Secondary Logo

Journal Logo

Economics, Education, and Policy: Research Reports

Coordination Patterns Related to High Clinical Performance in a Simulated Anesthetic Crisis

Manser, Tanja PhD*; Harrison, Thomas Kyle MD†‡§; Gaba, David M. MD†‡§; Howard, Steven K. MD†‡§

Editor(s): Dexter, Franklin

Author Information
doi: 10.1213/ane.0b013e3181981d36
  • Free
  • Appendices

Within different fields of acute patient care that are characterized by highly dynamic work processes, such as surgery, anesthesia, and intensive care medicine, the competent and coordinated management of crisis situations is essential to patient safety. Incident analyses show that many crisis situations in health care involve difficulties in teamwork, specifically in communication and coordination.1–3

Psychological research on teamwork distinguishes different types of work groups. For example, a crew has been defined as a “group of expert specialists each of whom have specific role positions, perform brief events that are closely synchronized with each other, and repeat these events across different environmental conditions.”4,5 Following this definition, most work teams in the context of acute patient care are crews. Several crews may work together closely, for example, in the operating room.6

Research in various high-risk industries found that high-performing crews exhibit different coordination patterns than low-performing crews, especially during the management of critical events. One crucial characteristic of high-performing crews is “adaptive coordination,” i.e., the adaptation of their coordination process to the situational requirements.7–9 For example, the occurrence of a critical event increases the coordination demands because it may require the crew to change priorities, set new goals, change task allocation, and restructure the crew. At the same time, several tasks have to be performed in parallel, increasing the demands for synchronizing activities, avoiding conflicts, and handing over tasks.10 At the same time, the crew’s ability to crossmonitor each other’s activities may be reduced. Under such circumstances, high-performing crews have been observed to increase their communication, specifically information exchange and verbalization of plans.7,9,11

In anesthesia, a few studies have investigated the structure and process of (adaptive) crew coordination.12–14 These studies show that anesthesia crews adapt their coordination process during anesthetic induction to the state of the patient12 and that a high level of standardization of the process was associated with more “implicit coordination” and less leadership.13 A recent study in cardiac anesthesia showed that anesthesia crews adapt their patterns of coordination to the changing situational requirements in the course of coronary artery bypass surgery.14 None of these studies provide empirical evidence on the relationship of coordination behavior and differences in clinical performance.

We conducted a simulation-based study investigating which coordination patterns support high clinical performance of two-person anesthesia crews treating a simulated episode of malignant hyperthermia (MH). We hypothesized that anesthesia crews adapt their coordination process to the occurrence of the simulated MH crisis. We expected to observe both structural changes (e.g., adding crew members, altering crew members’ roles) and procedural changes (e.g., proportion of time spent on different coordination behaviors). We also hypothesized that anesthesia crews with different levels of clinical performance exhibit different coordination patterns in response to the simulated MH crisis.


Study Participants and Simulator Scenario

We decided to study the relationship between coordination patterns and treatment performance in the management of MH for two reasons. First, MH is an anesthetic crisis known to be very resource intensive and requiring effective coordination under intense time pressure. Second, for MH a widely accepted treatment protocol is available which facilitates the evaluation of clinical performance.

The review of video material for this study was approved by the institutional review board. Written informed consent was obtained from all participants before recording.

We analyzed the work and coordination process and the clinical performance of two-person anesthesia crews in 24 MH scenarios videotaped during anesthesia crisis resource management (ACRM) courses in a full-scale patient simulator. This is a subset of data used in a previous study on cognitive aid use.15 Study participants were all first year anesthesia residents at Stanford University, participating in ACRM courses between 1998 and 2003. Each crew consisted of two first year residents; the total number of study participants was 48. Cases in which the number of anesthesia crew members exceeded two because of logistics of running the training course were not included in this study. Study participants had received no specific training in the management of MH before the simulation scenario. However, we could expect that all residents (even CA1s) have heard of MH and have some idea of how it is treated. Moreover, some participants may have read about MH crisis management before attending the ACRM course.

In the simulator scenario, care was handed-off to the primary anesthesia provider by a confederate anesthesiologist after anesthesia induction was complete. During the scenario, one other anesthesiologist of equal level of training (first responder) could be summoned for help in the management of the patient. The role of the surgeon and the circulating nurse were played by an experienced anesthesiologist instructor and by a retired operating room nurse. The role of the “scrub tech” was played by one of the participants. These role players supported the anesthesia crew in the management of the simulated MH crisis when asked to do so by the study participants. None of the anesthesia crews had help from any personnel except those mentioned above, although they could make phone calls to outside consultants.

The case scenario involved a general anesthetic on a healthy adult female undergoing elective orthopedic surgery. During surgery, the patient experienced an episode of MH and became progressively unstable over 15–20 min. For each case, two investigators (TM and TKH) jointly determined the transition from routine patient care and the management of a simulated crisis situation (i.e., Phase I and Phase II) using two behavioral indicators: (a) the participant(s) verbally stated a diagnosis of MH or (b) the participant(s) called for the MH cart.

Coding of Clinical and Coordination Activities

Clinical and coordination processes were coded by a psychologist experienced in conducting observational studies in operating room environments (real and simulated) using a personal digital assistant-recording device.16 Video recordings of real anesthesia inductions were used to train the coding of coordination behavior. During this training period, clinicians were available to discuss questions concerning the categorization of specific coordination activities. The observation system used in this study has been developed and piloted by our research group in a field study in cardiac anesthesia.14 It consisted of 36 mutually exclusive observation categories that were grouped into clinical activities, coordination activities, teaching, and other communication. The main categories for coordination activities were information management, task management, coordination via the work environment (i.e., coordination activities based on acoustic or visual coordination cues, such as crew members’ actions), and metacoordination (i.e., the “coordination of coordination”: coordination activities team members use to plan or organize how they would coordinate with each other). Each of these categories is subdivided into as many as 12 codes (for a comprehensive list of observation categories and codes including examples see web addendum 1 [available at]14). Interactions among the anesthesia crew, as well as between the anesthesia crew and other members of the perioperative team (i.e., surgeon, circulating nurse and other personnel, such as anesthesia technician) were recorded. Coordination activities by other members of the operating room team were only recorded when directed toward the anesthesia crew.

Coding Reliability

Data recording included concurrent events, such as activities of multiple people and multiple activities performed by one person simultaneously. Thus, we assessed the reliability of the coding of coordination activities using Cohen’s kappa statistics for concurrent event coding based on second-by-second comparisons at the level of observation codes.16 A single trained observer coded clinical and coordination activities. Six months later, the same observer rescored 10% of the Phase I and Phase II segments selected at random to assess coding reliability (i.e., intraobserver agreement over time). After visual inspection of the time-series of codings, they were compared both with an exact second-by-second comparison and with a comparison that included a lag of 2 s for onset and offset times.17

Clinical Performance Scoring

Because this study focused on the relationship of coordination patterns and clinical performance during the simulated crisis and not during the pathway to the diagnosis, clinical performance was assessed during Phase II only. Clinical performance scores in treating the episode of MH (unidimensional score 1–25, with 25 being the highest possible score) were calculated using a time-based scoring system for critical treatment steps (Table 2 and Ref. 15). Points were assigned for the various treatment steps needed to successfully treat an episode of MH as outlined by the Malignant Hyperthermia Association of the United States. For each treatment step, the time from articulating the diagnosis of MH until the successful completion of that step was measured by two of the authors (TKH and TM) and assigned between 1 and 3 points based on (a) the criticality of the treatment step and (b) the time to completion as defined in Table 1.

Table 1:
Malignant Hyperthermia Treatment Score Used for Clinical Performance Scoring15
Table 2:
Intraobserver Agreement Over Time for the Video-Based Coding of Coordination Activities Before (Phase I) and After the Declaration of a Simulated Anesthetic Crisis (Phase II) Given in Cohen’s kappa

Data Analysis

Observational data were prepared for statistical analysis in SPSS (Version 15.0) using customized software.16 The proportion of time spent on each observation category (percentages) was calculated for Phase I (average duration 21 min 52 s [±4 min 17 s]) and Phase II (average duration 20 min 40 s [±5 min 01 s]) separately. All aggregate data are shown as mean with standard deviations in parentheses unless otherwise specified.

To investigate whether anesthesia crews adapt their work and coordination patterns to the occurrence of the simulated MH crisis we (a) compared the average of concurrent activities between Phase I and Phase II (t-test for repeated measures) and (b) performed a multivariate analysis of variance (MANOVA) for repeated measures to compare the proportion of time spent on clinical work and the four main coordination categories (i.e., information management, task management, and coordination via the work environment and other communication) across phases.

To investigate the impact of coordination patterns on clinical performance in the management of the simulated MH crisis, we ran a series of hierarchical regression analyses with MH treatment scores as outcome variable. Hierarchical regression analysis allows for testing setwise partial association, i.e., testing whether the variance in the dependent variable (treatment score) can be explained by one variable alone (cognitive aid use, see first step), or if the new set of independent variables (coordination activities; see second step) explaining additional variance in the dependent variable.18 The order in which the predictors were entered was based on theoretical considerations and on the results of a previous empirical study indicating that cognitive aid use is positively related to treatment performance.

Only data from Phase II (i.e., after the declaration of the MH crisis) were considered in this analysis. In the first step of each hierarchical regression analysis, we included the variable “cognitive aid use” to statistically control for the previously demonstrated influence of cognitive aid use on clinical performance15 (i.e., anesthesia crews that use the cognitive aid more frequently during the management of MH yielded better treatment scores).

In a first hierarchical regression analysis, the three main coordination categories information management, task management, and coordination via the work environment were used as predictors. Metacoordination, other communication (only trace amounts observed during Phase II), and teaching were excluded from this step of statistical analysis. The observation system included teaching, but this was not observed during simulations (as opposed to real operating room cases) probably because in the simulations the anesthesia crew members had equal levels of training and no teaching obligation was imposed on them. In subsequent hierarchical regression analyses, we used the specific coordination behaviors detailing each of the three main observation categories (i.e., information management, task management, and coordination via the work environment) as predictors. For example, task management was detailed by the coordination behaviors planning, task distribution, clarification, initiate an action, and assistance (see also web addendum 1 available at

To investigate whether, during the initial management of a simulated crisis, different coordination patterns predict high performance, we ran separate hierarchical regression analyses using data from the first 5 min after declaration of the crisis only. In a final hierarchical regression analysis, we used the proportion of time spent on coordination activities among the anesthesia crew and between the anesthesia crew and other team members as predictors.

One case was excluded from all hierarchical regression analyses based on the results of an outlier analysis. First, visual inspection of partial regression diagrams showed that the same case was an outlier for all information management categories. Second, we confirmed the results of this visual inspection by testing whether this case fits the regression model when included. The outlier did not fit the model for three coordination behaviors within the information management category (i.e., situation assessment, decision making and feedback/acknowledgment). Excluding this outlier reduced the sample size for all hierarchical regression analyses to n = 23.

In addition, qualitative analyses of the video material were conducted to aid the interpretation of the quantitative results. We selected six cases for qualitative review: the three highest and the three lowest scoring crews. Each simulation scenario was reviewed with an experienced anesthesiologist and ACRM instructor (SKH) who commented on all episodes that showed examples of good or poor coordination discussing the context, the content, and any observable effects of the coordination behavior. The coordination episodes highlighted by the expert reviewer were then transcribed verbatim and sorted into thematic groups. The results of this qualitative analysis were then used to understand the patterns of clinical and coordination activity over time in each case.


Coding Reliability for Clinical and Coordination Activities

Overall, the reliability of clinical and coordination activity coding was good to excellent19 (Table 2). Applying the 2 s tolerance rule for onset and offset times,17 the fair or poor kappa values for some observation categories in the second-by-second comparison (i.e., those categories with a mean duration of 2–6 s) improved to good or excellent agreement.19 This improvement indicates that disagreement was mostly because of slight phase shifts in the onset and offset times and not because of content.

Changes in Work and Coordination Patterns Before and After the Declaration of a Simulated Anesthetic Crisis (Phase I Versus Phase II)

Our results show that anesthesia crews adapted their work and coordination patterns to the simulated crisis. A t-test for repeated measures comparing the average of concurrent activities between Phase I and Phase II showed a statistically significant increase in concurrent activities (t [23] = −18.50, P < 0.001) from an average of 1.11 (±0.09) concurrent activities during Phase I to 2.14 (±0.29) during Phase II (Note: Concurrent activities include activities of multiple people and multiple activities performed by one person simultaneously). Clinical activities increased from an average of 0.77 (±0.09) in Phase I to 1.42 (±0.31) in Phase II indicating concurrent clinical activities performed by the anesthesia crew during Phase II. Figure 1 summarizes the proportion of time spent on the various coordination activities during Phases I and II. The most dramatic increase (fivefold) was observed in task management. Coordination via the work environment and other communication (i.e., social communication not related to the task) were both decreased in response to the crisis situation.

Figure 1.:
Changes in coordination patterns before (Phase I) and after the declaration of a simulated anesthetic crisis (Phase II) (mean and sd) (n = 24).

This increase in activity is also illustrated in Figures 2 and 3 depicting the distribution of clinical and coordination activity in the course of the MH scenario for one top and one bottom scoring anesthesia crew. Whereas clinical and coordination activities were performed sequentially during Phase I, these activities were performed concurrently for most of Phase II. This pattern was more pronounced for the higher scoring crew (Fig. 2).

Figure 2.:
Distribution of clinical activities (black line) and coordination activities (gray line) for a high-performing anesthesia crew (Case 4; malignant hyperthermia treatment score 25 of 25 possible points). Clinical activities (moving average 60 s); coordination activities (moving average 60 s).
Figure 3.:
Distribution of clinical activities (black line) and coordination activities (gray line) for a low-performing anesthesia crew (Case 3; malignant hyperthermia treatment score 8 of 25 possible points). Clinical activities (moving average 60 s); coordination activities (moving average 60 s).

To investigate whether anesthesia crews adapt their clinical and coordination activities to the occurrence of the simulated MH crisis, we performed a MANOVA for repeated measures (i.e., Phases I and II), including the proportion of time spent on clinical activities as well as on information management, task management, and coordination via the work environment as dependent variables. The MANOVA yielded a statistically significant effect of Phase (F [5,19] = 162.81, P < 0.001, ηp2 = 0.98). Subsequent univariate analysis showed that this effect was statistically significant for all dependent measures (Table 3).

Table 3:
Results of the Univariate Tests for the Dependent Measures Included in the MANOVA for Repeated Measures Comparing Clinical Activities, Task Management, Information Management, Coordination via the Work Environment, and Other Communication Across Phases

When interpreting these changes in work and coordination patterns (i.e., procedural adaptation), changes in anesthesia crew composition (i.e., structural adaptation) have to be considered. Once the crisis situation was declared, all primary anesthesiologists requested and received assistance from a second anesthesiologist (first responder). However, increased resources alone cannot explain the shift in the proportion of time spent on coordination categories because, for example, the proportion of time spent on information management more than doubles and the proportion of time spent on task management increases fivefold (Fig. 1).

Coordination Patterns Associated with Higher Clinical Performance Scores (Phase II)

To investigate the relationship between coordination activities and clinical performance in the management of the simulated MH crisis, we ran a series of hierarchical regression analyses with MH treatment scores as outcome variables using data from Phase II only (i.e., after the declaration of the MH crisis). In this step of statistical analysis, one case was excluded based on outlier analysis. The average MH treatment score yielded by the 23 anesthesia crews that were included in the hierarchical regression analyses was 20.35 (±3.02). Three anesthesia crews reached the highest possible score of 25 points and the lowest performing crew scored 14.

In a first hierarchical regression analysis, the three main coordination categories information management, task management, and coordination via the work environment were used as predictors and the variable “cognitive aid use” was included in the first step. This hierarchical regression analysis showed that the proportion of time spent on task management predicted clinical performance during the management of the simulated MH episode (Phase II); the proportion of time spent on task management was significantly lower for anesthesia crews with higher clinical performance scores (Table 4).

Table 4:
Results of the Hierarchical Regression Analysis Including the Three Main Coordination Categories as Predictors for Malignant Hyperthermia Treatment Scores (n = 23) (One Outlier Excluded from the Analysis)

The qualitative video analysis supported the results of the quantitative analysis of our observational data (Figs. 2 and 3 for the distribution of clinical and coordination activities for one of the top and the bottom scoring crew). The qualitative analysis revealed that anesthesia crews with higher clinical performance scores were more specific about work roles (in some crews we observed an explicit transition of the leadership role) and prioritized clinical tasks effectively. Efficient coordination thus decreased task management requirements and contributed to less task distribution especially with nonanesthesia team members. Low-performing anesthesia crews, however, failed to prioritize clinical tasks clearly, thereby increasing workload and task management requirements. As illustrated in Figure 3, we observed an upward trend in coordination activity (mainly task management) for the bottom-scoring crew, whereas clinical tasks were performed only periodically and clinical activity finally decreased below one with two anesthesia providers present.

In subsequent hierarchical regression analyses, we used the specific coordination behaviors detailing each of the three main observation categories as predictors. These analyses showed that (a) within the task management categories, the proportion of time spent on task distribution was the best predictor for clinical performance (Table 5) and that (b), although the proportion of time spent on information management overall did not predict clinical performance, situation assessment showed a statistically significant positive relationship with clinical performance (Table 6). None of the observation categories detailing coordination via the work environment showed a statistically significant relationship with clinical performance. Although not statistically significant, it seems important to point out that the proportion of time spent on monitoring of other anesthesia crew members, a coordination behavior aiding in the maintenance of a shared mental model, showed a positive correlation with MH treatment scores (β = 0.251, ns).

Table 5:
Results of the Hierarchical Regression Analysis Including the Five-Task Management Categories as Predictors for Malignant Hyperthermia Treatment Scores (n = 23) (One Outlier Excluded from the Analysis)
Table 6:
Results of the Hierarchical Regression Analysis Including the Four Information Management Categories as Predictors for Malignant Hyperthermia Treatment Scores (n = 23) (One Outlier Excluded from the Analysis)

Further hierarchical regression analysis using the proportion of time spent on coordination activities among the anesthesia crew and between the anesthesia crew and other team members as predictors revealed that situation assessment within the anesthesia crew was especially associated with higher performance scores (β = −0.369, P < 0.05). This result was also supported by the qualitative video analysis. Higher performing crews focused on the coordination of the anesthesia work process, keeping other team members informed, but not getting distracted by extensive conversation. They showed more situation assessment and provided more unsolicited information (in part by “thinking out loud”), which helped them to maintain a shared mental model.20 In contrast, the lower scoring anesthesia crews were more likely to split into independent subcrews (i.e., each anesthesiologist worked with nonanesthesia team members) that did not coordinate effectively the resources that they mobilized.

Also, when analyzing the first 5 min after declaration of the crisis separately, hierarchical regression analysis showed that, in contrast to the results based on data from the whole crisis management period (Phase II), information transfer was the only observation category predicting high clinical performance scores (β = 0.359, P < 0.05). The critical importance of information transfer in an early stage of crisis management was also illustrated by the qualitative analysis of Case 4 (Fig. 2). Once the crisis was declared and the first responder joined the anesthesia crew (Phase II), the top scoring crew exhibited a marked increase in information transfer followed by a short episode of task management (i.e., planning and task distribution). After this initial coordination episode (approximately 3 min) during which almost no clinical tasks were performed, the anesthesia crew focused on clinical tasks (both anesthesia providers working simultaneously) while keeping each other informed, then reassessed the situation, and initiated another cycle of clinical tasks.


This full-scale simulator study provides empirical evidence that anesthesia crews dynamically adapt to the situational task requirements. Once the simulated anesthetic crisis was declared, we observed (a) structural adaptation by mobilizing additional resources and in some anesthesia crews a transition of leadership from the primary anesthesiologist to the first responder and (b) procedural adaptation through increased clinical and coordination activity and new patterns of coordination. Moreover, this study for the first time identified coordination patterns that are associated with higher and lower clinical performance scores in the management of a simulated anesthetic crisis.

Despite the difficulty of comparing findings among studies that use different systems to describe coordination behavior, we believe that our study is consistent with other research on coordination processes and (clinical) performance in (medical) crews or teams. For example, research in various (mostly nonclinical) settings has highlighted the importance of verbalizing situation assessments to maintain good coordination.21,22 We therefore believe that this coordination mechanism should be a focus during team training addressing the role of the sender and the receiver of verbal communication.

