The Coordination Problem

For both care provider teams and clinics, the issue of coordinating care for patients is problematic. Clinics' focus on function and aggregate clinical outcomes impedes coordination efforts. Similarly, communication and coordination among care provider team members are likely to be minimal given the different roles of the team members. Furthermore, the differences in expertise and goals between care provider teams and clinics contribute to the difficulty in simultaneously achieving high performance for both. Thus, the organizing challenge is to coordinate activities among all providers to offer seamless care specifically tailored to the needs of each patient.

There are three broad solutions to the coordination problem. The first is for patients and caregivers to perform the coordination role. This occurs for patients who seek to control their health care, can assemble the information about their health conditions, are aware of their needs, and can adequately communicate such information with all providers. Although coordinating care may be burdensome, it can ensure that care matches the patient's preferences and biographies, and it can also be rewarding for caregivers. In this situation, the role of providers is to support the patient and caregiver in care coordination.

The second solution is to use interorganizational networks to facilitate coordination, such as mental health networks (Provan and Milward, 1995) or palliative care networks (for example, Bainbridge et al., 2010). Coordinating through interorganizational networks can be accomplished by investing in tools that increase the ability of organizations to share information and by establishing boundary-spanning roles (Faraj and Yan, 2009) and objects such as care plans (Bechky, 2003; Nicolini, Mengis, and Swan, 2012). Interorganizational coordination can also be accomplished by standardizing provider roles and care pathways across organizations (Deneckere et al., 2012). Accomplishing coordination in interorganizational networks is likely to require significant investment in information technology as well as the willingness of providers and organizations to adapt their practices to network standards. Justifying this investment requires a high number of clients seen in common in order to achieve scale economies.

The third solution is to implement roles or teams that act as a patient's agent in coordinating care. As the need for both service provision and coordination increases, care management teams are favored, providing services and coordinating care tailored to patients' circumstances with teams of interprofessional members. Examples of care management teams include carve-out programs for specific conditions (Blumenthal and Buntin, 1998), assertive community treatment (ACT) teams for individuals with severe mental illness (Monroe-DeVita, Morse, and Bond, 2012; Stein and Santos, 1998), chronic care teams (Bodenheimer, Wagner, and Grumbach, 2002; Wagner, 2000), and PCMHs (Alexander and Bae, 2012; Stange et al., 2010) for individuals with multimorbidity (Valderas et al., 2009). A strength of coordinating with a care management team rather than an interorganizational network is that this team provides patients with a relational home and an expert adviser in navigating the care provider network. This is likely to be beneficial in situations where the effectiveness of delivery depends on a strong personal relationship such as in motivational interviewing (Hettema, Steele, and Miller, 2005; Miller and Rose, 2009).

Defining Care Management Teams

A theory of care management teams requires a definition that clearly distinguishes them from care provider teams and clinics. These teams are often imprecisely defined (Alexander and Bae, 2012; Vest et al., 2010). In fact, Wagner's (2000) definition of patient care teams can be interpreted as either a care provider team or a care management team. For example, the following segment of the definition suggests a care-management-team interpretation: “A patient care team is a group of diverse clinicians who communicate with each other regularly about the care of a defined group of patients and participate in that care” (Wagner, 2000, p. 569). However, this next segment of the definition suggests a care-provider-team interpretation:

Effective team care for chronic illness often involves professionals outside the group of individuals working in a single practice; it may involve multiple practices—for example, primary and specialist care—or it may involve multiple organisations, such as a general practice and a community agency. Teams that cross practice or organisational boundaries may create communication and administrative nightmares but are essential for optimizing care for many patients. (Wagner, 2000, p. 569)

We define care management teams as “real” teams designed to provide and coordinate patient-centered care. Wageman and colleagues (2005) identify three core features of real teams: clear membership boundaries, interdependence that fosters collective responsibility for assessable outcomes, and membership stability. Care management teams should be designed to have well-defined boundaries, be held collectively accountable for patient care and outcomes, be highly interdependent, and have stable membership (Monroe-DeVita et al., 2012; Monroe-DeVita, Teague, and Moser, 2011).

