Distributed leadership in health quality improvement... : Health Care Management Review (2025)

Studies of public sector leadership have grown considerably in number, and the area of public leadership research has moved from a primary focus on traditional hierarchical, political, and administrative leadership to also including leadership in interorganizational collaboration and networks (^{Chapman et al., 2016}; ^{Croft et al., 2022}). This is particularly the case in health care research, where the complexity of modern patient care has increased awareness about collaboration between mutually interdependent actors and organizations. Many health care organizations have thus started to work systematically with collaborative and network-based approaches to quality improvement (QI) (^{Currie & Lockett, 2011}; ^{Mainz et al., 2015}). A recent example is ^{Atkinson et al. (2022)} studying collaborative patient safety learning laboratories to support interdisciplinary team innovation in the United States.

The QI collaborative (QIC) model represents one such approach to network-based QI that has become widely used across different clinical areas and organizational settings. Improvement areas include (among others) adult medical and surgical admissions, obstetric patients, intensive care patients, organ donors, management of chronic conditions, fall prevention, and pain management. The majority of the QICs are implemented within acute hospital settings and ambulatory care, but organizational settings also include general practice, nursing homes, and ambulance services (^{Carstensen et al., 2022}; ^{de Silva, 2014}). Within a QIC, a learning network of teams of health care professionals from hospitals and other health care provider organizations share methods, ideas, and data within and across teams to improve health care processes and outcomes for a targeted area of care (^{Zamboni et al., 2020}). However, despite the widespread use of QICs, empirical research is lacking on how they are implemented, and which leadership practices are exerted within them to support QI (^{Carstensen et al., 2022}, ²⁰²³; ^{Dückers et al., 2009}; ^{Zamboni et al., 2020}).

To describe patterns of influence and leadership practices in settings where activities and services are performed across organizations and professions and largely outside the boundaries of formal, hierarchical leadership, we argue for the usefulness of a distributed leadership perspective. Distributed leadership can be described as a phenomenon whereby influence is distributed among organizational members to organize concerted action within and across organizational levels (^{Gronn, 2002, 2009}). A distributed leadership perspective is therefore particularly suitable for analyzing how leadership is practiced in health care and QI teams, where many activities have a highly conjoined production requiring extensive coordination and the inclusion of different competencies across organizational levels and boundaries (e.g., ^{Chreim & MacNaughton, 2016}; ^{Jakobsen et al., 2021}; ^{White et al., 2014}). Previous distributed leadership research shows positive effects on, for instance, patient satisfaction and innovative behavior (cf. systematic reviews by ^{Bolden, 2011}; ^{Tian et al., 2016}); however, very few studies have examined distributed leadership practices in interorganizational and cross-professional settings (^{Boak et al., 2015}; ^{Chreim & MacNaughton, 2016}; ^{McKee et al., 2013}). Here, an important condition for expecting positive outcomes of distributed leadership is whether the leadership practices in regard to the aims, roles, and scope of the collaborative work are aligned within and across organizational entities (^{Harris et al., 2007}; ^{Jakobsen et al., 2021}; ^{Leithwood et al., 2007}; ^{Thorpe et al., 2011}). This may be particularly challenging when leadership is exerted in interorganizational collaborative arrangements.

To contribute to research investigating distributed leadership in interorganizational collaboratives and the literature on QIC leadership, we therefore examine the following research questions: How is leadership within health QICs characterized by aligned distributed leadership practices? And how do these practices relate to team members’ experiences with progress and achievements in the QI work? These research questions are studied in a qualitative multicase study of two nationwide Danish QICs implemented within a national health care quality program. Involving multiple professions and collaboration across organizational and sector boundaries, this case offers a more complex setup than most existing distributed leadership research. Hence, this study contributes with important insights into the interrelated objectives of (a) how aligned distributed leadership practices can support QI efforts in health care and (b) how leadership is practiced in complex, interorganizational, and cross-sector collaboratives without clear hierarchical anchoring.

Literature Review and Theoretical Framework

Leadership in QICs

The literature on QIC leadership is sparse, typically with a perspective limited to leadership support of QICs in hospital settings. Studies have noted how supportive hospital management is positively associated with improved QIC outcomes and constitutes a critical factor for professional engagement in QIC implementation (^{de Silva, 2014}; ^{Dückers et al., 2009}; ^{Hulscher et al., 2013}; ^{Lowther et al., 2021}; ^{Williams et al., 2022}; ^{Zamboni et al., 2020}). Similarly, several studies have documented the negative effects of lacking leadership support on QIC results and team participation in the QI work (e.g., ^{Hulscher et al., 2013}; ^{Lowther et al., 2021}). However, these studies devote little attention to addressing how leadership supports or hinders QIC implementation and which types of leadership practices are associated with these outcomes.

A small number of recent studies have begun to extend our insights into leadership needs and practices in QIC implementation. ^{Lowther et al. (2021)} report that regular communication with leaders was a successful tool to overcoming barriers of insufficient team member engagement and to obtain buy-in from the leaders to implement the QIC change initiatives. ^{De Silva (2014)} describes the importance of a present and supportive hospital executive for the success of the QIC project, as it facilitates staff recognition of a strong mandate for collaborative participation. Finally, ^{Zamboni et al. (2020)} point to the role of hospital leadership in promoting open dialogue and bottom-up problem-solving in QIC implementation, which may facilitate a collective sense of responsibility in the QI work. In terms of leadership practices, Zamboni et al. highlight how an important aspect of the leaders’ role is to pave the way for a shared vision of the QIC project. This requires a shift away from traditional authoritarian leadership models, which may “be more easily achieved where some of the habits for improvement already exist, or where organizational structures, for example, decentralized decision making or non-hierarchical teams, allow bottom-up problem-solving” (^{Zamboni et al., 2020}, p. 16). These findings suggest a need for a leadership perspective capable of encompassing the complex nature of interorganizational QI efforts in health care and identifying which leadership practices beyond “supportive/unsupportive” help QICs to achieve the desired improvement achievements. The present study examines how aligned distributed leadership can offer such a perspective.

A Distributed Leadership Perspective

Leadership can be defined as “the process of influencing others to understand and agree about what needs to be done and how to do it, and the process of facilitating individual and collective efforts to accomplish shared objectives” (^{Yukl & Gardner, 2020}, p. 26). Depending on the situation and task at hand, multiple actors may be performing generic leadership activities related to change-oriented, task-oriented, and relation-oriented behavior to improve services and support team functioning (cf. leadership behavior meta-categories of ^{Yukl & Gardner, 2020}). For instance, a QIC can include both administrative and clinical leaders together with different clinical specialists, all of whom partake in coordinating team activities, ensuring progress and desired quality achievements by simultaneously exerting horizontal and vertical influence in the QI work. In this sense, “the distribution of clinical and administrative responsibilities across individuals in the leadership constellation may not stem entirely from a formal description of the division of labor, but come about with the influence of circumstances such as team requirements” (^{Chreim & MacNaughton, 2016}, p. 201). Challenging the leader–follower dyad as the dominant leadership model, distributed leadership has at its core the conception of leadership as a collective phenomenon arising across individuals, positions, and levels (^{Gronn, 2002, 2009}). In other words, the focus is on the act of leading rather than the initial role positioning (^{McKee et al., 2013}, p. 12). As such, distinct features of distributed leadership in comparison with related collective leadership approaches, such as, for instance, shared leadership and democratic leadership, are the multiplicity of actors involved in exerting influence within and across organizational levels, and that leadership is conceived as more than (democratic) involvement of subordinates in decision-making processes (^{Bolden, 2011}; ^{Woods, 2004}). Thus, although democratic leadership can be a component of distributed leadership, the latter encompasses a broader organizational approach that encourages leadership emergence from various levels and functions. In contrast, nondistribution of leadership is characterized by clear identification of a formal, hierarchical leader and a (primary) vertical flow of authority and information rather than dispersed influence across actors.

Leadership practiced in collaborative teams can denote a particular distributed leadership configuration or constellation (^{Chreim & MacNaughton, 2016}, p. 202; ^{Gronn, 2009}, p. 383); that is, the specific pattern of leadership actions performed by multiple actors in the team and the observed division of roles and influence among them (^{Denis et al., 2001}). Drawing on ^{Gronn’s (2002)} emphasis on concerted action as a core element in distributed leadership, distributed leadership practices can materialize as spontaneous collaboration; that is, intuitive working relationships and division of leadership that emerge over time or as more formal and institutionalized practices of influence (p. 431). Hence, a specific pattern of distributed leadership practice within a team can be relatively more emergent or planned (^{Thorpe et al., 2011}).

An important point of awareness when studying the contributions of a distributed leadership approach to public service improvements is, however, whether distributed leadership practices are aligned within and across organizational entities. Existing research in distributed leadership (^{Harris et al., 2007}; ^{Jakobsen et al., 2021}; ^{Leithwood et al., 2007}; ^{Thorpe et al., 2011}) emphasizes how a precondition for positive distributed leadership outcomes is the willingness and competence of team members to assume leadership and exert influence in accordance with stated goals and priorities of the collaborative work. We argue that this implies shared understandings of the aims and methods of the performed activities as well as the scope and structures within which service delivery and desired improvements take place. Moreover, it is important that there is a clear and mutually recognized role and task distribution (^{Chreim & MacNaughton, 2016}). Similar points regarding the importance of alignment between interdisciplinary team activities, organizational context, and project objectives have also been identified in other studies of collaborative arrangements to foster health care innovation (^{Atkinson et al., 2022}).

In our study, we thus focus on different domains of aligned distributed leadership practices (cf. Figure 1). These aligned practices can come about through informal processes (particularly when the involved actors already know each other well) or more formal processes of deliberate alignment efforts (^{Harris et al., 2007}; ^{Leithwood et al., 2007}). Whereas informal, dynamic alignment processes are often characterized by being tacit and spontaneous, formal processes are more institutionalized practices of working out agreements among the sources of leadership (managers and employees) about which leadership functions are best carried out by which source (^{Harris et al., 2007}, p. 344). Hereby, these different alignment processes often correspond to emergent and planned practices of distributed leadership (^{Thorpe et al., 2011}). Whereas the former processes provide empirical evidence for short-term, positive changes, a formal process of establishing aligned distributed leadership practices is positively associated with more long-term results (^{Leithwood et al., 2007}). Conversely, misaligned distributed leadership is characterized by failure to establish effective coordination and consensus between tasks and functions or ambiguity in goals and roles and outright conflict (^{Thorpe et al., 2011}, pp. 245–246). ^{Jakobsen et al. (2021)}, for example, revealed how the degree of alignment conditions the impact of distributed leadership on individual performance and innovative behaviors in a hospital setting, whereas ^{Chreim and MacNaughton (2016)} found that ambiguous distributed leadership roles in health care teams harmed team functioning.

The achievement of aligned distributed leadership practices in the domains of scope and structure, aims and methods, and role and task division in the collaborative work may be especially challenging in collaborative arrangements spanning sectors and organizational boundaries—often with no clear hierarchical anchoring and including multiple professions that must coordinate their efforts while potentially holding conflicting values and norms. Here, ^{Currie and Lockett (2011)} have stressed how “the health and social care context creates a paradox for distributed leadership” (p. 287); that is, distributed leadership is regarded as inevitable when complex public services require change and improvement, but it is this very complexity that renders attempts to practice distributed leadership difficult. This study therefore explores how distributed leadership practices are actually enacted and aligned in a context of interorganizational and cross-sector health care QICs. In line with our theoretical arguments, the identification of aligned distributed leadership practices in QIC work focuses on three different domains of aligned distribution of leadership and the planned or emergent processes whereby these practices come about. Based on the emphasis in previous research on the importance of aligned distributed leadership practices (^{Harris et al., 2007}; ^{Jakobsen et al., 2021}; ^{Leithwood et al., 2007}; ^{Thorpe et al., 2011}), we expect more aligned practices within these domains to positively influence how team members experience progress and achievements in the QIC work. This is illustrated in Figure 1.

Methods

Study Design

The analysis was based on a qualitative case study of two nationwide QICs being implemented within the Danish national program for health care quality. Our unit of analysis is local QI teams including different health professionals—and, if relevant, different hospital departments and/or municipal health care provider organizations—within each QIC (see Supplemental Digital Content A, https://links.lww.com/HCMR/A131, for a detailed description of the study settings and structure of Danish QICs). The study applies a multicase design, as it allows for a systematic and in-depth exploration of aligned distributed leadership practices across QICs with different degrees of organizational complexity and the relation of these practices to how the teams experience the progression and achievements in the QIC work (^{Yin, 2018}).

Our two cases are “QIC on children diabetes” (QIC diabetes) and “QIC on upper femur hip fractures among 65+-year-old patients” (QIC fractures). These choices were based on a diverse case selection strategy (^{Seawright & Gerring, 2008}) to allow for the exploration of different characteristics associated with the QICs that could be significant in relation to the exercise of distributed leadership in the local QI teams and for the potential alignment of distributed leadership practices between team members. As evident from Table 1, which summarizes the characteristics of and data collected in the two QICs, the variation related to differences in organizational complexity and the composition of the local QI teams, according to which QIC fractures is substantially more complex in terms of cross-professional and interorganizational/cross-sector collaboration than QIC diabetes.

The selection of local QI teams within the two QICs sought to ensure organizational and geographical variation to increase the analytical generalizability (^{Brinkmann & Kvale, 2015}; ^{Yin, 2018}). In both QICs, we thus included QI teams from all five Danish regions, and we selected teams from different-sized hospitals (university hospitals and smaller hospitals) in QIC diabetes and from hospitals and municipalities in QIC fractures. The selection of QI teams was also based on pragmatic factors, however, such as being able and willing to participate.

The study was conducted in accordance with the Declaration of Helsinki and registered by the register of public research projects in Central Denmark Region (File No. 1-16-02-285-19). The Central Denmark Region Committees on Health Research Ethics deemed our study not to be a health research study, according to the Consolidation Act on Research Ethics Review of Health Research Projects, Consolidation Act number 1083 of 15 September 2017 Section 14, meaning that it does not require ethics committee approval. Prior to all observations, written and oral information about the study were provided to all participants, and oral, informed consent was obtained. For the interviews, all participants were informed about the study, and written consent was obtained. In the analysis and presentation of the collected data, all participating QI teams and individual participants were anonymized.

Data Collection

The data collection took place from spring 2020 to fall 2021 and was based on a combination of qualitative interviews, participant observation, and documentary material.

Qualitative interviews involved group and single-person interviews with local QI teams and local and regional coordinators in both QICs. We conducted group interviews with 18 QI teams—10 fracture teams and eight diabetes teams (66 informants in total). The groups had two to six participants with mixed professional backgrounds. In the group interviews with QIC fractures teams, the local coordinators also participated, whereas they were individually interviewed in QIC diabetes (two single-person interviews, three group interviews with nine informants in total). This difference in mode of interviewing local coordinators is explained by COVID-19 restrictions and developments in the organization of the QICs from the early QIC fractures to the later QIC diabetes. Finally, we conducted single-person interviews with regional coordinators from each of the five regions in both QICs (10 interviews). Thus, in total, we conducted 12 single-person and 21 group interviews with a total of 85 informants (Table 1 provides an overview of the conducted interviews; in addition, Supplemental Digital Content B, https://links.lww.com/HCMR/A132, provides more detailed insights into the interviews and informants).

The interviews were all conducted either by videoconference (n = 29) or in-person at a location of the informant’s choosing (n = 4). As data collection took place during the COVID-19 pandemic, most of the interviews had to be held by videoconference using Microsoft Teams or a regional virtual meeting platform according to availability of the informants. The group and single-person interviews lasted 45–90 minutes and were all conducted by the first author, who is an experienced social scientist and unaffiliated with the QIC implementation. The interviews were all guided by a semistructured interview guide with open-ended questions (^{Brinkmann & Kvale, 2015}). The guide was informed by our theoretical framework and concerned issues related to the organization of the QI team and the collaboration between team members, decision-making processes within the team, the degree of aligned practices between team members, and experienced challenges and possibilities in the collaboration. The interviews were all digitally recorded, transcribed verbatim, and anonymized.

Participant observations consisted of observations at different meetings central to the QI teams’ work within the QICs (e.g., national learning sessions, regional network meetings, and local improvement team meetings; cf. ^{Spradley, 2016}). In total, we observed 34 meetings of approximately 60 hours (cf. Table 1). The observations provided insights into the collaboration among team members, decision-making processes, and distributed leadership practices in the QI teams as they unfolded in practice, as well as the underlying tacit knowledge. Field notes consisted of both descriptive information (factual data) and reflective information (thoughts, ideas, and questions) to support the analysis of the empirical data (^{Spradley, 2016}).

Finally, we collected documentary material (e.g., administrative and policy documents and role descriptions) to provide context information regarding the distributed leadership practices within the QI teams, for example, in relation to the teamwork structure, and the formal collaboration practices and task and leadership role distribution within the teams. The documentary material also contributed to corroborating and augmenting the interview and observation data (^{Yin, 2018}).

Analysis

Thematic analysis (^{Braun & Clarke, 2019}) was used to identify and analyze practices of distributed leadership in the transcribed data, field notes, and documents. Initially, a set of theoretical, closed codes derived from the operationalization of the central concepts and subconcepts of our theoretical framework were applied. These were (a) aligned distributed leadership in relation to (1) the scope and structure of the quality work, (2) the aim and content/tools of the quality work, and (3) collaboration practices in terms of task distribution and leadership roles; (b) experienced progress and achievements in the QI work; and (c) context regarding the QIC organization (e.g., the interorganizational structure) and relationships with the hierarchical hospital/municipal management, regional coordinators, and national project management. A preliminary test coding of seven randomly selected group interviews was conducted by two of the authors to form an initial validation of the theoretically informed closed codes. The results of this initial coding were then discussed among all authors, which led to the addition and omission of some codes (for instance, a context code on amalgamation of wards) and more detailed definitions and minimization of overlap in the coded content regarding the subcodes of aligned distributed leadership. Additional codes involved a distinction between formal and informal achievements in relation to experienced progression in the quality work (see Supplemental Digital Content C, https://links.lww.com/HCMR/A133, for the final thematic coding scheme). Subsequently, the interviews were all coded using NVivo 12 software (QSR International, Melbourne, Australia).

The authors then further analyzed and discussed the coded material using the theoretical framework to inform the identified processes of aligned distributed leadership practices as being relatively planned or emergent and more/less aligned. Specifically, regarding distributed leadership practices in relation to each subtheme, we focused on statements about who initiated, coordinated, and followed up on team activities and how such decisions were made in terms of patterns of horizontal and vertical influence. Alongside this interview content coding, findings were triangulated with field notes from the observations and collected documents to improve the credibility of the analytical insights (^{Miles et al., 2014}). The triangulation particularly evolved around validating the described meeting structures and participants as well as observing the implications of the context influence, for instance, regarding the significance of the complexity of the QIC organization in relation to the distributed leadership practices within the QI teams.

Results

The theoretical introduction to aligned distributed leadership practices (^{Harris et al., 2007}; ^{Leithwood et al., 2007}) and the thematic coding of the qualitative material resulted in three subthemes that constitute different domains of aligned distributed leadership in the local QI teams: (a) scope and structure of the teamwork, (b) collaboration practices: task distribution and leadership roles, and (c) understanding and deciding on QIC aims and methods. In relation to each of these three subthemes, this section presents our findings on how leadership in QI work is characterized by aligned distributed leadership practices within the local QI teams and how this relates to the team members’ experiences with progression and formal/informal achievements in the QI work. Whenever relevant to the interpretation of the displayed findings, conditional context factors regarding the QIC implementation and relationships with the hierarchical hospital/municipal management, regional coordinators, and national project management are included.

Scope and Structure of the Teamwork

Regarding the scope and structure of the teamwork, most QI teams across the two QICs held regular, formal meetings to discuss their QI work. The frequency and duration of these meetings varied across teams, usually lasting roughly 60 minutes and held biweekly or monthly. The frequency and duration of meetings were generally an informally negotiated practice among the team members, as the following quote indicates:

We initially held monthly meetings. Finding a time that suited everybody was difficult, but we felt it important that we all participated to be able to coordinate our improvement work. We decided to change the frequency of our meetings to bi-monthly, however—it’s simply too hard to uphold monthly meetings where everyone can participate. (Physician, hospital team, QIC fractures)

The quote also illustrates how the meeting structure in several teams was renegotiated and adjusted during the implementation period in relation to what was needed and possible. As such, aligned practices regarding the teamwork structure arose through informal, emergent processes. The field notes made of observations from local QI team meetings verified this alignment practice and showed how meetings were often held without all team members being present or with some coming and going as their work schedule allowed.

In several QI teams, the QIC meetings became embedded in existing meeting structures. This was particularly the case in QIC diabetes, where most QI teams used existing monthly meetings in their clinical units to discuss QIC matters, as the members of the clinical unit and the QI team typically coincided. Among the QIC fractures teams, a few teams also embedded discussions of their QIC work into existing work practices addressing QI (e.g., weekly whiteboard meetings). Only one QI team reported having no formal meeting structure because of scheduling difficulties. Deciding on a meeting structure generally did not create tensions among team members, and there appeared to be general satisfaction across QI teams with the negotiated structure. However, in a few teams (mainly within QIC fractures), different geographical locations of team members or misalignment between team member work schedules challenged the possibilities for regular meetings. This created frustration, as it was considered detrimental with regard to experienced progress and the achievement of local priorities and the formally stated goals for the QIC work (listed in Supplemental Digital Content D, https://links.lww.com/HCMR/A134).

As with the meeting structure, decisions on the size and composition of the QI team (which a formal manager typically decides) were also a negotiated and emergent practice distributed within the teams. The team size generally varied from small, two- to three-member teams to larger teams of up to 14 professionals. In both QICs, several teams described how their team size and composition had developed over time in line with the development of their local QIC project:

We’ve joined the QI team continuously, one-by-one. First, it was us here in the physical rehab unit, then staff from the nursing unit came on board together with occupational therapy. We then invited the nutrition staff to join, as we became aware of the importance of their perspective for the success of our improvement project. It’s all interrelated—we need the different professional groups. (Physiotherapist, municipal team, QIC fractures)

The quote illustrates how a guiding principle of the negotiations was to ensure a sufficiently broad representation of professionals and departments to match the QIC project aims and scope. Consequently, the QIC fractures teams were typically larger compared to QIC diabetes. As the QIC fractures teams span across wards and organizational units and had no existing clinical fora for this type of improvement work, “the more the merrier” sometimes became the easy (and only feasible) solution in the absence of clear, hierarchical anchoring that had planned and institutionalized the team composition. Moreover, the team composition sometimes also relied on prior knowledge of each other and personal connections, which naturally contributed to the intuitive and tacit alignment of practices—but at the cost of transparent criteria for team inclusion. The size and composition of the QI team were not generally a source of conflict. However, the analysis revealed how the team size and composition can both enable and disable well-functioning distributed leadership practices in relation to achieving positive QI results. Although distributed leadership in teams with a broad representation of professionals across wards and organizational units can support the achievement of informal gains of improved communication and knowledge regarding conjoined tasks in the patient care programs, it can also harm the progression toward the formal goals of the QIC work because of prioritization difficulties.

As regard the issue of time and prioritization of the QIC work, the analysis showed how, across QI teams, team members shared the understanding that they lacked the time and resources to prioritize the work as highly as wanted:

It’s difficult. There are no resources for this work. We just have to include it alongside other tasks. On the funny side, nobody ever talks about what we should then stop doing. It’s add on, add on, add on…. The clinicians find energy in this, and they want to make a difference. But it’s demotivating when the resources are lacking. (Local coordinator, QIC diabetes)

The quote highlights how the lack of time and resources for QI work impacts the experienced progression and achievements of QI. Moreover, there is a general call for hierarchical and explicit leadership prioritization from the hospital and regional management of the QIC work regarding decisions on what the professionals should stop doing to have time for the QI work. However, the professionals clearly strive to achieve results for the patients despite how the contextual conditions for the QI work could have been more supportive.

Although most teams reported an aligned understanding of sparse time and resources, the teams differed as to whether prioritization and time for the implementation work had been explicitly discussed or if alignment was more a result of tacit, intuitive processes. Most teams reported having mutual respect despite all team members not necessarily contributing equal amounts of time but with what was possible given the individual workload and schedule. Perceptions of prioritization appeared to be misaligned in only a few teams. A physiotherapist in a QIC fractures team, for instance, explained that when she returned from maternity leave, where her replacement had not prioritized the QIC work as highly, several improvement measures had deteriorated. This exemplifies how lower goal attainment can result when the team composition changes and previously aligned practices are not institutionalized and fail to be renegotiated. A QIC diabetes team nurse provides another example of misalignment of prioritization, explaining how she felt she had been forced to take on extra work because, in her opinion, other team members did not assume their responsibility in the joint improvement project:

I feel I’ve taken on several tasks in our QIC project, because no one else in the team handled them or thought they should, because they weren’t able or willing to do it. The responsibility we take for the project really differs…. Maybe we haven’t sufficiently discussed and defined from the beginning who is responsible for what—and then somebody just ended up taking the lead, because no one else did. (Nurse, QI team, QIC diabetes)

This quote also points to another important aspect of aligned distributed leadership practices regarding who decides and takes the lead on the scope and structure of the QIC work, namely, the practices of collaboration and distribution of tasks within the QI teams, expanded on below.

Collaboration Practices: Task Distribution and Leadership Roles

In relation to the distribution of tasks in the local QIC projects, there generally appeared to be a clear and mutually recognized process for task distribution in the teams—a very democratic process experienced as meaningful for all team members and supportive of the QIC work progression. The process was characterized by practices of intuitive alignment that emerged bottom-up through informal negotiations among the team members. Most teams described how they used their biweekly or monthly QIC meetings as a platform for distributing the different QIC tasks at hand. As the quote below demonstrates, available time and resources as well as competencies worked as guiding principles for the bottom-up distribution of the tasks:

Nurse 1: It [distribution of tasks] is mainly based on where the task is located and who has the competencies or experience from similar work to handle it.

Physician: But it isn’t necessarily related to our professional background—“I do this because I’m a physician, and you do this because you’re a nurse.” We rarely work like that, actually.

Nurse 2: No, it’s more, “Who has the resources to handle the task?” Right? It’s typically something we talk about and agree on in our team meetings.

Nurse 1: Yes, and I actually think that we take the tasks ourselves more often than they’re distributed to us. (QI team, QIC diabetes)

In some cases, the guiding principle of competencies rested on the team members’ professional backgrounds. As the quote illustrates, however, competencies stemmed evenly from experiences with similar tasks in previous projects or concurrent work. Furthermore, characteristic for most teams was that the perceptions of who had the time or competencies generally appeared to be aligned and self-evident in the team and interestingly it seemed to overrule the otherwise profound professional hierarchies between, for instance, doctors and nurses.

However, there were some exceptions from this general pattern. The first examples were the misaligned prioritization of time for the QIC work mentioned at the end of the previous section. Second, the local coordinator in two other teams assumed a more dominant role, distributing tasks to other team members without further deliberation. Still, the team members experienced this as a respectful distribution whereby aligned distributed leadership practices were achieved. However, these exceptions raise awareness that the informally negotiated and emerging distributed leadership practices can be vulnerable configurations. To avoid this pitfall, there was only one team where team members reported having worked out a formal collaboration contract specifying roles and tasks in the QIC work. The following quote illustrates how they experienced that this deliberate aligning of their distributed leadership practices advanced their QIC work:

I think that it has been important for the success of our QIC project that we’ve written down what bothers us and what bothers the others [departments affected by the QIC project]. And then, together with our managers, we’ve worked out a formal collaboration contract specifying how to work together with our project. And we’ve continued to rewrite the contract until everybody was satisfied and we could make a difference. (Physiotherapist, hospital team, QIC fractures)

The quote highlights how alignment must be ensured not only within the QI team but also horizontally into the departments in which the QI work is supposed to result in changed work practices. Furthermore, the quote points to the importance of a formal (clinical) leader taking active part in the QI work, either as a team member or by giving explicit support from the sidelines. This is further substantiated by the following quote, which explicitly links leadership support to the team’s clinical achievements:

In terms of patient treatment, we’ve achieved better quality during the QIC project period because we’ve focused on it and worked with it in the team, but also because we’ve had leadership support for it on the ward. (Physician, hospital team, QIC fractures)

The analysis generally reveals how one team member in each QI team was typically entrusted as team leader. In many of the cases, team members agreed that the local coordinator was the team leader. In some teams, however, the overall leadership responsibility was either (a) assigned to a formal clinical manager participating in the QI team, such as a ward manager; (b) placed with the leader of an existing clinical unit in which the QI team was embedded (mainly the case for QIC diabetes and typically a physician); or (c) the leadership responsibility was divided between the leader of the clinical unit and the local coordinator. Regardless of the specific leadership constellation, in most teams, the decision for who was designated team leader materialized as an emergent practice characterized by intuitive alignment without any formal decision-process. Moreover, in a few teams, nobody would even claim to possess the team leader role; nevertheless, the other team members would all surprisingly point at the same person as the one they recognized as the team leader. This testifies to a distributed leadership practice, which is very informal and tacit yet widely aligned.

Besides being viewed as the team leader, team members usually shared a highly aligned understanding of the local coordinators as facilitators of the local QIC implementation, both in terms of process facilitation and methodological support (largely corresponding to how the formal role descriptions provided by the national and regional levels specified the coordinator’s role). Core tasks involved keeping track of the implementation process and timeline; planning, facilitating, and following up on team meetings and agreements; providing methodological support; ensuring data measurement and monitoring the progress of the QIC project; and acting as a link between the local QIC level, regional coordinators, and national project management. This was also evident in the field notes from the observed local QI team meetings. Despite the highly aligned understanding of the coordinator’s role, the analysis also revealed variation in the enactment of the role across coordinators, particularly in terms of the degree of active involvement in the QI team’s work. Thus, where some coordinators mainly supported the team’s QI work from the sidelines, other coordinators assumed a more active role in the team, acting as a driving force or even as team leader, as described above. This was particularly evident in the more complex QIC fractures teams, where this role distribution regarding the local coordinator also seemed more necessary for experiences of progression in the QI work. Importantly, regardless of the specific role enactment of the coordinator, the understanding of the coordinator’s role came out as mutually recognized and aligned in the teams.

