Psychology by BogdanovaThe Encyclopedia of Psychology for the Artificial Era

RESEARCH DOSSIER

Human Team Coordination and Crew Resource Management

Abstract. Human team coordination and crew resource management are examined as an integrated psychological and sociotechnical research object involving interdependence, information exchange, collective monitoring, role allocation, adaptive leadership, error management, and operational learning. This critical narrative monograph traces the historical development of CRM in aviation, distinguishes process theories from field interventions, reviews observational studies, training evaluations and healthcare meta-analyses, and examines the limits of transferring practices across high-consequence settings. It synthesizes evidence on shared mental models, closed-loop communication, authority gradients, situation awareness, handoffs, behavioral marker systems and simulated and real operational outcomes. The principal synthesis treats effective coordination as a dynamic control process in which compatible expectations are maintained, discrepancies become detectable, and corrective actions remain possible. Contemporary implications include the design of human–Artificial operational teams. The review makes no claim to an exhaustive protocol-based systematic review or new pooled estimates.

Keywords: crew resource management; CRM; team coordination; human factors; safety-critical teamwork; shared mental models; closed-loop communication; situation awareness; mutual performance monitoring; team training; nontechnical skills; threat and error management; handoffs; psychological safety; human–Artificial cooperation.

Introduction: Coordination as an Operational Achievement

Human team coordination is the situated organization of interdependent human actions toward a jointly consequential goal. It comprises the distribution of responsibilities, timely exchange of consequential information, mutual monitoring, adjustment of activity, and recovery when assumptions or plans become invalid. A crew is coordinated when its members can construct compatible expectations about the task, act within understood boundaries of responsibility, discover discrepancies, and revise the collective course of action before those discrepancies propagate into harm. The core scientific object is an unfolding pattern of team behavior within a technical and institutional work system. It cannot be reduced to interpersonal rapport, the excellence of the most experienced operator, or the existence of a written procedure.

Crew resource management (CRM) denotes a family of training and operational practices developed in aviation to improve the use of available human, informational, and material resources. Its historical emphasis on cockpit communication, decision making, leadership, workload distribution, and error management has expanded to interprofessional teams in other high-consequence settings. The historical review by Helmreich, R. L., Merritt, A. C., & Wilhelm, J. A (1999) documents this development, including the transition from cockpit to crew resource management and the limits of generalizing training across cultures. Salas, E., Wilson, K. A., Burke, C. S., & Wightman, D. C (2006) ask the central evaluation question: which aspects of CRM training improve attitudes, knowledge, observable behavior, or operational outcomes, under which conditions?

This monograph treats coordination as a multilevel process connecting individual cognition, interpersonal action, team structure, technological affordances, and organizational policy. It is a critical narrative synthesis of identifiable original studies, theories, regulatory guidance, and research reviews; it involves no new participants, original effect-size pooling, or claim of a protocol-based systematic search. The guiding questions concern how teams synchronize under uncertainty, how specific interventions change performance, how coordination can be measured validly, and when transfer across work environments is justified. These questions organize the following chapters and distinguish established associations from causal claims and from theoretical proposals.

Defining Teams, Crews, and Coordination

A team is a bounded or functionally recognizable set of people whose work is interdependent and whose activity contributes to a consequential common objective. Membership may be stable, as in a flight crew on a long-haul operation, or temporary and rapidly reconfigured, as in a trauma resuscitation. A crew is a particularly operational form of team in which roles and duties are embedded in equipment, procedures, time-critical decisions, and an accountability structure. Coordination is the set of processes through which those dependencies are actively managed. Cooperation, teamwork climate, and collective identification may support coordination, but they are distinct constructs and require different measures.

A useful distinction separates taskwork—performing technical operations—from teamwork—organizing how members' operations fit together. A technically correct action may arrive too late, be based on a superseded plan, or fail because another person did not receive a critical update. Conversely, apparently fluent discussion may conceal missing data or unjustified consensus. Marks, M. A., Mathieu, J. E., & Zaccaro, S. J (2001) provide a temporally based account of team process, while Salas, E., Sims, D. E., & Burke, C. S (2005) specify interlocking behavioral components and coordination mechanisms. Both frameworks direct research toward observable episodes rather than global labels such as a 'good team.'

For operational research, team coordination can be defined at the event level using who holds information, who is responsible for a task, how an instruction is acknowledged, when a discrepancy is noticed, and what corrective action follows. At the team level, the construct may concern stable routines or the distribution of adaptive capacity. At the organizational level, it concerns staffing, policy, technology, training, and opportunities to report defects. These levels interact but must remain analytically identifiable to prevent a positive culture survey from being mistaken for evidence of safe performance during a specific emergency.

Historical Origins of Crew Resource Management

Modern CRM developed from aviation human-factors work investigating accidents in which technical capability existed but the coordination of information and authority failed. NASA's 1979 workshop, Resource Management on the Flightdeck, is a widely documented milestone. The NASA Aviation Safety Reporting System recounts how research at Ames and investigation of accidents highlighted communication, leadership, and decision making; the workshop stimulated structured training in the airline industry (National Aeronautics and Space Administration, Aviation Safe (2014)). The historical source requires precision: antecedent work in European aviation and earlier crew-training practices existed, so 1979 marks an institutional consolidation rather than a sudden invention of every relevant idea.

Helmreich, R. L., Merritt, A. C., & Wilhelm, J. A (1999) describe successive generations of CRM, including an early focus on cockpit interpersonal behavior, later attention to organizational culture and whole crews, and an error-management orientation. This change mattered conceptually. The operational question became how to avoid threats when possible, detect errors after they occur, and mitigate consequences before an unsafe state becomes irreversible. CRM also developed alongside line-oriented flight training, realistic simulations, and procedures defining the duties of both pilots. The Federal Aviation Administration (2004) and its Federal Aviation Administration (2017) are institutional documents, not experiments demonstrating a single numerical treatment effect.

Historical explanations of accidents should be read with care. Investigative narratives can identify coordination vulnerabilities, but they do not identify counterfactual outcomes under a different training program. A disciplined history compares contemporaneous research, subsequent organizational revisions, changes in pilot monitoring, and the evolution of evaluation methods. The expansion of CRM beyond flight decks further illustrates that a shared training label does not guarantee that equivalent mechanisms, risks, team boundaries, or measures are involved.

Task Interdependence and the Architecture of Joint Work

Coordination becomes necessary because a member's action depends on another person's knowledge, timing, approval, or output. Pooled work aggregates separable contributions; sequential work requires ordered handoffs; reciprocal work requires continuing mutual adjustment. These arrangements create different vulnerabilities. A missed handoff can disrupt sequential interdependence, while delayed feedback can destabilize reciprocal control. The laboratory and field literature on group effectiveness emphasizes that task design and reward design shape behavior. Wageman, R (1995) examined how patterns of interdependence in organizational groups relate to their effectiveness and illustrates why exhortation alone is an incomplete intervention.

In high-consequence systems, dependencies are partly engineered. Checklists, communication protocols, cockpit displays, physical room layout, and authority rules alter which person can observe which change. The distribution of authority affects whether an anomaly is treated as information, insubordination, or a request for collective reassessment. Consequently, the same training content can function differently when delivered to a well-staffed crew with stable instrumentation and to an understaffed unit operating with frequent interruptions and multiple competing priorities.

A rigorous work analysis maps the task, the roles, the timing of transitions, and the location of information before attributing difficulties to personality or a generalized teamwork deficit. An intervention should identify the dependency it intends to change: for example, a standardized read-back closes an acknowledgment loop; a named coordinator resolves ambiguity about prioritization; a brief changes the distribution of knowledge before the action phase begins. This mechanism-first approach improves explanatory clarity and permits comparison among interventions that share a name but operate through different causal paths.

Temporal Dynamics of Team Processes

Team processes vary across an operational episode. Marks, M. A., Mathieu, J. E., & Zaccaro, S. J (2001) distinguish transition processes, action processes, and interpersonal processes. Transition processes include mission analysis, planning, and strategy formulation. Action processes include progress monitoring, systems monitoring, team monitoring, and coordination of operations. Interpersonal processes, such as conflict management and motivational regulation, operate during both preparation and execution. This temporal framework accounts for why an after-action debrief cannot replace a pre-operation shared plan and why a plan alone cannot repair an unexpected failure during action.

The later LePine, J. A., Piccolo, R. F., Jackson, C. L., Mathieu, J. E (2008) supports a multidimensional organization of teamwork processes and finds positive relations with performance and satisfaction. The magnitude and structure of those associations vary with such features as team size and task interdependence. This evidence supports the importance of processes while leaving open the question of which process, changed by which intervention, produces better operational outcomes in a particular domain.

Temporal sampling is consequently central to measurement. A crew might demonstrate excellent planning and weak cross-monitoring, or poor early shared understanding followed by rapid recovery. One summary rating after the exercise conceals these trajectories. Event-based observation can identify when a team's operating model diverges, whether the divergence becomes visible, and how long corrective coordination takes. The timing of the same communication behavior can determine its utility: a callout before a critical transition may avert a mistake, while the identical callout after a system state changes may add distraction.

The Five Teamwork Components and Their Coordination Mechanisms

Salas, E., Sims, D. E., & Burke, C. S (2005) organize teamwork around team leadership, mutual performance monitoring, backup behavior, adaptability, and team orientation. Their synthesis also emphasizes shared mental models, closed-loop communication, and mutual trust as coordinating mechanisms. This framework is especially productive for CRM because it separates the behavioral work of teams from the cognitive and relational resources that make that work possible. It also discourages treating a single variable, such as assertiveness, as a comprehensive surrogate for competent crew coordination.

Leadership organizes goals and resource allocation; monitoring makes relevant deviations detectable; backup behavior reallocates assistance; adaptability changes strategies when conditions change; and team orientation orients members toward shared rather than isolated results. The mechanisms are interdependent. Monitoring without clear communication may fail to influence the operator. Communication without a shared interpretation of the work may be precise but irrelevant. Trust without independent verification may increase premature acceptance of another member's estimate.

The framework provides constructs for training objectives, behavioral marker design, and experimental manipulation. Its propositions should be examined as a model, with the distinction between a theoretically appealing component structure and a validated measure of each component preserved. In particular, the relevant performance criterion should be specified before interpreting a high teamwork rating: correct decision, time to correct an error, avoided hazard, patient outcome, or endurance over repeated missions are scientifically different outcomes.

Shared Mental Models and Team Knowledge

A shared mental model is an organized overlap in members' representations of the task, equipment, roles, or likely sequences of events. Its functional value arises when one member can anticipate another's information requirements and actions. Shared representation does not require every crew member to know everything. Effective teams also rely on differentiated expertise and on accurate knowledge of who possesses which information. The central requirement is sufficient compatibility of expectations to coordinate work, with room for specialized knowledge and independent judgment.

The evidence base includes the DeChurch, L. A., & Mesmer-Magnus, J. R (2010) and their separate DeChurch, L. A., & Mesmer-Magnus, J. R (2010). The latter synthesized 23 independent studies and showed that relationships depend partly on how knowledge representations are elicited and aggregated. Measures attentive to the structure of knowledge were particularly informative for understanding team processes. Treating a collection of individually answered quiz items as a direct observation of a working crew's joint cognition would therefore be a validity error.

Shared models can be assessed before and after training through concept sorting, paired comparisons, scenario judgments, role-expectation probes, and event-based prediction tasks. The measurement design should distinguish agreement on correct information from agreement on a common error. A team could be highly aligned around a mistaken assumption. Research should therefore assess both the configuration of knowledge and the quality of its updating when discrepant evidence appears.

Situation Awareness and Teamwide Monitoring

Situation awareness is concerned with the perception, comprehension, and projection of relevant environmental states. Endsley, M. R (1995) describes the role of attention, working memory, goals, mental models, automation, and system complexity in dynamic operations. In team settings the problem expands: members may have access to different displays and distinct portions of the operating environment. Teamwide awareness depends on communicating what other members need to know, aligning interpretations, and maintaining a workable distribution of monitoring responsibilities.

A crew may divide monitoring by subsystem or professional role. Such division saves attention but makes boundary failures possible: one member assumes a condition is being watched while another assumes the first member is responsible. Effective coordination therefore requires explicit assignment of critical observations, visible state changes, and a route for urgent discrepancies to reach the person with decision authority. A shared display can assist, but the mere presence of identical information does not establish common comprehension.

Methods should distinguish awareness as a cognitive state from communication as a behavior and from operational success as an outcome. Querying members during simulator pauses may capture knowledge but interrupt task performance; retrospective reports risk reconstruction after the event. Behavioral observations capture whether relevant updates were made, though silence can sometimes reflect accurate anticipation rather than ignorance. Mixed measurement strategies and careful timing provide stronger inference than a single questionnaire.

Explicit Communication, Closed Loops, and Read-Back

Communication is a mechanism for changing the distribution and interpretation of operational knowledge. In a closed loop, a sender communicates relevant information or an instruction, a receiver acknowledges or repeats the substance, and the sender confirms alignment when appropriate. Closed-loop communication is especially important when misunderstanding would be costly and the message concerns a critical system state, a dose, a position, a time, or a planned action. Salas, E., Sims, D. E., & Burke, C. S (2005) identify it as a central coordination mechanism rather than a stylistic preference.

A communication episode can fail through omission, ambiguous referent, incorrect recipient, delayed transmission, or misunderstood content. Lingard, L., Espin, S., Whyte, S., Regehr, G., Baker, G. R (2004) used direct operating-room observation to classify recurrent communication failures and their effects. The study anchors a concrete empirical approach: observe actual teams in their environment, identify what information was needed, classify the failure, and examine whether a process was disrupted. This is more informative than using the absolute amount of speech as an indicator of team effectiveness.

Communication density can be excessive. During a critical maneuver, redundant statements may compete with visual monitoring and procedural execution. Research should measure message relevance, timing, addressee, acknowledgment, correction, and downstream behavior. An efficient, quiet phase can be highly coordinated when roles and state are already aligned; when uncertainty rises, a deliberately explicit communication mode may become necessary. The empirical target is functional information transfer at the right time.

Implicit Coordination and Anticipatory Adjustment

Experienced teams often anticipate one another's needs before a request is spoken. This is implicit coordination: timely adjustment based on shared knowledge of tasks, roles, and the evolving environment. Rico, R., Sánchez-Manzanares, M., Gil, F., & Gibson, C (2008) propose a team-knowledge-based account in which situation models and contextual conditions support anticipatory behavior. The model helps explain how teams can remain efficient without describing every routine step aloud.