In contrast, care provider teams are not real teams because they are unlikely to have clear boundaries, high interdependence among all members of the network, collective responsibility, or membership stability. Care provider teams also differ from care management teams in terms of team identity coordination roles, and coordination effectiveness. Whereas care provider teams lack a team identity, exhibit variable effectiveness in care coordination, and assign coordination roles to patients, caregivers, and primary care providers, care management teams maintain a clear team identity, exhibit high effectiveness in care coordination, and assign coordination roles to the care management team.

In contrast to clinics, which have a functional focus, care management teams have a patient focus. Furthermore, these teams differ from clinics in their composition of providers, team identity, coordination roles, and coordination effectiveness. Clinics are composed of providers within the same unit, clinic, and organization; maintain professional and organizational identities rather than a team identity; exhibit low effectiveness in care coordination; and lack clear coordination roles. Care management teams are composed of providers from potentially different units, clinics, and organizations; maintain a clear team identity; are highly effective in coordinating care; and rely on the care management team as a whole to coordinate care (table 6.1).

The next section discusses care management teams as a mechanism affected by the context they operate in using a context-mechanism-outcome configuration (CMOc) approach (Pawson, 2003; Pawson and Manzano-Santaella, 2012; Pawson and Tilley, 1997). This section is followed by a discussion of key issues in designing care management teams and then by a conceptual model of care management team implementation examining determinants of implementation outcomes.

The Role of Fidelity

Fidelityfidelity is the assurance that the theory implicit in an intervention has been implemented as intended (Mowbray et al., 2003). The promise that a specific team design will result in desired outcomes is warranted only when the intervention is explicitly laid out in theory and is faithfully implemented in practice. In other words, designing care management teams involves developing explicit fidelity measures to assess how well those teams work as intended. Fidelity measures can be grouped into structural-process characteristics and clinical aspects (Bond et al., 2005). Using ACT teams as an example, structural-process fidelity measures include characteristics such as the mix of health professionals a team should have, workload defined by the number of patients per team member, information management such as regular care planning and having care plans on file, coordination mechanisms such as daily team meetings, and structures that support individualized care and quality improvement (Bond et al., 2009; Monroe-DeVita et al., 2011). Clinical fidelity measures include specifics about services and interventions that the teams should provide, such as motivational interviewing (Burke, Arkowitz, and Menchola, 2003; Madson and Campbell, 2006; Madson, Loignon, and Lane, 2009) and supported employment (Bond, Becker, and Drake, 2011).

Developing fidelity measures is beneficial for a number of reasons:

• It forces a clear statement of each of the care management team's design features, which increases the rigor of the design.
• It supports standardization and provides concrete guidance for implementation.
• It describes what should occur if the team is implemented correctly.
• It supports the assessment of how well each design feature is implemented.
• It supports the assessment of the relationship between the design features and team performance.
• It identifies gaps in implementation and suggests solutions to fix the gaps.
• It supports the long-term diffusion of the care management team.

The fidelity measures for a care management team need to be developed to fit its specific theoretical underpinnings (Schoenwald et al., 2011). The specification of fidelity measures should be based on a strong theoretical foundation so that it allows the examination of causal mechanisms as recommended by realistic evaluation, an approach that asks “what works for whom in what circumstances … and why” (Pawson and Manzano-Santaella, 2012, p. 178). Such a foundation reduces the chance of applying ad hoc definitions of an intervention in different studies and implementations and it supports cumulative research. A review of PCMHs, for example, found that “multiple organizations and individuals have notable variations in their definitions of the medical home” (Vest et al., 2010). This variation suggests that research on PCMHs may be difficult to accumulate because researchers are not studying the same intervention.

The fidelity specification for care management teams should start with the definition of a team as

(a) two or more individuals who (b) socially interact (face-to-face or, increasingly, virtually); (c) possess one or more common goals; (d) are brought together to perform organizationally relevant tasks; (e) exhibit interdependencies with respect to workflow, goals, and outcomes; (f) have different roles and responsibilities; and (g) are together embedded in an encompassing organizational system, with boundaries and linkages to the broader system context and task environment. (Kozlowski and Ilgen, 2006, p. 79)

The discussion below focuses on specifying fidelity measures for care management teams along with key team design features (Wageman et al., 2005).

A Real Team

As noted, a care management team should be a real team with clear boundaries, high interdependence, and membership stability (Wageman et al., 2005). The clear boundaries are necessary to identify who is jointly accountable for the team's outcome. Having clear boundaries does not mean that an individual cannot be a member of multiple teams. A primary physician may be a member of a dozen PCMHs. In ACT teams, a psychiatrist is usually a member of two teams. The important point is that a provider should see herself as a distinct member of a specific team and understand her role on that team despite multiple team membership. Furthermore, every team member should be able to identify who is and is not on the team. The risk of not making membership clear is that as multiple team membership increases, team members are less likely to share a common identification (O'Leary, Mortensen, and Woolley, 2011), which attenuates the effect of accountability.

Interdependence among team members means that they need to work closely together in performing tasks (Thompson, 1967; Van de Ven, Delbecq, and Koenig, 1976), which differentiates teams from individuals who are colocated in the same space performing independent tasks. Interdependence creates a sense of shared purpose, a motivational factor for information sharing (De Dreu, 2007). Interdependence facilitates coordination through both process mechanisms, such as team meetings, information sharing, and decision making, and cognitive mechanisms, such as the team mental model and transactive memory. Stability of membership implies that care management teams should be more than temporary ad hoc teams assembled for the duration of a specific activity, such as surgical teams or flight crews. Stability fosters the understanding of patients and their contexts over time that is necessary for situational awareness and patient-centered care. Stability also contributes to cognitive coordination mechanisms because such mechanisms usually take a long time to develop.

Team Size and Workload

Ideally a care management team should have relatively few members, from four to twelve people. This limited range allows key skills and services to be included while avoiding scope diseconomies caused by larger teams fragmenting into subgroups. The limited team size facilitates high levels of overall team interdependence by limiting the number of possible work relationships in the team. It also makes it feasible to define clear team boundaries and assign accountability. In practice, larger teams, such as those incorporating entire clinics, often fragment into subgroups, which may impede team functioning by reducing interdependence and blurring team boundaries and responsibilities.

The workload of a care management team—the number of patients per team member—will be a function of the care demands posed by its target patients. For example, the workload for ACT teams is designed to be around ten clients per team member (Teague, Bond, and Drake, 1998). This workload reflects the high complexity and uncertainty of the clients (e.g., individuals with severe mental illness) and the nature of the services (e.g., services delivered in community or clients who are visited frequently). The workload for PCMH teams is likely to be larger because the patient population will be a mix of patients in both fairly stable condition and in crises. In Aetna's embedded case managers (ECM) program (Hostetter, 2010), a patient-centered primary care program, an ECM can support a population of around two thousand Medicare patients because relatively few patients, around seventy to eighty, are in a crisis mode that requires intensive attention at any given time. The number of patients a palliative care team can support will similarly be determined by the varying demands of the patients and caregivers in the population served.

Task Scope and Team Composition

Three steps determine the task scope for a care management team. First, all the tasks that need to be performed to create value for the target patients should be specified. Second, the task mix can be used to identify the occupations necessary to perform those tasks. Third, the team has to determine which services it should directly provide and which should be provided externally. Because tasks are the functions that a team must perform to achieve its goals, they are a logical starting point for team design. Once the tasks are determined, the skills required by the individuals to perform tasks and the occupations of those individuals can be determined. The last step is necessary to limit team size and focus the team on the most important tasks.

Initially focusing on tasks rather than occupations strengthens team design because it focuses on the functions that must be accomplished rather than the person performing the functions. Focusing on occupations—the types of skills that providers have—can result in designing the team in terms of provider skills rather than the types of services that patients need. In the case of PCMHs, framing the team as being led by a primary care physician may lead to an ineffective design. In contrast, focusing on tasks and patient needs can result in identifying the occupation best able to perform the tasks and meet patient needs. The design of a coordinated care clinic (CCC) by Hennepin County Medical Center (HCMC) reflects a task-driven team design (Johnson et al., 2012). HCMC is a safety-net hospital in Minnesota that serves a population of high-risk patients who are medically complex and have significant mental health and social issues, including unstable housing and lack of social support. The need to address mental health and social issues is reflected in the CCC team design, which includes a full-time nurse practitioner, a Registered Nurse care coordinator, a social worker, a half-time physician, a pharmacist, a chemical dependency counselor, and a one-tenth-time clinical psychologist. The team's occupational mix is designed to fit the complex set of tasks required for patients.

The decision about which services the team should supply versus those supplied by external providers will ultimately determine a care management team's composition. Choosing among occupations is difficult because patients with complex conditions often need a wide variety of services. Three factors may influence this decision: team size, task importance, and opportunity cost. An appropriate team size can help a care management team maintain a high level of interdependence among key occupations needed for patient care. It should not be excessively large because the size of the team is negatively related to the level of interdependence. It should not be too small either so as to exclude key occupations. Fidelity measures for ACT teams, for example, suggest that team size should be in the range of eight to twelve team members and include key occupations such as psychiatrist, nurse, substance abuse specialist, and vocational specialist. The importance of a task can be characterized in terms of the prevalence of it. In principle, occupations performing more important tasks should have priority for inclusion on the team. This principle, however, needs to be conjoined with the consideration of opportunity costs of not including other occupations. The accessibility of external services and the losses in coordination efficiency both contribute to the opportunity cost.

Division of Labor

Division of labor determines which team member, group of members, or entire team perform which task. The result is a matrix of the assignment of individuals to tasks. Tasks can be classified by the levels of skill required to perform them: generic, enhanced, and specialized (National End of Life Care Programme, 2012). Generic tasks are those that can be performed by any member of the care management team. In ACT teams, for example, a generic task is “eyes-on-meds,” where a team member observes and ensures that a client has taken the prescribed medications. Enhanced tasks are those that any team member can perform with appropriate cross-training from a team member with greater expertise. Specialized tasks are those that can be performed only by an expert. Care management teams usually have a mix of generic, enhanced, and specialized tasks. Although a task may be generic or enhanced, specific individuals still need to be assigned to the task. The assignment of individuals from different occupations to perform generic and enhanced tasks may strengthen a team by encouraging team members to communicate, learn from each other, and develop a team identity based on common tasks. Tasks can also be classified by their functionalities: clinical, team coordination, and team improvement. Clinical tasks are those with activities directly related to interacting with and caring for patients. Team coordination tasks are those performed to exchange information and synchronize actions among team members, such as team meetings, huddles, and setting schedules. Team improvement tasks include activities that are intended to improve task performance by measuring, monitoring, and reflecting on performance metrics and team structures and processes (Schippers, Homan, and van Knippenberg, 2013).

Table 6.2 presents a tool that can be used to determine the division of labor in a care management team. Using the nine cells in the table to guide the division of labor will ensure that:

• Specialized tasks, especially specialized clinical tasks, are assigned to team members with the appropriate expertise
• Generic tasks are rotated among team members if possible, which builds team identity by sharing tasks
• Enhanced tasks are assigned to facilitate cross-training
• Team coordination tasks are performed by team members with appropriate skills
• Team improvement tasks ideally involve the entire team

We recommend using vignettes or real situations, or both, to identify specific tasks, which will reduce the likelihood of framing the team design solely from a clinical perspective and will increase the likelihood of identifying team coordination and improvement tasks that are necessary to support clinical tasks. Describing tasks using specific terms such as “assess the housing needs for a client” rather than generic terms such as “conduct a needs assessment” is preferable because it allows the tasks to be illustrated with specific examples and ensures that what is meant to be done is clearly stated. Once the tasks have been identified and classified, the process of selecting care management team members can be addressed.

