Organizations are increasingly becoming dynamic and unstable. This evolution has given rise to greater reliance on teams and increased complexity in terms of team composition, skills required, and degree of risk involved. High-reliability organizations (HROs) are those that exist in such hazardous environments where the consequences of errors are high, but the occurrence of error is extremely low. In this article, we argue that teamwork is an essential component of achieving high reliability particularly in health care organizations. We describe the fundamental characteristics of teams, review strategies in team training, demonstrate the criticality of teamwork in HROs and finally, identify specific challenges the health care community must address to improve teamwork and enhance reliability.
Keywords: High-reliability organization, teams, teamwork, health care, patient safety, training
A healthy 38-year-old woman was admitted to a major medical center to deliver her first child. Although she was a low-risk patient with only mildly elevated blood pressure, her admission ended tragically when she underwent an emergency cesarean after a failed forceps delivery. Once inside the abdominal cavity, the uterus was found to have ruptured, and the placenta was in the abdomen. She delivered a stillborn fetus. After an unsuccessful attempt to repair her uterus, she received a full hysterectomy, underwent blood transfusions, and endured endless complications resulting in a 3-week hospital stay, including 18 days in intensive care. What went wrong? According to root cause analyses, lack of teamwork played a significant role. Specifically, communication was poor; there was a lack of mutual performance cross-monitoring, inadequate conflict resolution, poor situational awareness, and work overload. A major response to the tragedy was the initiation of team training at the medical center (Sachs 2005).
Safety is a fundamental patient right, though not a certainty (Knox and Simpson 2004). When patients arrive at a health care organization, they expect to leave that institution in equal or better health. Patients and their families do not expect physicians, nurses, and other hospital staff to make mistakes, or worse yet cover up as opposed to communicate errors. The publication of To Err Is Human by the Institution of Medicine (IOM) highlighted the fact that the delivery of care is not error free. The report concluded that medical errors cause up to 98,000 deaths annually. The IOM report brought national focus to this important issue and has since spawned significant research on the causes of medical errors and the effectiveness of different strategies for making health care a more reliable system (Kohn, Corrigan, and Donaldson 1999).
The IOM issued a number of recommendations designed to move health care institutions toward high reliability. HROs are institutions that operate in complex, hazardous environments making few mistakes (i.e., medical errors) over long periods of time. Recommendations related to voluntary error reporting, systems changes, safety systems design, and standard for health care professionals were presented in To Err Is Human. The IOM also pointed toward the need for enhanced teamwork. Historically physicians, nurses, and other health care professionals have functioned as discrete parts. The IOM recommended that interdisciplinary team training programs be established, based on sound principles of team management, to improve coordination and communication among health care staff (Kohn et al. 1999).
The Agency for Healthcare Research and Quality (AHRQ) is the lead federal agency in supporting and implementing the recommendations of the IOM in its effort to reduce medical error and improve patient safety. As part of this agenda, AHRQ established the HRO network to support patient safety leaders by providing them with a forum for learning about promising practices and identifying new and innovative ways to implement research findings. AHRQ's goal is to create high-reliability health care organizations. In support of that goal, AHRQ will launch Team Strategies and Tools to Enhance Performance and Patient Safety (TeamSTEPPS) during 2006 and distribute this team training curriculum to members of the HRO network (Alonso et al. 2006).
The purpose of this paper then is to demonstrate that teamwork is an important component of HROs. While moving health care toward team-based work will not automatically result in high reliability, there are many parallels between teams and HROs (Knox and Simpson 2004; Wilson et al. 2005). To justify our argument, we begin by providing an overview of teams, teamwork, and strategies for promoting team effectiveness. Much has been written about these topics in domains where high reliability is critical because the consequences of error are great (e.g., the military, commercial aviation, and air traffic control) (Salas and Cannon-Bowers 2000; Davies 2001; Baker et al. 2003; Salas Sims and Kleien 2004). Later in this document, we describe the key characteristics of HROs. The concept of HROs has been around for more than 20 years, but has only recently begun to take hold in health care with the publication of To Err Is Human and AHRQ's patient safety agenda. In this section, we compare the fundamental features of teamwork and the critical characteristics of HROs and demonstrate how these characteristics are interwoven as well as how and why the HRO environment demands teamwork. Finally, although the science of teamwork has been around for over 30 years, it is only recently that this concept has begun to take hold in health care. In the last section we present a series of challenges researchers and practitioners need to address to advance the health care community's understanding of team performance and instantiate these practices as part of health care's quest to achieved high reliability.
TEAMS, TEAMWORK, AND TRAINING
Teams and Teamwork
There is a general consensus in the research literature that a team consists of two or more individuals, who have specific roles, perform interdependent tasks, are adaptable, and share a common goal (Salas et al. 1992). To work effectively together, team members must possess specific knowledge, skills, and attitudes (KSAs), such as the skill in monitoring each other's performance, knowledge of their own and teammate's task responsibilities, and a positive disposition toward working in a team. Such KSAs comprise teamwork (Cannon-Bowers et al. 1995;, Sims, Salas, and Burke 2004).
Based on its definition alone, it is easy to see how teamwork is critical for the delivery of health care. Physicians, nurses, pharmacists, technicians, and other health professionals must coordinate their activities to deliver safe and efficient patient care. As specified in our definition of a team, health care workers perform interdependent tasks (e.g., a surgeon cannot operate until a patient is anesthetized) while functioning in specific roles (e.g., surgeon, surgical assistant, anesthesiologist) and sharing the common goal of safe care. However, despite the importance of teamwork in health care, most clinical units continue to function as discrete and separate collections of professionals (Knox and Simpson 2004). This is partially due to the fact that members of these teams are rarely trained together; furthermore, they often come from separate disciplines and diverse educational programs.
Given the interdisciplinary nature of the work and the necessity of cooperation among the workers who perform it, teamwork is critical for ensuring patient safety. Teams make fewer mistakes than do individuals, especially when each team member knows his or her responsibilities, as well as those of other team members (Smith-Jentsch, Salas and Baker 1996;Volpe et al. 1996, Sims et al. 2004,). However, simply installing a team structure does not automatically ensure it will operate effectively. Teamwork is not an automatic consequence of co-locating people together and depends on a willingness to cooperate for a shared goal. Teamwork does not require that team members work together on a permanent basis. Teamwork is sustained by a commitment to a shared set of team KSAs rather than permanent assignments that carry over from day to day (Morey et al. 2002).