Implicit coordination becomes vulnerable when a team is newly assembled, the task departs from normal conditions, role expectations differ, or an unexpected automation mode changes what is possible. Training may support anticipation through exposure to likely contingencies, but scripts must be flexible enough to recognize that a surprising state demands explicit communication. Research can test this through controlled manipulations of familiarity, workload, or procedural novelty while coding requests, spontaneous backup, timing, and errors.

A mature CRM system therefore combines implicit adjustment during stable operations with deliberate verbal confirmation at specified transitions and when assumptions become uncertain. The appropriate balance is task-dependent. Research should avoid scoring every unspoken action as a communication deficit, just as it should avoid interpreting apparently effortless performance as proof that all members possessed accurate information.

Authority Gradients, Speaking Up, and Psychological Safety

High-consequence teams often require clear authority for final decisions. That authority can coexist with reliable mechanisms for subordinate expertise to correct a mistaken interpretation. An authority gradient becomes a safety concern when status differences prevent relevant information from entering the decision process. CRM interventions have historically addressed the capacity to question, challenge, and escalate without ambiguity about who makes a final operational choice. The scientific issue is whether the organization gives a legitimate route for discrepant evidence to be heard and resolved.

Edmondson, A (1999) of 51 work teams examined psychological safety, learning behavior, and performance, reporting an association between shared beliefs about interpersonal risk and team learning. This work provides important organizational evidence but does not establish that a psychological-safety survey score directly causes lower accident rates in every aviation or medical crew. Safety for voice needs to be examined alongside expertise, actual challenge behavior, escalation protocols, and institutional responses.

Assertiveness training is more specific than encouragement to 'speak up.' It can teach how to state a concern, explain the operational evidence, obtain acknowledgment, repeat it if unresolved, and invoke a defined escalation path when necessary. Researchers can manipulate authority cues and urgency in simulations, then measure the probability and timing of a challenge and the correction of an unsafe plan. The goal is an accountable channel through which warranted dissent can change collective action.

Leadership, Delegation, and Role Clarity

Crew leadership establishes mission priorities, makes decisions when coordination requires central authority, and preserves a clear account of role boundaries. The appropriate pattern of leadership can change as a task unfolds. Klein, K. J., Ziegert, J. C., Knight, A. P., & Xiao, Y (2006) investigated extreme action medical teams and described dynamic delegation: senior leaders temporarily transfer active leadership to developing members while remaining able to intervene. Their qualitative observations illustrate how hierarchical responsibility and flexible execution can be integrated under demanding conditions.

Role clarity concerns both responsibility and authority. A named team leader may fail to coordinate if no one knows who monitors the equipment, communicates with outside personnel, or completes the handoff record. Conversely, members can perform smoothly under distributed leadership when the transfer of authority is visible and decision rights are shared in advance. The distinction matters for crisis design: a team can require one person to make a time-critical call while empowering others to surface conditions that change that call.

Leadership research should measure actual allocation and transfer of responsibilities, not merely follower approval or personal charisma. Observable indicators include initiation of a brief, confirmation of priorities, recognition of overload, solicitation of discrepant information, and explicit delegation and recovery of control. Dynamic delegation also creates tradeoffs between current mission speed, novice learning, and risk. The causal value of a particular pattern must therefore be tested against the complexity and consequences of the operational setting.

Mutual Performance Monitoring and Backup Behavior

Mutual monitoring is the capacity of team members to recognize when another person's action or the shared system state diverges from the agreed plan. Backup behavior is an adaptive response that provides help, redistributes work, or compensates for reduced capacity. They are related but distinct. A person may notice an error yet lack a safe way to intervene, or may offer unnecessary assistance that displaces monitoring of a different hazard. The teamwork model of Salas, E., Sims, D. E., & Burke, C. S (2005) makes their relationship explicit.

In aviation, the pilot monitoring role gives formal status to checking aircraft trajectory, guidance modes, procedures, and deviations. Federal Aviation Administration (2017) emphasizes the importance of well-defined monitoring duties and training. This guidance describes an operational control policy; its issuance is not itself a randomized assessment of monitoring efficacy. The underlying mechanism is a detection-and-response loop: an independent observation changes a decision before the system crosses a dangerous threshold.

Experimental research could vary workload allocation, information redundancy, and the accessibility of override action, then measure detection latency, false alarms, recovery time, and total task performance. A team-level outcome must include the work left undone by a backup action. Effective assistance is contingent, timely, and compatible with continuing responsibility for the rest of the system.

Workload, Stress, Fatigue, and Time Pressure

High workload narrows opportunities for monitoring and communication, yet low workload can also reduce vigilance. Coordination provides a means to redistribute demands, but redistribution itself consumes attention: someone must notice overload, identify available capacity, request help, and synchronize the handoff. Stress and fatigue may alter attention, memory, tolerance of ambiguity, and willingness to challenge authority. These processes are widely relevant to CRM, but their effects depend on task characteristics and should not be converted into an unsupported universal decline function.

The mechanisms connecting workload to team failure include delayed message acknowledgment, more frequent interruption, omission of a monitoring role, and oversimplified decision rules. Some can be measured behaviorally in simulators; others require physiological or self-report measures. A workload scale, a fatigue report, and a recorded lapse in critical cross-checking are separate measurements. Their convergence is informative; none should automatically be interpreted as a validated proxy for operational safety.

Training can establish preplanned redistribution rules, concise brief formats, and conditions that trigger a second check. Work design remains fundamental: scheduling, staffing, break opportunities, interface salience, and realistic task allocation determine whether a crew can enact the behavior it has been taught. Any study claiming improvement must account for changes in these conditions as well as the training itself.

Threat and Error Management

Threat and error management treats hazardous events as a process in which adverse conditions, human actions, system defenses, and opportunities for recovery interact. CRM contributes behavioral countermeasures: an anticipatory brief, independent cross-check, timely callout, explicit challenge, or coordinated response after a deviation. Helmreich, R. L., Merritt, A. C., & Wilhelm, J. A (1999) articulate the error-management orientation in aviation training. Its explanatory value is that operational robustness depends on detecting and containing error, including errors made by experienced individuals.

An error is not identical to a bad outcome, and a good outcome does not prove that no error occurred. A study can distinguish threat exposure, error occurrence, error detection, error management, and realized harm. These are related stages with different opportunities for intervention. The failure of a monitoring loop can turn a recoverable mistake into a persistent unsafe condition; an effective response can avert consequences even after an earlier control has failed.

Threat and error management also demands a systems account. If a team is repeatedly surprised by the same automation state or overloaded by a recurrent scheduling pattern, that hazard should be addressed in interface design or policy rather than confined to crew discipline. Operational research can use incident data to identify patterns, but reports are selective and shaped by reporting culture. Combining them with direct observation and simulator interventions improves inference about countermeasures.

Standard Operating Procedures, Checklists, and Briefs

Standard operating procedures encode institutional decisions about who should perform an action and how others should verify it. Checklists externalize memory requirements and create shared moments for confirmation. Briefings establish a prospective model of the mission: goals, constraints, roles, expected changes, and contingency plans. These are coordination tools with distinct functions. A checklist can support a critical state verification while leaving an unanticipated anomaly unexplained; a brief can make a plan explicit but cannot ensure it will remain valid.