Choosing specialists is difficult because complex patients often require a wide variety of specialized skills. Criteria for choosing specialists include the importance of the skill for patient care, the importance of maintaining interdependence of the specialist with other team members, and opportunity cost. For opportunity cost, the number of patients associated with a team may not be able to support a particular specialist full time. For occupations where the opportunity cost is high, such as psychiatrists on ACT teams, primary care physicians in PCMHs, or palliative care physicians in later-life supportive care management teams, one solution is to employ a provider part time on the care management team. With this arrangement, it is important to make the part-time provider very clear about her membership and role on the team. Because certain providers may not be present full time on the team, it may be important to create a separate role for team lead. On ACT teams, for example, the team leader is a formal role often occupied by an individual with a social work background while the psychiatrist is a care management team member.

Coordination

In an extensive review of the coordination literature, Faraj and Xiao (2006, cited in Okhuysen and Bechky, 2009, p. 469) define coordination as the “temporally unfolding and contextualized process of input regulation and interaction articulation to realize a collective performance” (Faraj and Xiao, 2006, p. 1157). Okhuysen and Bechky (2009) classify coordination mechanisms in five categories:

1. Plans and rules
2. Objects and representations
3. Roles
4. Routines
5. Proximity

Clearly describing the mechanisms that will be used by a care management team to coordinate care is an essential part of team design. Designing coordination mechanisms should achieve the following objectives: maximizing information sharing, minimizing process costs, and maximizing mechanism modality across different types of information. For example, electronic health records and registries are optimal for sharing codified information, whereas high levels of interdependence are needed for sharing tacit information. Specifying “boundary objects” (Nicolini et al., 2012) such as care plans or care pathways (Deneckere et al., 2012; Gittell, 2002) to support interprofessional coordination within the care management team will create a common, usable information platform that reduces ambiguity (Bechky, 2003; Gurses et al., 2008). This is true as well for specifying artifacts used in coordinating with external providers.

In implementing coordination mechanisms, one objective is implementing them in a standardized fashion (Spear and Bowen, 1999) in order to minimize errors and ambiguity (Gurses et al., 2008; Spear and Schmidhofer, 2005). In ACT, for example, team members are encouraged to record the outcome of every client encounter using the goal, intervention, result, and plan scheme, focusing on how the intervention reached its goal and developing a plan for the next encounter based on this evaluation. Another objective is implementing coordination mechanisms in a situationally aware or mindful manner (Endsley, 2000; Levinthal and Rerup, 2006). This means that coordination should flexibly adapt to a patient's emerging condition and context. For example, when ACT team members meet a client for “eyes-on-meds” and detect the presence of evidence of drug or alcohol use, this could alert the team to the reoccurrence of substance abuse. For an elderly patient, the cue could be a more disorderly living situation or unexpected symptoms. Situational awareness or mindfulness implies that team members are alert to the cues that a routine is not functioning as intended so that they can adjust the routine or reassess the care plan.

At its root, coordination is an information processing activity. Effective coordination ensures that the right individual knows the right information at the right time to perform a task effectively. For care management teams, the essence of coordination is managing like encoding, storing, and retrieving information about the patient's conditions, goals, and needs and about all the services that have been provided or are planned for the patient (Argote and Miron-Spektor, 2011; Haas, 2006).

Although the literature on coordination mechanisms is important for team design, it does not provide any tool that supports assessing the need for coordination. Table 6.3 presents a tool for assessing coordination needs by characterizing the relationship between patient information and the information acquired by the provider in terms of information availability. The need for coordination occurs only when the provider is unaware of information about the patient that he or she should reasonably know, as in the “unknown/available” cell in the table. Examples include providers not knowing a lab test result or a significant change in the patient's preference for treatment. In situations where the provider knows the patient's information and what needs to be done, as in the “known/available” cell, the patient outcome relies on the provider's performing the task at hand correctly. In situations where the provider knows that certain information about the patient is unavailable, as in the “known/unavailable” cell, an assessment is called for. An example is when a patient's care needs to shift from activities of daily living to spiritual issues as death approaches. An effective palliative care team will anticipate this shift and be watching for it. In situations where the provider does not know the patient's situation and the information is unavailable, as in the “unknown/unavailable” cell, reconsidering and reframing the patient's situation is necessary. The distinction among the cells is important because each represents a different type of cause of performance shortfalls and thus calls for a different solution. For instance, if the provider knows the patient's needs and preferences but could not do an adequate job in treating the patient, it is a task performance issue, not a coordination issue.

