Working together in emergency care? How professional boundaries influence integration efforts and operational performance

Patient flows in emergency care

Our focus in this research is on a specific process and patient flow within the hospital, a patient's journey through the ED. In this patient journey, there are several process steps that a patient must go through to be assessed, diagnosed, receive interventions and, ultimately, be discharged from the ED (see Nugus et al., 2014, for a detailed depiction of this non-linear process). Discharge can occur either via admission to the hospital, discharge home or to an external provider (see Figure 1 for flow chart). Importantly, there are several key decision moments throughout this process when professionals from different specialties, departments and functions must integrate rapidly to determine the trajectory of a patient's journey through and, ultimately, out of the ED. As patient care becomes increasingly complex, and patients suffer from multiple morbidities, the process becomes even more complicated (e.g. more decision-making moments, ambiguity and need for consultation) and time consuming (see also Mintzberg, 1993) as the prevalence of interdisciplinary interfaces increases. However, literature has shown that the involvement of multiple specialists and delays in consultations and slow decision-making are key factors in long lead times (Driesen et al., 2018; Vegting et al., 2015).

The particular challenge of these interdisciplinary interfaces in emergency care is that they are often ad hoc and dependent upon specific patient needs (pathology, comorbidities and emergent symptoms) and specialist's availability. Ad hoc teams are characterized by “rapid formation, an abbreviated lifespan and often limited experience working together previously” (Weaver et al., 2014, p. 363), which are factors that can create barriers to integration. Complications at these interfaces can create delays that impact lead times (e.g. by increasing time to treatment or consultant signoff) and can contribute to crowding (Morley et al., 2018). While literature on professional service operations emphasizes the importance of knowledge exchanges and collaboration between actors (Lewis and Brown, 2012; Ellram et al., 2004), our theorizing often falls short of considering the unique nature of interprofessional interactions and interfaces, leaving us “blind to leverage points where the right pressure, adroitly applied at the right time and place, could bring about desired change” (Harvey et al., 2016, p. 7). In such knowledge-intensive operational processes, it is, therefore, important that we give more attention to this level of interaction (see also Frangeskou et al., 2020), as it is precisely these “human” processes that contribute to the level of complexity involved (Ellram et al., 2004).

A focus on professional integration

Research has shown that integration between professionals is particularly difficult to achieve due to specialization and organizational compartmentalization (Ferlie et al., 2005; Hewett et al., 2009; Wright et al., 2017), which can create “professional inward-directedness” (Saltman et al., 2006) that restricts integration. Hospital organizations remain, for the most part, functionally and structurally differentiated, forming “pragmatic boundaries” (Carlile, 2002) between professional groups and sub-disciplines (specialties). Differentiation and specialization function to provide a valuable division of labor and can promote efficiency (Lillrank, 2012), yet they also encourage the creation and reinforcement of boundaries between professionals and require integrative strategies to overcome fragmentation, break down functional barriers (Zhao et al., 2011) and optimize flow (Drupsteen et al., 2013).

In this study, we focus on internal integration between medical professionals, defining internal integration in this context as “the process of interaction and collaboration in which medical professionals from different disciplines work together in a cooperative manner to arrive at mutually acceptable outcomes for their healthcare organization” (adapted from Pagell, 2004). Following Leuschner et al. (2013), we view integration as multi-dimensional, requiring the fostering of trusting relationships (relational integration), sharing of information and knowledge (information integration) and joint working (operational integration). We apply these dimensions to the level of individual interactions between professionals, whereas they are normally studied at the level of departments (e.g. purchasing, manufacturing) or organizations (e.g. external). While each dimension is often tested separately in the literature, scholars have argued that all dimensions are highly interrelated and play a key part in determining performance (Leuschner et al., 2013) and should, therefore, be considered together (Van der Vaart and van Donk, 2008). For instance, studies have shown that the interpersonal elements are a key part of ensuring information flow (Power, 2005), and that communication and clarity about roles and responsibilities enable joint coordination of actions and decisions (Meijboom et al., 2011).

In knowledge-intense interactions, professionals work jointly at different parts of the process and in different ways, to share their tacit knowledge and expertise. The relational elements in such interactions can play an important role in speeding up flow, or in the willingness to share information and expertise. In healthcare, professionals are interdependent and rely on one another to support the patient journey (Senot et al., 2016). Integration in healthcare is critical because poor integration across providers can lead to delays, resulting in poor quality and unsafe care (Meijboom et al., 2011), and research shows that as a result of poor coordination, medical errors are higher for patients who are treated by multiple doctors (Schoen et al., 2007). The necessity and importance of interdisciplinary consultation is only increasing as patient care has become more complex (e.g. patients with co- or multi-morbidities) (Nolte and McKee, 2008), and lead times for these patients increase (Sampalli et al., 2015), making integration essential. Patient complexity creates situations of ambiguity, e.g. what specialty the patient belongs to, or can create the need for more consultation or diagnostic use to determine a course of action for patients, increasing the potential for delays and requiring a high level of integration across all dimensions of integration. To deliver safe, efficient and effective care, medical professionals need to communicate and share information about patients' symptoms and diagnosis (information integration) (Hewett et al., 2009), coordinate care processes and make joint decisions about treatment (e.g. whether to admit or discharge a patient) (operational integration) and foster and maintain relationship attributes like trust and commitment (relational integration) (Leuschner et al., 2013; see Table 1).

Challenging integration and performance: the role of professional boundaries

Despite the importance of integration in healthcare processes, the current division of labor and organization of care delivery continues to form and reinforce boundaries between groups (Lillrank, 2012; Carlile, 2002). “Boundaries” are defined as demarcations that help to distinguish groups from one another (Comeau-vallee and Langley, 2019) and act as tools “by which individuals and groups struggle over and come to agree upon definitions of reality” (Lamont and Molnar, 2002, p. 168). However, boundaries are not self-sustaining; they exist within and emerge from interactions (Giddens, 1984) and are supported by organizational and individual efforts (Comeau-Vallee and Langley, 2019). Prior empirical work has shown that medical professionals experience and work to establish, maintain and defend professional boundaries as part of daily work (Nancarrow and Borthwick, 2005; Powell and Davies, 2012).

In practice, boundaries are multilayered, and may result from [and be reinforced by] status differences, socialization processes (Finn, 2008), functional differentiation, diverse approaches to work and localized norms and values (Hewett et al., 2009). For example, a large body of OM literature has focused on the challenges of internal integration between functional departments such as purchasing, manufacturing, operations and sales (e.g. Ellegaard and Koch, 2012) due to boundaries between these groups. In healthcare, medical professionals undergo a long educational and socialization process that institutionalizes divisions and hierarchical relations between roles (Finn, 2008) and within professional groups (Powell and Davies, 2012). These divisions are further enforced at the organizational level as physicians are spatially separated and organized into specialty departments, working in specialty silos (Spee et al., 2016), with limited coordination between providers (Ferlie et al., 2005; Thijssen et al., 2013). Such a scenario can create significant barriers to improving internal integration as “the tendency is towards conflict and contestation to the detriment of professional integration” (Finn, 2008, p. 108).

To manage cross-boundary connections, actors may engage in attempts to “close” (e.g. reinforce distinctions) or “open” (e.g. expand, redraw) boundaries (Cregard, 2018), inhibiting or enabling integration. Such shifts are most likely to occur when prompted through organizational change (Hernes, 2004). This makes our current context an interesting case in which to view how boundaries come to be negotiated, shifted or created as the result of efforts to improve care processes, and to examine the resilience of boundaries at important interfaces in the patient care process. To demonstrate how interprofessional dynamics can influence important operational outcomes, we explore the influence of professional boundaries on integration and operational performance, defined here as flow and speed.

In acute settings, performance concerns the speed in which a patient is seen, assessed and a treatment pathway is determined and executed. Integration plays a key role in improving care processes and may lead to shorter lead times and better service to the customer (Power, 2005), better care and improved patient flow (Meijboom et al., 2011). Yet, the task of integrating disparate professionals remains an ongoing challenge for healthcare systems worldwide, and there remains a lack of attention to the individual level in the general OM literature (Dai and Tayur, 2020). We suggest that a focus on the factors at the professional level that impede or support integration (i.e. boundaries) may help us to offer insights that may be overlooked, with a focus at the level of firms or functions, and allow us to attend to the unique challenges facing integration in professional service contexts.

Case selection and setting

We purposively selected our case organization (Stake, 1995) based on its ongoing efforts to improve acute care performance and a high demand for services. Haaglanden Medical Center (HMC) is an inner-city, Level 1 trauma center in The Netherlands. The ED at HMC is one of the busiest in The Netherlands, with over 50,000 patients visiting annually and a 20% admission rate (van der Linden et al., 2018). During the study period, HMC was trialing a six-month reconfiguration (October 2017–March 2018) to co-locate medical specialists from five key specialties (cardiology, neurology, surgery, internal medicine, radiology) to the ED. While the reconfiguration is not the focus of the present study, this context provided a rare opportunity to examine how structural changes influence integration over time. During the trial period, specialists were physically located in the ED during peak hours (12–8 pm) and worked alongside emergency physicians (EPs) and residents.

Our unit of analysis is professional dyads involved in emergency care delivery (i.e. the interactions between professionals), although in a final stage of analysis, we also aggregate findings to consider insights and make comparisons at the group level (i.e. medical specialists versus EPs, specialty groups). We were provided with a list of internal contacts from the fourth author and management involved with the ED interventions. Contacts were invited in the first and third round via email, and in the second round, in addition to email invitations, a researcher (first author) was onsite for observations and introduced the morning meeting held between specialists and EPs. In place of scheduling, physicians could stop by the researcher's designated room on the ED for interviews at a convenient moment before, during or after a shift, reducing the burden on respondents. Specialists were selected to represent a variety of opinions toward ED involvement and level of experience, and we only sampled specialists from specialties who have frequent interactions with and responsibility for the ED and ED patients. We sampled from the entire pool of EPs (N = 12) to get a variety of views, and we selected respondents across levels of experience, age and gender. The study was granted ethical approval by the internal review board of the faculty of redacted for blind peer review.

Data collection

Interviews were conducted at three time points (at the beginning, during and after the colocation trial period) to see if and how professional dynamics and boundaries shifted over time and in response to structural change. Two rounds of interviews were conducted during the trial period from November 2017–March 2018, and follow-up interviews with a small subset of participants were conducted in October 2019. In total, 28 in-depth interviews were conducted onsite and were recorded, transcribed and coded by two researchers (see Table 2 for a list of interviewees). In total, 25 interviews were conducted with EPs and specialists from the key specialties that interact most frequently with the ED. One emergency nurse practitioner and two managers were also interviewed for context. Interviews lasted between 30 and 60 min, and respondents gave verbal permission to be audio recorded. Respondents voluntarily engaged in interviews and provided informed consent. Respondents were informed that identifying information would be anonymized, so they could speak openly.

In line with an inductive approach, we used a semi-structured interview guide. We asked respondents about their work and collaboration with other professionals in the ED, their responsibility for emergency care and to reflect upon care delivery processes. We also gave respondents space to reflect upon how they felt collaboration functioned before, during and after the reconfiguration, and discussed ongoing challenges and suggestions for improvement. Key issues discussed included the sharing of knowledge (e.g. consults, supervision), making diagnostic and treatment decisions, and relational dynamics between specialists and between specialists and EPs, such as the quality of their interactions and collaborations. In the final part of interviews (if it had not emerged naturally), we asked respondents to comment on how their collaboration with other professionals impacted performance specific to the ED, namely, patient flow and speed.

In addition to interviews, a day of observation with staff was conducted in the first phase of interviews, and a week of observation was conducted at the ED by the principal researcher during the second phase of interviews. