referring to tonsil swab he has].

Site 7

The trainee goes to assist with retraction but the surgeon says ‘I’ve got that, don’t worry’.

Site 7

Again, we did not identify any examples of assistance being rejected in this way in the laparoscopic operations and the laparoscopic phases of the robot-assisted operations, owing to the consistent nature of the tasks the first assistants were undertaking.

In relation to assisting with retraction, it seems that in the robot-assisted phase of an operation it can be challenging for the first assistant to know whether or not the surgeon needs assistance, because of the change in the division of labour RAS introduces, whereby, with a four-arm robot, the surgeon typically uses two of the arms for retraction (with the camera held by one arm and scissors held in the remaining instrument arm). This provides another motivation for offering, rather than providing unprompted, assistance. While some surgeons talked about the first assistants still being on the learning curve, so they were still learning what assistance to provide, it also seemed the surgeons were still learning what they could manage on their own and when they needed assistance, as revealed in the following instances:

Shall I retract up for you, Prof?

Yeah, maybe. Let me just see. I might be able to do something myself. It’s the beauty of having two left hands, yeah? Actually I think it would be helpful if you could just lift this bowel up here.

Site 1

Would you like better traction from me or are you happy?

I would but it seems like we’re making some decent progress so I just want to wait.

Site 6

Do you want me to hold that out the way?

Erm, should be OK [sounding uncertain]. Just trying to work out which plane I’m going to start with.

Site 7

The surgeon in the first example had experience of over 100 colorectal robot-assisted operations, making him one of the surgeons with most experience of RAS in our sample. Such examples suggest that it is only if and when the surgeons establish more routinised ways of working with the robot, in terms of what they undertake alone and where they require assistance, that the first assistant could be expected to have knowledge of what assistance the surgeon is likely to want.

To determine whether or not, despite limited unprompted actions, using an experienced first assistant led to the first assistant’s actions being completed in a timely manner, as hypothesised in the theory, we used the video data to analyse the time taken by first assistants to respond to requests to provide suction and to assist with retraction (see Appendix 11 for further details of the method). A total of 65 instances were identified: 32 requests for suction and 33 requests for retraction. The median time to respond was 11 seconds (range 3 seconds to 69 seconds). In all of the eight responses that took less than 7 seconds, the first assistant had pre-empted the request, either picking up the instrument, and sometimes starting to insert it but outside the field of view of the camera, or asking the scrub practitioner for the instrument. However, again, the ability to pre-empt the request in this way did not seem to require significant experience of colorectal surgery, with the first assistants in these instances varying substantially in their level of experience, one first assistant being a FY2. Although all first assistants showed engagement, in terms of monitoring the screen, there seemed to be an additional factor of keenness to demonstrate to the surgeon their ability to respond quickly.

Equally, the first assistants in the longer responses varied in their level of experience, and some of those first assistants who provided the very quick responses were also those who provided some of the slower responses. In the 11 responses that took more than 20 seconds, there were a variety of causes, including being slowed down by the need to first remove another instrument and the speed of the scrub practitioner’s response. Yet some of the causes did appear to be related to experience. For example, in the longest response, the scrub practitioner had to first show the first assistant how to use that particular grasper, as the first assistant had not used it before:

Do you have a fan retractor?

Yeah.

The scrub practitioner walks to tray [placed a metre or so away from the patient].

Ask him to, can you take the [. . .] suction off, and then put the fan retractor please.

Yeah.

You need to show him how it works, I don’t think he knows how it works.

The scrub practitioner shows the trainee how to open and close the fan retractor. The trainee then takes the retractor and inserts it.

Site 6

There were also several examples, among those responses that lasted 20 seconds or more, in which the first assistant was quick to move the instrument to the port but then struggled to move the instrument into the camera’s field of view. The ability to move an instrument into place could be expected to vary according to the level of experience, and this was a challenge experienced more by inexperienced first assistants, although instances of experienced first assistants undergoing this same challenge were also witnessed.

Does experience of robot-assisted surgery support co-ordination?

As described above, in all operations in which the first assistant had experience of RAS, they also had experience of colorectal surgery. Therefore, we are not able to assess whether or not experience of RAS without experience of the particular procedure is enough to support co-ordination between the surgeon and first assistant. Instead, we compared those cases in which the first assistant had experience of RAS with those in which the first assistant did not have that experience, in order to understand what experience of RAS contributes. The SURG-TLX questionnaire gathered data on participants’ experience of RAS and so this was used to help analyse the OTAS data to determine if the first assistant’s level of experience of RAS impacted the OTAS co-ordination score for the surgical subteam. A one-way independent ANOVA for just the robot-assisted operations, with first assistant’s experience as an independent variable, was performed. There was no significant effect of the first assistant’s experience on the surgical subteam co-ordination score [F(5,15) = 0.68; p > 0.05].

Using the qualitative data, we sought to explore further how and in what contexts the first assistant’s experience of RAS contributed to effective co-ordination. The tasks a first assistant may undertake in RAS that they do not undertake in laparoscopic surgery, and therefore where the impact of experience of RAS could be expected to be noticeable, are changing the robotic instruments and moving the robotic arms externally in order to move the instruments inside the patient, although these tasks did not happen in all of the robot-assisted operations we observed. Certainly, first assistants with experience of RAS were more likely to be asked to change the robotic instruments, whereas the scrub practitioner may be asked to do this if the first assistant does not have experience of RAS (discussed further below).

We might also anticipate that increased experience in RAS means that the first assistant better understands the challenges of communication and co-ordination in RAS and how to manage these challenges. Because in RAS the surgeon rather than the first assistant controls the camera, the first assistant is not able to position the camera in ways that will help them complete their actions. This appears to present a challenge for first assistants and, in response to this, across difference sites and with different first assistants, we observed first assistants asking the surgeon to move the camera:

The ODP, acting as first assistant, is very confident communicating with the surgeon e.g. ‘Move that camera, I can’t see’.

Site 1

The trainee says ‘Can you show me where I am?’ The surgeon draws back the robotic camera so the trainee can see her instruments on the screen.

Site 7

The surgeon tells the trainee to ‘go down a bit’ to find a swab left inside the abdomen . . . The trainee asks the surgeon to move the camera so he can see it.

Site 1

However, asking the surgeon to move the camera in this way does not seem to be the result of experience of RAS but rather just a response to the challenge experienced. For example, whereas two of the field notes extracts above involve first assistants experienced in both colorectal surgery and RAS, the third extract involves a first assistant experienced in colorectal surgery but not in RAS. Although it might be anticipated that first assistants who are more experienced in colorectal surgery would have the confidence to speak up because they would be sure about what they needed, whereas a less experienced first assistant may be hesitant, we also identified instances of first assistants with limited experience of both colorectal surgery and RAS asking the surgeon to move the camera.

What appears to be more significant in encouraging first assistants to change the way they communicate in a robot-assisted operation is how the surgeon encourages them to communicate and, related to this, their relationship with the surgeon. There was variation in this between sites. Surgeons at two of the sites repeatedly encouraged the first assistants to tell them what they were doing and to speak up if they were unsure about something. One surgeon, at site 7, said he did this partly to ensure that the first assistant felt comfortable speaking up when necessary. Another surgeon, at site 5, referring to all staff, said he deliberately engaged in general conversation before an operation to encourage people to talk to him, demonstrating awareness of potential problems arising from staff feeling unable to speak up.

Another strategy one surgeon would use to encourage the first assistant to communicate was to ask the first assistant how they were doing generally, that is, without reference to a particular task. During one operation, for example, there was a pause in the operation while a report was found and the surgeon asked the first assistant, ‘are you OK?’. The surgeon and the first assistant conversed a lot during this operation; afterwards, the surgeon said that the first assistant had limited experience of assisting in RAS but the two of them had worked together often and he perceived this familiarity as enabling the first assistant to feel confident in speaking up.

Does an experienced scrub practitioner support co-ordination?

Comparison of occasions on which there was an inexperienced first assistant also highlighted the role of the scrub practitioner in supporting an inexperienced first assistant. We observed a number of operations with first assistants at different stages of their training and without experience of RAS where, thanks to the scrub nurse’s guidance and assistance, the actions were performed correctly and safely, although not necessarily in a timely manner.

We also observed scrub practitioners assisting first assistants by changing robotic instruments, and sometimes the surgeon explicitly asked the scrub practitioner rather than the first assistant to do this. There was uncertainty among the theatre nurses and ODPs who we spoke to about whether they should be doing this when they had not undertaken training to be a first assistant, with putting instruments inside a patient being perceived as being outside the scope of their role. Despite this, when asked to do it, all scrub practitioners did.

With first assistants inexperienced in colorectal surgery, guidance did not relate only to those aspects that are specific to RAS. Such guidance mainly related to how to use particular laparoscopic instruments, such as how to apply a clip or the fan grasper described above. The scrub practitioner provides this guidance because the surgeon is not there at the patient’s side to be able to demonstrate, and is often asked by the surgeon to help the first assistant in this way, as illustrated by this field note extract:

11.10 a.m. The surgeon says to the scrub practitioner: ‘Have some hem-o-loks ready for me if necessary.’ The camera is dirty – the surgeon asks the first assistant to take it out. He guides her and tells the scrub practitioner to help her – the surgeon moves his head out of the console. They clean the camera and put it back in again but the surgeon says it’s still dirty, so they clean it again . . . The surgeon tells the first assistant to check with him anything she’s unsure about. The scrub practitioner explains to the first assistant about how the suction works. The surgeon asks the scrub practitioner to explain to the first assistant how the hem-o-lok works – she does . . . 11.55 a.m. The surgeon asks for a hem-o-lok for the first assistant to practice. The scrub practitioner gets a hem-o-lok and she and the circulating practitioner guide the first assistant on how to close it.