The Federal Aviation Administration (2017) emphasizes the integration of monitoring into clear procedures. In healthcare, the Agency for Healthcare Research and Quality (2023) describe role assignment, shared goals, workload, resources, and changes of plan. Haynes, A. B., Weiser, T. G., Berry, W. R., Lipsitz, S. R (2009) documented improvements associated with implementation of a surgical safety checklist across eight hospitals, but the before–after design warrants caution in attributing all outcome changes to the checklist alone.

Effective deployment depends on execution fidelity, local adaptation, and the reason for a step. A 'tick-box' response can produce a record without the intended information exchange. Researchers should directly observe whether a check produces mutual confirmation, whether exceptions trigger meaningful discussion, and whether the tool fits operational timing. The question is what control function the procedure supplies, not simply whether a paper or digital form was completed.

Handoffs and Continuity Across Team Boundaries

Handoffs transfer responsibility, information, and expectations between people or units. They are especially vulnerable because the receiving team may inherit an operating condition without the history that explains it. A handoff succeeds when the receiver knows the current state, outstanding hazards, anticipated changes, actions already taken, and who is accountable for subsequent decisions. It also requires a protected interval for clarification and explicit confirmation that responsibility has passed.

Catchpole, K. R., de Leval, M. R., McEwan, A., Pigott, N., E (2007) examined a structured handover intervention in pediatric surgical care, drawing on aviation and Formula 1 practices. Their prospective before–after study observed 23 handovers before and 27 after implementation and found fewer recorded technical errors and information omissions. This is a valuable demonstration of process redesign, with the limitations of a modest, single-context, nonrandomized comparison. It should not be generalized automatically to all clinical handoffs or interpreted as a controlled estimate of mortality effects.

A robust handoff design specifies an information minimum, the timing and presence of essential participants, an opportunity to ask questions, and a visible transfer of accountability. Researchers can evaluate completeness, accuracy, interruptions, recall, subsequent discrepancies, and the ability to resolve an unexpected change. Strong measures track what the receiving team actually understood and did, rather than assuming that delivery of a standardized message implies successful transfer.

Briefs, Huddles, Debriefs, and Operational Learning

Briefs, huddles, and debriefs create different temporal opportunities for coordination. A brief builds an initial common plan. A huddle reorients the team when circumstances change. A debrief reconstructs an episode, compares behavior with intended controls, and identifies a change in future practice. Agency for Healthcare Research and Quality (n.d.) treats debriefing as a short, structured reflection on goals, communication, responsibilities, resources, and error management.

The causal chain from debrief to safety is longer than the activity itself. A team may identify a pattern yet lack authority to change staffing or equipment; a debrief may become ritualized or blame-oriented and suppress later disclosure. For learning to persist, recognized problems must be documented, assigned, corrected, and retested. Evaluation should distinguish participant satisfaction from changes in behavior and downstream operational performance.

After-action research can compare video-assisted and discussion-only reviews, facilitator styles, immediacy of feedback, and the durability of behavior in later scenarios. It can also examine whether corrective action addresses a person-specific skill deficit or a repeatable system defect. The relevant unit of analysis is a learning cycle in which observation leads to revision and subsequent verification.

Nontechnical Skills and Behavioral Marker Systems

CRM is commonly operationalized through nontechnical skills: cognitive and social behaviors that complement technical proficiency. The major challenge is that constructs such as leadership, situation awareness, and teamwork must be translated into reproducible behavioral ratings. An unanchored global impression is vulnerable to halo effects, the observer's expectations, and knowledge of a successful or unsuccessful outcome.

O'Connor, P., Hörmann, H.-J., Flin, R., Lodge, M., Goeters (2002) describe the NOTECHS European behavioral marker program and report an assessment exercise involving instructors from 14 airlines. The study identifies issues of rater training, accuracy, and inter-rater consistency. In anesthesia, Flin, R., Patey, R., Glavin, R., & Maran, N (2010) describe the development and application of the ANTS taxonomy. Such frameworks translate broad psychological language into observable, job-relevant actions while requiring continuous construct validation.

A behavioral marker system should specify the observation window, the opportunities for a behavior to occur, the definition of an adequate response, and procedures for resolving rater disagreement. Absence of a challenge during a scenario with no contradictory information is not equivalent to failure to challenge. Strong assessment combines multiple episodes, independent observers, scenario comparability, and criteria that do not reward unsafe 'heroic' improvisation. Validity belongs to the instrument in a particular use, not to the prestige of the acronym printed on its rating form.

Measurement of Coordination as a Dynamic Construct

Coordination measurement must match the scientific question. Self-reported teamwork climate measures beliefs and perceptions. Communication coding measures who said what, to whom, and when. Behavioral observation measures execution, correction, and adaptability. Team cognition instruments estimate shared or distributed knowledge. Operational indicators quantify delays, recovery, adverse events, or mission success. These measures may covary, but substituting one for another weakens interpretability.

Brannick, M. T., Prince, A., Prince, C., & Salas, E (1995) evaluated team process ratings in simulated military aircrew missions. Their findings illustrate both the promise of behaviorally anchored observation and the importance of multiple observations for stable assessment across scenarios. DeChurch, L. A., & Mesmer-Magnus, J. R (2010) further demonstrate that the operational definition of shared cognition changes observed relations to performance. Reliability and construct validity require explicit evidence rather than intuitive acceptance of an apparently relevant scale.

For event-based analysis, a useful record includes a timestamp, initiator, recipient, task state, message function, acknowledgment, resulting action, and outcome. At the team level, analysts must account for clustering of individuals in crews and repeated episodes in crews. Time-to-detection, proportion of critical messages acknowledged, false-escalation rate, and successful recovery can illuminate separate mechanisms. The unit and denominator for each rate must be declared to prevent incorrect comparisons between environments with different hazard exposure.

Communication Networks and Distributed Information

Operational information travels through patterned relationships. A cockpit has a small and relatively stable communication network; emergency medicine or disaster response may combine several professions, external units, remote specialists, and changing leadership. Network structure affects whether relevant updates reach decision-makers and whether an isolated expert can correct the team's operating picture. Communication frequency alone is inadequate: a heavily connected member may be overwhelmed, and a peripheral member may possess the only consequential observation.

Network coding can identify senders, recipients, directional ties, and latency across successive task phases. Analysis should distinguish task-relevant edges from routine social exchanges, and routine coordination from exceptional escalation. Temporal network methods can examine whether responsibility migrates as the task changes. In fast-changing teams, member turnover and variable boundaries create an additional measurement problem: the network at the beginning of an episode may no longer be the team that makes the final decision.

Network interventions include redesigning the escalation pathway, supplying a shared situation display, or assigning a liaison at a boundary between specialties. The causal hypothesis is that a shorter, more reliable path for critical information improves recognition and response, subject to attention and workload constraints. This proposition can be tested by manipulating information availability and network roles while measuring actual decision and recovery outcomes.

Experimental Simulation and Ecological Validity

Simulation is central to CRM research because researchers can reproduce a critical condition without exposing passengers or patients to that risk. A carefully designed scenario can vary authority gradients, misleading information, environmental workload, unexpected role changes, or opportunities for mutual monitoring. The simulation can provide a repeatable observation window and a known correct response against which behavior can be assessed. Nevertheless, a scenario is a sample of operating situations, not a complete representation of the domain.

Control conditions matter. If a team receives additional training time, new communication scripts, a more favorable instructor, and a redesigned checklist, the resulting difference does not isolate the effect of any one component. Randomizing teams, standardizing scenarios, and evaluating maintenance over time strengthen inference. Research should report task difficulty, crew experience, scenario realism, rater blinding, attrition, and potential contamination when trained and untrained teams work in the same organization.

Transfer requires evidence that trained behavior survives changes in workload, colleague composition, equipment, and routine practice. Success in a classroom exercise may indicate learning; observed application in subsequent work indicates behavioral transfer; fewer consequential errors or improved patient outcomes indicate a more distant result. These stages correspond to a broader training-evaluation logic used in Hughes, A. M., Gregory, M. E., Joseph, D. L., Sonesh, S. C (2016), and each requires its own measurement design.

Evaluation of CRM Training

CRM training can improve awareness of human-factors problems, provide behavioral scripts, and offer repeated practice in adaptive coordination. Evidence should be evaluated by training content, fidelity, comparator, follow-up duration, and outcome level. Salas, E., Wilson, K. A., Burke, C. S., & Wightman, D. C (2006) reviewed the CRM evaluation literature and identified continuing needs in measurement and research design. The literature contains positive findings, but different interventions and outcome definitions prevent one universal effectiveness estimate from being scientifically defensible.

An evaluation architecture distinguishes reactions to a course, acquired declarative or procedural knowledge, behavior demonstrated in a training scenario, transfer to actual operations, and objective results. Favorable participant reactions have limited evidentiary value for the most consequential endpoint. A simulated improvement in error detection is stronger mechanistic evidence but still requires a credible link to field performance. Longitudinal assessment is necessary because practices decay without reinforcement, changes in staffing, and supportive procedures.

Training must also be integrated with operational governance. A crew may learn to challenge a decision but work in a setting that punishes such action; it may learn closed-loop communication but lack workable channels for urgent information. Training-only studies should therefore document organizational supports, instructor competence, and post-course opportunities to practice. Strong inference requires a defined intervention and an explicit mechanism rather than treating the CRM label as an active ingredient.

Evidence from Aviation Operations

Aviation supplies the historical institutional base for CRM and extensive opportunities for simulation, line-oriented evaluation, and incident analysis. The National Aeronautics and Space Administration, Aviation Safe (2014), Helmreich, R. L., Merritt, A. C., & Wilhelm, J. A (1999), and Federal Aviation Administration (2004) establish the development and operational content of training. The strength of this tradition lies in connecting social and cognitive skills with repeatable procedures and formally assigned monitoring roles.

It is harder to estimate an isolated causal effect of CRM on rare catastrophic outcomes. Air transportation safety has evolved simultaneously through aircraft design, weather information, maintenance, navigation, automation, procedural standardization, reporting systems, and regulation. Before–after trends in accident counts therefore cannot be assigned wholesale to CRM training. An incident or simulator finding can demonstrate an error-management mechanism, whereas an industry-wide outcome claim demands a design capable of separating simultaneous changes.

Further aviation work should target observed monitoring quality, mode awareness, authority challenge, and recovery time in matched events. Reporting and simulator studies become more informative when they map behavior to the particular threat and state transition instead of treating the absence of an accident as a score. This distinction permits legitimate conclusions about improved coordination without overstating what can be inferred about very rare events.

Evidence from Healthcare and Surgery

Healthcare provides an important transfer test because clinical teams are often multidisciplinary, have rapidly changing membership, and combine incomplete information with time pressure. Lingard, L., Espin, S., Whyte, S., Regehr, G., Baker, G. R (2004) documented communication failures in operating rooms; Manser, T (2009) reviewed teamwork research in dynamic healthcare settings. These sources establish plausible coordination vulnerabilities rather than uniform proof that every aviation intervention transfers unchanged.

Two well-known outcome studies require a careful separation of design and inference. Haynes, A. B., Weiser, T. G., Berry, W. R., Lipsitz, S. R (2009) found improvements following checklist introduction in an international before–after study. Neily, J., Mills, P. D., Young-Xu, Y., Carney, B. T., West (2010) reported lower surgical mortality associated with medical-team training in a Veterans Health Administration implementation study with comparison facilities. The latter included 74 participating and 34 nonparticipating facilities; changes in risk-adjusted rates and the observational implementation design should be reported together. Neither study is a randomized isolation of a single CRM behavior from all other features of care.

At the evidence-synthesis level, Hughes, A. M., Gregory, M. E., Joseph, D. L., Sonesh, S. C (2016) conducted a meta-analysis reporting benefits across reactions, learning, behavioral transfer, and results, with standardized effects reported across criteria. Buljac-Samardzic, M., Doekhie, K. D., & van Wijngaarden, J (2020) reviewed a broad array of healthcare team interventions and emphasized heterogeneity of designs, settings, and outcomes. Accordingly, the well-supported practical conclusion is that structured team interventions can improve relevant outcomes under studied conditions, while the exact clinical effect and preferred package depend on implementation and context.

Handover, Operating-Room, and Clinical Team Evidence

The handover experiment by Catchpole, K. R., de Leval, M. R., McEwan, A., Pigott, N., E (2007) illustrates how a process intervention can be tied to direct observations of information omission, technical errors, teamwork, and duration. It is especially instructive because the intervention changed the joint workflow rather than focusing exclusively on interpersonal training. The small comparison groups and context-specific design also illustrate the limits of interpreting a successful local improvement as a universal treatment effect.

Weaver and colleagues (2010)36022-3) examined operating-room teamwork and performance with a multilevel evaluation. Mazzocco, K., Petitti, D. B., Fong, K. T., Bonacum, D., Broo (2009) connected observed surgical team behavior with patient outcomes in a field study. These papers show that observational measures can be linked to outcomes, yet team and case characteristics, institutional differences, and rater bias must be considered. A poor outcome may influence observers' retrospective impressions, while teams treating more complex cases may encounter more opportunities for coordination failures.

Clinical research benefits from reporting a sequence of intermediate endpoints: whether the brief occurred, whether the relevant information was acknowledged, whether the action followed the changed plan, whether errors were detected, and whether adverse consequences were prevented. This mechanism-focused account produces more useful evidence than a generic declaration that teamwork improved.

Rapidly Reconfigured and Extreme Action Teams

Some teams assemble around a crisis, perform intense interdependent work, and disband rapidly. Their coordination challenge includes immediate role recognition and transfer of authority across changes in membership. Klein, K. J., Ziegert, J. C., Knight, A. P., & Xiao, Y (2006) studied emergency medical teams and developed the concept of dynamic delegation. Their qualitative evidence is valuable for identifying processes that may be overlooked by fixed-leader models, while causal claims require further controlled comparison.

Rapidly reconfigured teams depend on recognizable professional routines and shared artifacts. A short admission or incident brief can instantiate a workable plan among people who have never previously worked together. The cost is possible overreliance on assumed role knowledge; membership changes may invalidate earlier assumptions about capability, expertise, or responsibility. Observation should distinguish apparent role stability from the actual changing network of decision rights.

In disasters and other distributed emergencies, hierarchical command structures and local adaptive action must interact. This suggests research designs that compare the reliability of handoff across changing leadership, rather than proposing that centralized command or universal shared leadership should prevail in every context. The most important question is whether the structure makes the current state, task owner, and escalation route knowable when conditions change rapidly.

Culture, Professional Hierarchies, and Transfer Between Settings

CRM emerged within institutional arrangements that vary across national aviation cultures, professional hierarchies, and organizations. Helmreich, R. L., Merritt, A. C., & Wilhelm, J. A (1999) explicitly identify limits of cross-cultural generalization. A speaking-up intervention assumes particular norms about authority, status, and acceptable challenge. Transferring the same words into a different language or profession does not guarantee identical interpretation or behavior.

A serious transfer study begins with the local work system: how authority is assigned, how errors are reported, where crews can pause work, how handovers occur, and which resources are available. Then the intervention is adapted in a documented manner. Fidelity should be assessed at the level of function—successful acknowledgment of a critical discrepancy, for example—while permitting culturally and operationally appropriate language and role structure. A protocol that works only when delivered in an unrealistic manner has limited ecological validity.

Cross-cultural evidence also needs measurement invariance. If respondents in different settings use teamwork rating scales differently, apparent differences in scores may represent changes in interpretation rather than genuine coordination. Combining translation checks, qualitative observation, behavioral markers, and task-level outcomes can distinguish adaptation that preserves causal function from adaptation that removes it.

Implementation, Organizational Support, and Safety Culture

Training operates within a wider organizational ecology. Procedures, staffing, scheduling, psychological safety, interface design, leadership accountability, and access to debriefing can all determine whether coordination practices are possible. An intervention with a well-conducted classroom component may fail when teams cannot practice, when senior staff model incompatible behavior, or when system redesign is needed to resolve a chronic hazard. The causal pathway from individual training to outcome is therefore conditional on institutional implementation.

Buljac-Samardzic, M., Doekhie, K. D., & van Wijngaarden, J (2020) organize healthcare team interventions across training, tools, and organizational redesign, underscoring the inadequacy of treating all approaches as one intervention. Agency for Healthcare Research and Quality (n.d.) supplies practical protocols that must be evaluated for fidelity and outcomes in particular organizations. Local process redesign, audit, reinforcement, and feedback are parts of implementation, not cosmetic additions.

A useful evaluation traces adoption, reach, fidelity, sustainability, and unintended effects. Observed improvements should be compared with contemporaneous changes in workload, case complexity, equipment, and organizational priorities. Because implementation can vary among wards or units, cluster designs, interrupted time-series methods, or phased rollout analyses may be appropriate when randomized individual assignment is impractical. The unit at which policy is introduced should inform the unit of statistical analysis.

Team Coordination in Remote and Technology-Mediated Work

When crews communicate through radios, software systems, video, or remote workstations, timing and visibility become partly technological. A delay, muted channel, ambiguous status indicator, or failure to update a shared record can create a coordination fault even when each participant understands the mission. The relevant psychological processes include attention allocation, common-ground maintenance, message acknowledgment, trust in remote expertise, and the capacity to establish who presently controls a resource.

Remote teams require analysis of both the communication network and the represented work object. An asynchronous system may be effective for routine planning but unsuited to an urgent, evolving hazard. A shared dashboard can improve access to a common situation picture while creating new risks if stale information appears current. Researchers should specify latency, provenance, version changes, and the meaning of an acknowledgment in the medium being used.

The empirical program is straightforward: manipulate channel latency, access asymmetry, interface feedback, or handoff protocols; measure misunderstanding, recovery time, workload, and task success; and test transfer to authentic operations. These studies extend CRM's coordination questions without presuming that the historical flight-deck training package is sufficient for all remote work.

Artificial Systems as Operational Team Participants

The Artificial Era raises a consequential extension of crew coordination: artificial systems can supply recommendations, monitor streams, allocate tasks, or act through tools within human-supervised operations. Their behavior changes the structure of team interdependence even when formal accountability remains with designated humans. The research object is the interaction among machine outputs, human understanding, decision authority, and error recovery. The terms artificial agency or artificial teammate require explicit operational definitions linked to observable capacities, permissions, and conditions of control.

The relevance of historical CRM is functional. The same questions arise: who detects an anomaly, who acknowledges it, which evidence supports a decision, when is authority transferred, and how is a misleading system output corrected? A machine-generated recommendation may be fast and precise in format while incompletely calibrated to a situation. Teams need visibility into data currency, confidence limits, operating mode, and the route for rejecting or reversing an action. These mechanisms should be measured directly rather than inferred from anthropomorphic descriptions.

The conceptual horizon From Homo to Artificial motivates new research into coordination across biological and nonbiological participants. It does not supply empirical proof that a particular system has humanlike awareness, competence, or responsibility. Experimental designs can compare human-only teams with carefully specified hybrid configurations, independently manipulating automation role, authority, transparency, and failure rate. Such studies can reveal changes in communication burden and error detection while preserving the distinction between a philosophical interpretive framework and validated psychological outcomes.

Research Design, Confounding, and Causal Identification

CRM research draws on laboratory experiments, field observations, simulator trials, qualitative ethnography, implementation studies, regulatory practice, and evidence synthesis. Each method answers a different question. An observed association between communication quality and outcomes can support a process hypothesis but cannot by itself exclude confounding by team experience, case complexity, leadership quality, or equipment reliability. Randomization can improve causal identification for training packages when the unit of assignment and contamination are handled correctly.

Simulation trials can manipulate one coordination mechanism at a time. Field interventions may require cluster randomization or a credible comparison series because clinicians and pilots share practices within organizations. Before–after comparisons require attention to secular trends and changes in reporting. Uncontrolled improvements may arise from increased observation, concurrent safety initiatives, or regression toward the mean. When outcomes are rare, proximal indicators can be measured more precisely, though they should be explicitly described as proxies.

A strong evidentiary record reports eligibility, sample construction, intervention content, follow-up, denominators, missing data, inter-rater reliability, effect estimates, and uncertainty. It declares the level of analysis—individual, team, shift, facility, or mission—and tests whether conclusions survive alternative explanations. For qualitative studies, it documents sampling, field context, analytic decisions, and negative cases. Scientific synthesis integrates these forms of evidence according to their strengths rather than forcing unlike designs into a single hierarchy without regard to the research question.

Contradictory Findings and Limits of Generalization

The most persistent difficulty in CRM evaluation is that a common label covers training programs with different contents and durations, delivered in different institutional environments and measured with incompatible outcomes. Positive findings for knowledge acquisition cannot establish identical effects on real-world safety. Some studies evaluate individual attitudes, others team behavior, and others patient or mission outcomes; pooling these indiscriminately risks constructing a result that belongs to no well-defined intervention.

Salas, E., Wilson, K. A., Burke, C. S., & Wightman, D. C (2006) call for stronger evaluation. Hughes, A. M., Gregory, M. E., Joseph, D. L., Sonesh, S. C (2016) provide quantitative evidence in healthcare while their training and outcome categories require careful interpretation. Buljac-Samardzic, M., Doekhie, K. D., & van Wijngaarden, J (2020) highlight uneven evidence across interventions and settings. A defensible synthesis therefore recognizes improvements that have been observed and evaluates whether those observations apply to an intended intervention in an intended population.

Another tension concerns standardization and adaptability. Procedures reduce preventable variation, yet rigid compliance can be maladaptive during an unanticipated state. Likewise, distributed decision-making can improve local responsiveness while obscuring final accountability. The productive scientific question is which configuration provides reliable control under specified task demands. Future studies should directly test interaction effects rather than framing each practice as a universally beneficial personality trait.

Evidence Quality and the Interpretation of Safety Outcomes

The significance of a finding depends on the outcome measured. A statistically reliable rise in teamwork ratings is informative about ratings; its practical significance depends on whether those behaviors are consequential for the task. A reduction in errors is meaningful only with an explicit definition of error, a denominator, and comparable opportunities for error to occur. Changes in adverse events may reflect detection and reporting practices as well as underlying risk. The strength of inference must match the design.

The Neily, J., Mills, P. D., Young-Xu, Y., Carney, B. T., West (2010) implementation study provides an illustrative contrast between a mortality endpoint and experimental control. It associated an intervention with a lower rate, but the intervention was not a double-blind randomized manipulation of communication alone. The Catchpole, K. R., de Leval, M. R., McEwan, A., Pigott, N., E (2007) study provides direct evidence of changed handoff processes with a smaller scope and a less distal outcome. Neither is scientifically inferior in all respects: each gives different information along a proposed causal chain.

Publication bias, selective outcome reporting, missing follow-up, and differences in organizational incentives must be considered when synthesizing favorable findings. Where original full texts have not been accessible, this monograph confines numerical or procedural descriptions to independently identifiable abstracts or publisher records. It does not claim that all included studies were assessed with a formal risk-of-bias instrument or that the narrative constitutes a PRISMA systematic review.

Design Principles for Operational Coordination

The evidence supports several mechanism-oriented design principles. First, make the current task state and critical changes visible to those responsible for acting. Second, assign monitoring duties explicitly and make independent cross-checking feasible. Third, establish unambiguous acknowledgment for critical communications. Fourth, define a workable escalation path that permits important discrepancies to alter decisions. Fifth, provide opportunities to rehearse unusual conditions and to inspect performance after the event. These are principles whose effectiveness depends on the work system in which they are implemented.

Design begins with a task and hazard analysis, proceeds to specification of the desired behavior, and then to a measure capable of detecting whether the behavior occurred. A practice may be retained, adapted, or abandoned based on those results. An apparent conflict between speed and communication may be resolved by selecting a shorter, better-timed message rather than eliminating a confirmation loop. Likewise, task allocation can reduce workload if monitoring obligations remain explicit.

Operational design requires testing for unintended effects. A new checklist can increase interruption, a new communication channel can distribute responsibility ambiguously, and a dashboard can display outdated information with misleading authority. A mechanism-oriented evaluation tests both the intended gain and the tradeoff. This approach links CRM to engineering psychology without conflating human motivation with the full set of conditions that determine safe performance.

Research Agenda and Open Problems

The strongest research agenda moves from broad claims that teamwork matters toward specific hypotheses about the conditions under which coordination mechanisms work. One line should investigate critical thresholds: when does workload make closed-loop communication too slow, and what alternative confirmation design preserves the same safety function? Another should examine team composition and turnover, including how quickly newly assembled members can establish functional common ground without falsely assuming identical expertise.

A second line concerns measurement invariance and causal mediation. Studies should test whether behavioral markers have comparable meanings across aviation, healthcare, emergency response, and remote operations. They should also evaluate whether training changes team knowledge, whether changed knowledge alters monitoring and challenge behavior, and whether those behaviors precede observable error recovery. The same research can identify failures of transfer instead of presenting any unsuccessful intervention as a mysterious exception.

A third line examines human–Artificial systems. Work is needed on authority transfers among humans and machines, observability of automated recommendations, error trapping when multiple agents share a data source, and the durability of human monitoring skill when routine decisions are automated. The most informative trials would preregister critical outcomes, use representative tasks and failure modes, test follow-up, and compare plausible alternative designs. Longitudinal research is especially important because teams, technologies, and institutions learn together over time.

Critical Scientific Synthesis

Taken together, the literature portrays coordination as the production and maintenance of compatible operational expectations under changing constraints. Teamwork theories identify component behaviors and cognitive mechanisms; aviation CRM provides a historically developed intervention tradition; clinical research supplies observations of communication failures and studies of training, handoffs, and safety outcomes. The convergence is strongest around the practical importance of well-timed information exchange, explicit responsibilities, mutual monitoring, and opportunities for corrective action.

The evidence is more differentiated when a particular practice is asked to explain a distant outcome. Observational associations, simulator improvements, and implementation effects have distinct inferential strengths. Organizational changes are frequently bundled with training, making it difficult to isolate a single ingredient. The scientifically appropriate synthesis is therefore conditional: coordination practices can improve observable team processes and sometimes outcomes, while effects depend on task interdependence, institutional implementation, measurement validity, and the possibility of responding to unexpected events.

Human team coordination is best understood as a temporally organized control achievement across individual cognition, interpersonal communication, collective responsibility, and sociotechnical design. Its successful application requires continuous interaction between knowledge of what should happen and evidence of what is happening. This account provides a stable research structure for aviation and other safety-critical domains, and it supplies testable questions for the increasing participation of artificial systems in operational teams.

Conclusion

Human Team Coordination and Crew Resource Management constitute a coherent psychological research object centered on interdependent action in high-consequence work. The historical CRM program expanded the operational understanding of safety from isolated individual competence to the collective management of information, authority, workload, and error. Teamwork-process theory, shared-mental-model research, behavioral observation, and implementation studies together explain how coordination is organized and how it can be evaluated.

The fundamental findings concern mechanisms that can be specified and observed: role assignment, explicit acknowledgment, anticipation of another member's needs, monitoring, timely challenge, backup action, structured handoff, and learning from operational events. Their effects depend on the task, team structure, technology, and institutional environment. Credible conclusions separate acquisition of knowledge, transfer of behavior, error management, and ultimate safety outcomes.

The emerging scientific frontier is the architecture of coordination across people, digital infrastructure, and increasingly capable artificial systems. Its central question is how a collective operational process remains intelligible, correctable, and accountable as the distribution of information and action changes. Progress will come from mechanism-specific experiments, reliable field measurement, and cautious integration of heterogeneous evidence. The enduring purpose of CRM is to make coordinated action reproducible and recoverable under uncertainty.