Embedding: Care Management Team Networks

One of the defining characteristics of teams is that they “are together embedded in an encompassing organizational system, with boundaries and linkages to the broader system context and task environment” (Kozlowski and Ilgen, 2006). Care management teams are embedded in two types of networks that have to be managed to solve the care coordination problem: networks that involve relations with patients and caregivers and with other health care providers and community organizations.

A Conceptual Model

Figure 6.2 shows the conceptual model for care management team implementation. It builds on current research on care coordination and team effectiveness. Research on system improvement indicates that three key principles for designing high-performance teams are clearly specified tasks and roles, unambiguous communication, and simple and direct work flows (Ghosh and Sobek, 2006; Gurses et al., 2008; Spear and Bowen, 1999). Research on team effectiveness reveals that the relationship between team design and team outcomes is moderated (Kenny, 2011b) and mediated (Kenny, 2011a) by team processes and emergent states (Cohen and Bailey, 1997; Ilgen et al., 2005; Kozlowski and Ilgen, 2006). Team processes are interactions among team members, while emergent states describe cognitive, motivational, and affective states of teams (Marks, Mathieu, and Zaccaro, 2001). We see care coordination as a dynamic, emergent, and adaptive process operating in specific clinical contexts (Allred, Burns, and Phillips, 2005; Lemieux-Charles and McGuire, 2006) with the relationship between team design and team effectiveness (i.e., achieving desired patient and staff outcomes) being moderated and mediated by team processes. Consistent with research showing that well-designed teams perform better than poorly designed teams (Cohen and Bailey, 1997) and that leadership and quality improvement activities work better in well-designed teams (Wageman, 2001), research on ACT teams shows that fidelity to team design can improve client outcomes, but important variation exists across teams and contexts (Burns et al., 2007). Such variation can be attributed to team processes occurring during team implementation. Leadership, for example, facilitates team learning to identify problems of wasted effort; develop, implement, and test interventions to fix the problems; and adopt effective interventions. Leadership similarly facilitates the implementation of evidence-based practices (Substance Abuse and Mental Health Services Administration, 2013).

For care management teams, we list the key moderating and mediating team processes in figure 6.2 and discuss their roles in translating team design into team outcomes (Kenny, 2011a 2011b). First, research suggests that a constructive context moderates the relationship between team design and team performance by promoting growth, development, and the performance capabilities of team members (Wholey et al., 2013). A constructive context is a safe environment where information and ideas can be freely exchanged and task processes refined and includes team processes such as team learning—training and quality improvement activities (Levitt and March, 1988; Tucker, Nembhard, and Edmondson, 2007); constructive controversy—team processes with which team members express their opinions directly and explore opposing positions open-mindedly when resolving conflicts (Shah, Dirks, and Chervany, 2006); and (lack of) conflict—perceived incompatibilities, opposing interests, or discrepant views among team members (Jehn, 1995). The constructive context also includes emergent states such as psychological safety (a shared perception that the team is a safe place for admitting errors, asking for assistance, or discussing difficult issues without the fear of negative consequences to individuals; Edmondson, 1999), and goal agreement (a shared perception of a team's priorities and objectives; Zohar, 2000). A constructive context is likely to be particularly important for teams focusing on information management, such as care management teams. It is worth noting that while the moderating processes and states appearing in figure 6.2 are distinctive constructs, they are interrelated. Psychological safety, for example, is a moderating state for care management teams because it improves the implementation of evidenced-based practices (Tucker, Nembhard, and Edmondson, 2007), and it is likely associated with constructive controversy.

