Safer Surgery 224 Yule et al 2006) and systematic behavioural task analysis (Manser and Wehner 2002, Weinger et al. 1994). Team Performance and Patient Safety The process of providing healthcare is inherently interdisciplinary involving physicians, nurses and allied health professionals from different specialties. In the patient safety literature, it has been widely recognized that team performance is crucial to providing safe patient care and that many of the factors contributing to adverse events in healthcare originate from awed teamwork rather than from a lack of clinical skills. Thus, teamwork has become a key factor addressed by many system-based interventions to improve patient safety and medical education standards (Carthey et al. 2001, Chassin and Becher 2002, Davies 2001, Donchin et al. 1995, Lingard et al. 2004, Manser 2009, Sutcliffe et al. 2004, Wilson et al. 1995). The quality and efciency of patient care as well as the condence of patients in healthcare providers were shown to be affected by poor coordination among providers at various organizational levels (Edmondson 1996, Gerteis et al. 1993, Gittell et al. 2000, Young et al. 1998). It has been argued that there is much rhetoric about how to set up medical teams while the process of teamwork in patient care has not been studied systematically (Cott 1997). Given the importance of teamwork to safe and efcient patient care, both clinical competence and the ability to work in teams need to be assessed when studying the professional behaviour of healthcare providers (Fletcher et al. 2002, Gaba et al. 1998, Wilson et al. 2005). However, many studies of teamwork in healthcare still centre around the main research themes identied by Nagi (1975): (a) status, power, authority, and inuence, (b) roles and professional identities, and (c) decision-making and communication (Manser 2009). In recent years signicant progress has been made in research on healthcare teams (Manser 2008, Manser 2009) and this progress is clearly documented by the contributions in this book. Aspects of teamwork that were found to correspond to the quality and safety of patient care are, for example, perceived team climate, leadership behaviour furthering open communication, and adaptive coordination and leadership (Manser, 2009). Adaptive Coordination as a Key Characteristic of Successful Team Performance in Complex Work Environments In general, teamwork provides the opportunity for pursuing multiple goals and performing multiple interdependent tasks simultaneously. Thus, teams have to share information and resources, plan and synchronize task execution; they have to coordinate and anticipate errors likely to occur (Brannick et al. 1995, Brannick and Prince 1997, Dickinson and McIntyre 1997, Hackman and Morris 1975, Kontogiannis and Kossiavelou 1999). A multitude of denitions regarding Identifying Characteristics of Effective Teamwork 225 the term ‘coordination’ has been proposed, such as ‘the process by which team resources, activities, and responses are organized to ensure that tasks are integrated, synchronized, and completed within established temporal constraints’ (Cannon- Bowers et al. 1995, p. 345) or simply ‘some kind of adjustment that one or more of the team members make so that the goal is reached’ (Brannick and Prince 1997, p. 4). Studies in many complex work environments show that successful teams adapt their coordination process to the coordination requirements that vary depending on task characteristics (e.g., task complexity, task interdependence), team characteristics such as familiarity of its members (Foushee et al. 1986, Kanki and Foushee 1989), and the situation (e.g., time pressure, routine vs. non-routine procedure) (Kontogiannis and Kossiavelou 1999). For example, effective teams have been described to increase their communication in emergency situations – specically information exchange and verbalization of plans (Brehmer 1996, Orasanu 1990, Orasanu and Fischer 1992). However, team adaptation is not limited to coordination behaviour. Integrating evidence from empirical studies conducted in various safety critical domains (Burke et al. 2006) have developed a comprehensive model of team adaptation that includes adaptations concerning a) the input into the teamwork process such as a mobilization of additional resources or a structural reconguration of the team as well as b) process adaptations, i.e., changes in coordination mechanism, decision making, and communication patterns in response to unexpected events (Brehmer 1996, Entin and Serfaty 1999, Serfaty et al. 1994, Serfaty et al. 1993). Investigating Adaptive Coordination in Anaesthesia Crews This chapter addresses the issue of adaptation and adaptability of coordination processes in anaesthesia in the face of changing situational requirements. We pursued the question of adaptive coordination (i.e., the situational management of coordination requirements) initially via an interview study and then in two observational studies. The main results of the interview study will be summarized briey because they were used in the development of the taxonomy of coordination behaviour used for the two observational studies. Collaborative Management of Unexpected Events in Anaesthesia Little empirical evidence exists about how anaesthetists perceive changes in coordination requirements related to different phases of a procedure or different surgical procedures and which coordination strategies they use to maintain good clinical and team performance. Semi-structured interviews focusing on coordination requirements in cardiac anaesthesia were conducted in a tertiary care hospital in the USA with all six anaesthesia attendings who perform cardiac anaesthesia, ve third-year anaesthesia residents who had just completed their cardiac rotation, and Safer Surgery 226 the only two cardiac surgeons for the hospital (Manser 2006, Manser et al. 2006). Participants were asked to describe the coordination needs within the anaesthesia crew as well as to other members of the perioperative team (a) during different phases of cardiac anaesthesia, (b) for two different surgical procedures, and (c) how these coordination needs change, for example, in case of unexpected clinical events. A qualitative content analysis of the verbatim transcripts was performed (ATLAS.ti Software, Version 5) deductively applying categories derived from the conceptual framework for a team’s response to unexpected events proposed by Wehner and colleagues (2000). Their model of adaptive collaborative practice (see Figure 14.1) emphasizes that within work processes, initial coordinatedness has usually evolved over a long period of time based on division of labour. Unexpected events or disturbances within work processes demand cooperation. Corrective cooperation (e.g., workarounds) can help to restore the original coordinatedness by means of collective process-related and situational action. However, if too much corrective cooperation is needed, a remediation of the initial coordinatedness may be necessary. This remediation process involves either expansive cooperation or co-construction. Expansive cooperation takes place if the problem requires an expansion of the range of interaction or the number of interaction partners (e.g., solving problems that occur at organizational boundaries). Co-construction is an attempt to generate organizational solutions Figure 14.1 Conceptual framework of adaptive collaborative practice (by Wehner et al. 2000) Identifying Characteristics of Effective Teamwork 227 that transcend single cases (e.g., change projects). In addition to this ideal type of progression, Wehner et al. (2000) outline self-regulatory processes, i.e., a short- paced cycle consisting of corrective cooperation and remediative coordination. In this process, the coordinatedness is renewed by means of active coordination behaviour of the local actors, which normally leads to a sufcient adaptation to the situational requirements. In summary, the results of our interview study show that work processes in cardiac anaesthesia are characterized by interlinked tasks in a multidisciplinary and multiprofessional team that acts within narrow temporal constraints within a complex and highly dynamic work environment. Depending on the surgical procedure and the segment of the case, the task interdependence between anaesthesia crew and surgical crew changes and so do the coordination requirements. Accordingly, there is a high degree of situational variability, not all of which can be planned for in advance but most of which has to be corrected for, swiftly and effectively. This variability and situational exibility is, to a certain extent, already implemented within the initial coordinatedness. Anaesthetists rely on a combination of initial coordinatedness, anticipation and response or reaction to events (i.e., adaptation to situational requirements is achieved mainly through self-regulation as a sub-process of corrective cooperation). All interviewees highlighted the importance of observing the surgical eld as it supports the anticipatory handling of unexpected events and of potential breaks in the initial coordinatedness. For example, one anaesthesia attending pointed out that: For most procedures [.] you have this kind of ideal course, a template in your brain [.] And then if things start deviating from that template [.] I mean I’m kind of running this program of what should be going on. I’m always comparing it to things that are happening left and right. (A2-1:35) The adaptation to ‘unexpected’ events, which not only bear a potential for the breakdown of coordination, but also of harm to the patient, is primarily performed through self-regulatory processes along situational and case-specic peculiarities. The central importance of self-regulation processes is seen in the adaptation of role distribution to situational requirements or in the anticipation of possible breaks in coordinatedness. As the following quote shows, anticipation and adaptation require a detailed knowledge of surgical interventions and therefore a profound understanding of the task procedure linking all team members. I think it’s really important to recognize that cardiac surgery, cardiac anesthesia is a team sport. It’s a team effort and the analogy I frequently make is to ballet. There is a choreography. Everyone has studied it. Everyone knows it. Everyone knows that each has a different part; there’s the underlying music. There’s the set. Okay, let’s go. And then stuff happens. And you know, somebody stumbles or a piece of the set doesn’t go where it’s supposed to go or whatever. And Safer Surgery 228 so some of what you do is response or reaction. A lot of it is, more of it is anticipation. (A6-2:66) Development of an Observation System for Adaptive Coordination The development of this observation system for coordination behaviour in anaesthesia was guided by four aims. Firstly, the observation system should be comprehensive to allow for continuous behaviour coding. Secondly, to allow for a systematic comparison of coordination behaviour and technical as well as non-technical performance ratings in future studies, one key requirement for the observation system was to include only descriptive categories of coordination activities. Thirdly, we decided to include clinical activities performed by the anaesthesia crew in the observation system, because balancing the efforts to coordinate and the efforts to perform the task is a critical aspect of team performance. Finally, the dynamic composition of the anaesthesia crew (i.e., the absence of any of the anaesthesia crew members from the operating room) was recorded because it inuences work and coordination patterns. In developing the observation system we used three information sources: literature review, data from the interviews summarized above and eld notes from ethnographic observations. In a rst step, a literature review was conducted focusing on systems to describe and measure coordination behaviour in complex work settings. In our review we included studies from different domains using a spectrum of methodological approaches such as observational studies, verbal protocol analysis, and questionnaires. Ten lists of descriptive categories relevant to coordination behaviour were identied from the literature and analysed to establish which categories they contained and how they were structured (Bowers et al. 1998, Bowers et al. 1993, Grote 2003, Kanki and Foushee 1989, Leedom and Simon 1995, Mackenzie et al. 1993, Marks et al. 2001, Risser et al. 1999, Urban et al. 1995, Xiao et al. 1998, Xiao and LOTAS 2001) (for an overview see Manser et al. 2008). The main difculties in analysing these systems were that (a) the same concepts were listed on different levels of abstraction in different systems and (b) descriptive categories and evaluation criteria were often summarized within the same system. We integrated the descriptive categories identied from the literature review into a list and then sorted them into groups with common themes at equivalent levels of abstraction. In a second step, interviews were conducted focusing on strategies for the situational management of coordination requirements in anaesthesia (Manser et al. 2006). The coordination strategies and behaviours derived from this interview study – complemented with eld notes prepared during ethnographic observation in the operating room – were used to rene the list of coordination behaviours derived from the literature review and dene a prototype of the observation system. In a nal step, this prototype was tested and rened using video recordings of anaesthesia inductions for cardiac surgery. Identifying Characteristics of Effective Teamwork 229 The resulting observation system consisted of 36 mutually exclusive codes that were grouped into observation categories at two levels of abstraction (see Figure 14.2). The category labels for coordination activities were information management, task management, coordination via the work environment, and meta-coordination (i.e., coordination activities team members use to coordinate about their coordination process). To get an accurate picture of the overall communication, we also recorded communication related to teaching (within the anaesthesia crew) and ‘other communication’ such as coordination activities not related to the actual case but to the overall work process, social communication not related to the task, and non-codable utterances. To allow for an analysis of the distribution of clinical tasks within the anaesthesia crew, the observation system contains three codes for clinical activities: clinical activities performed by anaesthesia crew member A (e.g., the anaesthesia resident), by anaesthesia crew member B (e.g., the anaesthesia attending), or by the (usually) two members of the anaesthesia crew working simultaneously. Figure 14.2 Overview of the observation system for coordination behaviour in anaesthesia crews Safer Surgery 230 Data Recording In both observational studies on adaptive coordination in anaesthesia crews, we used a PDA-based data recording system that allows for the recording of concurrent events such as activities of multiple people as well as multiple activities carried out by one person simultaneously (Held and Manser 2005) (see Figure 14.3). Previously, this data recording system had been used successfully in observational studies of individual anaesthesia care providers (Manser et al. 2007a, Manser and Wehner 2002). The main reason for using this technology was that it can be used for participant observation in the operating room, either in hospital or in a simulation centre, and for behaviour coding using video material. However, the FIT-system software was not originally designed to record as many combinations of behaviour codes including the person performing this behaviour and its target, as needed in this study. As a result, a lot of time- consuming data reformatting and careful checking for any error that may have occurred during this procedure was required. Coding Reliability Due to having only one trained observer, a full assessment of the reliability of the coding system has not yet been assessed. We did, however, assess intra-observer agreement over time using randomly selected 10 per cent- segments of videotapes from a full-scale simulator study that were coded a second time six months after the initial coding by the same observer (Manser et al. in press, Manser et al. 2007b). In computing Cohen’s Kappa, we: (a) used weights to correct for any inuence of the frequency of occurrence, (b) calculated separate values for routine anaesthetic care and crisis management, and (c) allowed for a lag of 2 seconds for onset and offset times as suggested (Bakeman and Gottman 1997) for second by second coding of complex behavioural processes (Held and Manser 2005, Manser et al. 2007a). Good to excellent intra-observer agreement over time was shown for the various observation categories with an overall Figure 14.3 System for data recording: FIT-system (left) and template with observation codes including ‘buttons’ for members of the operating room team (right) Identifying Characteristics of Effective Teamwork 231 agreement of .85 for routine anaesthetic care and .88 during the management of a simulated anaesthetic crisis (see Table 14.1). Adaptive Crew Coordination in Anaesthesia: Two Observational Studies In the following, we summarize the results of two observational studies investigating patterns of adaptive coordination in anaesthesia crews: one eld study in the highly complex clinical setting of cardiac anaesthesia and one video-based study of adaptive coordination in the event of a simulated medical emergency. The aims were (a) to test whether the new observation system is sensitive enough to capture adaptation processes in anaesthesia crews, (b) to uncover patterns of adaptive coordination in relation to task characteristics inuencing coordination processes (e.g., task interdependence within the perioperative team), and (c) to identify coordination behaviours associated with high clinical performance. In both studies, we focused on the anaesthesia crew. Thus, interactions within the anaesthesia crew, as well as between the anaesthesia crew and other team members (e.g., surgeons, circulating nurse), were recorded. Coordination activities by other members of the operating room team, however, were only recorded when directed towards the anaesthesia crew. Observation categories During routine anaesthetic care During the management of a simulated anaesthetic crisis Information management .80 .86 Task management .65 .77 Coordination via work environment .69 .92 Other communication .86 .57 Metacoordination .59 N ot observed Teaching Not observed Not observed Clinical work .92 .92 Overall .85 .88 Note : Intra-Observer Agreement given in Cohen’ s Kappa. Values of .40 to .60 = fair; .60. to .75 = good; over .75 = excellent (Fleiss 1981) Table 14.1 Intra-observer agreement over time for the observation system reported at the level of observation categories Safer Surgery 232 Study 1: Adaptive Coordination in Cardiac Anaesthesia Study procedure Over a period of four months in a tertiary care hospital, a trained observer coded the coordination process during 24 cases of cardiac surgery including 16 cases with and 8 without cardiopulmonary bypass (i.e., 123 hours 21 minutes of second-by- second coding of coordination and clinical work processes). In this eld study, the anaesthesia crew was equipped with wireless microphones that transmitted to headphones worn by the observer who was positioned adjacent to the anaesthesia workspace. This allowed for accurate coding without disturbing the clinical work process. Data were analysed using the proportion of time spent on each observation category calculated for the overall case duration as well as for each phase of the operation. Multivariate analysis of variance was used to investigate whether different patterns occur during different phases of and between different types of surgical procedures. Main results In the following we provide an overview of the results of our eld study in cardiac anaesthesia which are reported elsewhere (Manser et al. 2008). In summary, our results indicate that anaesthesia crews dynamically adapt to task requirements. We observed different coordination patterns in response to varying task requirements (a) during different phases of a cardiac anaesthesia case and (b) depending on the type of procedure. We observed a statistically signicant increase in both clinical and coordination activity during case segments when the surgical and the anaesthetic task are tightly coupled. Specically, we found increased coordination activity during the surgical phase including coronary artery bypass grafting. This increase was stronger during coronary artery bypass grafting procedures on the beating heart during which the highest level of interdependence exists between the anaesthesia and the surgical crews due to the criticality of this phase of the procedure (Manser et al. 2006). Interestingly, the highest level of ‘coordination via the work environment’ (i.e., an implicit coordination mechanism) was recorded during these periods whereas during other case segments, most time was spent on information management. This result is in line with a key nding of team adaptation research; that teams choose different coordination mechanisms depending on the situational requirements (e.g., explicit vs. implicit coordination mechanisms) (Serfaty et al. 1998, Serfaty et al. 1993, Xiao and LOTAS 2001). Implicit coordination refers to a shared understanding of the task requirements that has usually been established through explicit coordination before engaging in the actual task or during periods of low workload (Orasanu 1990, Orasanu 1993, Wehner et al. 2000, Wittenbaum et al. 1998) and does not necessarily involve verbal communication (MacMillan et al. 2004, Serfaty et al. 1998, Serfaty et al. 1993, Xiao and LOTAS 2001). Implicit coordination is often assumed to be especially effective in high workload situations where resources for explicit coordination may be limited (Entin and Serfaty 1999, Grote et al. 2004). However, empirical evidence on implicit coordination and clinical performance in healthcare teams is scarce and studies show mixed results. Future research needs to systematically include factors potentially mediating these effects. For example, the importance of experience and expertise in effectively Identifying Characteristics of Effective Teamwork 233 using implicit coordination is highlighted by our result that anaesthesia attendings spent more time on coordination and particularly on coordination via the work environment than anaesthesia residents. Study 2: Adaptive Coordination in the Management of a Simulated Anaesthetic Crisis The goal of this simulator-based study was to systematically analyse the relationship between coordination patterns and clinical performance in the management of a medical crisis situation (Manser et al. 2009). Study procedure Using the same observation system that was successfully piloted in the eld study described above, we coded the coordination processes of anaesthesia crews in 24 scenarios videotaped in a full-scale patient simulator. Study participants were all rst-year anaesthesia residents participating in anaesthesia crisis management courses at the Patient Simulation Center of Innovation, VAPAHCS, between 1998 and 2003. The simulator scenario involved a general anaesthetic with a routine phase before the simulated patient experienced an episode of Malignant Hyperthermia (MH). Clinical performance of the anaesthesia crews was rated against the treatment protocol for MH by two independent observers using a time-based scoring system for critical treatment steps (Harrison et al. 2006). Main results The results of this study are described in detail elsewhere (Manser et al. 2009, Manser et al. 2007b) and are summarized here, with the most important ndings highlighted. Firstly, we observed statistically signicant changes in clinical work and coordination patterns before and after the declaration of the simulated anaesthetic crisis (i.e., adaptation to the occurrence of a medical crisis). Secondly, the coordination process of anaesthesia crews with higher clinical performance ratings was characterized by signicantly less task management, more situation assessment (especially within the anaesthesia crew) and higher levels of information transfer during the rst ve minutes after declaration of the crisis. Finally, the proportion of monitoring of other anaesthesia crew members – an implicit coordination behaviour aiding in the maintenance of a shared mental model – showed a positive correlation with clinical performance. Summary and Outlook The results of the observational studies summarized in this chapter highlight the importance of adaptive patterns of coordination in anaesthesia and complement existing empirical results on coordination in anaesthesia (Grote et al. 2004, Hindmarsch and Pilnick 2002). One aspect of adaptive coordination that has been largely ignored in previous studies (often conducted in work environments with less variation in team composition over time) is the fact that adaptation takes place at different levels of the organization and comprises both process adaptation and structural adaptation (Manser et al. 2008). The observation system described in this chapter has successfully been applied to investigate adaptive coordination behaviour . had just completed their cardiac rotation, and Safer Surgery 226 the only two cardiac surgeons for the hospital (Manser 2006, Manser et al. 2006). Participants were asked to describe the coordination. Safer Surgery 224 Yule et al 2006) and systematic behavioural task analysis (Manser and Wehner 2002,. somebody stumbles or a piece of the set doesn’t go where it’s supposed to go or whatever. And Safer Surgery 228 so some of what you do is response or reaction. A lot of it is, more of it is anticipation.