Site 7

By asking the scrub practitioner to support the first assistant in this way, the surgeon is acknowledging the expertise of the scrub practitioner. Discussing this with a surgeon and theatre nurse after an operation, they described the scrub practitioner as having a more important role and more responsibility during RAS than during laparoscopic surgery. They felt that this flattened the hierarchy between surgeons and the rest of the team.

The comparison of such instances suggests that the skill mix of those at the patient side needs to be considered, rather than just thinking about the experience of the first assistant. In all sites, thought was given to skill mix but there was variability in how this was put into practice. For example, in one operation in site 1, the first assistant was not scheduled to be assisting with that operation, but the surgical trainee who was scheduled to assist (and who had experience with RAS) had just been on a night shift so the surgeon sent him home. A theatre nurse who had experience of RAS was on the team, but it seemed the nurses had decided between themselves that the less experienced nurse would scrub while the other would circulate. Discussing this with clinical members of our team, it was felt that having the more experienced practitioner circulating would be a reasonable strategy if the circulating practitioner took on the role of guiding the scrub practitioner – the circulating practitioner is better placed to be able to take on a facilitative role, whereas the scrub practitioner cannot move outside the sterile field – but the extent to which the circulating practitioner guided the scrub practitioner in this particular operation varied.

The ability of the scrub practitioner to support an inexperienced assistant was dependent on the first assistant’s willingness to acknowledge the need for support and to accept such support. During one operation, the first assistant was experienced in colorectal surgery but it was only his fourth experience of assisting with a robot-assisted operation, and he said to the scrub practitioner:

You’ve done lots of these cases, haven’t you? This is only my fourth . . . I need to hang out with you. I think I’d learn a few things.

Site 6

In another operation, while the surgeon was docking the robot, the first assistant, who had not done a robot-assisted operation before, asked the scrub practitioner what she would be doing:

I’m not really sure what I’ll be doing.

Sorry?

I’m not really sure what I’ll be doing.

You’ll be er . . .

Retracting. I know that.

Retracting. Sucking, suction. Firing hem-o-loks onto vessels.

Nice.

Firing [staple] guns across the bowel potentially.

Cool.

Changing instruments.

She then asks the scrub practitioner to show her the robotic instruments.

Site 1

While the willingness to accept support varied between individuals, it seemed that the surgeon could also influence this willingness by, as described above, asking the scrub practitioner to support the first assistant, highlighting this aspect of the scrub practitioner’s role.

A revised theory

We sought to test the following theory:

Knowledge gained through experience of RAS and/or the particular procedure enables the first assistant to anticipate the surgeon’s requests and react to events without prompting, supporting co-ordination between the surgeon and first assistant and ensuring that the first assistant’s actions are performed correctly and in a timely manner. The success of this strategy may be dependent on their experience of working with the surgeon, as well as how engaged the first assistant is in the procedure (i.e. how carefully they monitor the screen).

On the basis of the analysis presented above, we can make a number of revisions to this theory, more clearly specifying the way in which experience contributes to a timely response and the type of experience that is important:

Knowledge gained through experience of the particular procedure enables the first assistant to work with less guidance, adjusting the position of the grasper and moving the suction instrument without prompting. Their experience also provides them with knowledge of what actions are likely to be required, which they make use of by making oral offers of assistance and/or preparing to act (e.g. by holding the instrument ready). Together, these behaviours support co-ordination between the surgeon and first assistant and increase the likelihood the first assistant’s actions are performed correctly and in a timely manner. The success of this strategy depends not only on engagement in the procedure but also, in preparing to act, the first assistant’s keenness to demonstrate to the surgeon their ability to respond quickly.

Whether or not experience enables first assistants to assist with retraction without prompting if and when surgeons develop more routinised ways of working with the robot remains to be established.

The analysis also suggests that experience of colorectal surgery is not sufficient to engage as a competent practitioner in RAS and, when the first assistant lacks experience of RAS, the experience of the scrub practitioner is an important resource. Thus, we need an additional theory:

Where the first assistant is inexperienced, in either RAS or the particular procedure, but the scrub practitioner is experienced in RAS, knowledge gained through experience enables the scrub practitioner to both guide the first assistant and undertake certain tasks (changing the robotic instruments) on their behalf, ensuring these actions are performed correctly. The success of this strategy depends on the first assistant’s willingness to acknowledge the need for support and to accept such support, which can be increased through the surgeon acknowledging this aspect of the scrub practitioner’s role.

Finally, RAS requires the first assistant to communicate more explicitly with the surgeon, a topic discussed further below (see Explicit communication), but the ability of the first assistant to do this depends not on their experience but on the encouragement of the surgeon. Thus, we add a second additional theory:

When the surgeon encourages the first assistant to communicate both actions and concerns, the first assistant will feel comfortable to speak up when necessary, supporting co-ordination between the first assistant and surgeon and helping to ensure the first assistant’s actions are performed correctly. The success of this strategy is likely to depend on the first assistant’s relationship with the surgeon.

In summary, we can state that experience of a particular procedure does offer benefits in relation to co-ordination between the first assistant and surgeon in RAS but that this alone is not enough. RAS requires different forms of communication, which the surgeon needs to encourage the first assistant to provide. With the surgeon unscrubbed and at a distance, the first assistant and scrub practitioner work more as a unit than in other forms of surgery, and the skill mix of this unit has important implications for seamless co-ordination and safe working practices in the OT.

More explicit pre-sequences

Whereas Mondada describes quite subtle visible preparation prior to a request, we noticed a great deal of vocal and verbal work that preceded the production of an explicit request. For instance, consider these examples:

.pt (pause) OK and then: (pause) (name) if you can take this now. (pause) just gras:p the artery.

Site 1

Alright, can you clean the camera for me?

Site 7

(Name), can you get a gras:per, and grab where I’m holding please?

Site 1

Notice that often there is vocal or verbal action work prior to the production of the request. In the first instance, there is a lip smack (‘.pt’), a micro-pause, an ‘OK and then’, another micro-pause and the use of the first assistant’s name before the request is produced. In the second instance, the word ‘alright’ is produced prior to the production of the request. In the third instance, the name is used prior to the request being delivered.

So, unlike the laparoscopic sequence format outlined by Mondada, the surgeon seems to do vocal work prior to the delivery of the request. We noticed repeatedly how during these sequences the first assistant (and often the scrub practitioner) would look to the screen or to the surgeon in anticipation of a request. They might have fewer resources to assess how to assist, but these actions were seen as flagging an upcoming request. In one case, the use of the first assistant’s name by the surgeon following a period of silence encouraged the scrub practitioner to stop her conversation with the circulating practitioner and reach out for an instrument, ready to support the first assistant with the request.

It was noticeable that the surgeons’ requests in robot-assisted phases of the operations often included the use of the recipient’s name. In the laparoscopic phases, the surgeon has a range of resources to identify the recipient (volume, gaze orientation, etc.) and they are able to observe whether or not all members of the team are focused on the surgical site. Obviously, such practices for indicating who an instruction is aimed at do not work in the robot-assisted phase of an operation and surgeons are less able to assess what colleagues are doing. The strategy of using the recipient’s name was used even when it was clear from the nature of the request who it was aimed at; so, the use of the recipient’s name may not be necessary to clarify who the intended recipient is but acts as a device to alert the intended recipient’s attention. Indeed, often the surgeon would switch between questions about anatomy (directed to surgical trainees or others in the room) and requests for assistance (directed to the first assistant): these would routinely be marked by the use of a name.

Sometimes, surgeons also used the intended recipient’s name when their initial question did not receive a response, as in these examples:

Is that the artery there? (pause) (Name)?

Don’t know.

Oh, that’s not very helpful. Have a wild guess.

Site 1

Is there anything in that? (pause) Do you think? (pause) What do you think, (Name)?

Site 1

Although taken out of context, these field notes extracts may be interpreted as suggesting that the first assistant did not know or was not sure how to answer the surgeon’s questions; however, this was in fact an experienced first assistant. The first assistant did appear engaged in terms of watching the screen but it seemed it was difficult for the surgeon to gain the first assistant’s attention after there had been a period without communication.

Multipart requests

Whereas Mondada’s account of requests in laparoscopic surgery reveals their terse character (often simply the name of an instrument or a simple deictic), within our corpus from RAS we found that these requests often had multiple components. Aside from the pre-request work discussed above, the requests themselves often had at least two distinct phases. For instance, consider this case from site 1:

(o)k(ay)- (0.5) .hhhhhh (1.3) now you can take it here again er: (.) nhhhh (Name)

(3.2)

.pt (0.5) and kind of- (1.5) pull it to the left.

(.)

Yeah. (0.3) that’s it.

Site 1

These requests would often follow long sequences of little participation from the first assistant. The first assistant might be waiting to assist or they might simply be holding back some part of the internal anatomy. However, when the request comes, they often have to reposition themselves, insert an instrument or switch to a new instrument. The surgeon’s requests are routinely built to allow for that. So, the first part of the request is routinely followed by more specific instructions. In this case the request to ‘take it’ is followed by a 3.2-second pause while the assistant inserts an instrument. As the instrument appears on screen, the request continues (‘and kind of-’) and as it nears the bowel the requests specifies further (‘pull it to the left’).