Furthermore, our finding that different coordination patterns were most effective during different stages of crisis management (e.g., information transfer predicted high clinical performance scores only during the first 5 min of Phase II) is in line with research highlighting the importance of time in the analysis of adaptive teamwork.23,24 However, our observation that lower performing crews spent a larger proportion of time on task management, specifically on task distribution (Table 5), partly contradicts research on trauma team performance that showed that a lack of explicit coordination (especially in directions and task delegation) frequently lead to coordination breakdowns during nonroutine situations.9 This contradiction points to a critical area of team research: the optimal balance between time spent on coordination and time spent on task-related activities. Insufficient coordination can result in duplication of effort or counterproductive actions but, on the other hand, if all crew members “know what to do,” excessive time spent coordinating will only delay the performance of critical tasks. Future studies also need to address the role of the composition of the anesthesia crew and the experience level of crew members as potential moderators of the effectiveness of coordination patterns. In a previous study, we showed that anesthesia attendings relied more on implicit modes of coordination than anesthesia residents, even during critical episodes of cardiac anesthesia.14 It is possible that the lower performing crews in this study had to spend more time on explicit task management because they lacked effective implicit coordination strategies, or individually they were less sure of what to do.

The in-depth analysis of outliers is a promising approach to further investigate characteristics of poor performance or the difference between “imperfect but good enough” and “unsafe performance.” Unfortunately, this question cannot be answered based on the single outlier identified in this study. However, the outlier analysis in this study showed that a crew can be different from others in one or two aspects of their coordination behavior without a dramatic effect on performance, whereas the outlier was differed from the other crews in many aspects, including the two that were statistically significant in our study.

This study has a number of limitations. First, we assessed clinical performance based on adherence to and timely implementation of a treatment protocol for MH, although we do not know the exact effects of different levels of nonadherence or delays in treatment on patient outcome. Better linkage between crew performance and outcome would be helpful but can only be measured in real cases, which for most life-threatening situations are both rare and unscheduled, making such studies at this time impossible to perform.

Second, success in a complex activity such as anesthesia can be achieved in different ways and is dependent on many variables. In this study, we statistically controlled use of a cognitive aid that has been shown to impact MH treatment performance15 (i.e., anesthesia crews that use the cognitive aid more frequently during the management of MH yielded better treatment scores). However, we did not systematically investigate how cognitive aid use affected coordination patterns. Additional factors, such as individual clinical knowledge and skill or the perception of the primary anesthesiologists’ competence by the first responder, may affect both clinical performance and coordination patterns.

Another limitation is that we studied coordination only in the management of one clinical event, MH, for which a clear treatment protocol is available. At first sight, our finding that higher scoring crews focused on coordinating among the anesthesia crew seems antithetical to ACRM principles, highlighting the importance of keeping all team members “in the loop.”6 However, MH is primarily an anesthesia-related crisis; successful coordination strategies for other types of critical incidents may be different. Including a spectrum of medical crises with different characteristics in future studies may clarify the generalizability of our findings.

Finally, we studied artificial anesthesia crews of first year residents instead of anesthesia crews containing clinicians of varying levels of experience. Thus, study participants were inexperienced at managing a whole team during a medical crisis and were not fully trained professionals. As such, they were not experienced at taking the top leadership role. In this retrospective study, however, it was not possible to include participants with different levels of experience working together in “real” crews. Future research on true combined teams with all levels of team members may show different coordination patterns, especially when focusing on leadership25,26 and on implicit forms of coordination, such as coordination via the work environment that is an important coordination mechanism for experienced clinicians.14

Comprehensive studies of team performance during complex simulations or real crisis situations, especially studies focusing on the transition from a routine anesthetic patient care to the management of a crisis situation, will contribute to a better understanding of the coordination-performance relationship and finally to the development of specific coordination training to further improve performance.