Critical Components of Teamwork
Extensive research on teamwork during the past 20 years (McIntyre, Salas, and Glickman 1989;Howard et al. 1992; Helmreich and Foushee 1993;Holzman et al. 1995)suggests that teamwork is defined by a set of interrelated KSAs that facilitate coordinated, adaptive performance (Cannon-Bowers et al. 1995;, Salas, Bowers and Cannon-Bowers 1995;Baker et al. 2003). Teamwork is distinct from taskwork (e.g., surgical skill) but both are required for teams to be effective in complex environments (Morgan et al. 1986). Furthermore, in health care, knowledge and skill at the task are not enough. Teamwork depends on each team member being able to anticipate the needs of others; adjust to each other's actions, and have a shared understanding of how a procedure should happen (e.g., knowing the steps in an appendectomy
Recently, researchers have begun to identify skills that define team performance in health care. This line of research began with the work of Gaba et al. (2001) who developed Anesthesia Crisis Resource Management (ACRM). ACRM was designed to help anesthesiologists effectively manage crises by working in multidisciplinary teams that include physicians, nurses, technicians, and other medical professionals (Howard et al. 1992; Gaba et al., 1998, 2001). ACRM uses patient simulators to provide training in specific technical and generic teamwork skills. The simulated anesthesia environment consists of a real operating room with standard equipment and situations requiring actual performance of clinical interventions. A life-like mannequin with appropriate breath and heart sounds permits team members to perform clinical procedures such as endotracheal intubation and infusion of intravenous drugs. Scenarios presented include overdose of inhalation anesthetic, cardiac arrest, and complete power failure (Holzman et al. 1995). The team skills trained in this simulated environment include making inquiries and assertions, communicating, giving and receiving feedback, exerting leadership, maintaining a positive group climate, and reevaluating actions.
In addition to anesthesia, a number of researchers have recently begun to identify the KSA requirements of teamwork in other health disciplines. For example, Healey, Undre, and Vincent (2004) have developed the Observational Assessment for Teamwork in Surgery (OTAS) to assess cooperation, leadership, coordination, awareness, and communication in surgical teams (Healeyet al. 2004). Thomas, Sexton, and Helmreich (2004) developed 10 behavioral markers for teamwork in neonatal resuscitation teams and Flin and Maron (2004) have identified nontechnical skill requirements for teams in acute medicine.
These studies encapsulate the core KSA requirements for physicians, nurses, and other health care professionals to function effectively in a wide variety of health care teams. Although different researchers use different terminology to define these KSA requirements (e.g., Thomas et al. 2004) identify one of their behavioral markers as “Information Sharing,” while others identify a requirement for “Communication” (Leonard, Tarrant 2001; Flin and Maron 2004; Healey et al. 2004), we argue, much as have Salas and colleagues, that these generic KSAs can be clustered into eight broad competencies of teamwork (Sims et al. 2004). These competencies must be possessed by health care professionals so they can perform (1) in the variety of teams of which they are part and (2) a variety tasks requiring coordination in day-to-day practice. Table 1 presents each KSA, its definition, and behavioral examples.
Team KSA Competencies and Outcomes
Characteristics of Effective Teams
Teams whose members possess a shared commitment to the KSAs presented in Table 1 have been shown to out perform teams whose members do not possess these attributes (Smith-Jentsch et al. 1996; Leonard, Tarrant 2001; Salas et al. 2001; O'Shea et al. 2003). One important fact to note about these KSAs is that they are all individual as opposed to team-level competencies. In other words, team members bring these KSAs to each team task they engage in; the competencies are not unique to the task or the team. When team members work together on a more permanent basis, these competencies are refined over time (they are tailored within the team) and some additional competencies emerge (e.g., knowledge of teammate characteristics).
Regardless of whether a team has consistent membership or not, when team members possess the KSAs in Table 1 they are able to perform as a highly reliable and efficient system. Table 2 presents a list of characteristics of effective teams that make them reliable and efficient. These characteristics are similar to the properties that embody HROs, which we describe in the next section. However, before our review of HROs, we describe the mechanisms by which effective team performance can be achieved.
Characteristics of Effective Teams
How to Promote Teamwork
There are three basic strategies by which effective teamwork can be achieved. First, specific individuals, who have the correct KSAs, can be selected to participate in a team or to perform team-based work (Klimoski and Mohammed 1994). This strategy requires precise measurement of individual-level team competencies and a correct balancing of task-oriented and team-oriented KSAs among team members. Second, teamwork can be enhanced by modifying tasks, workflow, or structure (Campion, Medsker and Higgs 1993;). In other words, one can examine the environmental conditions in which team-based work occurs and reengineer these conditions accordingly. Finally, individual team member competencies can be developed through training (Cannon-Bowers et al. 1995, 1989;, Cannon-Bowers and Salas 1997; Leonard, Graham, and Bonacum 2004,). Team training has been the most widely applied strategy to improve team performance.
Team training is defined as applying a set of instructional strategies that rely on well-tested tools (e.g., simulators, lectures, videos) (Salas et al. 1999;, Salas, Rhodenizer and Bowers 2000). Effective team training reflects general principles of learning theory, presents information about requisite team behaviors, affords team members the opportunity to practice the skills they are learning, and provides remedial feedback.
A great deal of research has been devoted to the most effective strategies and techniques for training specific team KSAs. A comprehensive review of this research has presented an extensive collection of principles and guidelines concerning the design and delivery of team training. For example, guidelines exist for assertiveness training (Smith-Jentsch et al. 1996), cross-training (Volpe et al. 1996), stress management training (Driskell and Johnston 1998), and team self-correction (Smith-Jentsch et al. 1998).
Team training programs have been an essential component of the airline industry's efforts to achieve high reliability. For over 30 years, crew resource management (CRM) has been a critical part of most airlines' efforts to improve their margin of safety. Recent research suggests that CRM training results in heightened safety-related attitudes; improved communication, coordination, and decision-making behaviors; and enhanced error-management skills (Wiener, Kanki and Helmreich 1993;). Helmreich and Merrit 1998;). CRM training has also demonstrated consistently positive results across a wide range of team structures, including pilot crews, maintenance crews, dispatch crews, and air traffic control teams (Helmreich, Foushee 1993; Oser et al. 2001; Smith-Jentsch et al. 2001
Interestingly, CRM's effect on the ultimate criterion—a reduction in the number of accidents—has yet to be empirically established (Salas et al. 2001). However, accidents represent a poor criterion methodologically because they exhibit an extremely low base rate (Helmreich and Foushee 1993). Thus, researchers have relied on surrogate measures—like improvements in team-related knowledge and skills, behavioral demonstrations of CRM skills on simulated flights, instructor evaluations of trained versus untrained crews, and changes in an organization's safety culture—to demonstrate the effectiveness of CRM training Helmreich and Foushee 1993; Hansberger and Holt, Boehm-Davis 1999; Ikomi et al. 1999; Incalcaterra and Holt 1999; Holt, Boehm-Davis, and Hansberger 2001).
