Implementing Coordinated Care Networks: The Interplay of Individual and Distributed Leadership Practices

Abstract

How does leadership emerge and function when multiple health care organizations come together to form a network? In this qualitative comparative case study, we draw on distributed leadership theory to examine the leadership practices that manifested during the implementation of three coordinated care networks. Thirty leaders and care providers participated in semistructured interviews. Interview data were inductively analyzed using thematic analysis. Although established in response to the same policy initiative, each case differed in its leadership approach and implementation strategy. We found that manifestation of distributed leadership was contingent on the presence of an individual leader who acted as a unifying force across their respective network. Our findings suggest that policies to encourage the development of interorganizational networks should include sufficient resources to support an individual leader who enables distributed leadership.

Introduction

Health care systems designed to treat acute care needs are under increasing pressure to replace fragmented and costly hospital-based care with care that is more coordinated and community based (Tinetti et al., 2012). Coordinated, community-based care has the potential to improve quality of care, enhance the patient experience, and reduce costs (Rocks et al., 2020). To facilitate this transformation, jurisdictions around the world are mandating or encouraging the formation of networks that bring together health care organizations to jointly coordinate patient care (Baxter et al., 2002; Suter et al., 2009).

However, many models of coordinated care delivery exist, numerous influencing factors have been identified, and evidence of effectiveness is mixed and of moderate quality (Baxter et al., 2018; Rocks et al., 2020). Thus, how to best design and implement a coordinated care network remains unclear. Furthermore, the establishment of a network does not necessarily result in collaboration among network members and in the delivery of coordinated care (Evans et al., 2013; Singer et al., 2011). Networks are “complex, messy, dynamic, and difficult to manage” (Goodwin, 2008), requiring committed leadership that supports collaboration across professional and organizational boundaries and moves the network toward its goals (Cunningham et al., 2019; Huxham & Vangen, 2000). Leadership is frequently cited as a key enabler in the literature on coordinated care but is rarely the focus of empirical studies (Aunger et al., 2021; Edgren & Barnard, 2012; Evans, Daub, et al., 2016; Mitterlechner, 2020; Suter et al., 2009). We thus lack a nuanced understanding of how leadership functions when multiple organizations—each with their own leadership structures—come together as a network.

Leadership in networks differs fundamentally from leadership in organizations. In networks, formal and informal leaders interact across organizational boundaries. Accountability mechanisms are often unclear and the role clarity that typically underlies organizational and team-based leadership is absent or contested, particularly during the early phases of network implementation (Denis et al., 2012). As such, the source of leadership influence in networks may sometimes be individual and other times distributed or shared (Gronn, 2009; Günzel-Jensen et al., 2018).

Research on coordinated care networks suggests that traditional individual or “heroic” leadership approaches may hinder network implementation and performance (Grudniewicz et al., 2018; Tsasis et al., 2012). Distributed leadership in these networks is thought to be effective because influence is “dispersed across multiple actors, with no single agent having full authority, resources or expertise to lead the change” (Mitterlechner, 2020, p. 8). However, distributed leadership could also inadvertently result in a leadership vacuum where no one has the authority and resources to manage the network (Buchanan et al., 2007). In general, we know little about how distributed leadership manifests during the implementation of coordinated care networks and with what impact on network functioning. The aims of this study are to examine (a) how distributed leadership manifested in the implementation of three coordinated care networks and (b) the role and influence of individual leaders within these networks.

New Contributions

This study contributes to two bodies of literature: (a) distributed leadership and (b) coordinated/integrated care delivery. Regarding distributed leadership, we critically examine the role of formal and informal individual leaders in newly formed interorganizational distributed leadership configurations. In so doing, we heed calls from scholars regarding the need for research on distributed leadership to better reflect the increasing complexity of organizational forms and leadership practices over time (Bolden, 2011; Currie & Spyridonidis, 2019; Harris & Gronn, 2008).

This study also contributes to the literature on coordinated and integrated care delivery. In a recent review on leading integrated care initiatives, only 5% of included papers applied leadership theories (4/73) and most studies involved a broad sweep of organizational factors rather than an in-depth examination of leadership structures and practices (Mitterlechner, 2020). Our study involved a deep empirical analysis of leadership informed by distributed leadership theory. Our findings add nuance to recommendations in the literature promoting distributed or shared leadership by demonstrating that a hybrid configuration of individual and distributed leadership is more likely to succeed during the network implementation stages.

From a policy perspective, networks are increasingly a cornerstone of health system reform efforts aimed at reducing costs and better coordinating health and social care services (Flieger et al., 2021; Rocks et al., 2020; Vickery et al., 2020). These time- and resource-intensive reforms (Lebina et al., 2020; Maruthappu et al., 2015) demand effective leadership (Mitterlechner, 2020). The results of our study can be used to inform network design and implementation and possibly to assess network implementation based on the extent of distributed leadership.

Theoretical Framework: Distributed Leadership

Traditional theories of leadership focus on the traits, behaviors, and attitudes of individual formal leaders (Bass, 1990; Conger & Kanungo, 1998). Distributed leadership theory focuses on how leadership is spread among multiple individuals over time and across one or more hierarchical levels (Denis et al., 2012; Günzel-Jensen et al., 2018). For example, in a quality improvement study, change leadership manifested not only through senior leaders and middle managers but also through informal opinion leaders and champions at other hierarchical levels over time (Fitzgerald et al., 2013). Their respective leadership actions were often concurrent, rather than sequential, and had a cumulative impact on change efforts. Distributed leadership goes beyond the number of individuals involved in a leadership activity. The unit of analysis of distributed leadership, therefore, is not the individuals aggregated in a leadership activity, but rather the shared leadership actions or practices, which “stretch” over multiple individuals and sometimes across hierarchical and/or organizational boundaries over time (Gronn, 2002, 2008; Spillane, 2012).

Distributed leadership configurations can be distinguished based on the extent to which distributed leadership is enacted from the top-down by formal leaders versus the bottom-up by widely dispersed individuals (Currie & Lockett, 2011; Gronn, 2002). When distributed leadership manifests as a top-down phenomenon, the focus is on alignment of leadership action across actors in a common direction (referred to as “conjoint agency”; Currie & Lockett, 2011; Gronn, 2002). Movement in a common direction is achieved when actors synchronize their individual leadership actions with those of their peers. When distributed leadership manifests more as a bottom-up phenomenon, the focus is on collaboration among diverse actors to complete tasks, which contributes to pooling of expertise, role interdependence, mutual understanding, and efforts to institutionalize collaboration (referred to as “concertive action”; Currie & Lockett, 2011; Gronn, 2002). Distributed leadership requires both conjoint agency and concertive action (Currie & Lockett, 2011; Gronn, 2002). Without conjoint agency, there is divergence in the direction of leadership actions, fragmentation of leadership influence, and potential for inertia. Without concertive action, there is broad disunity, discord, and disengagement.

Figure 1 presents a simplified version of Currie and Lockett’s (2011) 2 × 2 matrix of distributed leadership configurations based on the degree of conjoint agency and concertive action. When leadership is neither conjoint nor concertive, leadership is individualistic (bottom right). When leadership is both conjoint and concertive, leadership is purely distributed (top left quadrant). When leadership is concertive, but not conjoint, leadership is broadly distributed and highly collaborative with potential for a “nobody in charge” approach (top right quadrant). When leadership is conjoint, but not concertive, leadership is concentrated among a select group, most often consisting of formal leaders (bottom left quadrant). This matrix reinforces the argument that the presence of multiple leaders does not necessarily imply a distributed leadership approach. Conversely, the concentration of leadership in a single individual does not necessarily imply an individualistic leadership approach. In addition to the number of individuals involved in leadership, it is necessary to consider their position type and hierarchical placement in the organization or network, as well as their leadership actions. Therefore, individualistic and distributed leadership are not mutually exclusive; they exist on a continuum.

Figure 1.

Matrix of Distributed Leadership Configurations.

Source. Adapted from Currie and Lockett (2011).

Distributed leadership has particular relevance to interorganizational networks in health care. The notion of one formal organizational leader at the top of the hierarchy does not apply to networks (Huxham & Vangen, 2000). In networks, multiple leaders exist and are physically distributed across member organizations. As such, leadership is both vertically distributed within each organization and horizontally distributed across organizations in the network (Chreim et al., 2010). In the health care context, the limited control of administrative leaders over physicians reinforces the vertical distribution of leadership and renders clinician engagement necessary for change (Nembhard et al., 2009; Nieuwboer et al., 2019; Roberson, 2019). The practice of leadership in health care networks is thus intrinsically distributed with no single leader or organization in a position to unilaterally direct the activities of network members through hierarchical decree.

However, distributed leadership does not replace or suppress the role and contribution of individual leaders. The source of leadership influence may sometimes be an individual and other times distributed among multiple individuals, or a dynamic interplay between the two (Gronn, 2009; Günzel-Jensen et al., 2018). Distributed leadership may influence individual leadership action by limiting the power and control of any one leader, thus compelling collaboration. Individual leaders can also influence distributed leadership. For example, an empowering leadership style by a formal leader can support employees’ perceived agency in enacting distributed leadership practices by creating an “atmosphere of trust” (Günzel-Jensen et al., 2018). Similarly, formal leaders can create a culture predicated on collaboration and power sharing that allows distributed leadership to unfold (Leithwood et al., 2007). A hybrid individual-distributed model of leadership may thus be a more effective approach and a more accurate representation of practice (Gronn, 2009).

We theorize that a hybrid individual-distributed model of leadership may be particularly important during the early stages of coordinated care network implementation when members are getting to know one another, and the network’s strategic goals and plans are nascent. A central individual leader may be needed to bridge the gap between those who are involved in leadership across the network (typically organizational leaders or individual representatives of the partner organizations, i.e., horizontally distributed leadership) and those involved in leadership within a given partner organization (typically frontline care providers, coordinators, and other clinical and nonclinical staff, i.e., vertically distributed leadership). Without intentional intervention to support communication and goal alignment during the early stages of implementation, these dispersed leaders may be working toward different visions—a phenomenon that is common in coordinated care networks (Suter et al., 2009). The role of the individual leader, then, may be to not only facilitate distributed leadership through empowerment as previous studies have found (Günzel-Jensen et al., 2018; Leithwood et al., 2007) but also to manage the inherent tension between concertive action and conjoint agency, for example, by ensuring that as leadership becomes more distributed (concertive action), leadership action remains aligned to a common direction (conjoint agency).

No research has empirically studied the role of individual leaders within distributed leadership configurations in newly formed coordinated care networks. Leadership discretion is higher in a flexible policy context, which may generate more variation in leadership practices across networks, offering a rich opportunity to study the role and influence of individual and distributed leadership. In this qualitative case study, we explore leadership practices that manifested during the implementation of three coordinated care networks that were formed in response to a flexible government policy.

Method

In a previous multimethod comparative case study evaluation, three coordinated care networks known as Health Links in one region in Ontario, Canada, were studied using semistructured interviews, surveys, document review, and administrative data (Gutberg et al., 2017; Mondor et al., 2016). The three cases were purposefully selected because they were “early adopters” of a government policy aimed at stimulating the development of coordinated care networks. This previous evaluation was guided by the Context and Capabilities for Integrating Care (CCIC) framework, a conceptual framework that presents 18 organizational and network factors that support integrated care organized into three categories: basic structures, people and values, and key processes. The CCIC framework was developed and validated through a literature review and qualitative study (Evans, Grudniewicz, et al., 2016; Evans et al., 2017) prior to the Health Links evaluation and has been applied in international evaluations of coordinated care networks (Asthana et al., 2020; Wodchis et al., 2018). In this article, we present the results of a secondary analysis of leadership using the interview data from the Health Links evaluation. Below we describe the study setting, interview methods, and our secondary analysis of the interview data pertaining to leadership.

Study Setting

The Health Links initiative was established in the province of Ontario, Canada, which has a population of 14.8 million (Statistics Canada, 2021). Ontario consists of 14 geographic regions, each governed by a regional body known as a “Local Health Integration Network” (LHIN). At the time of analysis, LHINs funded all health services within the regions. The Health Links initiative aimed to transform existing relationships between organizations into more fulsome networks. Each Health Link was composed of organizations within a geographic area (e.g., primary care practices, hospitals, community care agencies) that voluntarily partnered to form a network. Partner organizations worked together to coordinate care with the help of an appointed care coordinator. Voluntary partnerships did not involve bundled payments or any similar financial accountability mechanisms between partners; rather, partnership in each network entailed the coordination and delivery of existing services. Start-up funding of approximately US$175,000 was provided to each approved Health Link by the Ministry of Health and Long-Term Care (MOHLTC). Further funding was at the discretion of the LHINs depending on their own regional priorities. However, funding was limited and was not used to provide additional health services. Enrolled populations for all Health Link networks range in size from 13,000 to 76,000 (Mondor et al., 2016). Coordinated Care Plans were used to establish a patient’s care team, document patient goals, and facilitate communication among a diverse team of providers.

A “low-rules approach” was used by the MOHLTC for the implementation of Health Link networks. When first announced, the only formal rules were as follows: (a) Each Health Link must have a lead or colead organization; (b) performance measurement must be a key component, focused on the number of completed Coordinated Care Plans; and (c) primary care involvement must be a cornerstone (Angus & Greenberg, 2014; Ontario Ministry of Health and Long-Term Care, 2015). Beyond appointing a designated lead organization, leadership structures in the Health Links were adaptable. Each Health Link decided how much control would be held by the lead organization, which organizations would be involved in strategy setting, and to what extent frontline care providers would be able to “lead from below.” Each Health Link designated a project manager in the lead organization. They were responsible for working with partners, convening meetings, and establishing operating budgets. Lead organizations convened regular steering committee meetings for partner organizations to jointly set the strategic direction of their Health Link. Although there was variation in lead organization type, the three cases presented herein were all hospital-led Health Links to render differences in leadership practices more comparable. The role of the LHINs varied across the province with most adopting the Ministry’s “low-rules approach,” whereas a few others took a prescriptive approach, opting to standardize the Health Links within their region (e.g., deciding who would be the lead organization or targeted patient populations; Grudniewicz et al., 2018).

Interview Participant Sampling

Health Link project managers were asked to identify individuals who could describe Health Link implementation over time. Interview participants were classified as either organizational “leaders” (individuals involved in managing a Health Link) or “providers” (front-line clinicians, e.g., family physicians and nonclinical providers, e.g., care coordinators, delivering or coordinating care as part of a Health Link). Snowball sampling was used by asking interviewees to suggest others who met the inclusion criteria. Representation was sought from at least one informant per organization in each Health Link. Health Link business plans were used to identify organizational partners in each network.

Interview Data Collection

Interview data were collected between February and June 2016. The Health Link networks included in the study were within the first 2 years of implementation at the time of data collection (i.e., “early adopters”). One-on-one interviews were conducted primarily over the telephone, although a small subset of interviews was conducted in person. Ethics approval for the study was obtained.

Three research team members conducted interviews using a semistructured interview guide (Online Supplement, Appendix A). Questions inquired about the experience of integrating care, including the process of partnering with other organizations (i.e., building the network) and implementing Health Links processes such as the care plans. Near the end of the interview, participants were presented with a handout on the CCIC framework and asked to rank the factors in the framework based on their relative importance in shaping the implementation and functioning of their Health Link (Online Supplement, Appendix B). Participants were then asked to discuss the role and influence of their top six factors and probing questions were asked regarding these factors. Interviews continued until thematic saturation and sufficient stakeholder representation (between leaders and providers and among partner organizations) were achieved. All interviews were audio recorded and transcribed verbatim.

Interview Data Analysis

Interviews were initially analyzed for the Health Links evaluation using a descriptive, deductive coding process to identify participants’ perceived importance of each factor in the CCIC framework, as well as inductive coding of any additional implementation factors discussed. The descriptive coding process was done by three research assistants and overseen by the project lead. The team began by each coding the same transcript independently and meeting to discuss discrepancies and reach consensus on the coding framework and definitions. Team members then proceeded with individually coding the remaining interview transcripts, meeting regularly to address concerns. Based on the deductive analysis, key factors were identified as critical to implementation of the three Health Links. “Leadership Approach” and “Clinician Engagement and Leadership” were among the most important factors influencing network implementation and functioning.