1. Donchin Y, Gohper D, Olin M, Badihi Y, Biesky M, Sprung C, Pizov R, Cotev S. A look into the nature and causes of human errors in the intensive care unit. Crit Care Med 1995;23:294–300
2. Lingard L, Espin S, Whyte S, Regehr G, Baker GR, Reznick R, Bohnen J, Orser B, Doran D, Grober E. Communication failures in the operating room: an observational classification of recurrent types and effects. Qual Saf Health Care 2004;13:330–4
3. Sutcliffe KM, Lewton E, Rosenthal ME. Communication failures: an insidious contributor to medical mishaps. Acad Med 2004;79:186–94
4. Sundstrom E, De Meuse KP, Futrell D. Work teams: applications and effectiveness. Am Psychol 1990;45:120–33
5. Webber SS, Klimoski RJ. Crews: a distinct type of work team. J Business Psychol 2004;18:261–79
6. Gaba DM, Howard SK, Fish K, Smith BE, Sowb YA. Simulation-based training in anesthesia crisis resource management (ACRM): a decade of experience. Simul Gaming 2001;32:175–93
7. Brehmer B. Dynamic and distributed decision making. J Fire Service Coll 1996;1:17–36
8. Entin E-E, Serfaty D. Adaptive team coordination. Hum Factors 1999;41:312–25
9. Xiao Y, the LOTAS Group. Understanding coordination in a dynamic medical environment: methods and results. In: McNeese M, Salas E, Endsley MR, eds. New trends in collaborative activities: understanding system dynamics in complex environments. Santa Monica: Human Factors and Ergonomics Society, 2001:242–58
10. Kontogiannis T, Kossiavelou Z. Stress and team performance: principles and challenges for intelligent decision aids. Saf Sci 1999;33:103–28
11. Orasanu J, Salas E. Team decision making in complex environments. In: Klein G, Orasanu J, Calderwood R, Zsambok C, eds. Decision making in action: models and methods. Norwood, NJ: Ablex Publishing Corp., 1993:327–245
12. Hindmarsch J, Pilnick A. The tacit order of teamwork: collaboration and embodied conduct in anaesthesia. Sociol Quarterly 2002;43:139–64
13. Grote G, Zala-Mezö E, Grommes P. The effects of different forms of co-ordination on coping with workload. In: Dietrich R, Childress TM, eds. Group interaction in high risk environments. Aldershot: Ashgate Publishing, 2004:39–55
14. Manser T, Howard SK, Gaba DM. Adaptive coordination in cardiac anaesthesia: a study of situational changes in coordination patterns using a new observation system. Ergonomics 2008;51:1153–78
15. Harrison TK, Manser T, Howard SK, Gaba DM. Use of cognitive aids in a simulated anesthetic crisis. Anesth Analg 2006;103:551–6
16. Held J, Manser T. A PDA-based system for online recording and analysis of concurrent events in complex behavioral processes. Behav Res Meth Instru Comput 2005;37:155–64
17. Bakeman R, Gottman JM. Observing interaction: an introduction to sequential analysis. Cambridge: Cambridge University Press, 1997
18. Hayes AF. Statistical methods for communication science. Mahwah NJ: Erlbaum, 2005
19. Fleiss JL. Statistical methods for rates and proportions. New York: Wiley, 1981
20. Mathieu JE, Heffner TS, Goodwin GF, Salas E, Cannon-Bowers JA. The influence of shared mental models on team process and performance. J Appl Psychol 2000;85:273–83
21. Johannesen LJ, Cook RI, Woods DD. Cooperative communications in dynamic fault management. Proceedings of the Human Factors and Ergonomics Society 38th Annual Meeting. Santa Monica, 1994:225–9
22. Salas E, Prince C, Baker DP. Situation awareness in team performance: implications for measurement and training. Hum Factors 1995;37:123–36
23. McGrath JE. Time, interaction, and performance (TIP). A theory of groups. Small Group Res 1991;22:147–74
24. Tschan F, Semmer N, Nägele C, Gurtner A. Task adaptive behavior and performance in groups. Group Processes Intergroup Relations 2000;3:367–86
25. Tschan F, Semmer N, Gautschi D, Hunziker PR, Spychiger M, Marsch S. Leading to recovery: group performance and coordinative activities in medical emergency driven teams. Hum Perform 2006;19:277–304
26. Cooper S, Wakelam A. Leadership of resuscitation teams: “Lighthouse Leadership”. Resuscitation 1999;42:27–45
© 2009 International Anesthesia Research Society