This reveals the ways in which requests are designed with regard to the very different nature of tasks in the OT. The surgeon has much greater control over the number of instruments and, therefore, can conduct much more of the work in the surgical site him- or herself. When the first assistant is called on, the requests are designed to provide progressively more specific information as they move closer to a position to help.

Oral responses to requests

While our theory discussed in Chapter 5 concerns explicit communication by the surgeon, in looking at these request sequences we also identified more explicit communication practices by other members of the team than occurred in laparoscopic surgery.

Mondada notes that surgeon-initiated requests for immediate action are responded to non-vocally: the action is done and the surgeon can see the first assistant and the scrub practitioner doing it (or moving to do it). 86 In many of our cases, we saw requests carried out with no verbal component. However, in RAS it is necessary to make a distinction between requests that can be monitored by the surgeon through action on screen and requests that have a significant off-screen component that is invisible to the surgeon. The latter requests require verbal acknowledgement or the surgeon cannot tell if the request is being actioned. Consider the following examples from site 1:

(Name), can you get a gras:per, and grab where I’m holding please?

(0.5)

Yeah.

So you can take it here?

(6.5)

(sorry . . . hhhhh (1.0) m’okay.)

In the first case, the first assistant needs to remove his suction tube, take a grasper from the scrub practitioner and insert it before he can complete the request. The ‘yeah’ seems to mark that that work will be done; the surgeon, after all, will not be able to see much of these preparatory actions. In the second case, we see a request for the first assistant to grasp the bowel. The first assistant starts to do this, but struggles with his instrument. The ‘sorry . . . hhhhh (1.0) m’okay’ seems to make available his struggles with his instrument and account for the delay: these struggles are not visible to the surgeon. Thus, at times the first assistant (or, indeed, the scrub practitioner) will use the vocal channel to compensate for the limited visual channel open to the surgeon.

This expectation of verbal acknowledgement at particular times can be seen in this case:

Can I >have an< hem-o-lok, please?

(3.0)

Where is it?

(0.5)

It’s coming.

The surgeon has asked for a new robotic instrument, but after a pause of 3 seconds he asks ‘where is it?’. Interestingly, the scrub practitioner turned to get the instrument on the word ‘hem-o-lok’ and indeed had one loaded with a clip and ready to be inserted. So, in many ways she was fully prepared and acting swiftly. However, none of this embodied work was visible to the surgeon. Indeed, he could not have expected the scrub practitioner to complete the task within 3 seconds. However, he does seem to pursue a response. It is also interesting that, following the pursuit of a response, it is the first assistant and not the scrub practitioner who notes ‘it’s coming’, again showing how they work as a team, a very different working relationship from that in laparoscopic surgery, where each works more directly with the surgeon rather than with each other. Discussing this instance with clinical members of our team, the nature of the task and, associated with it, the surgeon’s sense of time pressure – the surgeon has the vessel displayed and so will want to clip it as quickly as possible – was highlighted as a relevant contextual factor in his desire for a response.

More generally, practices of making the surgeon aware of what is happening were most common in sites 6 and 7, where the surgeons encouraged this from the team. When this does not happen, it can cause problems, in terms of both co-ordination and safety. For example, in one operation, the scrub practitioner was asked to change one of the robotic instruments. She did not tell the surgeon when she was about to reinsert the robotic instrument and, at the time, the surgeon had his head outside the console, so that the instrument was not inserted under the surgeon’s vision. The role of oral responses in supporting the surgeon’s situational awareness is discussed further in Chapter 8.

When explicit communication did not work

In line with our theory, the engagement of team members affected the success of the strategy of using more explicit communication. For example, on one occasion, although the scrub practitioner was generally engaged throughout the operation, at the point that a surgeon made a request for the first assistant to help with retraction, the scrub practitioner was talking to the circulating practitioner and so did not hear the request, leading the first assistant to have to ask the scrub practitioner for a grasper. Because of this, there was a 20-second lag between the surgeon’s request and the grasper being visible on the screen. In the laparoscopic operations and laparoscopic phases we observed, a failure to respond to a request at all only occurred when the request was aimed at the wider team and not to anyone in particular (e.g. a request to turn the lights off).

Even when the intended recipient was engaged, the strategies of explicit communication did not always succeed in supporting co-ordination. Many team members felt that the sound quality of the microphone and speakers made it difficult to hear and understand the surgeon, and this seemed to be affected by the speed or volume with which the surgeon spoke. The surgeon at site 5 chose not to talk through the robot microphone at all and took his head out of the console every time he needed to talk to anyone. We observed many instances in which the first assistant or scrub practitioner had to ask the surgeon to repeat his request several times. Although not unique to RAS, this occurred more frequently in the robot-assisted phases of operations.

Examples of requests that were not explicit

We have been discussing a tendency towards request sequences that are more explicit than those outlined in previous research on laparoscopic surgery. However, this is a tendency or, in realist terms, a demi-regularity. When actions are requested in a series – suction over a period of a minute or so, or a number of clips to be applied to a vein – the first assistant is more able to anticipate what might be required next. The action is available on the shared screen, and the tasks are closely coupled, and thus the practices for co-ordination are much more subtle and delicate.

Another way in which surgeons’ requests were less explicit was in communication with the scrub practitioner. When the surgeon makes a request of the first assistant, such as helping with retraction or providing suction, the scrub practitioner hears this and immediately responds by providing the necessary instrument if the first assistant does not already have it (or, if necessary, asking the circulating practitioner to get it), so that the scrub practitioners across all sites and all operations largely worked without direct prompting. However, exceptions to this were witnessed when the first assistant was inexperienced, so that the surgeon would direct his or her instructions to the scrub practitioner instead, as in the example above when the surgeon asked the scrub practitioner if she had a fan retractor and then asked her to show the first assistant how it worked.

A revised theory

We sought to test the following theory:

By using explicit communication, the team can hear and understand the message, which enables them to act accordingly and complete requests accurately and in a timely manner, supporting effective co-ordination. This strategy will work if the team are engaged and awaiting the surgeon’s requests.

Our analysis revealed some systematic differences in the organisation of surgeon-initiated requests for immediate action during the robot-assisted phase of the operation and, thus, we have unpacked various ways in which communication while the surgeon was working in the console can be seen as more ‘explicit’, as indicated in the theory. In contrast to Mondada’s request format from laparoscopic operations, in robot-assisted phases of surgery:

• There is a more explicit ‘preparation’ or ‘pre-request’ work.

• Requests are routinely less terse and are built over a series of utterances.

• Recipients verbally acknowledge requests for which off-screen work is required before the action is visible on screen. When such acknowledgements are absent, the surgeon may pursue a response.

These communicative practices attend to (and reveal) distinctive qualities of RAS:

• There are more tasks under the surgeon’s control, with less involvement from others and longer periods in which they are waiting to assist.

• The roles of the first assistant and scrub practitioner have changed in RAS, with requests now often involving both parties in their successful accomplishment.

• There are asymmetries in the visibility of conduct – action off-screen is no longer visible to the surgeon.

These differences in the division of labour have implications for the surgeon’s situational awareness, and OT teams are increasingly attentive to these implications in and through their communicative practices, a topic we will return to in Chapter 8.

On the basis of this analysis and the clarification of the ways in which communication is more explicit, it seems that explicit communication is not a single strategy. Thus, we revise our theory into several theories, each addressing the different ways in which the strategies support co-ordination and the contexts in which they do so:

Through oral preparation and/or pre-request work by the surgeon, particularly the use of the intended recipient’s name, the attention of the first assistant and scrub practitioner is secured, enabling them to hear the message, supporting effective co-ordination. The success of this strategy in securing attention depends on the engagement of the first assistant and scrub practitioner.

By using longer oral requests, built over a series of utterances, the surgeon provides the first assistant with time to take any necessary preparation (e.g. reposition themselves, insert an instrument, change an instrument), so they are ready for the more specific instructions that follow. This supports them to complete requests accurately, supporting effective co-ordination. This strategy is more likely to be successful if the first assistant is engaged or the surgeon has already secured the first assistant’s attention. However, the ability of the first assistant to understand the request may be negatively impacted by issues of sound quality and the speed and volume of the surgeon’s voice.

Oral responses to the surgeon’s requests confirm that work will be done and make apparent any challenges in completing the request, supporting effective co-ordination. Use of this strategy by the first assistant and scrub practitioner is more likely when the surgeon encourages the first assistant and scrub practitioner to communicate their actions and its success is dependent on the volume and clarity with which the information is communicated.

Does robot-assisted surgery lead to less tiredness for the surgeon?

Beyond the self-reports provided via the SURG-TLX, it was difficult from the observations and video data of the robot-assisted phases of operations to assess the surgeons’ levels of tiredness, particularly given their position within the console. However, noticeable when observing the laparoscopic operations and the laparoscopic phases of robot-assisted operations was that the surgeons do not exhibit many obvious signs of physical fatigue. They stand squarely and on two feet, and operate for hours without a break or a drink. During the laparoscopic phase of one operation, a surgeon instructed a relatively inexperienced trainee in how to place his arms to minimise physical fatigue:

Put your elbows in next to your sides or you will get too tired . . . use both hands so you don’t have to bend.