Secondary Analysis on Leadership Practices

Leadership was prioritized by participants as an important factor in both the evaluation of these three Health Links (described above) and in a previous broader evaluation of Health Links across the province of Ontario (Evans, Daub, et al., 2016). Leadership was not explored in depth in these evaluation studies beyond stating that leadership framing, commitment, and support influenced partnership building and staff engagement in the networks. It is unclear, based on these evaluations, how leadership manifested in Health Links and what influence different leadership configurations had on the networks. As a result, we decided to undertake a secondary, inductive analysis of leadership using the data already coded during the evaluation to understand leadership practices in the implementation of these coordinated care networks. In addition to the “Leadership Approach” and “Clinician Engagement and Leadership” factors mentioned above, we also pulled data coded under CCIC factors on “Governance,” “Organizational and Network Culture,” and “Partnering” because these codes contained content that reflected leadership practices. We conducted an inductive thematic analysis (Braun & Clarke, 2006) of these data with a focus on characterizing leadership practices. Through an iterative process involving coding the data, perusals of the literature on network leadership, and discussions with the research team on emergent findings, the role of individual and shared leadership emerged as an overarching issue and distributed leadership theory as the most appropriate lens with which to interpret the results. The first author thus undertook an analytic mapping exercise to map coded instances of leadership practices along a continuum of individualized to distributed leadership. This process was informed by the work of Currie and Lockett (2011), using the dimensions of concertive action and conjoint agency to map out a “spectrum of leadership variants” (p. 288). Members of the research team (JG, RA, SK, AG) met 3 times during the analytic mapping process to discuss the classification of coded chunks of data to reach a consensus. This mapping exercise validated and helped refine the initial inductive coding of leadership practices. Finally, we also coded for contextual factors influencing leadership.

Findings

Twenty-one leaders and nine providers participated for a total of 30 participants (Table 1). Some participants were members of organizations that spanned multiple Health Links, although they were only counted once, and classified herein based on their self-identified primary Health Link. For example, there was only one home and community care access center (CCAC) within the region; as a result, the CCAC’s mandate spanned all Health Link networks in the region, such that all of our cases were partnered with the same CCAC.

Table 1.

Participant Breakdown per Health Link.

Participant role	Case 1	Case 2	Case 3
Leaders/managers	6	10	5
Providers	4	2	3

For each case, we describe the network and focus on two contextual factors: information technology and clinician engagement. In our analysis, these factors were prominent and may help explain variation in leadership practices between cases. We then describe the network’s leadership approach, drawing from the concepts of conjoint agency and concertive action in our summative remarks. Figure 2 presents a mapping of the three cases on to Currie and Lockett’s (2011) 2 × 2 matrix of distributed leadership configurations.

Figure 2.

Three Health Link Cases Mapped on to Matrix of Distributed Leadership Configurations.

Source. Adapted from Currie and Lockett (2011).

Case 1

Health Link Description

Case 1 implemented a structure of “core” organizations across the network. In addition to the lead organization, three other organizations were designated as “core” partners. All four organizations were listed on the business plan, including the lead organization (a hospital), a primary care practice, a CCAC, and paramedic services. At the time of the study, CCACs were responsible for providing and coordinating home and community care services through designated care coordinators.

This Health Link’s vision was to strengthen clinical relationships to facilitate the care coordination process—particularly between family physicians and care coordinators. This explains the Health Link lead organization’s decision to invite the CCAC and primary care practice as core partners. It resulted in family physician engagement and efforts to develop a strong operational working relationship with the CCAC and care coordinators. Information technology intentionally played a minor role in their implementation efforts. Senior leaders opted not to participate in an MOHLTC pilot initiative of a system for electronic Coordinated Care Plans (known as the Coordinated Care Tool; see Case 2 for further details). This was prompted by two factors. First, the lead organization had recently hired a new chief information officer, so the organization decided “consciously not [to push] that much on IT” (Case 1, Leader 1). Second, given the strong focus on clinician engagement, the lead organization recognized that the Coordinated Care Tool would have required additional time and training, particularly privacy training for physicians. Given that family physicians had access to an existing Electronic Medical Record (EMR) tool that allowed them access to hospital records, a designated electronic tool specifically for Health Links patients was seen as not worth the investment in training for physicians.

Individual and Distributed Leadership Practices

How do we influence . . . Like how do we get others to change their behaviours to meet the needs that we think [are] better for the system? (Case 1, Leader 1)

This quotation demonstrates Case 1’s leadership approach. In this case, the Health Link project manager was the central leadership figure. The lead organization decided what was best for their network and brought others on board with that vision. Leadership was distributed in terms of the involvement of multiple core partner organizations, however, leadership stemmed first and foremost from strong internal direction by the project manager and lead organization. Thus, leadership was distributed only to the extent that it served to further the well-defined vision of this formal leader.

Within the lead organization, senior management was very supportive of the Health Link and committed resources to advance its implementation. As one participant noted when referring to their efforts in pushing forward the Health Links agenda:

We have a steering committee that meets at the hospital. It’s co-chaired by myself and a senior leadership vice-president . . . as highly placed in the hospital structure as you can get. Remember that we meet once a month with our CEO directly, and the Health Link team. So you can’t get more highly placed access to the leadership. (Case 1, Leader 4)

The core partners shared strategic and operational decision-making responsibilities and worked together to advance their Health Link vision, including prioritizing input from clinicians. It was the leadership of the hospital, however—and particularly the Health Link project manager—that ultimately dictated the design of the Health Link. This included decisions on who should be invited to join as a core partner and when, how, and to what extent noncore partner organizations should be involved. In this sense, the decision-making power rested with the lead organization. This leadership approach was supported by most partner organization leaders who perceived that it furthered the implementation goals of the network. For instance, it was the lead organization that decided not to participate in the Coordinated Care Tool pilot, as described above. Although this decision directly affected all partner organizations, partners seemed to recognize the barriers to meaningful IT adoption in the Health Link and that the Coordinated Care Tool itself would not be enough to achieve system-wide integration. Moreover, given that clinician—and particularly physician—engagement was such a strong pillar of implementation in this case, physicians themselves believed that IT should have been the purview of the provincial government, rather than left to each Health Link to address. As stated by one family physician, “Well, it’s probably wider than the hospital. Because it is the province that licensed different EMRs . . . We’re not all on the same EMR . . . This is an e-Health Ontario problem to solve.” (Case 1, Leader 4) This physician goes on to state,

The hospital as a matter of leadership and culture supports Health Links very, very strongly. So, if they had the hundred million dollars, they’d be spending it on [IT]. They know it’s a priority. I believe we’re talking about resources. The will, the leadership, and the culture for this kind of integrated health is all here. (Case 1, Provider 4)

Clinician engagement also extended into direct decision making in this Health Link, and the resulting distributed leadership practices were perceived as particularly effective in achieving physician buy-in. Family physicians were actively involved in decision making and broader leadership in this Health Link due to several factors, including (a) the primary care practice was named a “core” partner, which meant that senior leaders from this practice were actively involved in strategic meetings and decision-making processes; (b) the Health Links steering committee was cochaired by an administrative lead and a family physician; and (c) a substantial number of family physicians who worked in the partnering practice were also credentialed at the hospital’s Department of Family Medicine, meaning they spent at least a portion of their time at the hospital, which facilitated conversations and buy-in for the Health Link. Again, the influence of the Health Link project manager’s leadership could be felt here, as the intentional decision to enable family physicians to share in the Health Link network’s leadership stemmed directly from the project manager’s vision of strengthening clinical relationships.

Family physicians were also involved in patient care planning, and enacted leadership at the front lines through case conferencing. Case conferences were meetings of the full patient care team (e.g., care coordinator, family physician, social worker) to discuss the care for one patient, and often included patients and caregivers. Although case conferences were too time consuming for many physicians, they found workarounds to stay engaged in the process. For example, they called the care coordinator directly, who could then bring points of discussion back to the rest of the care team. Although many family physicians were indeed frontline leaders and champions of Health Links, leadership at the level of care provision varied across individuals, as reflected by a Health Link leader:

. . . We have some family physicians in that [primary care practice] that are probably the best family physicians we’ve ever worked with, and then some that may be some of the worst . . . in terms of coordinating care for patients. (Case 1, Leader 1)

Furthermore, although distributed leadership extended to frontline provider involvement, this was perceived by participants to focus largely on family physicians with lesser involvement of other clinicians or nonclinical providers.

Overall, the leadership approach in Case 1 blended a strong focus on alignment from the top-down (high conjoint agency), driven by the lead organization and project manager, with more widely distributed leadership across core partners (moderate concertive action), placing this case within the bottom left quadrant of Figure 2 as an example of concentrated group leadership. The interplay of these leadership practices facilitated the implementation of the Health Link: The network had a clear vision, built by the network project manager and maintained by the core partners, for how they wanted to coordinate care, especially through involvement of family physicians. The project manager’s strong vision could have reflected an individualistic leadership approach; however, they galvanized the hospital’s senior management into offering resources and support to advance the network and meaningfully engaged core partners and family physicians in decision making. This leadership approach ultimately allowed the Health Link to advance a vision that partners both believed in and had the knowledge to implement.

Case 2

Health Link Description

Case 2 focused on medically complex patients with mental health challenges. It was also the only Health Link of the three cases who participated in the Ministry-led pilot of an electronic version of the Coordinated Care Plan, known as the “Coordinated Care Tool.” The pilot was designed to facilitate both the creation and sharing of digital Coordinated Care Plans among Health Link providers. However, not all partners had access to the digital tool, resulting in challenges in sharing information. These challenges were compounded by two issues. First, one of the major partner organizations (the CCAC, responsible for assigning designated Health Link care coordinators from within their organization) mandated that Coordinated Care Plans be entered in their own internal electronic medical record system (rather than the Coordinated Care Tool), resulting in duplication and care plans that could then only be shared by fax. Second, only some partners had authority to create and edit the plans. Many of the “non-author” partners were community organizations, focused on mental health or social services, resulting in the perception that they did not have an active role in the Health Link, despite its explicit mental health focus.

Although efforts were made to engage family physicians, physician engagement was not a core component of the network’s activities, as they were in Case 1. Participants perceived that physician engagement was insufficient and occurred too late, particularly regarding the Coordinated Care Tool. Because family physicians were not included in the decision making to participate in the pilot, they had to be convinced of the value of spending time learning and adopting the new technology.