Second, interpersonal processes such as team members helping each other to perform tasks and states such as transactive memory and social capital—knowing who has what expertise and knowing who would help if requested—moderate the relationship between team design and team performance by cultivating a highly interdependent work system and facilitating information sharing. A key function of care management teams is information management. Not only do care management teams need to assess and manage information about patients' circumstances, they also need to monitor and coordinate services provided by providers outside the team. Because less expensive coordination mechanisms are available for patients with less complex conditions, care management teams usually focus on patients with greater complexity and uncertainty. For complex, dynamic, evolving, and emergent patient circumstances, information management requires a significant amount of patient-specific knowledge that has to be interpreted and integrated by multiple professionals. Teams can invoke two types of information management and coordination mechanisms. First, rules, protocols, standardized roles, and information media (traditional media like paper forms, or electronic ones such as electronic health records) can be used to coordinate actions for routine issues and manage codified information (Okhuysen and Bechky, 2009). Second, more organic mechanisms and a highly interdependent work system are required to coordinate actions and manage information for complex and uncertain circumstances. In this situation, a large amount of information, often tacit or not codifiable, has to be gathered during task performance to complete the task effectively (Galbraith, 1974).

Interpersonal processes and states can be characterized and studied from a social network perspective. Social networks within care management teams are a key determinant of information management and performance (Balkundi and Harrison, 2006; Flap, Bulder, and Völker, 1998). High levels of interdependence, for example, facilitate the exchange of tacit information and promote cohesion and helping behavior among team members (Lawler, Thye, and Yoon, 2009). Explaining the role of social networks in care management teams promises significant improvement in our understanding of the variation in outcomes across superficially similar teams. Even in teams of ten to twelve individuals, organizational processes are likely to create subteams (Simon, 1962; Thompson, 1967), hierarchy, and peripheral actors. These varying relational structures affect teams' cohesion, effectiveness in information management, and performance. Incorporating social networks in team research and implementation is important because social networks are manageable. Team leaders can change networks relatively simply by managerial actions, such as by assigning different individuals to work together or through arranging new interaction opportunities.

Third, the mediating team processes and states in care management team implementation include information accessibility, encounter preparedness, and patient-centered care. In care management teams, team design, constructive context, and interpersonal processes all contribute to ensure that the right information is available to the right individual at the right time. Information accessibility characterizes the process by which team members acquire task-relevant information efficiently, with minimal waste (Spear and Bowen, 1999; Shah and Ward, 2007). The consequence of having information available is that team members can be more attentive to or improve situational awareness of patients' circumstances (Endsley, 2000). Situational awareness has two components: encounter preparedness and patient-centered care. Encounter preparedness entails being sufficiently prepared for patient encounters by having knowledge of a patient's current conditions, needs, and goals and knowing the schedule and specific tasks. Patient-centered care entails observing elements of an encounter that requires adapting care to a patient's diagnosis, needs, and goals. Encounter preparedness is a necessary condition for patient-centered care because it enables care management team members to recognize when and how an encounter differs from expectations and to adapt responses to the emergent situations. We anticipate that encounter preparedness will reduce team member stress because it clarifies the situation and expectations and will improve patient-centered care by making departures from plans that result in more appropriate adjustments. Patient-centered care will ultimately improve patient outcomes. Encounter preparedness and patient-centered care are reasonable end points for the analysis of care management team implementation because the linkage between patient-centered care and patients' medical or social outcomes is likely to be attenuated by a wide variety of exogenous factors.

Relational Aspects of Team Implementation

This section provides some hypotheses for relational aspects of care management teams based on the emergent states that affect their functioning. These states reflect four types of relationships:

• The effect of interdependence on team climate and encounter preparedness, having the necessary information for providing effective patient care at patient encounters
• Social capital, that is, being able to turn to other team members for help
• Transactive memory, that is, knowing the knowledge, skills, and abilities of other team members
• The effect of interdependence and sharing common tasks on the development of team identity, based on the argument that individuals from differing occupations performing generic tasks facilitate the adoption of a team identity while specialists performing specialized task facilitate professional identification