None of the interviewed QI teams considered the coordinator to be responsible for participating in the definition of the local QIC project content, including aims, focus areas, and change initiatives. The analysis revealed how this reflected a distribution of clinical and more administrative leadership responsibilities in the team between the health care professionals and coordinator, as the following quote from a local QIC diabetes coordinator illustrates:

It’s me who ensures that the local goals are defined, but the team decides on the content of the specific goals and defines the change initiative. I can’t contribute to that, as I don’t know what makes sense clinically or what’s realistic…I can help facilitate the process. (Local coordinator, QIC diabetes)

As the quote illustrates, the distribution of clinical and administrative responsibilities was guided by who had the professional competencies for the respective tasks. Generally, this decision on role division appeared emergent and intuitively aligned within the teams without there having been explicit discussions around it. Given the similarities across all teams, the aligned role perceptions and tasks of the local coordinator exemplify a distributed leadership practice that has become institutionalized over time.

Understanding and Deciding on QIC Aims and Methods

As regard the QIC project objectives, the analysis showed that team members within most QI teams agreed on the aims, focus areas, and change initiatives on which to concentrate their local QIC project. For most teams, the agreed understandings of the focus areas and change initiatives came about through a rather informal, unproblematic process. However, the analysis showed how, for a few teams, the alignment stemmed from processes of more deliberate efforts and negotiations:

Leading physician: I clearly remember our first meetings—they were characterized by very low levels of agreement, including a huge discussion between anesthesia, acute medicine, and orthopedics in relation to how we should focus our project….

Physiotherapist: We had some initial meetings where we talked and talked without getting anywhere—frustrating to spend time on. It really helped when we decided to have a manager at the meetings—then I really think we made progress.

Leading physician: I think I realized that we needed a stricter structure and I needed to be present as leading physician. At some point, you just need to decide on a direction instead of discussing it endlessly. I think that made a great difference for us. (Hospital team, QIC fractures)

Other teams also highlighted the importance of such explicit alignment efforts for the achieved results, emphasizing that the alignment of practices in relation to scope and task distribution cannot stand alone—aligning the local aims and priorities constitutes a foundation for the others. The quote also pinpoints how the alignment of aims in this complex QIC fractures team was resolved by the presence of formal managers, who took charge of the decision-making process regarding the local aims of the QIC work. In another instance, ending meetings with pizza was a deliberate alignment strategy for a very large and diverse team within QIC fractures:

It’s the thing about bringing people together for a meeting—put them in the same room, lock the door, and tell them to reach an agreement [about the aims of the local QIC project]. And promise them pizza when they reach agreement [laughs]. It actually worked for us. (Physician, hospital team, QIC fractures)

This reveals different push-and-pull strategies for team members to agree on the aims and action plans of the QIC work where—in these cases—the presence of a formal leader and a potential reward (pizza) acted as catalysts for achieving alignment. This also testifies to the finding that deliberate alignment efforts were more often required for reaching a prioritization of local aims in QIC fractures teams involving more departments, professions, and geographical locations and where team members did not necessarily know each other beforehand.

Finally, as regard the QIC methodology, the analysis showed how, across teams, the processes aligning distributed leadership practices revolved around decisions regarding the degree and type of use of the different tools and methods more than how they were interpreted. Few teams applied the QIC methodology systematically and consistently. More commonly, the teams reported a pragmatic or critical use of QIC methodology, where they carefully selected the methods and tools to apply and how to do so. Regardless of the specific degree and type of use, the teams were characterized by an internally shared understanding emerging out of rather intuitive processes:

We haven’t talked much about it, but I feel we agree that the QIC methodology generally makes sense. At the same time, I think we’ve also agreed that it’s more important to focus on the change initiatives than to be systematic and consistent in the methodology. That we haven’t really been. It has probably mainly been our local coordinator who has held on to “this is a QIC project, we need to make our PDSAs and assess our initiatives.” (Nurse, QI team, QIC diabetes)

The quote further testifies to the fact that in many of the QI teams that applied the methodology more pragmatically or critically, the local coordinator was tasked with following up on the use of the methods; something that appeared to be mutually recognized and highly aligned in the teams.

Discussion

This study has investigated how leadership in local QI teams participating in interorganizational and cross-sector health QICs can be characterized by aligned distributed leadership practices. We studied the relationship between these practices and how the QI team members experienced the QI work progress and achievements. Based on a qualitative case study of two different QICs, our empirical findings demonstrate how leadership in local QI teams are characterized by aligned distributed leadership, with activities being widely distributed based on emergent and negotiated perceptions of the roles, scope, and desired results of the QIC work. We summarize our findings in the following and position them in relation to the existing research on the implementation and leadership of QICs and distributed leadership in public service delivery more generally. Furthermore, we consider the methodological strengths and limitations of our study as well as practice implications.

Emergent and Widely Aligned Distributed Leadership Practices

In relation to the scope and structure of the teamwork, the decisions on meeting structure, team composition, and prioritization of the QIC work, which are typically formal managerial decisions, were distributed within the team and continuously renegotiated and adapted throughout the QIC implementation process. Where possible, QIC meetings were embedded in existing meeting structures and work processes, and team composition mainly relied on ensuring that complementary skills and knowledge were present in the team. The latter resulted in the complex QIC fractures teams generally being much larger and experiencing more difficulties in establishing aligned practices than in the less complex QIC diabetes teams—although outright conflict was not an issue. A similar pattern of emergent and negotiated practices was identified in relation to the collaboration practices, task distribution and leadership roles within the teams, and the understanding and decisions on the aims and methods of the QIC work. Although rarely formal or explicitly discussed, there was general agreement within the teams regarding the task and role distribution and decisions on where the QI work should be heading. In line with existing studies of distributed leadership and service improvements in health care (^{Boak et al., 2015}; ^{Fitzgerald et al., 2013}; ^{Jakobsen et al., 2021}), such aligned distributed leadership practices were identified as important determinants for experiences of progression in the QIC work and directly linked by the informants to the achievement of desired results.