Site 7

This indicates that the surgeon is aware of strategies to overcome fatigue. The first assistants and scrub practitioners, however, often show signs of physical fatigue in laparoscopic operations and the laparoscopic phases of operations, such as moving their weight from one foot to the other and bending and stretching their backs. The first assistants also got into more awkward positions than the surgeons. When we interviewed the surgeons about this, they agreed that they had learned to work in ways that were less tiring. This suggests that the ergonomic benefit of being able to sit down that RAS provides offers less benefit to experienced surgeons, in terms of levels of tiredness, than might be initially expected, because they have already learnt to manage this challenge of laparoscopic surgery. However, in the interviews, the surgeons also pointed to the fact that the spatial configuration in laparoscopic surgery is designed to support the consultant surgeon, which is another reason why first assistants might find themselves in more awkward positions:

The primary surgeon usually has the best position because he needs to have it. So it’s often that the assistants around this person have to adapt because obviously the person who is operating needs to have the best kind of comfortable position, and also usually the position looking straight forward onto the operating field, whereas the assistants are all coming from the side.

Site 1, surgeon

Although it did not make a significant difference, the surgeons still perceived that working at the console was more comfortable and could be less tiring:

You know, once you’re used to laparoscopic surgery, once you’re used to standing and operating, which is what we are used to, then to do it sitting down and then operating yes it helps, but I don’t think that it is a big enough . . . it is difficult to quantify but it is certainly advantage is what I would say.

Site 6, surgeon

It’s not just the fact of sitting down, but also the fact that you can relax in between you can . . . because you’re not scrubbed you can walk away and have a drink and all these kind of things, look at your phone [laughs].

Site 1, surgeon

Stress can also impact on the perceived physical and mental demands of an activity, which we consider in the following section.

Does robot-assisted surgery lead to less stress for the surgeon?

The analysis of the field notes and video data indicated a number of factors that could affect the surgeon’s level of stress, which could be both positively and negatively impacted by the use of the robot, and in some cases strategies were established for reducing the negative impact. For example, time pressures can be a source of stress. ROLARR found that the mean operative time (defined as time from first incision to time of wound closure completion) for robot-assisted operations was 4 hours and 59 minutes, while for laparoscopic operations it was 4 hours and 21 minutes. 194 The mean theatre time (defined as time between the patient first entering the anaesthetic room and the patient leaving theatre) for robot-assisted operations was 6 hours and 31 minutes, while for laparoscopic operations it was 5 hours and 53 minutes. Similarly, we found that the mean operative time (using the same definitions as ROLARR) for the robot-assisted operations we observed (6 hours and 32 minutes, range 4 hours and 55 minutes to 8 hours and 50 minutes) was longer than for the laparoscopic operations (5 hours and 29 minutes, range 3 hours and 42 minutes to 6 hours and 25 minutes). OT teams in our sites recognised this fact and, therefore, if at all possible, would put a single robot-assisted case on a list, to reduce the time pressure. If more than one operation was scheduled, the robot-assisted case would ideally be placed at the beginning so that the set-up could be completed in good time. In many other cases the initial lack of a bed caused delays to the start of robot-assisted operations. This sometimes appeared to cause stress to the team, and to the surgeons, as they were left in a state of uncertainty. Some team members became annoyed about the possibility of the list over-running. Although such delays occurred for both laparoscopic and robot-assisted operations, there was particular concern when it happened for a robot-assisted operation because of the longer duration of robot-assisted operations. It was not clear how the team’s feelings about this situation affected the stress levels of the surgeon during the operation. Some concerns were vocalised in the surgeon’s hearing, however, and may have had an adverse effect on some occasions. In one operation at site 7, the robot was required by another surgeon at a specified time, and this may also have caused some stress.

Unfamiliar equipment can also be a source of stress, and some of the surgeons were relatively inexperienced with the robot. The stereo viewer and the touchpad in the armrest can display more than 300 different messages and icons for the surgeon to interpret and act on. During several operations, the surgeon did not know what the messages meant. In one operation at site 6, for example, a message came up on the screen saying ‘move master grip to match the instruments’ but the surgeon did not know what this meant, first asking the scrub practitioner what it meant and then telling the first assistant to ‘make a note of that and I will ask [the representative]’. However, even for more experienced surgeons, the robot could be a source of confusion. During an operation at site 1, a message appeared on the screen that the surgeon did not seem to understand, and two yellow warning lights appeared on the arm. He remarked ‘it has a life of its own’, suggesting that he was not sure what was going on at that particular moment, despite having experience of over 100 robot-assisted operations.

A frequent problem concerning robot-assisted equipment was the arms clashing, and this happened at least once in the majority of the operations we observed. Ideally, the ports should be positioned in such a way as to prevent this problem but that was not always possible and clashes often occurred. When this happens the surgeon cannot move his instruments. If this happens at a crucial point it could be dangerous and so may cause the surgeon stress.

The complexity of a procedure and patient factors will also affect the level of stress, and the robot can interact with this in multiple ways. For example, if the case is complex, a surgeon may choose to use the robot to take advantage of the precise and dexterous EndoWrist instruments and enhanced vision. In one operation, the tumour was low in the rectum and the surgeon warned that an abdominoperineal resection may have to be performed. The surgeon remarked to the researchers that ‘things would be tense’ in the OT because of this, indicating that he may have been feeling some stress. Another surgeon observing told the researchers that the tumour was ‘as low as a rectal tumour can get’ and that an anterior resection might be easier to achieve using the robot. However, patient factors can compound problems related to lack of confidence with RAS. In one operation at site 7, the patient was unusually tall. Despite a relatively high level of experience with RAS, having completed over 100 robot-assisted colorectal operations by the end of the period of data collection, the surgeon admitted having difficulty remembering where to place the ports for the robotic instruments and always used a leaflet supplied by the manufacturer to guide him. In this case the length of the patient’s abdomen was longer than normal and the usual positioning of the ports did not work as well. This may have caused stress during the operation, as the surgeon remarked that the instruments could not be placed in ‘ideal positions’ during the surgery. Similarly, in an operation at site 1, the patient had an unusually narrow abdomen, which meant that, despite the surgeon’s experience with RAS, it was difficult to avoid the robot arms clashing.

Issues related to teamwork could also be a source of stress and, as described in the previous chapter, RAS introduces a number of challenges for teamwork and communication. For example, having to repeat requests can be a source of frustration for the surgeon. Whereas tiredness was difficult to judge, frustration was often easy to detect in the surgeon’s voice and, for those operations we video recorded, the microphone the surgeon wore allowed us to capture their deep sighs and cursing. Related to this is the experience of the team, and it seemed that working with inexperienced staff can be stressful for the surgeon. This was apparent during an operation at site 6, where the first assistant was inexperienced and the surgeon’s words and tone of voice revealed that he was stressed on several occasions. At one point, for example, when the first assistant was failing to adequately help the surgeon with retraction, the surgeon said, ‘stop poking around there . . . If you can’t see, I can’t see’. Later on, when there was a possibility that the first assistant would do harm to the patient, the surgeon said loudly, ‘hold this – don’t pull until I have let go of it or it will tear. Jesus Christ man, don’t scrape the sacrum or you will cause a sacral bleed!’. In an operation at site 1, both the first assistant and the scrub nurse were inexperienced with the robot and problems occurred, which appeared to cause stress for the surgeon.

The surgeon’s position in the console could have both positive and negative impacts on their level of stress. It has the potential to aid the focus of the surgeon, decreasing the surgeon’s stress by reducing distractions, although the extent to which surgeons stayed in the console varied considerably, again influenced by the experience of the team (discussed further below). The fact that the surgeon was obliged to have their head in the console could also be a cause of stress if they did not trust the team to let them know what was happening.

Noise within the OT can also be a cause of stress, particularly in RAS. In part, it is because of the speakers within the console, which increase the surgeon’s awareness of noise within the OT. As a result, whereas music was typically played in laparoscopic operations, music was not played in any of the robot-assisted phases we observed. Additionally, the difficulties in communication between the surgeon and the scrub practitioner and first assistant could be exacerbated if there was additional talking by circulating, anaesthetic or other staff. Surgeons at all sites recognised this problem. During an operation at site 7, for example, the surgeon asked the staff to be quiet at one point so he could hear what the first assistant was saying. In another example, during an operation at site 6, the anaesthetist was speaking loudly on his mobile phone. This obviously distracted the surgeon at the console, and he politely asked the anaesthetist to leave the OT. The surgeon at site 5 agreed that if staff talked too loudly it could be difficult for the first assistant to hear the surgeon on the console. One of the surgeons at site 1 explained that, if extra staff came into the OT and wanted to chat before he had finished his case, he asked the scrub practitioner to send them out as he found it distracting.

Decisions related to the implementation of RAS could also impact on the level of stress in a robot-assisted operation. For example, there was general agreement among the staff we observed that a large enough OT was essential to the smooth running of a robot-assisted procedure. The surgeon at site 7 felt very strongly about this matter, as it could compromise patient safety:

This theatre is too small for a robotic case.

Just chill.

I can’t just chill, I want to do a good operation and the circumstances are not good. It won’t be your name that goes on the list, it will be mine.

The surgeon at site 6 was also unhappy if a large OT was not available for his robot-assisted cases and after one operation said the operation was difficult to perform because of the small OT. At this site a number of large OTs were available but were sought after by several surgeons. At site 7 a large, dedicated robotic OT had been established, while at site 5 the OTs were not very large but there was no choice about using them.

The consideration of the above issues suggests the robot does not, necessarily, lead to less stress for the surgeon. If the surgeon was able to use the robot with an experienced assistant and an experienced team, and in a large enough, quiet OT, the potential stress-reducing benefits of the system (good vision, focus, control of the camera and precise and ergonomic instruments) might be realised.