Individual and Distributed Leadership Practices

Health Links means a hundred different things to everybody. Some people think Health Links is just [Name of Health Link project manager] all by [themselves] . . . They don’t realize they’re a Health Link, you know? (Case 2, Leader 1)

Similar to Case 1, there was strong support for Health Links from senior leadership of partner organizations in this network. Many partner organizations had a representative sitting on the Health Link’s steering committee, including paramedic services and smaller community partners, some of whom reported that their involvement was facilitated by the “respectful” engagement of community partners: “[hospital] leadership there has great knowledge of the community and great respect for the work we do in the community. And I think that’s everything” (Leader, Cases 1, 2, 3). However, even though senior leadership of the lead organization supported the initiative in principle, the actual work of the Health Link was managed entirely by the Health Link project manager, with support from a care coordinator and administrative assistant. It appeared as though inadequate resources were dedicated to the Health Link. The project manager was managing the network’s operations in addition to the full-time responsibilities associated with their formal position at the hospital. They were also responsible for the Coordinated Care Tool pilot. This left little time for the project manager to act as a leader, engaging in critical leadership practices such as championing the network to nonaffiliated partners and clinicians in the region, securing resources for providers to coordinate care, and establishing a strategic direction at the steering committee table.

The absence of a centralized figure to guide the network resulted in confusion among partners. Several participants reported that details around implementation of the Health Link had not been well communicated to staff, resulting in confusion around the goals of the Health Link and a lack of basic understanding of its operations. This was particularly an issue for those care providers involved in the completion of Coordinated Care Plans with patients. The consequences of a lack of vision and leadership in this Health Link are evident in the following quote, from a provider trying to describe their understanding of the Health Link leadership:

So that’s not my understanding that [hospital] is the lead. My understand[ing] is that as far as the Coordinated Care Plan goes, [mental health organizations] are the lead . . . But we have CCAC as the overall . . . I guess the lead. Once we make the referral then [the CCAC] will act as the lead. (Case 2, Provider 8)

This quote demonstrates that the frontline providers involved in the care of Health Links patients had little clarity about network operations, which limited the perceived value of the Health Link. Another contributing factor was the lack of family physician engagement. Several participants mentioned wanting to get primary care buy-in through “small wins” and clear outcomes, but this was largely not achieved:

I would have thought that [by now the] Health Link would be well known by the primary care community because of the successes, and in fact, that they would be searching us out to help with patients and make their life easier. And that I don't think has happened . . . Some individual physicians had good results . . . Others have had involvement and keep asking the question—So where is there any difference? (Case 2, Provider 7)

By the time the steering committee attempted to create a strategy for engaging family physicians, there were already doubts about the value of Health Links. Other frontline providers had similar experiences. One supervisor at a partnering mental health organization reported that although their director was regularly attending steering committee meetings and disseminating information down to supervisors, this information had not translated into frontline action (e.g., taking initiative to identify patients that might be well-suited for enrollment in the Health Link).

Nevertheless, frontline providers created new relationships with their counterparts in partnering organizations, thus enacting local leadership at an operational level. This occurred in “Community Rounds,” a regular opportunity for frontline providers from the Health Link to meet and discuss anonymized patient cases. Community Rounds included frontline providers and managers, as opposed to management-only steering committees. In these rounds, attendees learned about the services offered by other partner organizations, allowing frontline providers to make real-time decisions around the care offered to their Health Link patients.

So if somebody comes up with an idea, their manager will say yes, we can do that. Because that’s often a barrier, right . . . But we’ve found in the rounds, talking about these complex patients . . . makes it more concrete . . . to the partnership. It illustrates what can be done when you bring everyone together. (Case 2, Leader 4)

Overall, Case 2 demonstrates an absence of top-down alignment of leadership (low conjoint agency) and an overreliance on frontline leadership (low to moderate concertive action), placing it into the bottom right quadrant of Figure 2 as an example of weak distributed leadership, sitting closer on the continuum to “individualistic leadership.” Although there were numerous partners on the steering committee, the lack of direction from the Health Link project manager seemed to limit the perceived effectiveness of this group in setting a strategic agenda and vision, and in disseminating information about the network. Leadership was enacted through frontline provider interactions, particularly through Community Rounds. However, these leadership practices were not distributed consistently across the network nor did they engage all partners. As a result, Case 2 represents a paradox in which the data suggest both an “individualistic” leadership approach that resulted in ineffective network implementation according to stakeholders, and the need for a stronger individual leader.

Case 3

Health Link Description

Case 3 had a “false start,” which involved a change in senior leadership and resetting of the direction and aims of the entire Health Link, as reflected by the new project manager:

I was asked to provide interim coverage for the director on a short-term basis. But then shortly after, the project manager . . . she left also. So I was left sort of covering all of it. Now I’m in a joint position . . . I am actually the one armed paper hanger here because there isn’t anybody. (Case 3, Leader 1)

In the early iterations of this Health Link, the hospital’s vision was to adopt a primary care focus. This involved early engagement of primary care stakeholders, as well as attempting to develop a patient identification and referral process that was based in a primary care clinic (as opposed to, for example, emergency departments [EDs] of local hospitals). Although this seemed to be a reasonable approach to establishing the Health Link, it did not consider the perspective of family physicians themselves, who—similar to Case 2—struggled to see the value of Health Links.

So our goal was to try and get referrals from primary care. But it was very hard . . . to engage with the diverse population of primary care providers in the community when you have a “product” that you’re trying to sell to them and they don’t understand the concept and there's no real net gain to them. (Case 3, Provider 3)

This sentiment was prominent among family physicians associated with team-based interdisciplinary primary care practices, as they perceived Health Links as a duplication of services already offered by their organizations. Like Case 1, this network did not participate in the Coordinated Care Tool pilot. In fact, the role of information technology was not emphasized by participants in this network.