Similar to the airline industry, some type of formal team training is now a major component of training in most branches of the United States Armed Forces. For example, all branches of the Armed Forces give their aircrews a military version of CRM, ranging from Fighter Resource Management (FRM) for single-seat fighter pilots to CRM training for the large crews that staff transport and patrol aircraft (Spiker et al. 1998). In addition, many sailors, soldiers, airmen, and marines receive team training. For example, the Navy, having tested several team-training approaches (Serfaty, Entin, and Johnston 1998), has adopted an approach called Team Dimensional Training (TDT), which resulted from the TADMUS program (Cannon-Bowers and Salas 1998). TDT addresses team-related knowledge and skills, provides practice in briefing and debriefing, and trains trainers and team leaders to evaluate and critique team skills (Tannenbaum, Smith-Jentsch, and Behson 1998).
TEAMS AND HROs
The characteristics of HROs dictate that teamwork is an essential component of such organizations. HROs will not achieve high reliability unless its members are able to effectively and efficiently coordinate their activities. In the previous section, we have spent considerable time clarifying what we mean by “teams” and “teamwork.” In this section we define the concept of high reliability, review the key characteristics of HROs, and demonstrate why teamwork is so critical in such organizations.
HROs are defined by their potential for causing failures that lead to catastrophic consequences. If the potential is high (thousands of dramatic failures), but the actual number of failures is low, the organization is an HRO (Roberts 1990a). For example, a nuclear power plant failure could result in horrific consequences for its surrounding community, although such failures are extremely rare. The same can be said for many U.S. hospitals—there are thousands of opportunities for major accidents everyday. Although the IOM estimated that 98,000 preventable deaths occur per year, the actual occurrence of medical error resulting in deaths is extremely low (Kohn et al. 1999).
HROs are those organizations that function in hazardous, fast-paced, and highly complex technological systems essentially error-free for long periods of time (Roberts 1990a, b). Roberts and Rousseau (1989) identified eight characteristics of HROs: (1) hypercomplexity, (2) tightly coupled, (3) extreme hierarchical differentiation, (4) many decision makers working in complex communication networks, (5) high degree of accountability, (6) frequent, immediate feedback regarding decisions, (7) compressed time factors, and (8) synchronized outcomes (Roberts and Rousseau 1989). Below we review each of these characteristics, demonstrate how teamwork is an essential component of effective performance in such organizations, and provide a health care example, where appropriate.
Hypercomplexity is defined as an extreme variety of components, systems, and levels, each having their own standard procedures, training routines, and command hierarchy (Roberts and Rousseau 1989;). Based on its definition alone, successful performance in hypercomplex environments relies upon multiteam systems and teamwork is an essential component of such environments. For example, Roberts and Rousseau describe aircraft carrier operations is indicative of hypercomplexity. Pilots, air traffic controllers, dispatchers, ground crews, and many others must work collectively to launch and recover aircraft. These interdependent teams (e.g., air traffic control team, aircrews, maintenance teams, etc.), must coordinate their activities and efficiently monitor each other's performance.
Similarly, the delivery of health care occurs in a hypercomplex environment that is dependent on multiteam systems. Even though health care workers have historically operated in distinct silos and have been trained in separate professions and possess distinct expertise, these individual must coordinate to deliver safe care. At the most basic level, physicians must accurately communicate treatment information to the nurse based, in part, on information the nurse presents to the physician regarding the patient's condition. Orders are written on the basis of this discussion and the physician's examination of the patient. These orders are distributed to the pharmacy, X-ray, labs, physical therapy, etc., so that other health care professionals can collect additional information to provide insight regarding the patients or initiate treatment.
Tight coupling is defined as reciprocal interdependence across many units and levels. Tight coupling relates to task interdependence, which is the defining characteristic of teams. That is, tasks performed by one member of the team are dependent on tasks performed by other members of the team and the performance of these tasks must be coordinated among team members for effective team performance (delivery of safe care). For example, in health care an emergency C-section is a tightly coupled event that involves several different members of the labor and delivery team. The nurse handling the case is typically the first to observe fetal distress and must communicate this information to the attending physician. The doctor must decide if a C-section is necessary based upon the information the nurse provides and review of information collected from the fetal heart rate monitor. If the attending decides to operate, appropriate staff must be notified (anesthesiologist, neonatologist, or pediatrician), and the patient must be moved to surgery. Before making the initial incision, the patient must be properly anesthetized and the staff should be briefed as to the status of the patient and the baby. This process, which involves a series of interdependent steps, can take place in a matter of minutes depending on the history of the case and the level of fetal distress observed. For such a sequence to run smoothly, teams must use effective communication and have a shared understanding of the mother's and baby's condition. The Joint Commission (JCAHO) reported that ineffective communication resulted in 70 percent of all preventable errors involving death or serious injury from 1995 to 2003 (JCAHO 2004).
Extreme hierarchical differentiation is defined as an organizational structure in which levels and roles are clearly differentiated. This characteristic is also true of most health care teams. Physicians tend to be at the top of this hierarchy with the case or treatment resulting from their directions. Therefore, a great deal of coordination is necessary to keep physicians, nurses, and technicians working together as a cohesive unit. Unfortunately hierarchy often makes it more difficult for medical teams to achieve this level of coordination and cohesiveness. In fact, research suggests that the extreme hierarchical difference between physicians and nurses in particular can contribute to dysfunctional communication yielding less than optimal patient care Keenan Cooke and Hillis,1998; Knox and Simpson 2004).
Although most medical teams are hierarchical, high-reliability teams trained in teamwork exhibit characteristics such as assertiveness and mutual trust, which reduce the negative effects of hierarchy. Mutual trust, an essential teamwork KSA, involves a shared belief that team members will protect and support the interests of their team (Sims et al. 2004). Team members with mutual trust are willing to admit to mistakes and accept and appreciate feedback Bandow 2001; Webber 2002). This allows team members to firmly assert their concerns even to a higher-ranking team member without fear of reprisal.
Another key characteristic of HROs is that they contain many decision makers working in complex communication networks (Roberts and Rousseau 1989). This characteristic personifies most health care teams. First, team members continually need to make important decisions concerning patient care (e.g., start an IV, induce labor, administer narcotics, admit patient). Consequences of these decisions clearly have implications on the ultimate well being and safety of patients. As most teams in health care are comprised of four to six unique individuals, however, decisions are not always unanimous. Second, as different team members are trained separately in their respective professions (e.g., medical school and nursing school), they have learned to communicate differently and have varying styles of conveying information depending on their role. Fortunately, new and emerging techniques like the Situation Background Assessment Recommendation (SBAR) strategy have been used in health care to overcome such communication difficulties with positive results (Leonard et al. 2004). This particular strategy facilitates clear and concise communication among members of health care teams by providing an easy-to-remember acronym used for framing critical conversations.