A revised theory

We started with the following theory:

When a surgeon is experienced in undertaking RAS, the ergonomic benefits of RAS reduce the surgeon’s levels of stress and tiredness, enabling them to persist longer with the operation and thereby reducing the rate of conversion to open surgery.

We sought to test the first part of this theory, that for surgeons experienced in RAS, the ergonomic benefits of RAS result in lower levels of stress and tiredness than in laparoscopic surgery. Our analysis suggests that, although surgeons find working at the console more comfortable, this does not significantly impact their level of tiredness because they have developed strategies to manage the physical demands of laparoscopic surgery. A number of contextual factors can affect a surgeon’s stress levels in an operation, and the resources introduced and taken away by RAS interact with these factors in both positive and negative ways. The most significant of these factors appeared to be the experience of the team, particularly the first assistant and the scrub practitioner, tying in with the theories ‘experienced assistant’ and ‘scrub practitioner support’ presented in Chapter 7. Therefore, it is appropriate to revise these theories to include reduced levels of stress for the surgeon as an additional impact of these strategies:

Knowledge gained through experience of the particular procedure enables the first assistant to work with less guidance, adjusting the position of the grasper and moving the suction instrument without prompting. Their experience also provides them with knowledge of what actions are likely to be required, which they make use of by making oral offers of assistance and/or preparing to act (e.g. by holding the instrument ready). Together, these behaviours support coordination between the surgeon and first assistant and increase the likelihood that the first assistant’s actions are performed correctly and in a timely manner. This in turn reduces the surgeon’s level of stress. The success of this strategy depends not only on engagement in the procedure but also, in preparing to act, the first assistant’s keenness to demonstrate to the surgeon their ability to respond quickly.

Where the first assistant is inexperienced, in either RAS or the particular procedure, but the scrub practitioner is experienced in RAS, knowledge gained through experience enables the scrub practitioner to both guide the first assistant and undertake certain tasks (changing the robotic instruments) on their behalf, ensuring these actions are performed correctly and reducing the surgeon’s level of stress. The success of this strategy depends on the first assistant’s willingness to acknowledge the need for support and to accept such support, which can be increased through the surgeon acknowledging this aspect of the scrub practitioner’s role.

Does robot-assisted surgery reduce the surgeon’s situational awareness?

The mean surgical subteam intraoperative monitoring score across all operations observed was 3.68 (range 2–5, SD 0.72), suggesting that, in most cases, the situational awareness of the surgeon and first assistant either moderately enhanced team functioning or had neither a positive nor negative impact. To make an initial assessment of the impact of RAS on the surgeon’s situational awareness, the correlation between the percentage of the operation that was robot-assisted and the OTAS intraoperative monitoring score for the surgical subteam was calculated using the Pearson product-moment correlation. There was no significant relationship between the percentage of the operation that was robot-assisted and the intraoperative monitoring score for the surgical subteam [r = 0.21; p > 0.05 (two-tailed)].

This was reflected in our observations, where we found that, in general, the surgeons did seem to be aware of what was going on within the OT, despite their position in the console. This was demonstrated, for example, by joining in with discussions with the rest of the OT team. However, this did vary over the course of the operation, with the surgeon becoming more focused on what was happening in the surgical site and less likely to join in conversations or to communicate with the rest of team as the operation progressed. For example, on one occasion there were two surgical trainees assisting with the operation, one observing and the other acting as first assistant. At one point in the operation, the trainees swapped roles and did not inform the surgeon. He became aware of this change only when he directed a question at one of the surgical trainees and was told that the other surgical trainee was now acting as first assistant. Although such examples serve to illustrate the way in which the surgeon’s awareness of the broader OT varies over the course of the operation, we did not observe any problems resulting from the surgeon lacking awareness of what was happening outside the console.

It is also important to note that, while the theory assumes that the surgeon’s situational awareness is reduced because of their position in the console, there was significant variation in the extent to which the surgeon stayed in the console. For example, in one totally robot-assisted operation, the surgeon brought his head out of the console only twice, both times when a member of staff approached him to ask him a question. When the camera was being cleaned, meaning that surgeon would not be able to see or do anything, he kept his head inside the console, appearing to use this as an opportunity to rest. Resonating with our theory, the extent to which the surgeon came out of the console appeared to vary according to the experience of the first assistant, suggesting that when the surgeon trusts the first assistant to carry out the requests successfully they feel more comfortable remaining in the console. Related to this may be the ability of a more experienced first assistant to tell the surgeon what is happening, as suggested by this quotation from a surgeon:

If there is anything which doesn’t look right, I just need to see what is happening, and that is only possible by taking your eyes off the console and seeing what is happening. [. . .] If you have an experienced assistant then at least okay you don’t need to take your eyes off but at least you can ask them, you can talk to them and they will communicate things back to you.

Site 6, surgeon

The extent of the trust a surgeon has in the team that are supporting him or her can change over the course of an operation. For example, following an incident involving a first assistant and scrub practitioner with limited experience of RAS where a robotic instrument was incorrectly inserted, the surgeon then regularly brought his head out of the console to see what the first assistant and scrub practitioner were doing, and on numerous occasions left the console in order to help them. This also suggests that the level of distraction experienced by the surgeon will vary according to the experience of the first assistant and the scrub practitioner.

There was variation across the sites in terms of where the robotic console was positioned, which impacted the ability of the surgeon to be able to observe what the first assistant and the scrub practitioner were doing. In site 1, the console was placed parallel to the patient and the team, enabling the surgeon to see the first assistant and scrub practitioner simply by moving his head out of the console. In the other sites, the console was placed further away, making it harder for the surgeon to see the first assistant and the scrub practitioner without leaving the console.

We also observed instances in which, drawing on cues provided through the robotic console, the surgeon’s situational awareness appeared to be better than the team’s. Some support for this is provided by the OTAS data, in which the mean surgical subteam intraoperative monitoring score in robot-assisted operations was 3.69 (range 2–5, SD 0.7), slightly higher than the mean nursing subteam intraoperative monitoring score of 3.62 (range 2–5, SD 0.81). For example, the surgeon might experience difficulties in moving the robotic instruments and ask the team if the robotic arms are clashing, if the team has not already noticed it and alerted the surgeon. There were multiple instances in which the surgeon asked if the gas that was inflating the patient’s abdomen was leaking, leading the team to check and correct the situation. OT staff we spoke to thought the surgeon might be more aware of that because of the magnified 3D image that the robot provides, but the surgeons understood the situation differently. Potentially, RAS makes it harder to perceive this, because they can only draw on cues from within the surgical site:

I think laparoscopic you see it easier because you are standing there, you know. You’re standing there, you know that the tube is kinked or the anaesthetic muscle relaxant has worn off so the patient is coughing, so that it is pressing, or the suction is being operated, so you probably pick it up earlier with laparoscopic surgery.

Site 6, surgeon

The reason that the surgeons notice the gas is leaking is because of their role, because it affects the ease with which they can operate:

I mean the surgeon is acutely aware that there’s a loss of volume and that everything gets more difficult, so the surgeon is certainly aware that things are becoming difficult and the reason is there’s no gas. But he’s always going to be the first one who is aware of it because he gets the feedback that things have suddenly become more difficult.

Site 1, surgeon

This again points to the implications of a changed division of labour; because the first assistant is doing less within the surgical site, they may be less likely to notice such changes.

Where the surgeons did seem to lack situational awareness related to the positioning of instruments within the surgical site, in terms of both instruments held by the robotic arms and laparoscopic instruments inserted by the first assistant. There were a number of occasions when surgeons would say that they had ‘lost’ an instrument. Sometimes this was in reference to the robotic arms not moving in the way the surgeon wanted, usually because the robotic arms were clashing. More often, such statements were made when the surgeon had positioned one instrument to assist with retraction, moved the camera away, and then was struggling to find it again. This was not a significant problem because if they pull back the camera they can find the instrument, but it was an issue that appeared to disrupt the flow of the operation, and there are safety risks if the surgeon tries to move the instrument when it is not under vision.

This problem did not appear to be due to lack of experience, as it was a challenge encountered both by surgeons experienced in RAS and by those who had less experience. That this problem occurred frequently led us to consider whether the changed division of labour, with the surgeon now controlling the camera and up to three instruments, increased the cognitive load for the surgeon. Some support for this was provided by the analysis of the SURG-TLX mental demand data presented above. However, the surgeons did not perceive the situation in that way, emphasising that the cognitive demands of monitoring and co-ordinating with the first assistant were reduced:

I don’t think it impacts my cognitive load because when I’m operating laparoscopic, if I have anyone junior to middle grade inclusive, junior to middle grade I’m forever having to correct them on their assistant positioning and camera holding. So actually it’s allowed me to concentrate more on what I’m operating on.

Site 7, surgeon

Another common occurrence relating to the positioning of instruments within the surgical site, and one that appeared to cause significant frustration for both the surgeon and the first assistant, was when the surgeon asked the first assistant to insert an instrument and the first assistant did it, but the surgeon did not realise. This results in delays to the procedure as the surgeon has to ask what is happening. We did not witness such instances in the laparoscopic operations or in the laparoscopic phases of the robot-assisted operations that we observed. In a laparoscopic operation, the surgeon is standing by the first assistant and, even if they cannot see the instrument on the screen (the intracorporeal view), they can observe that the first assistant has inserted the instrument (the extracorporeal actions) and would be able to determine from how the assistant is holding the instrument where the end of that instrument is. But in RAS, the surgeon does not have these cues.