Individual and Distributed Leadership Practices

Senior leadership reflects the direction of the organization. So if the organization has changed [its] emphasis . . . it’s very difficult and [Health Links] will fall apart . . . And if you don’t have that support, it’s like sort of working with Jell-O. There's nothing to really hold onto. (Case 3, Leader 1)

Turnover was a significant issue in this case. At the time of the interviews, it had only been 3 months since the turnover of the Health Link project manager. Turnover also included hospital senior management and leadership representation on the steering committee. As stated by one leader, “there’s not a single person that was attached to that from the top down that’s still there” (Case 3, Leader 4). However, the most significant turnover was the replacement of the Health Link project manager. Its impact can be best understood by comparing the leadership approach before and after the new project manager took over.

The first iteration of this Health Link under the prior project manager reflected uncoordinated and disjointed leadership at the steering committee table. Participants discussed having upward of 20 partners invited to the steering committee, with no strategic aim for their participation. Although in theory this would have been a worthwhile approach to incorporate multiple perspectives—creating an opportunity for strong distributed leadership across partners—the lack of overall direction from the Health Link project manager limited the steering committee’s ability to create and execute meaningful strategy. The manager of one community organization stated that community partners were not purposefully engaged in these meetings, “because it was more of a CCAC and hospital kind of interchange. [We] were there and they could offer support or other resources but . . . [we] didn’t need to be there” (Case 3, Leader 4). This suggests that although governance may have been structurally distributed (in terms of the number of strategic partner organizations), leadership itself was not distributed, and may not have been present at all.

This lack of direction was compounded by the ongoing turnover of the hospital’s senior leadership team, particularly in terms of inconsistency regarding their internal championing of the Health Link. This reduced access to resources and resulted in having to complete their Health Link work “off the side of their desk,” an expression used by several participants to describe how Health Links had been added to their existing full-time responsibilities.

Conversely, when the new Health Link project manager took over the role, one of the first course corrections was to reengage select partners on the steering committee. They made a concerted effort to reengage in a manner that captured the expertise of providers, by seeking out and leveraging ideas already implemented in these organizations. This was done to ensure community organizations felt like equal partners, and to set a strategic direction that reflected the way on-the-ground care was organized. Participants were more selectively invited to join the committee to offer representative membership that allowed for active engagement. To this end, organizational partners who had formerly participated in the network’s steering committee were asked to submit an expression of interest to participate on the reformed steering committee. The outcome of this shift was a stronger sense of shared purpose, where partners were able to understand the value of the Health Link network. A paramedic service representative described this:

Really we’re trying to stay at the table because we feel we have lots of links to Health Links in terms of identification of patients [and] in terms of potentially having an intervention or when we’re responding to somebody that has a Coordinated Care Plan in place. We can get authority for paramedics to actually follow the Coordinated Care Plan. (Cases 2 and 3, Leader)

There was also a perception among participants that the new project manager had reenergized and revived what some had considered a failing implementation:

. . . [Health Link project manager is] a feet-on-the-ground kind of person. There was a resurgence of interest by organizations, individuals, doctors, the community services, all kinds of stuff, when [they] took over and said [they were] just going to make it really simply about providing better coordinated care to the client. (Case 3, Leader 7)

The new project manager championed two key initiatives upon taking the leadership role. First, the project manager established a “Health Link Alert” system where CCAC coordinators received an alert in their electronic medical record when a Health Links patient came to the hospital’s ED, enabling rapid coordination of care and improved awareness of Health Links among ED nurses. Second, the Health Link project manager reintroduced Community Rounds, of which there had only been three at the time of interviews. Similar to Case 2, the Community Rounds were valued by providers, who reported feeling “they were finally doing something” (Case 3, Leader 1).

Overall, Case 3 demonstrates the importance of attention to network leadership dynamics over time. The earlier implementation of this Health Link lacked vision and direction (low conjoint agency) and lacked meaningful engagement of physicians or of partners at the steering committee table (low to moderate concertive action). The latter implementation phases were supported by a clearer top-down approach, and also the absence of an explicit strategy, choosing instead to focus on fostering an environment where partners could cocreate goals and strategies (moderate conjoint agency, high concertive action). We note this shift in leadership practices in Figure 2 with an arrow to demonstrate the movement of Case 3 within the top right quadrant. As a result, Case 3 may be the most successful example of distributed leadership, approaching the “pure distributed leadership” quadrant in Figure 2.

Discussion

Our study examined the role of individual and distributed leadership practices in the implementation of flexible coordinated care networks. We found that distributed leadership was essential to the networks’ implementation. Yet, manifestation of distributed leadership was contingent on the presence of a single individual leader who acted as a unifying force, promoting both conjoint agency and concertive action.

Our findings demonstrate that networks cannot be effectively implemented without the clear direction of an individual leader who cocreates and spreads a vision across individuals and organizations. For instance, although each network had a governance model in the form of steering committee tables, the enactment of leadership across these governance tables was highly context dependent. For example, in the revised structure in Case 3, the steering committee table distributed leadership across the entire network. However, this governance approach was perceived as largely ineffective in Case 2. Although their steering committee structure was the same as other networks, without clear direction on how Health Links should be implemented, from an individual leader or the lead organization, gathering partner organizations together had few results.