A high degree of accountability in HROs is characterized by the severe consequences that can result from errors (Roberts and Rousseau 1989). Although serve consequences may be characteristic of all teams (e.g., project teams), in health care, the consequence of a mistake can often be death the patient. Preventable medical errors that result in loss of human life eternally affect the patient's family, the staff that tended to the patient, the community, and the hospital's reputation. However, even small mistakes resulting in patient harm yield grave consequences, yet not all medical team members are held equally accountable when errors do occur. More often then not, it is the physician in charge of the patient's care that gets the brunt of the blame for any mistakes made. Malpractice lawsuits or the possibility of losing one's license are very real outcomes that add to pressures felt by already stressed physicians. Consequences may also be present for the hospital such as loss of accreditation, and negative media attention.
HROs are also characterized by “immediate feedback” resulting from their decisions; the plane crashes; there is a nuclear disaster; the patient is injured (Roberts and Rousseau 1989). In other words, there is an identifiable, measurable outcome associated with HRO performance. Such outcomes are typically an indicator of poor team and/or system processes within the HRO. For example, in aviation 60–80 percent of all accidents are attributed to human error as opposed to anything technically wrong with the aircraft. Similarly, the IOM report points to human error as a major contributor in patient deaths (Foushee 1984; Kohn et al. 1999).
Immediate feedback is also a characteristic of effective team performance. Team members must monitor each other and provide each other feedback to maximize team functioning. However, feedback here focuses on team process and its improvement rather than solely on team outcomes. To ensure that feedback occurs, team members must be trained to deliver timely, behavioral, and specific feedback to one another (using such strategies as TDT). The ability to monitor each other's performance and effectively provide feedback to other team members is a critical facet of achieving higher reliability in health care and elsewhere.
Major HRO activities often occur under compressed time, as in the case of naval flight operations where aircraft are launched and recovered in 48–60-second intervals (Roberts and Rousseau 1989). Somewhat different than the other characteristics, the extent to which this variable is related to team performance is a function on the environment in which the team operates as opposed to the team itself. Some teams operate under compressed time while others do not. The same is true of health care, with a slightly different twist. Routine procedures like childbirth can quickly become a stressful, time compressed situation should a problem arise with the mother or baby. In such cases, teams need to be able to quickly adapt. Team members may have to be quickly added and integrated into the team and task—anesthesiologist, emergency response team—or existing members may have to take on new roles—OB-GYN converts from coaching the mother through a normal delivery to conducting an emergency C-section.
The last characteristic of HROs is that critical outcomes occur simultaneously (Roberts and Rousseau 1989). As discussed earlier, team members work together on interdependent tasks. This is what separates teams from groups or individuals working in isolation. Interdependency creates the need for synchronization of activities and outcomes. For instance, when delivering a baby, each member of a labor and delivery team is actively engaged in different aspects of the process yet their actions are synchronized.
As can be seen from the description above, the HRO environment demands teamwork. Teamwork, or the KSAs that comprise it (refer to Table 1), are critical for successful performance in organizations that are hypercomplex, tightly coupled, hierarchical, time compressed, and rely upon synchronized outcomes (Sims et al. 2004). Aviation, the military, and now health care acknowledge the criticality of teamwork in achieving high reliability despite data that show a direct relation between team training and the ultimate criterion, a reduction in errors (e.g., accidents, deaths, etc.) (Salas et al. 2001). However, the science of teamwork and team training is still evolving, particularly in health care. The IOM report did much to stimulate research on health care teams, but much of this early work relied upon direct transitions from the commercial airlines to health care. Despite the work of Gaba and colleagues, it is only within the last 3 years that the science of health care teams has really begun to emerge and take hold (Baker et al. 2003). As a result, a number of questions remain that health care must address to have a firm understanding of teamwork and its relation to patient safety and high reliability. Direct transitions from aviation without additional study are insufficient. In the next section, we outline a series challenges the health care community must address to better understand health care team performance, how to maximize this performance, and ultimately improve patient safety.
Throughout this paper we focused on three basic themes. First, the delivery of health care, by its nature, requires that organizations providing such services act as HROs. Patients expect error-free care (Knox and Simpson 2004). Second, teamwork is an essential component of HROs. Although not the sole determinant of high reliability, HROs are typically comprised of teams embedded in multiteam systems and effective teamwork is critical for success in environments that demand high reliability (Wilson et al. 2005). Finally, the easiest way of improving teamwork is through training. Team training has been effectively implemented in the commercial airlines and the military with positive results (Salas et al. 2001). Such training programs are now emerging in health care with potentially similar benefits (Baker, Beaubien, and Holtzman 2003).
Nonetheless, teamwork in health care is an emerging science. To move this science forward so that findings can be transitioned and implemented, we recommend that health researchers, quality improvement specialists, regulatory bodies, and others seek to address the following challenges.
Challenge 1: A Theoretical Model of Team Performance in Health Care Should Be Developed
To date, research has not developed a comprehensive model of team performance in medical settings; consequently, existing and emerging team training programs are not grounded in a scientific understanding of what comprises effective teamwork in health care (Baker et al. 2003, 2003). In our review of the team literature, we recommend the teamwork framework advocated by Salas and colleagues and this framework serves as the foundation for AHRQ's TeamSTEPPS training program. However, the Salas model needs to be tested in health care to determine (1) the relations among predictors of performance (team KSAs) and (2) the relations between predictors (KSAs) and outcome criteria (e.g., quality of care, error management, efficiency, etc.).
Challenge 2: Proven Instructional Strategies Should Be the Basis for Team-Training Programs in Health Care
Team training is the most practiced strategy for enhancing team performance and improving team outcomes. Most HROs provide some form of team training and the science of team training has developed and validated numerous training strategies. Through a variety of formats and objectives, these strategies extend beyond CRM training. However, health care seems to be focused on adapting CRM programs derived directly from aviation (Baker et al. 2003) and not implementing other strategies which have been shown to be effective and may by more appropriate (i.e., team dimensional training). Therefore, we challenge health care to (1) move beyond CRM and look to other validated methods, (2) use these strategies wherever possible as the foundation for team training programs, and (3) test and refine these strategies to ensure that they effectively generalize to health care teams. It is only through implementation, testing, and refinement that health care will be able to understand and demonstrate how to enhance teamwork.
Challenge 3: Team-Training Strategies Must Be Further Adapted to Specific Health Care Needs
Similar to our previous recommendation, we are convinced that no single model of team training, like CRM, can be applied across all health care services and contexts when attempting to achieve high reliability. For purposes of this discussion, we define a “service” as a medical specialty or subspecialty, such as emergency medicine, general or family medicine, intensive care, general surgery, obstetrics, etc. Medical services differ dramatically across a variety of criteria: size, purpose, duration, redundancy of expertise, decision time, and consequence of error, to name but a few.
In addition, services operate in a number of diverse contexts. As an example, emergency medicine providers function in hospital emergency departments, in emergency-response mobile units, and on battlefields. Similarly, urban and rural providers operate in independent or multipractitioner offices, as well as in community walk-in clinics. Neither the competencies that impel successful teamwork nor an optimal team-training strategy can be expected to generalize across all these contexts. And, of course, not all members within the same team will necessarily need the same KSAs.