Despite the widely aligned distributed leadership practices in relation to the daily QI work, our analysis also revealed how the very bottom-up, emergent, and often tacit distribution within the teams constitutes a vulnerable practice. In some instances, we identified misaligned practices in relation to lacking responsibility for designated QI tasks, missing the renegotiation of role allocations after a change in the team composition and difficulties in agreeing on a course of action for the QI work (often also because of a lack of time to prioritize the work). The interviewed team members consistently linked these instances of misaligned distributed leadership to experiences of less progression in the QI work and deficiencies regarding the formally stated goals and locally decided priorities of the QIC project. These findings are in line with ^{Leithwood et al. (2007)}, who point out that emergent practices of distributed leadership have great potential for positive, short-term changes, but long-term, sustainable changes require deliberate alignment efforts, which were rarely present in our cases. For instance, in our case, there was only one QIC fractures team in which they had worked out a collaboration contract to avoid some of these misalignment pitfalls and ensure legitimacy and eased implementation of QI initiatives in horizontal collaborations across wards and sectors. Although a formalized contract can also be rigid and prevent adaptations to changed circumstances (in contrast to the flexibility typically highlighted as an advantage of a distributed leadership approach), such deliberate alignment efforts are well in line with ^{Gronn’s (2002)} emphasis on institutionalized distribution of leadership. Caution toward the sustainability of QIs via emergent distributed leadership practices is therefore warranted, and this calls for further research into the advantages/disadvantages of formalizing distribution of leadership.

In line with these notions, another main finding in our study was that attempts to establish aligned distributed leadership practices were challenged and had to be handled differently in the larger and more complex QIC fractures teams spanning sectors, professions, and locations compared to the less complex QIC diabetes teams. Although not a consistent result, these teams often experienced a meandering path toward achieving their QI priorities. At the same time, however, they also experienced more informal gains of distributed leadership in the interorganizational QIC work in terms of improved communication, awareness of complementary competencies across departments and sectors, and joint understandings of patients and their care pathways. This duality was also highlighted in a recent study of QIC implementation by ^{Williams et al. (2022)}, who concluded that “There are obvious benefits in creating a team from a wide constituency as they would bring different perspectives and networks to the team, but this can delay the norming, storming and performing elements” (p. 1000). As such, our results contrast previous findings emphasizing that QI work can constitute an arena for interprofessional conflicts between, for instance, doctors and nurses (^{Eriksson et al., 2016}). Rather our results illustrate how distributed leadership practices can help overcome such barriers of professional hierarchies—particularly when time and resource constraints also spur team members to pool their QI relevant competencies.

Other scholars evaluating the implementation of QICs have also brought up the positive influences of QIC participation on interprofessional collaboration (^{de Silva, 2014}; ^{Kotecha et al., 2015}; ^{Williams et al., 2022}). For example, in their study of the influence of a QIC program on team functioning in primary health care, ^{Kotecha et al. (2015)} found that QIC participation provided the opportunity to improve communication, trust, and promote collaboration across involved disciplines, as the QIC structure provided opportunities to work together and learn from each other. Drawing on these findings and the results of our study, we add that besides the structures provided by the QICs, the widespread distributed leadership practices within the QI teams constitute an important facilitator of these informal gains by allowing for relevant competences to be put into play whenever it was needed to advance the teamwork. Taken together, our results regarding the contextual impact of the complexity of QIC fractures therefore testify to the paradox of how distributed leadership can simultaneously be regarded as a necessity and an impossibility when dealing with complex public service delivery (^{Currie & Lockett, 2011}).

The Roles and Importance of Formal Managers and Local Coordinators

To unlock this Catch-22 (^{Currie et al., 2009}), our analysis in several instances identified the presence of formal managers in the QI team (or in close proximity to the team) and a hierarchical anchoring of the QIC project as important factors for experienced alignment of distributed leadership and progression of the QI work. Particularly in relation to the prioritization of the QIC work in terms of time and resources, the team membership of formal managers sped up the process of establishing aligned practices. This happened without compromising distribution of leadership within the team as these participating formal managers were rarely regarded leaders of the QI teamwork, nor did they interfere much with the content of the QI work. Even more importantly, our analysis also revealed how the active participation of formal managers in the QI work paved the way for horizontal alignment with other departments as well as increased legitimacy and eased access to implement QI initiatives in the involved departments/organizational units. The active participation of formal managers in the distributed leadership practices of the QI teams thus became directly linked to the teams’ opportunities to achieve the desired improvements.

These results correspond with the findings of existing studies in the QIC literature pointing to the positive association between supportive hierarchical leadership and improved QIC outcomes (e.g., ^{de Silva, 2014}; ^{Lowther et al., 2021}; ^{Williams et al., 2022}; ^{Zamboni et al., 2020}). Likewise, few studies of distributed leadership in health care have directly or indirectly explored the support of hierarchical and formal leadership (e.g., ^{Günzel-Jensen et al., 2018}; ^{Jonasson et al., 2018}; ^{McKee et al., 2013}). ^{Günzel-Jensen et al. (2018)}, for instance, found that both transformational, transactional, and empowering leadership exerted by ward managers positively impacted distributed leadership agency among the clinical staff during a hospital merger. However, whereas these studies largely examine the role of formal leaders detached from the distributed leadership practices, our study contributes with insights into how the managers directly engage in the distributed leadership practices—most notably by consolidating the emergent aligned practices within the teams and providing leeway for horizontal alignment across departments and sectors. Besides contributing positively to the QI achievements, the presence of formal leaders in close proximity to the local QI teams hereby also seemed to remedy concerns of lacking hierarchical accountability and legitimacy of the team’s work—factors that are often mentioned as disadvantages of a distributed leadership approach (^{Jakobsen et al., 2021}).

However, the role of formal managers in the distributed leadership practices cannot in itself explain the progression and achievements of the local QI teams. Across teams, a pivotal role was ascribed to the local coordinators. The specifics of how this role was agreed upon within the team and exerted by the coordinators proved decisive for the functioning and alignment of the distributed leadership practices within teams. Across QI teams, the analysis revealed a highly emergent and intuitively aligned understanding of the local coordinators as process and methodological facilitators of the local QIC implementation, yet with variation across coordinators regarding their active involvement in the QI team’s work. Generally, the coordinators assumed a more active role in the more complex QIC fractures teams than in the QIC diabetes teams, and the role of the local coordinator also seemed more necessary for experiences of progression in the QI work in these teams. Conversely, the absence of formal managers in some teams also unexpectedly seemed to weaken the mandate and impact of the local coordinators on the QI work. This observation underlines a duality in the coordinator role as being a focal point for distributed leadership to advance the local QIC projects while at the same time struggling to occupy a strong role unless formal managers also participate in the team or provide direct support on the side. This adds an important nuance to existing studies emphasizing the role of QI specialists as local coordinators and change agents in advancing the local QI work in QICs (^{Carstensen et al., 2023}; ^{Dückers et al., 2009}). Furthermore, it underlines how the emergence of distributed leadership in health care is contextually conditioned and often hindered by the complexity of professional and policy institutions (^{Currie & Lockett, 2011}) and preexisting hierarchies (^{McKee et al., 2013}).

However, a downside of the (active) participation of the local coordinator in the QI team as crucial for the functioning of aligned distributed leadership practices and the progression of the team’s QI work is the potential missing sustainability of the QIC outcomes and methodology in clinical practice when the QIC project, and therefore the support of the local coordinators, end. Still, the analysis also showed that the health care professionals generally engage in the QIC implementation and have a strong commitment to the local QI work despite the (sometimes) difficult alignment attempts and general lack of resources. We encourage future research to further examine the relationships between the participation of local coordinators in local QI teams and the sustainability of the QIC results and activities.

Study Limitations and Methodological Considerations

To evaluate the findings of the present study, some methodological considerations must be taken into account. First, we based our analysis on a qualitative case study, and the potential for transferring the results to other settings must therefore be considered. However, by providing a detailed description of the QICs alongside contextual information, the relevance and applicability of the results should be assessable. Moreover, for the use of two complementary cases, the two QICs, our results span across very diverse settings, allowing for an awareness of the more specific contextual conditions when inferring our findings. Second, our study reported on the relationship between the QI teams’ distributed leadership practices and their experienced progress and achievements of the QI work. However, we did not collect quantitative data to assess the relationship with the actual progress and achievements of the QI teams’ work. Collection of such quantitative data constitutes an important area for future studies. Third, an important point of awareness in terms of credibility of the reported findings is whether the participants have been able to speak openly in the group interview setting. Given that some of the displayed quotes can be interpreted as controversial (for instance, the disclosure of serving pizza as a strategic alignment tool), this concern does not seem to have affected our results. Moreover, we have no indications that the varying number of informants in the group interviews influenced our results given that the content coding focused on statements regarding experiences of aligned distributed leadership (rather than coding their social interactions as in typical focus groups; ^{Halkier, 2010}). Finally, a key strength of the present study is the comprehensible and triangulated data, combining interviews, participant observation, and documents, which provided a thorough basis for the analysis and for addressing the research aims. This is particularly in relation to providing accurate and nuanced insights into how QIC leadership unfolds in practice.

Practice Implications

By integrating empirical insights from a qualitative case study of the use of QICs for QI in health care with the literature on distributed leadership in public service delivery, this study has answered the call for a more nuanced understanding of the role of leadership in QIC implementation. The findings point to practice implications in relation to the enactment of distributed leadership in cross-professional and interorganizational/sector settings as well as the future planning and implementation of QICs.

First, the findings indicate that aligned distributed leadership practices are important for the experienced progression and achievement of desired results of the local QI work. Aligned distributed leadership practices ensure more agile QIC implementation processes and that the right competencies are continuously put into play in the QI work. The analysis points to a particular importance of aligned distributed leadership practices in relation to the (a) scope and structure of teamwork, (b) collaboration practices and task and leadership role distribution, and (c) understanding and deciding on QIC aims and methods. However, the findings also reveal how very bottom-up negotiated and emergent distributed leadership practices can constitute a vulnerable practice with clear drawbacks: lacking robustness, misalignments, and less progression and deficiencies in the achievement of results of the QI work. Thus, such emergent practices should be balanced by a more thorough and timely consolidation of the tacit distributed leadership practices to provide the best circumstances for ensuring robust and sustainable QI.

Second, our analysis shows that aligned distributed leadership practices within the local QI teams are not enough in themselves to experience progress and to achieve the desired results of the QI work. For the QIC project to succeed, these practices must be supplemented with the active participation of formal managers and local coordinators in the QI work, which should therefore be a clear investment priority of hospital/municipal managers. The analysis suggests a particular importance of the active participation of formal managers in relation to aligned practices of prioritization of the QI work and in creating legitimacy in the involved departments and horizontal alignment with other departments. Importantly, formal managers also play a crucial role in paving the way for the local coordinators to take on their role in advancing the more deliberate distributed leadership practices and QIC implementation and to ensure the sustainability of QI competencies once the local coordinator withdraws at the QIC termination. Furthermore, formal management support, for instance from the top hospital management, is also needed to ensure QI work is actually prioritized and appropriate amounts of resources are set aside so that these types of tasks do not just become an expected add-on crowding out time for regular work tasks.

Third and finally, the findings suggest that the establishment of aligned distributed leadership practices and achievement of QI priorities is more challenging in complex QICs, where QI teams span across departments, professions, geographical locations, and sectors. The possibility for embedding the QI work in existing hierarchies, organizational structures, and collaborations appears to facilitate the institutionalization of the distributed leadership practices, the organizational anchoring, and the success of the QIC project. This is more easily achieved in less complex QI teams. Thus, to increase the probability of long-term success and the sustainability of the QI work within complex QICs, hierarchical anchoring and support become decisive.

Acknowledgments

The authors thank Charlotte Jonasson, Aarhus University, and Gianluca Veronesi, University of Bristol and Aarhus University, for valuable comments to previous versions of the article.

References

Supplemental Digital Content

• HCMR_00_00_2023_09_15_CARSTENSEN_HCMR-23-031_SDC1.pdf; [PDF] (324 KB)
• HCMR_00_00_2023_09_15_CARSTENSEN_HCMR-23-031_SDC2.pdf; [PDF] (42 KB)
• HCMR_00_00_2023_09_15_CARSTENSEN_HCMR-23-031_SDC3.pdf; [PDF] (138 KB)
• HCMR_00_00_2023_09_15_CARSTENSEN_HCMR-23-031_SDC4.pdf; [PDF] (134 KB)