As described in Chapter 7, some surgeons were aware of this problem and overcame it by encouraging their first assistants to tell them what they were doing as they were doing it, leading to the use of more explicit communication than observed in the laparoscopic phases. For example, in one operation, the two first assistants were both inexperienced and the surgeon frequently asked them to tell him what they were doing. Waiting for one of them to insert an instrument, he said, ‘tell me when you are coming in’. However, this surgeon was also observed asking similar questions of very experienced assistants. In a postoperation interview this surgeon explained that he insisted on this frequent communication because it reduced his stress and frustration, and said that ‘a second can seem like a minute’ when he is at the console. He also said that he wanted to make sure the assistants were happy to speak to him.

This problem was also prevented by the first assistant asking the surgeon to move the camera so that they could see the instrument being inserted, so whether such a breakdown occurs will be affected by whether or not the first assistant is willing to direct the surgeon’s movement of the camera.

Related to this, while surgeons may have lacked confidence in the ability of an inexperienced first assistant to tell them what was happening with the patient, as described above, they did use the strategy of asking questions to maintain their awareness of what the first assistant was doing:

OK, are you putting traction with your left hand?

Yes, the left side.

OK, are you sucking, yeah?

Site 6

So where’s your grasper?

Site 6

And while you’re in there, suck, OK?

I am, yep, I’m sucking. Constantly.

Are you sucking?

Yep.

Site 7

That this was only observed when working with an inexperienced first assistant suggests the issue is not only that the surgeon cannot see what the first assistant is doing but also that they do not necessarily have confidence that the first assistant is completing their requests correctly.

Does the team contribute to the surgeon’s situational awareness?

In line with our theory, we observed several instances of members of the team, most often the scrub practitioner, alerting the surgeon to the robotic arms clashing:

The scrub nurse and first assistant tell the surgeon that the scissors and prograsp are clashing. The surgeon briefly pulls his head out of the console to look over to the robotic arms.

Site 1

You’re clashing.

Which way is arm 2 pointing?

They then move the robotic arms to different ports to overcome this problem.

Site 7

We also observed an occasion of the first assistant warning the surgeon that the gas was leaking and the scrub practitioner alerting the surgeon to an error message on the screen.

Beyond this, we did not observe the communication of other information to the surgeon, apart from relaying messages from staff and visitors outside the OT. This may be because there was no other information that it was necessary for them to communicate. However, more generally, we found that very little communication was initiated by the team, with nearly all communication with the surgeon being initiated by the surgeon. This was despite good relationships between the surgeons and the teams we observed. Generally, the surgeons were friendly and appreciative of the team’s help. Across the sites, staff seemed to appreciate that the surgeons might be stressed prior to an operation and that this might lead to a level of abruptness. Only at one site was a comment made by an ODP that she would be nervous about speaking up in front of one of the surgeons and that others felt the same, which she felt could affect patient safety if staff were reluctant to point out things that were going wrong.

The exception was site 7, where team members would initiate communication with the surgeon, for example asking him questions. Although this surgeon would express his frustration with the staff at times, it was clear that team members were not upset by this and often smoothed such incidents over with jokes. In the phase 1 interviews at this site, and in the brief interviews following the operations observed there, the surgeon and other members of the team talked about the need for a different type of communication within RAS, without the hierarchy, again pointing to differences of culture between sites.

The ability of the OT team to communicate information to the surgeon is also dependent on the team’s situational awareness. The mean nursing subteam intraoperative monitoring score across all operations observed was 3.64 (range 2–5, SD 0.79), suggesting that in most cases the monitoring behaviours of the scrub practitioner and circulating practitioner either moderately enhanced team functioning or had neither a positive nor a negative impact. The correlation between the percentage of the operation that was robot-assisted and the nursing subteam intraoperative monitoring score was calculated using the Pearson product-moment correlation. There was no significant relationship between the percentage of the operation that was robot-assisted and the nursing subteam intraoperative monitoring score [r = 0.01; p > 0.05 (two-tailed)]. However, as noted above, the leaking of gas from the abdomen was something the team did fail to perceive on repeated occasions. If such problems are not identified and resolved quickly, this does have the potential to negatively impact the surgeon’s decision-making because his or her view of the surgical site is hindered. If the surgeon does not immediately realise what is happening, he or she may base the decision on a restricted view or may simply be distracted by it.

There was also one occasion when the robotic arms were clashing but the first assistant and scrub practitioner did not appear to notice, becoming aware of it only when the surgeon asked them if the arms were clashing. This was not necessarily because they were disengaged but because they were looking elsewhere, the first assistant watching the screen and the scrub practitioner looking at the instrument trolley. This suggests a possible conflict between being engaged in the operation in terms of watching the screen and maintaining awareness of what else is happening. In other operations, we observed the scrub practitioner repeatedly moving their attention between the robotic arms and the screen, not only to see if the robotic arms were clashing but also to see if the arms were coming too close to the patient. This was a behaviour witnessed among those scrub practitioners who had greater experience of RAS.

The spatial configuration also appeared to impact the ability of the scrub practitioner to monitor both the screen and the robotic arms. For example, in one operation the scrub practitioner was seated with her back to the patient, facing the trolley and turning her head to the left to see the screen. In other operations, the scrub practitioner would sit facing the screen, able to look left to see the patient and the robotic arms and able to look right to see the instruments. However, such positioning is under the control of the scrub practitioner, and it was among the scrub practitioners more experienced in RAS that we observed them positioning themselves so as to be able to easily move their attention between the screen and the robotic arms.

A revised theory

We sought to test the following theory:

When the OT team communicate information to the surgeon to make them aware of problems with the patient or the robot, the surgeon’s situational awareness is maintained, enabling them to adjust their decision/course of action based on this information, avoiding complications during the procedure. Where the surgeon has trust in his team to communicate problems in this way, the surgeon will feel confident to remain within the console, resulting in reduced distraction and increased concentration, which in turn can positively impact decision-making. This strategy will be successful where the team has a positive relationship with the surgeon.

On the basis of the analysis presented above, we can make a number of revisions to this theory. Whereas the theory suggests the extent to which the surgeon stays in the console is based on the extent to which they trust the OT team to communicate problems with the patient or the robot, our observations suggest it is their trust in the first assistant’s ability to carry out requests successfully that impacts the extent to which they stay in the console, and that this trust in turn depends on the first assistant’s level of experience:

Where the surgeon trusts the first assistant to successfully carry out requests and to communicate any problems, the surgeon will feel more confident to remain within the console. This trust is likely to be present when the first assistant is experienced or where they are being supported by an experienced scrub practitioner. In this context, RAS results in reduced distraction and increased concentration for the surgeon.

A significant difference from the theory was that the surgeon’s overall situational awareness did not appear to be reduced. Where their situational awareness was reduced was in relation to whether or not the robotic arms were clashing, the position of the robotic instruments and whether or not a request had been actioned. The analysis provides insight into strategies used by scrub practitioners to maintain their awareness of the position of the robotic arms so as to be able to communicate this information to the surgeon:

Experience of RAS makes the scrub practitioner aware of the need to move their attention between the screen and the robotic arms and to position themselves in such a way that facilitates this. This enables them to notice when the robotic arms are clashing or about to clash and to alert the surgeon to this, maintaining the surgeon’s situational awareness and enabling them to adjust the positioning of their instruments.

In relation to the surgeon’s awareness of the position of the robotic instruments, we can add the following theory:

When the surgeon positions a robotic instrument to assist with retraction and then moves the camera to another area of the surgical site, they may lose awareness of the position of the instrument. This can disrupt the flow of the operation, potentially affecting the operation duration, and there are patient safety risks if the instrument is moved while not under vision.

Whether or not a request has been actioned is covered by the ‘oral response’ theory presented at the end of Chapter 7 and, therefore, we can revise this theory to reflect its contribution not only to co-ordination but also to the surgeon’s situational awareness:

Oral responses to the surgeon’s requests confirm that work will be done and make apparent any challenges in completing the request, supporting effective co-ordination and increasing the surgeon’s situational awareness. The use of this strategy by the first assistant and scrub practitioner is more likely when the surgeon encourages the first assistant and scrub practitioner to communicate their actions and its success is dependent on the volume and clarity with which the information is communicated.

Whether or not the team communicate information to the surgeon depends not only on the team’s situational awareness. Although the theory proposed that the team would communicate information when they have a positive relationship with the surgeon, our analysis suggests that this is not enough. Rather, the team need to be encouraged to communicate. This fits with the ‘license to speak’ theory presented at the end of Chapter 7. Although that theory focuses on communication by the first assistant, the analysis presented here suggests that this is a more general phenomenon and, therefore, we will revise that theory to reflect this:

When the surgeon encourages the OT team to communicate both actions and concerns, team members will feel comfortable to speak up when necessary. This supports co-ordination and helps to ensure tasks are completed correctly, as well as increasing the surgeon’s situational awareness. The success of this strategy is likely to depend on the team’s relationship with the surgeon.

Experienced assistant

Our interviewees agreed with the importance of having an experienced first assistant in RAS, and all emphasised the importance of having an experienced scrub practitioner. However, it seemed that in other types of robot-assisted operations an experienced first assistant would be more likely to act without prompting, rather than making an oral offer of assistance. It was felt there were several contextual factors that affected this. The first was the extent to which the steps of the operation were routinised. For example, at site 1, all interviewees agreed that robot-assisted prostatectomies were carried out in a very routinised way, making it easy for the first assistant to know what assistance was required:

Because we’ve always done prostatectomies, and both our surgeons are quite methodical, aren’t they, that it will be the same thing every single time. So that’s why we [the scrub practitioners] know exactly what we’re doing, and they [the surgical trainees] know exactly what they’re doing. It’s only when we have turnaround of [surgical trainees], say like in October, they might struggle again.