Therefore, in addition to the core competency taxonomy, the science of teamwork in health care must seek to develop service-specific taxonomies. These putative taxonomies would not be redundant with the generic, core competency taxonomy. Rather, a specific taxonomy would denote the specific KSA requirements that are central to teamwork in a given service, thus maximizing team performance within that service. The task content and procedures that define this service would drive the identification of relevant team competencies.
Virtually no previous research has addressed the manner in which differences within and among health care services should be reflected in service-specific taxonomies and customized training. Yet we find this issue sufficiently compelling, particularly with the context of HROs, because it suggests that customized solutions are warranted to achieve high reliability. Therefore, while a generic taxonomy of team KSAs is the foundation to teamwork in health care, we argue that these competencies must be refined and training must be tailored to a specific service to maximize team performance and safety.
Challenge 4: Team Training Must Be Institutionalized throughout Health Care and Professional Training
Finally, health care must work to integrate teamwork throughout every level of training and education of health care professionals. By “integrated” we mean instruction, measurement, and feedback on critical team KSAs occurs as part of a health care professional's technical education and training. Using this approach, team concepts become a part of everyday practice.
Several initiatives by physician education, certification, and licensing boards have already begun to move health care toward integrating team concepts. For example, the ACGME has identified several teamwork-related competencies that residents must master as part of the ACGME outcomes project (http://www.acgme.org). Similarly, AAMC funded a “critical incident” analysis to investigate the behaviors that result in successful and unsuccessful performance during medical school and residency (Adams et al. 2001). Although not originally targeted toward team performance, the results revealed the importance of a number of teamwork-related competencies.
Building off existing initiatives, we believe that the structure of health care, as currently conceptualized, offers appropriate junctures where teamwork skills could be evaluated. For example, like the examinations that are constructed for board certification in medical specialties, it might ultimately be useful to develop a board certification test for teamwork. Such an exam might combine a written test of knowledge and situational judgment with performance in a simulated scenario. Because the board examinations are practice-specific, their teamwork component could assess practice-specific teamwork competencies. In addition, the JCAHO currently evaluates hospitals on criteria that range from medical practices to managerial systems to facilities maintenance. At some point in the future, folding team competency criteria into the JCAHO evaluation might focus providers' attention on the importance of teamwork in medical settings, as well as yielding valuable research data.
In summary, we have argued that teamwork is an essential component of achieving high reliability for health care organizations. HRO environments demand teamwork and, as a result, the science of team training can provide great insights and proven techniques for improving performance within such organizations. In closing, we recommend that health care gain traction from the more than 20 years of research on team performance and training and that these principles be first tested and then integrated into the practice of health care and the training of health professionals. Although this will take considerable time, perhaps spanning a generation, this approach has been one of the key drivers in other industries achieving the highest reliability possible. We believe that the challenges we have presented here provide a roadmap with which health care can continue.
- Adams K A, Goodwin G F, Searcy C A, Norris D G, Oppler S H. Technical Report Commissioned by the Association of American Medical Colleges. Washington, DC: American Institutes for Research; 2001. Development of a Performance Model of the Medical Education Process.
- Alonso A, Baker D, Day R, Holtzman A, King H, Toomey L, Salas E. Reducing Medical Error in the Military Health System: How Can Team Training Help. Human Resource Management Review. 2006 in press.
- Baker D P, Beaubien J M, Holtzman A K. DoD Medical Team Training Programs: An Independent Case Study Analysis. Washington, DC: American Institutes for Research; 2003.
- Baker D P, Gustafson S, Beaubien J M, Salas E, Barach P. Medical Teamwork and Patient Safety: The Evidence-Based Relation. Washington, DC: American Institutes for Research; 2003.
- Bandow D. Time to Create Sound Teamwork. Journal for Quality and Participation. 2001;41:41–7.
- Barach P, Weingart M. Trauma Team Performance. In: Wilson W, Grande C, Hoyt D, editors. Trauma: Resuscitation, Anesthesia, Surgery, & Critical Care. New York: Dekker Inc; 2004.
- Campion M A, Medsker G J, Higgs A C. Relations between Work Group Characteristics and Effectiveness: Implications for Designing Effective Work Groups. Personnel Psychology. 1993;46:823–50.
- Campion M A, Papper E M, Medsker G J. Relations between Work Team Effectiveness: A Replication and Extension. Personnel Psychology. 1996;49:429–52.
- Cannon-Bowers J A, Prince C, Salas E, Owens J M, Morgan B B, Jr, Gonos G H. Proceedings of the Interservice/Industry Training Simulation and Education Conference 11th Annual Meeting. Arlington, VA: National Defense Industrial Association; 1989. Determining Aircrew Coordination Training Effectiveness; pp. 128–36.
- Cannon-Bowers J A, Salas E. Teamwork Competencies: The Interaction of Team Member Knowledge, Skills, and Attitudes. In: O'Neil H F Jr, editor. Workforce Readiness: Competencies and Assessment. Mahwah, NJ: Erlbaum; 1997. pp. 151–74.
- Cannon-Bowers J A, Salas E. Making Decisions under Stress: Implications for Individual and Team Training. Washington, DC: American Psychological Association; 1998.
- Cannon-Bowers J A, Tannenbaum S I, Salas E, Volpe C E. Defining Competencies and Establishing Team Training Requirements. In: Guzzo R A, Salas E, et al., editors. Team Effectiveness and Decision Making in Organizations. San Francisco: Jossey-Bass; 1995. pp. 333–80.
- Davies J M. Medical Applications of Crew Resource Management. In: Salas E, Bowers C A, Eleana E, editors. Improving Teamwork in Organizations: Applications of Resource Management Training. Mahwah, NJ: Erlbaum; 2001. pp. 265–81.
- Driskell J E, Johnston J H. Stress Exposure Training. In: Cannon-Bowers J A, Salas E, editors. Making Decisions under Stress—Implications for Individual and Team Training. Washington, DC: American Psychological Association; 1998. pp. 191–217.
- Driskell J E, Salas E. Collective Behavior and Team Performance. Human Factors. 1992;34:277–88.
- Flin R, Maron N. Identifying and Training Non-Technical Skills for Teams in Acute Medicine. Quality and Safety in Health Care. 2004;13:i80–4.[PMC free article][PubMed]
- Foushee H C. Dyads and Triads at 35,000 Feet: Factors Affecting Group Processes and Aircrew Performance. American Psychologist. 1984;39:885–93.
- Gaba D M, Howard S K, Fish K J, Smith B E, Sowb Y A. Simulation-Based Training in Anesthesia Crisis Resource Management (ACRM): A Decade of Experience. Simulation and Gaming. 2001;32:175–93.
- Gaba D M, Howard S K, Flanagan B, Smith B E, Fish K J, Botney R. Assessment of Clinical Performance during Simulated Crises Using Both Technical and Behavioral Ratings. Anesthesiology. 1998;89:8–18.[PubMed]
- Hansberger J T, Holt R W, Boehm-Davis D A. Instructor/Evaluator Evaluations of ACRM Effectiveness. In: Jensen R S, editor. Proceedings of the 10th International Symposium on Aviation Psychology, March 5. Columbus, OH: The Ohio State University Press; 1999. pp. 279–82.
- Healey A N, Undre S, Vincent C A. Developing Observational Measures of Performance in Surgical Teams. Quality and Safety in Health Care. 2004;13:i33–40.[PMC free article][PubMed]
- Helmreich R L, Foushee H C. Why Crew Resource Management? Empirical and Theoretical Bases of Human Factors Training in Aviation. In: Weiner E L, Kanki B G, Helmreich R L, editors. Cockpit Resource Management. San Diego: Academic Press; 1993. pp. 3–45.
- Helmreich R L, Merritt A C. Culture at Work in Aviation and Medicine: National, Organizational, and Professional Influences. Brookfield, VT: Ashgate; 1998.
- Holt R W, Boehm-Davis D A, Hansberger J T. Evaluation of Proceduralized CRM at a Regional and Major Carrier. Fairfax, VA: George Mason University; 2001.
- Holzman R S, Cooper J B, Gaba D M, Philip J H, Small S D, Feinstein D. Anesthesia Crisis Resource Management: Real-life Simulation Training in Operating Room Crises. Journal of Clinical Anesthesia. 1995;7:675–87.[PubMed]
- Howard S K, Gaba D M, Fish K J, Yang G, Sarnquist F H. Anesthesia Crisis Resource Management Training: Teaching Anesthesiologists to Handle Critical Incidents. Aviation, Space, and Environmental Medicine. 1992;63:763–70.[PubMed]
- Ikomi P A, Boehm-Davis D A, Holt R W, Incalcaterra K A. Jump Seat Observations of Advanced Crew Resource Management (ACRM) Effectiveness. In: Jensen R S, editor. Proceedings of the 10th International Symposium on Aviation Psychology. 5-3-1999. Columbus, OH: The Ohio State University Press; 1999. pp. 292–97.
- Incalcaterra K A, Holt R W. Pilot Evaluations of ACRM Programs. In: Jensen R S, editor. Proceedings of the 10th International Symposium on Aviation Psychology. 5-3-1999. Columbus, OH: The Ohio State University Press; 1999. pp. 285–90.
- JCAHO. Sentinel Event Statistics. 2004. Available at http://www.jcaho.org/
- Keenan G M, Cooke R, Hillis S L. Norms and Nurse Management of Conflicts: Keys to Understanding Nurse–Physician Collaboration. Research in Nursing and Health. 1998;21:59–72.[PubMed]
- Klein G, Pierce L G. Proceedings of the Sixth ICCRTS Collaboration in the Information Age Track 4: C2 Decision Making and Cognitive Analysis. 2001. Adaptive Teams.
- Klimoski R, Mohammed S. Team Mental Model: Construct or Metaphor. Journal of Management. 1994;20:403–47.
- Knox G E, Simpson K R. Teamwork: The Fundamental Building Block of High-Reliability Organizations and Patient Safety. In: Youngberg B J, Hatlie M J, editors. Patient Safety Handbook. Boston: Jones and Bartlett; 2004. pp. 379–415.
- Kohn L T, Corrigan J M, Donaldson M S. To Err Is Human. Washington, DC: National Academy Press; 1999.
- Kozlowski S W, Gully S M, Nason E R, Smith E M. Developing Adaptive Teams: A Theory of Compilation and Performance across Levels and Time. In: Ilgen D R, Pulakos E D, editors. The Changing Nature of Performance: Implications for Staffing, Motivation, and Development. San Francisco: Jossey-Bass; 1999. pp. 240–92.
- Leonard M, Graham S, Bonacum D. The Human Factor: The Critical Importance of Effective Teamwork and Communication in Providing Safe Care. Quality and Safety in Health Care. 2004;13:i85–90.[PMC free article][PubMed]
- Leonard M, Tarrant C A. Culture, Systems, and Human Factors—Two Tales of Patient Safety: The KP Colorado Region's Experience. Permanente Journal. 2001;5:46–9.
- Mathieu J E, Heffner T S, Goodwin G F, Salas E. The Influence of Shared Mental Models on Team Process and Performance. Journal of Applied Psychology. 2000;85:273–83.[PubMed]
- McIntyre R M, Salas E. Measuring and Managing for Team Performance: Emerging Principles from Complex Environments. In: Guzzo R A, Salas E, et al., editors. Team Effectiveness and Decision Making in Organizations. San Francisco: Jossey-Bass; 1995. pp. 9–45.
- McIntyre R M, Salas E, Glickman A S. Team Research in the 80s: Lessons Learned. Orlando, FL: Naval Training Systems Center; 1989.
- Morey J C, Simon R, Jay G D, Wears R, Salisbury M, Dukes K A, Berns S D. Error Reduction and Performance Improvement in the Emergency Department Through Formal Teamwork Training: Evaluation Results of the MedTeams Project. Health Services Research. 2002;37:1553–81.[PMC free article][PubMed]
- Morgan B B, Glickman A S, Woodward E A, Blaiwes A S, Salas E. Measurement of Team Behaviors in a Navy Environment. Technical Report No. NTSC TR-86-014. Orlando, FL: Naval Training Systems Center; 1986.
- O'Shea P G, Driskell J E, Goodwin G F, Zbylut M L, Weiss S M. Development of a Conditional Reasoning Measure of Team Orientation. Arlington, VA: U. S. Army Research Institute for the Behavioral and Social Sciences; 2004. (ARI Research Note RN-2004-10)
- O'Shea G, Driskell J E, Goodwin G F, Salas E, Weiss S. Assessment of Team Competencies: Development and Validation of a Conditional Reasoning Measure of Team Orientation. 2003.
- Oser R L, Salas E, Merket D C, Bowers C A. Applying Resource Management Training in Naval Aviation: A Methodology and Lessons Learned. In: Salas E, Bowers C A, Edens E, editors. Improving Teamwork in Organizations: Applications of Resource Management Training. Mahwah, NJ: Erlbaum; 2001. pp. 283–301.
- Porter COLH, Hollenbeck J R, Ilgen D R, Ellis APJ, West B J, Moon H. Backup Behavior in Teams: The Role of Personality and Legitimacy of Need. Journal of Applied Psychology. 2003;88:391–403.[PubMed]
- Roberts K H. Managing High Reliability Organizations. California Management Review. 1990a:101–13.
- Roberts K H. Some Characteristics of High Reliability Organizations. Organization Science. 1990b;1:160–77.
- Roberts K H, Rousseau D M. Research in Nearly Failure-Free, High-Reliability Organizations: Having the Bubble. IEEE Transactions on Engineering Management. 1989;36:132–9.
- Sachs B P. A 38-Year-Old Woman with Fetal Loss and Hysterectomy. Journal of the American Medical Association. 2005;294(7):833–40.[PubMed]
- Salas E, Bowers C A, Cannon-Bowers J A. Military Team Research: 10 Years of Progress. Military Psychology. 1995;7:55–75.
- Salas E, Burke S C, Bowers C A, Wilson K A. Team Training in the Skies: Does Crew Resource Management (CRM) Training Work. Human Factors. 2001;43:641–74.[PubMed]
- Salas E, Cannon-Bowers J A. The Science of Training: A Decade of Progress. Annual Review of Psychology. 2000;52:471–99.[PubMed]
- Salas E, Dickinson T L, Converse S A. Toward an Understanding of Team Performance and Training. In: Swezey R W, Salas E, editors. Teams: Their Training and Performance. Norwood, NJ: Ablex; 1992. pp. 3–29.
- Salas E, Rhodenizer L, Bowers C A. The Design and Delivery of Crew Resource Management Training: Exploiting Available Resources. Human Factors. 2000;42:490–511.[PubMed]
- Salas E, Rozell D, Mullen B, Driskell J E. The Effect of Team Building on Performance: An Integration. Small Group Research. 1999;30:309–39.
- Salas E, Sims D E, Klein C. Cooperation and Teamwork at Work. In: Spielberger C D, editor. Encyclopedia of Applied Psychology. Vol. 1. San Diego: Academic Press; 2004. pp. 497–505.
- Serfaty D, Entin E E, Johnston J H. Team Coordination Training. In: Cannon-Bowers J A, Salas E, editors. Making Decisions under Stress—Implications for Individual and Team Training. Washington, DC: American Psychological Association; 1998. pp. 221–45.
- Shamir B. Calculations, Values and Entities. Human Relations. 1990;43:313, 32.
- Sims D E, Salas E, Burke S C. 19th Annual Meeting of the Society for Industrial and Organizational Psychology. Chicago, IL: 2004. Is There a ‘Big Five’ in Teamwork?
- Smith-Jentsch K A, Baker D P, Salas E, Cannon-Bowers J A. Uncovering Differences in Team Competency Requirements: The Case of Air Traffic Control Teams. In: Salas E, Bowers C A, Edens E, editors. Improving Teamwork in Organizations: Applications of Resource Management Training. Mahwah, NJ: Erlbaum; 2001. pp. 31–54.
- Smith-Jentsch K A, Salas E, Baker D P. Training Team Performance-Related Assertiveness. Personnel Psychology. 1996;49:909–36.
- Smith-Jentsch K A, Zeisig R L, Acton B, McPherson J A. Team Dimensional Training. In: Cannon-Bowers J A, Salas E, editors. Making Decisions under Stress: Implications for Individual and Team Training. Washington, DC: American Psychological Association; 1998. pp. 271–97.
- Spiker V A, Silverman D R, Tourville S J, Nullmeyer R I. Tactical Team Resource Management Effects on Combat Mission Training Performance. 1998. Report No. USAF AMRL Technical Report AL-HR-TR-1997-0137. Brooks Air Force Base: U.S. Air Force Systems/Material Command.
- Stout R J, Cannon-Bowers J A, Salas E. The Role of Shared Mental Models in Developing Team Situational Awareness: Implications for Team Training. Training Resource Journal. 1996;2:85–116.
- Tannenbaum S I, Smith-Jentsch K A, Behson S J. Training Team Leaders to Facilitate Team Learning and Performance. In: Cannon-Bowers J A, Salas E, editors. Making Decisions under Stress: Implications for Individual and Team Training. Washington, DC: American Psychological Association; 1998. pp. 247–70.
- Thomas E J, Sexton J B, Helmreich R L. Translating Teamwork Behaviors from Aviation to Healthcare: Development of Behavioural Markers for Neonatal Resuscitation. Quality and Safety in Health Care. 2004;13:i57–64.[PMC free article][PubMed]
- Volpe C E, Cannon-Bowers J A, Salas E, Spector P E. The Impact of Cross Training on Team Functioning: An Empirical Investigation. Human Factors. 1996;38:87–100.[PubMed]
- Wagner J A. Studies of Individualism-Collectivism: Effects on Cooperation in Groups. Academy of Management Journal. 1995;38:152–72.
- Webber S S. Leadership and Trust Facilitating Cross-Functional Team Success. Journal of Management Development. 2002;21:201–14.
- Wiener E L, Kanki B G, Helmreich R L. Cockpit Resource Management. San Diego: Academic Press; 1993.
- Wilson K A, Burke C S, Priest H, Salas E. Promoting Health Care Safety through Training High Reliability Teams. Quality and Safety in Health Care. 2005;14:303–9.[PMC free article][PubMed]
Teamwork Is Key To Solving Complex Research Problems Author: ELIZABETH CULOTTA
Every team needs members who fill four basic roles, according to management consultant Glenn Parker, author of Team Players and Teamwork (San Francisco, Jossey-Bass Inc., 1990). One person can play more than one role, but most researchers interviewed by The Scientist agree that in successful teams, someone fills each of these niches.
The contributor: Contributors work on a piece of the project and deliver their expertise to solve problems. The majority of scientists fall into this category, says Parker. For example, in teams that model climate, a contributor might supply crucial temperature data. "They might not say much, but they'll contribute that one key piece of data, and that's what you need," says Jonathan Overpeck, a paleoclimatologist at the National Atmospheric and Oceanic Administration in Boulder, Colo.
The communicator: This person keeps track of social interactions, heals conflict, and keeps various team members informed. Communicators are sometimes scarce on scientific teams, and so may be especially valuable, says Parker. For example, computer scientist James Coggins of the University of North Carolina, Chapel Hill, is collaborating with ophthalmologists, physicists, and mathematicians on a new way to mathematically describe the shape of the cornea. One member, a research associate in ophthalmology, serves as the glue that keeps the group together. "She's the communications hub," says Coggins. "I go and tell her what I'm doing, and she'll tell the ophthalmologist, and she'll tell me what the physicist and mathematician are doing."
The challenger: This is the skeptic, who questions others' ideas and results. In Coggins's vision project, "the ophthalmologist is our challenger. The others are learning that even if their idea is a great mathematical idea, it's of no use if ophthalmologists can't understand how to use it." Perhaps reflecting the nature of the scientific method, scientific teams usually have no shortage of challengers, says Parker. That can lead to an overload of conflict. "There are lots of good questions being asked, but there's a tendency to perhaps enjoy the fight instead of what they're fighting for," says Parker.
The collaborator: This person articulates a shared vision for a project and pitches in wherever needed. Having a shared mission is crucial, scientists agree. "If the goal's not clear, then people may work very hard--in opposite directions," says Stephen Evola, a research director at Ciba Seeds Agricultural Biotechnology Unit in Research Triangle Park, N.C. For example, he helped lead an 18-member team that recently succeeded in creating a corn plant resistant to a troublesome pest, the corn borer. Researchers had to develop new technology to insert genes into corn--a tough problem that easily might have become a goal in its own right. Evola helped make sure everyone kept the larger team goal in mind.