Our findings demonstrated different, emergent approaches in how the individual leader interacted with and influenced the enactment of distributed leadership. In Case 1, the Health Link project manager’s leadership “radiated” throughout the network, fostering conjoint agency starting with hospital leadership, then expanding to hospital family physicians, and continuing outward toward core and peripheral partners. However, this “radial” model of leadership appeared less effective the further away it grew from the hospital. In Case 2, leadership was distributed at both the horizontal and vertical levels, but the two did not interact, and we suggest this is because of the lack of strategic direction and guidance from the lead organization and project manager. In Case 3, an individual leader influenced the enactment of distributed leadership across the network. The project manager took on the role of the transformational leader (Bass & Riggio, 2006): Participants offered overwhelmingly positive feedback regarding the Health Link project manager’s ability to turn around a seemingly sinking ship. Our results reinforce the argument that a strong individual leader does not necessarily imply an individualistic leadership approach; rather, a strong individual leader can be the driving force in achieving a distributed leadership model that maximizes both concertive action and conjoint agency (Günzel-Jensen et al., 2018; Leithwood et al., 2007).

These findings also point to how distributed leadership can be viewed as the outcome of effective network implementation, so long as it is underpinned by an effective individual leader. As networks develop, the emergence of distributed leadership may be an indicator of successful implementation. The movement of Case 3 from the “Broad Collaborative Leadership” quadrant toward the “Pure Distributed Leadership” quadrant in Figure 2 reflects this notion; the new project manager’s leadership practices, which enhanced conjoint agency and concertive action, contributed to higher perceived effectiveness of the network among participants. Conceptualizing distributed leadership as an outcome aligns with the work of Currie et al. (2011), who found that the emergence of leadership forms (particularly distributed leadership) could be examined as an outcome of policy initiatives aimed at creating networks. As a network matures, there may be a progression from fragmented or individualistic leadership toward leadership that is more distributed (Currie & Lockett, 2011). An indicator on the extent to which leadership is distributed could provide insight on network development and functioning and complement clinical process and outcome indicators.

This study has limitations. First, we conducted a retrospective secondary analysis, therefore the enactment of leadership was not a primary research question of the initial data collection. However, we believe this strengthens the validity of our findings. Interview studies focused on leadership as a primary aim are subject to impression management and other related biases (Alvesson & Einola, 2019). By drawing from a data set that focuses more generically on the Health Links and their implementation, we reduced potential for bias in data collection and were better able to ascertain both the enactment of leadership and its impact on the network. It was appropriate to explore leadership in these data given the inclusion of leadership in the CCIC framework (Evans, Grudniewicz, et al., 2016), which was used to design the interview guide and in the deductive analysis. The interviews were conducted during a single time point, and as a result, we are unable to capture implementation perspectives longitudinally. This was particularly relevant in the description of the third case, where our data only capture participants’ reflections of early implementation efforts and their experiences with and perceived impacts of leadership turnover. Nevertheless, we suggest this is buffered by participants’ ability to reflect on their own historical experiences with the implementation of Health Links. We attempted to further account for this in our design by selecting cases with comparably longer histories of implementation (i.e., “early adopters”), and by targeting individuals who had been part of the Health Link since its early implementation. Our data also have greater representation of organizational leaders than frontline providers, particularly family physicians. However, we would suggest this reflects more strongly the findings of our study, rather than a limitation per se. Given the challenges to clinician engagement addressed in our findings, as well as the noted lack of perceived value of Health Links, we expect that our recruitment efforts reflect the limited involvement of these providers.

Our results highlight several directions for future research. First, there is a need for longitudinal data collection to capture distributed leadership dynamics over time. Second, it remains unclear whether there should always be a dual emphasis on concertive action and conjoint agency. For example, should one be emphasized over the other depending on network stage of development or contextual factors? Third, we need to better understand how contextual factors influence the development of distributed leadership in coordinated care networks. For example, what influence might patient population, clinical and/or social foci, and nature of the organizations involved have on distributed leadership? Fourth, the notion that distributed leadership could serve as an indicator of an effective network requires further exploration.

Our results also have implications for policymakers and network leaders. First, it appears that coordinated care networks require a dedicated individual leader to achieve distributed leadership across organizations. As such, time and resources must be dedicated to the appointment and support of an individual leader. Second, the individual leader should strive to build both conjoint agency and concertive action as both are needed to achieve a pure distributed leadership model. Our results offer concrete examples of how leaders may promote and balance conjoint agency and concertive action. Third, provider engagement played a strong role in the enactment of distributed leadership. Where providers were engaged early and meaningfully, distributed leadership was more likely to develop. Finally, we recommend the development of formal or informal methods for assessing the extent to which leadership is distributed in a network. Periodic assessment can provide insight into network development and effectiveness and inform remedial action.

Conclusion

As interorganizational collaboration increasingly becomes the norm in health care delivery, understanding the leadership practices that underlie successful networks is essential. In this study, we examined how distributed leadership manifested in the implementation of three coordinated care networks operating in a flexible policy context, and the role and influence of individual leaders within these networks. We suggest that the growing support among scholars and practitioners for broad, collaborative models of distributed leadership emphasizing concertive action may be preemptive, or at the very least lacking nuance. Our findings show that distributed leadership is essential but insufficient on its own for successful network implementation in such highly flexible environments. Instead, these environments appear to require a unifying leader to direct network partners and promote distributed leadership by facilitating both conjoint agency and concertive action.