References

  • Agency for Healthcare Research and Quality. (2023). Sharing the care plan: Brief. TeamSTEPPS. https://www.ahrq.gov/teamstepps-program/curriculum/team/tools/briefs.html
  • Agency for Healthcare Research and Quality. (n.d.). Reviewing the team's performance: Debrief. TeamSTEPPS. https://www.ahrq.gov/teamstepps-program/curriculum/team/tools/debrief.html
  • Brannick, M. T., Prince, A., Prince, C., & Salas, E. (1995). The measurement of team process. Human Factors, 37(3), 641–651. https://doi.org/10.1518/001872095779049372
  • Buljac-Samardzic, M., Doekhie, K. D., & van Wijngaarden, J. D. H. (2020). Interventions to improve team effectiveness within health care: A systematic review of the past decade. Human Resources for Health, 18, Article 2. https://doi.org/10.1186/s12960-019-0411-3
  • Catchpole, K. R., de Leval, M. R., McEwan, A., Pigott, N., Elliott, M. J., McQuillan, A., MacDonald, C., & Goldman, A. J. (2007). Patient handover from surgery to intensive care: Using Formula 1 pit-stop and aviation models to improve safety and quality. Pediatric Anesthesia, 17(5), 470–478. https://doi.org/10.1111/j.1460-9592.2006.02239.x
  • DeChurch, L. A., & Mesmer-Magnus, J. R. (2010a). The cognitive underpinnings of effective teamwork: A meta-analysis. Journal of Applied Psychology, 95(1), 32–53. https://doi.org/10.1037/a0017328
  • DeChurch, L. A., & Mesmer-Magnus, J. R. (2010b). Measuring shared team mental models: A meta-analysis. Group Dynamics: Theory, Research, and Practice, 14(1), 1–14. https://doi.org/10.1037/a0017455
  • Edmondson, A. (1999). Psychological safety and learning behavior in work teams. Administrative Science Quarterly, 44(2), 350–383. https://doi.org/10.2307/2666999
  • Endsley, M. R. (1995). Toward a theory of situation awareness in dynamic systems. Human Factors, 37(1), 32–64. https://doi.org/10.1518/001872095779049543
  • Federal Aviation Administration. (2004). Crew resource management training (Advisory Circular 120-51E). U.S. Department of Transportation. https://www.faa.gov/regulations_policies/advisory_circulars/index.cfm/go/document.information/documentid/22879
  • Federal Aviation Administration. (2017). Standard operating procedures and pilot monitoring duties for flight deck crewmembers (Advisory Circular 120-71B). U.S. Department of Transportation. https://www.faa.gov/regulations_policies/advisory_circulars/index.cfm/go/document.information/documentid/1030486
  • Flin, R., Patey, R., Glavin, R., & Maran, N. (2010). Anaesthetists' non-technical skills. British Journal of Anaesthesia, 105(1), 38–44. https://doi.org/10.1093/bja/aeq134
  • Haynes, A. B., Weiser, T. G., Berry, W. R., Lipsitz, S. R., Breizat, A.-H. S., Dellinger, E. P., et al. (2009). A surgical safety checklist to reduce morbidity and mortality in a global population. New England Journal of Medicine, 360(5), 491–499. https://doi.org/10.1056/NEJMsa0810119
  • Helmreich, R. L., Merritt, A. C., & Wilhelm, J. A. (1999). The evolution of crew resource management training in commercial aviation. International Journal of Aviation Psychology, 9(1), 19–32. https://doi.org/10.1207/s15327108ijap0901_2
  • Hughes, A. M., Gregory, M. E., Joseph, D. L., Sonesh, S. C., Marlow, S. L., Lacerenza, C. N., Benishek, L. E., King, H. B., & Salas, E. (2016). Saving lives: A meta-analysis of team training in healthcare. Journal of Applied Psychology, 101(9), 1266–1304. https://doi.org/10.1037/apl0000120
  • Klein, K. J., Ziegert, J. C., Knight, A. P., & Xiao, Y. (2006). Dynamic delegation: Shared, hierarchical, and deindividualized leadership in extreme action teams. Administrative Science Quarterly, 51(4), 590–621. https://doi.org/10.2189/asqu.51.4.590
  • LePine, J. A., Piccolo, R. F., Jackson, C. L., Mathieu, J. E., & Saul, J. R. (2008). A meta-analysis of teamwork processes: Tests of a multidimensional model and relationships with team effectiveness criteria. Personnel Psychology, 61(2), 273–307. https://doi.org/10.1111/j.1744-6570.2008.00114.x
  • Lingard, L., Espin, S., Whyte, S., Regehr, G., Baker, G. R., Reznick, R., Bohnen, J., Orser, B., Doran, D., & Grober, E. (2004). Communication failures in the operating room: An observational classification of recurrent types and effects. Quality and Safety in Health Care, 13(5), 330–334. https://doi.org/10.1136/qhc.13.5.330
  • Manser, T. (2009). Teamwork and patient safety in dynamic domains of healthcare: A review of the literature. Acta Anaesthesiologica Scandinavica, 53(2), 143–151. https://doi.org/10.1111/j.1399-6576.2008.01717.x
  • Marks, M. A., Mathieu, J. E., & Zaccaro, S. J. (2001). A temporally based framework and taxonomy of team processes. Academy of Management Review, 26(3), 356–376. https://doi.org/10.5465/amr.2001.4845785
  • Mazzocco, K., Petitti, D. B., Fong, K. T., Bonacum, D., Brookey, J., Graham, S., Lasky, R. E., Sexton, J. B., & Thomas, E. J. (2009). Surgical team behaviors and patient outcomes. American Journal of Surgery, 197(5), 678–685. https://doi.org/10.1016/j.amjsurg.2008.03.002
  • National Aeronautics and Space Administration, Aviation Safety Reporting System. (2014, April). Three important aspects of CRM: Communication, communication, and communication. Callback, 411. https://asrs.arc.nasa.gov/publications/callback/cb_411.html
  • Neily, J., Mills, P. D., Young-Xu, Y., Carney, B. T., West, P., Berger, D. H., Mazzia, L. M., Paull, D. E., & Bagian, J. P. (2010). Association between implementation of a medical team training program and surgical mortality. JAMA, 304(15), 1693–1700. https://doi.org/10.1001/jama.2010.1506
  • O'Connor, P., Hörmann, H.-J., Flin, R., Lodge, M., Goeters, K.-M., & the JARTEL Group. (2002). Developing a method for evaluating crew resource management skills: A European perspective. International Journal of Aviation Psychology, 12(3), 263–285. https://doi.org/10.1207/S15327108IJAP1203_5
  • Rico, R., Sánchez-Manzanares, M., Gil, F., & Gibson, C. (2008). Team implicit coordination processes: A team knowledge-based approach. Academy of Management Review, 33(1), 163–184. https://doi.org/10.5465/amr.2008.27751276
  • Rosenman, E. D., Ilgen, J. S., Shandro, J. R., Harper, A. L., & Fernandez, R. (2015). A systematic review of tools used to assess team leadership in health care action teams. Academic Medicine, 90(10), 1408–1422. https://doi.org/10.1097/ACM.0000000000000848
  • Salas, E., Sims, D. E., & Burke, C. S. (2005). Is there a 'Big Five' in teamwork? Small Group Research, 36(5), 555–599. https://doi.org/10.1177/1046496405277134
  • Salas, E., Wilson, K. A., Burke, C. S., & Wightman, D. C. (2006). Does crew resource management training work? An update, an extension, and some critical needs. Human Factors, 48(2), 392–412. https://doi.org/10.1518/001872006777724444
  • Salas, E., Wilson, K. A., Burke, C. S., Wightman, D. C., & Howse, W. R. (2006). Crew resource management training research, practice, and lessons learned. Reviews of Human Factors and Ergonomics, 2(1), 35–73. https://doi.org/10.1177/1557234X0600200103
  • Wageman, R. (1995). Interdependence and group effectiveness. Administrative Science Quarterly, 40(1), 145–180. https://doi.org/10.2307/2393703
  • Weaver, S. J., Rosen, M. A., DiazGranados, D., Lazzara, E. H., Lyons, R., Salas, E., Knych, S. A., McKeever, M., Adler, L., Barker, M., & King, H. B. (2010). Does teamwork improve performance in the operating room? A multilevel evaluation. Joint Commission Journal on Quality and Patient Safety, 36(3), 133–142. https://doi.org/10.1016/S1553-7250(10)36022-3