All these roles are important, team experts say, but the most valuable team members are those who can identify what roles a team is missing--and pitch in to play that character.
The team included geologists, paleoecologists, marine scientists, glaciologists, and climate modelers and was called COHMAP, for Cooperative Holocene Mapping Project. Together, the researchers tested the computer model against data from the geologic record, and they found some areas of striking agreement. Their results, including a 1988 Science paper with 33 authors (COHMAP Members, 241:1043), are widely regarded as a seminal contribution to the field.
The secret to their success? Interdisciplinary teamwork, says Kutzbach. The team had access to different lines of independent evidence of past climatic conditions. "When you found someone in another discipline singing the same song, it was really nice," he says. "And even if there was disagreement, that told us where there were problems."
Indeed, researchers in many fields now recognize that no single person is able to contribute all the necessary expertise to solve increasingly complex problems. And so, from universities to corporate labs, scientists are signing on to teams.
But group dynamics among scientists are not always smooth. Scientists are trained to nurture an individual vision, and it's not always easy for them to harness their efforts to a team. How can managers and team members help make it work?
Research teams need a shared mission, a good organizational structure, and plenty of attention to interpersonal interactions, say Kutzbach and other experienced team leaders. And though few scientists get formal training in teamwork, successful teams offer lessons in collaboration.
For example, COHMAP served as a classroom for Jonathan Overpeck, now a paleoclimatologist at the National Oceanic and Atmospheric Administration in Boulder, Colo. Overpeck got his Ph.D. in 1985 while working on a piece of the COHMAP project. "I'm a COHMAP kid," he says. "I remember those annual meetings--in Madison, [Wis.] in the summer--as the best I've ever been to. Since then, I've been working my collaborations along the same lines as I saw at COHMAP."
One lesson Overpeck learned: the importance of social interactions. At COHMAP meetings, he talked with scientists from all over the world while sharing beers; today he finds that collaborations "with an element of fun" are most productive.
The social aspects of teamwork are even more important when the team is multinational, team leaders say. "I've seen it again and again at meetings in Europe: We'll socialize in the evening, and the next morning you go to the meeting and people are more friendly, more open--and more gets done," says Overpeck.
Overpeck has identified one of the key elements in successful teamwork--and one that scientific teams sometimes don't appreciate, says Lawerenceville, N.J.-based management consultant Glenn Parker.
Different cultural backgrounds can create working problems unless people take the time to get to know each other, says Parker. For example, he was called in to resolve conflict in a team that included an Asian-born chemist who was something of a loner. Parker asked team members to tell their life stories. The Asian chemist related how she had been brought up by her grandmother to be self-reliant, and how she had come, alone, to the United States for graduate study, and then stayed on. Knowing her history helped her colleagues understand and respect her independent style, says Parker.
Getting acquainted in the beginning may be good strategy, since the early stages of a collaboration--when the overall goal is defined and responsibilities assigned--may be especially tricky, team leaders say. For example, computer scientist James Coggins of the University of North Carolina, Chapel Hill, is part of a working group of the National Cancer Institute. The team's goal is to design new software to plan radiation therapy. Team members are scattered around the U.S. in three sites, on the East and West coasts and in the Midwest.
The initial plan for dividing up the job was for each site to write a version of software, and send it on to the next site. The receiving site would then manipulate the existing program. But that arrangement was ripe for conflict among sites, so Coggins stepped in with a different plan. All sites agreed on a minimal supporting software platform, and each site was responsible for writing programs to work with that platform. Naturally, the team spent a long time thrashing out the platform--but potential conflicts were settled early in the collaboration, rather than at the end, says Coggins.
Working with scientists hundreds of miles away, as Coggins's group does, creates its own problems and often changes the nature of the interaction, team leaders say. "Even if you're one building away, it makes a difference," says Alistair Glass, director of passive components research at AT&T Bell Laboratories, Murray Hill, N.J. "We have two facilities 35 miles apart, and it's very clear that it's much harder [to collaborate]. Even within hallways, geography makes a difference. It's important whose lab you're next to."
Physicist John Madey, director of the free-electron laser group at Duke University, agrees that "physical proximity is very important"--and he designed the new laser building at Duke to ensure such proximity. Physicists' offices in the new building are clustered in a group, so no one is left isolated.
Of course, scientists often have no choice: They must work with colleagues across town or even across the ocean. Such collaborations rely on E-mail, phone, fax, and, perhaps, video conferences. But the technology is still no substitute for face- to-face meetings.
Successful long-distance collaborations, therefore, tend to have a very focused goal. Specific tasks are parceled out among far- flung team members, each of whom tends to act as what Parker calls a "contributor" (see story on page 20). For example, an interdisciplinary group led by Lisa Cannon-Albright at the University of Utah School of Medicine in Salt Lake City recently narrowed the search for the gene for familial malignant melanoma, pinning the gene to a region on chromosome 9. The high-profile paper last November (L. Cannon-Albright, et al., Science, 258:1148, 1992) had 16 coauthors from all over the U.S.--but the closest collaborations occurred within a tight core of researchers at Utah, says Cannon-Albright. "Most of the collaborators did different pieces and were added in the last four weeks or so. They gave us information, and we took over from there." The whole team didn't meet together until after the paper was accepted.
So even in this larger team, the day-to-day experience of each scientist was that of working in a small group.
That's true for classic "big science" projects, too, says Gene Fisk, deputy spokesman for the D-zero project at Fermi National Accelerator Laboratory in Batavia, Ill. The D-zero experiment involves about 350 physicists who seek to understand the basic constituents of matter by studying the particles produced when protons and antiprotons crash together. The researchers are organized into groups.
Still, subgroups must rely on each other--and that can sometimes lead to conflict. For example, physicists who analyze data depend on the software that processes and manages data files. Fisk recalls one Saturday morning computer users' meeting, at which one physicist complained that he wasn't getting help from the computer division. Fisk had to serve as the liaison between the dissatisfied physicist and the computer specialists. He and others make sure that there are organizational structures--such as the computer users' meeting--to deal with such frustration.
Of course, it's possible to have too many meetings. If the only purpose of a meeting is to keep everyone informed, there may be a more efficient way to convey the information, says Parker. "We urge people to remember the cost of meetings. If you add up the time of everyone involved, they're quite expensive."
Working in a team does demand extra attention to administration, logistics, and personal concerns. But most experienced team researchers wouldn't even consider returning to a solitary mode of doing science. "We paid some overhead, in terms of time spent educating each other," says Kutzbach of the interdisciplinary COHMAP project. "But the payoffs at the end were magnificent."
Elizabeth Culotta is a science writer based in Durham, N.C.