Site 1, urology ODP

I suppose we’re a bit further down the line, and it’s almost protocol, it’s almost automatic what we do.

Site 1, urology surgeon

Related to this was the extent to which patient factors would affect the assistance required. For example, in colorectal surgery, variation in patient anatomy significantly impacts the extent of assistance needed for retraction. In our phase 2 observations, surgeons and first assistants would talk about the difficulty of achieving adequate retraction if a patient had a ‘floppy’ colon. In contrast, it was felt that the amount of retraction needed when undertaking a prostatectomy was more consistent across patients:

So maybe it works a bit different in prostates [. . .] compared to colorectal where there’s a lot of bowel and things like that hanging around.

Site 1, urology trainee

Another contextual factor was the frequency with which first assistants participate in robot-assisted operations and the variation in the operations that they do. At site 1, although the robot was used for a number of different colorectal operations, urology surgical trainees would participate in four robot-assisted prostatectomies a week, enabling them to quickly build up the experience to know what assistance was needed and to have confidence in providing that assistance without prompting.

Discussing this theory with the urology ODPs and the surgical trainee also revealed that, in relation to the changing of robotic instruments and monitoring the movements of the robotic arms to see if they are clashing, the responsibility was more shared, in part due to the spatial configuration in the OT. Whereas in most colorectal operations the robot would be brought in from the side, with the first assistant and the scrub practitioner standing on the other side of the patient, in prostatectomies the robot is brought in between the legs (see Figure 1), with the first assistant standing on one side and the scrub practitioner standing on the other side. The first assistant would take responsibility for the robotic arms on their side of the patient and the scrub practitioner would take responsibility for the robotic arms on the other side. Thus, there was less confusion about who was responsible for what.

Awareness of instruments

Our interviewees perceived that a surgeon ‘losing’ a robotic instrument, in terms of losing awareness of robotic instruments outside the camera’s field of view, was less likely to happen in other types of surgery because in those other types of surgery the surgeons are working in a smaller space:

More often than not all my instruments are in the field of view and I don’t need to change that field of view, I don’t need to pan in or out because I’m working in the pelvis.

Site 1, urology surgeon

I think this is mainly colorectal isn’t it because they have so much more than we would.

Site 7, upper GI surgeon

Situational awareness

In discussing the impact of RAS on situational awareness, a theme that came up among the comments of the surgeons we interviewed was that, although they cannot see what is happening in the OT, they hear it more clearly. Some said that this was because of the speakers inside the console: that although problems with the microphone in the console may mean that the OT team cannot hear the surgeon well, the speakers inside the console work well. However, others felt that they became more alert to noise because their ability to draw on their other senses was reduced:

But my theory is that you lose tactile feedback, so I think your other senses become enhanced and definitely sense of sound. And I’m very conscious that I can hear every single sound in the room, so if there’s something slightly different I pick it up.

Site 1, urology surgeon

I would also say I think I have a more heightened sense of awareness with my hearing in doing robotic case because you can’t see what’s going on in the room.

Site 7, urology surgeon

This could be problematic, as case-irrelevant discussions in the OT could be a distraction, a sentiment expressed by colorectal surgeons in phase 2.

The interviews also gave insight into the possible negative consequences of the surgeon not being able to see what is happening in the OT. Having participated in a greater number of robot-assisted operations, our interviewees were able to draw on that experience to provide examples of the rare occurrences we were unlikely to have witnessed in the sample of robot-assisted operations we observed and that the colorectal surgeons and OT teams may not have experienced. One urology surgeon described a robot-assisted operation in which the patient had a cardiac arrest. The anaesthetist was not part of the normal team and did not communicate to the surgeon what was happening. Although this is only one instance, we felt that its significance meant that it was important to include the description here, and it suggests that this is a topic on which further research is needed:

Back 2 or 3 years ago we had a death on table and it was a patient who had had previous cardiac surgery but had been passed fit, and we had a locum anaesthetist. And the case from a surgical perspective was absolutely fine; I was looking at the console, we were winding to a finish, I’d taken the prostate out and I was just starting the anastomosis. And the staff were getting worried in theatre, restless. And I could hear them having a conversation with the anaesthetist, and the anaesthetist wasn’t really . . . didn’t say anything at all. And then next I got one of my senior sisters who was not scrubbed telling me, we’re going to have to disconnect the robot, this man’s not well. And we disconnected and left him very rapidly and he arrested on the table with the cardiac arrest team and what have you. And he presumably had an MI, and I think, you know, it’s unfamiliarity with the stresses of the robotic surgery for an anaesthetist who is completely new to it, and for somebody who is not interacting with the team because they’re not a normal member of it. And it’s a lesson in communications. I mean probably there would have been a point earlier in the operation where the anaesthetist, if they’d been part of the team normally would have said, I’m not happy with this chap, is this normal. Or got somebody else who was more familiar with it. But there was absolutely nothing, zero, you know, from that. And if I was at the table and there were problems with the monitors I’d see it, and because my heads in the box, my back to everything, I just thought things were going swimmingly, absolutely no problem at all. And you know . . . but it’s a real lesson.

Site 1, urology surgeon

Implementation of robot-assisted surgery

During the interviews, we also asked about how RAS was introduced within the specialty, in order to understand if this was a relevant contextual factor. The interviewees’ comments reiterated many of the ideas about implementation reported in the phase 1 interviews. In relation to training, the importance of whole-team training was reiterated, as was the importance of understanding the impact of RAS on each other’s roles. Consequently, the two urology surgeons at site 1 felt that a surgeon should not operate at the console without previously having had experience of acting as a first assistant in a robot-assisted operation.

The interviewees also revealed differences within sites in the extent to which the OT team had been involved in the introduction of RAS and had been able to adapt their work processes around this new technology. For example, at site 1, where the colorectal OT team felt that they had little or no input into the introduction of RAS within their specialty, the urology OT team described the way in which they reviewed the instrument trays, cutting down the number of instruments in order to make their work easier:

So then we didn’t have lots of stuff on your trolley because you’ve got your lens, you know, the big lens, and that’s quite a big bit of kit, you know, you don’t want to be knocking stuff on the floor, getting yourself flustered. So we just said, right, let’s just . . . do we need this, do you think we need this . . .?

Site 1, urology ODP

Implications for the interpretation and reporting of ROLARR

The ROLARR trial found no significant difference between RAS and laparoscopic surgery in the primary outcome of conversion to open surgery, while the economic analysis found RAS to be more expensive than laparoscopic surgery due to the cost of consumables. 194 At the time of writing, an analysis of longer-term patient outcomes is not yet available. The phase 1 interviews were conducted at all six of the NHS ROLARR sites, enabling us to provide an account of the different ways in which RAS was implemented in these sites, in terms of components of the intervention (e.g. the amount and nature of training provided), and stakeholders’ perspectives of the value of those different approaches. The publication of these findings alongside the reporting of the ROLARR trial provides important information for health-care organisations that are considering introducing RAS, acting as a response to the call for better description of surgical interventions,195 while acknowledging that mandating the details of an intervention may increase surgeon reluctance to take up such interventions if they disagree with such details. 196 Below we have summarised the findings in the form of guidance for health-care organisations. We chose not to use variation in the components of the intervention to interpret differences between sites because, owing to the small number of operations undertaken by some sites, it is questionable how meaningful this would be.

We were also able to capture unanticipated consequences of RAS in terms of impacts on communication, teamwork and decision-making, along with the strategies used to counteract such unanticipated consequences. This has implications for the conduct of RAS, which we summarise below in the form of guidance for OT teams. These findings also act as an important reminder that surgical interventions are not static, suggesting that evaluations of surgical interventions need to track how the intervention changes over time as stakeholders respond to, and find ways to overcome, the challenges the intervention presents.

While a strength of the study is the methods of data collection, which allowed us to develop a detailed understanding of how and in what contexts RAS impacts communication, teamwork and decision-making, and how and in what contexts particular strategies work to support communication, teamwork and decision-making in RAS, the small number of operations observed could be perceived as a limitation. The account provided by the urology surgeon presented in Chapter 9, in which the surgeon was unaware of the patient’s deterioration, suggests that there may be negative impacts of RAS on the surgeon’s situational awareness that we have not captured. However, a predominantly qualitative study is not the appropriate design for capturing such rare adverse events. The IDEAL framework for surgical innovation recommends long-term surveillance through registries to capture rare end points relating to patient safety,197 and RAS has been presented as an example of why such systems of surveillance are needed. 151,198 In the USA, research has been undertaken into adverse events associated with RAS using the Food and Drug Administration’s Manufacturer and User Facility Device Experience (MAUDE) database,199 but there is known under-reporting of adverse events associated with RAS12 and it is questionable whether less indirect consequences of RAS, such as the example described above, which do not represent device malfunctions, would be reported. Thus, how to capture such impacts is a methodological question that remains to be explored.

This study started after the trial, which limited the impact it could have on the analysis of the trial. For example, in the discussion about emerging findings with the trial management group, there was interest in the impact of the skill mix of the first assistant and the scrub practitioner. However, although the trial case report form captured quantitative information about the experience of the surgeon, information about the experience of the OT team was captured as free text, making it difficult to use those data to support the testing and refinement of our theories. Therefore, there is a methodological issue here. Our experience suggests that, although realist evaluation can play a valuable role alongside RCTs, the elicitation of theories would ideally happen before the trial to ensure that relevant data are secured to support the testing of identified theories. This fits with recent calls for realist feasibility and pilot studies. 200

Implications for the implementation of robot-assisted surgery

Whereas phase 2 focused on the testing of theories related to communication, teamwork and decision-making, phase 1 captured a significant number of data regarding the different ways in which RAS was introduced across nine NHS hospitals, and stakeholders’ perspectives on the value of those approaches and the contexts in which they are effective. While acknowledging that the theories regarding implementation remain to be empirically tested, our findings suggest that, for health-care organisations seeking to introduce RAS or to increase the use of an already purchased da Vinci robot, the following strategies may be beneficial:

1. Engagement of staff at different levels of the organisation. Although board-level support is likely to be essential for the introduction of RAS, it is also important to engage team leaders, as they can assist in creating conditions that accommodate the introduction of RAS, such as organising training and ensuring that the right skill mix is available. Engagement of those surgeons who will not be using the robot is also important; if surgeons perceive that the introduction of RAS is supported by their colleagues, they are likely to be more willing to undertake an operation with robot-assistance despite the initial longer operation duration.

2. Handpicked dedicated robotic team. While unlikely to be feasible as a long-term strategy, a handpicked dedicated team can increase the speed with which experience is built up, increasing confidence and efficiency. However, care should be taken not to alienate those who are not part of that initial team.

3. Whole-team training. Ideally, the whole team should train together. This is beneficial in terms of understanding the impact of RAS on each other’s roles, supporting teamwork. It can work to increase trust in each other’s knowledge so the surgeon feels more confident in the team’s ability to support him or her.

4. A suitably sized OT. By having a suitably sized OT, operation duration is reduced, as staff are able to move quickly and the risk of desterilisation is reduced.

A key issue that arose throughout the study related to point 3: whole-team training for RAS. This was perceived as an important issue by the OT teams we interviewed and our study revealed substantial variation in the training received by OT teams. Our phase 1 interviews revealed that simply bringing together interdisciplinary teams of professionals can be useful to support team bonding and enhance trust across the team. However, our findings also suggest that the nature of the training can be further enhanced. Currently, training focuses on technical skills related to the manipulation and use of the robot: setting up the robot; docking and undocking the robot; inserting, using and removing instruments; troubleshooting; and so forth. Phase 2 revealed differences in the organisation of roles and issues for communication between RAS and laparoscopic surgery. Therefore, it makes sense for training to enable participants to learn more explicitly about the implications of the division of labour in RAS; the limitations and constraints on the surgeon’s field of view; and differences in the organisation of teamwork and communication in RAS. In sum, our findings suggest that there is value in providing whole-team training to reflect overtly on communication and other differences in teamwork in RAS.

Indeed, the project has already inspired preliminary work to consider the value of new forms of training for RAS. Following discussions of research within the project, one of the project team (AG) developed a pilot training simulation for whole-team training in RAS. The simulation enabled participants to experience and reflect upon a one-off robot-assisted emergency situation, and, in particular, to work together as a whole team and explore their respective roles and responsibilities. The findings from this study were presented at the phase 3 workshop in Leeds and point towards possibilities for further research on training. These issues and opportunities for training are highly relevant to organisations that are introducing RAS. They suggest the need not only for whole-team training, but also for training that moves beyond the technical through to training in teamwork and role distribution. This is an area in which further research is needed to develop training for the OT team and evaluate its impact.

Related to this is the issue of skill mix. Both organisations looking to introduce RAS and those already undertaking RAS should look not just at the training and experience of the surgeon but also at the training and experience of the first assistant and the scrub practitioner. The findings from phase 2 suggest that experience of assisting or undertaking a particular procedure openly or laparoscopically is not sufficient to assist competently when that procedure is being undertaken with robotic assistance. When the first assistant does not have the necessary experience, they should be supported by an experienced scrub practitioner.

A more general issue relates to the process by which RAS is introduced into an organisation. The implementation of RAS within NHS trusts has largely been surgeon led. This reflects what has been described as a more general pattern, whereby innovations are introduced into surgical practice through informal processes with an absence of quality control efforts, and some have argued that these informal processes put patients at greater risk of adverse events. 201 In none of the nine phase 1 sites did OT team members perceive themselves to have been involved in the introduction of RAS. When this is combined with a lack of training, it can create the sense that RAS is something that has been thrust upon the OT team, leading to feelings of resentment regarding this new technology. Although our phase 1 interviewees emphasised the importance of team leader support, it does not appear that those team leaders were involved in discussions prior to the introduction of RAS. Rather, it was only once the robot had been introduced that the support required became apparent. Our findings would suggest that there is potential benefit to be gained through involving team leaders earlier in the process, so issues of training for the OT team and skill mix can be properly addressed before the robot is introduced into practice.

Implications for communication and teamwork in robot-assisted surgery

Our findings reveal a number of strategies, implicit and explicit, used by surgeons and OT teams to encourage effective communication and teamwork in RAS. These are strategies that they have developed through experience, in response to the challenges RAS can present for teamwork and communication, and that could be employed by surgeons and OT teams elsewhere. As a consequence of the relationship between communication and teamwork in the OT and patient safety, there are already a number of interventions designed to improve communication and teamwork in the OT, such as the World Health Organization’s Surgical Safety Checklist, the use of which has been mandatory in NHS OTs since February 2010. 202 Evaluations of interventions that seek to make communication in the OT more explicit highlight the need to be sensitive to the cultural context, such as an emphasis on autonomy, a belief in individual excellence, or time pressures, which could lead to resistance,203 and the possibility of negative unanticipated effects, such as disrupting the dynamic flow of conversation, reinforcing professional divisions and/or creating tension. 204 Therefore, our intention is not to provide a rigid set of guidelines but instead to suggest strategies surgeons and OT teams can try, acknowledging the realist principle that not all strategies will work in all contexts and that use of one strategy may change the context, impacting on whether or not other mechanisms continue to fire.

For surgeons, our findings suggest the importance of:

• Encouraging the OT team to communicate both their actions and concerns. This helps OT team members feel comfortable to speak up, leading to improved co-ordination and increased situational awareness for the surgeon. This strategy is more likely to be effective when there is a positive relationship between the OT team and the surgeon.

• Alerting the attention of the first assistant before issuing a request, particularly after a period without communication. This supports the first assistant to hear the message, leading to improved co-ordination. This strategy is more likely to be effective if the first assistant is already engaged in the operation.

• Acknowledge the role of the scrub practitioner in supporting an inexperienced first assistant, so as to increase the first assistant’s willingness to accept that support. When the scrub practitioner is experienced, this increases the likelihood that actions are performed correctly.

For the OT team, our findings suggest the importance of:

• Providing an oral response to the surgeon’s requests. This reassures the surgeon that the request will be completed and makes them aware of any challenges in doing this. This strategy requires that team members communicate loudly and clearly.

• The scrub practitioner positioning themselves so they can easily move their attention between the screen and the robotic arms. This enables the scrub practitioner to notice when the robotic arms are clashing and alert the surgeon, increasing the surgeon’s situational awareness. More experienced scrub practitioners will be aware of the need to do this.

This guidance, and our theories relating to communication and teamwork, may have relevance for other areas of surgery in which technology leads to the separation of the surgeon from the rest of the team. For example, numerous surgical specialties use surgical microscopes. Although when using these the surgeon is not physically separated from the rest of the team, there are similarities in terms of the surgeon being focused on a different view of the surgical site and no longer able to see the wider OT, requiring a move from byproduct awareness to add-on awareness.

Implications for the design of technologies for robot-assisted surgery

In the phase 3 workshop, a number of suggestions for revisions to the design of the robot were made, largely focused on increasing the surgeon’s awareness of what is happening both in the surgical site and in the OT. To increase the surgeon’s awareness of the position of robotic instruments outside his field of view, a display on the console could indicate this. This could be extended to also show the position of laparoscopic instruments held by the first assistant. Related to this, a safety mechanism could be introduced that prevents the surgeon from moving a robotic instrument when it is outside their field of view. Although the surgeons in our phase 1 interviews did not see the lack of tactile information as a significant issue, workshop participants suggested that tactile information may enable the surgeon to better co-ordinate their movements with the first assistant, as they would be able to feel when their instruments were clashing.

To increase the surgeon’s awareness of the wider OT, the suggestion was made that the console could incorporate a video stream that provided a ‘window’ onto the OT, which could be turned on and off. Workshop participants also discussed issues related to the spatial configuration of team members and technology in the OT. Above we described differences in spatial configuration across our sites, which affected the ease with which the surgeon could see the patient and have face-to-face communication with the first assistant and the scrub practitioner. However, having discussed this with staff at the sites, it appears that this was not intentional but rather the result of aspects of the layout of the OT, with the console positioned to avoid doorways and to limit the number of cables across the floor. This is an area where in situ experimental studies could provide insight into the impact of different configurations. One suggestion was to have the console on a raised platform, allowing the surgeon to have a better view of the OT when they brought their head out of the console.

More generally, it appears that the design of RAS has focused on the needs of the surgeon. For example, although the control of the robotic instruments is generally perceived to be easy to learn, our observations suggest the colour-coding used to indicate when it is safe to insert a robotic instrument is not experienced as intuitive by first assistants and scrub practitioners. Although this may be overcome by better training, recognition that successful use of RAS depends not just on the surgeon but on the whole team suggests that there is value in considering the wider team in the design of the robot.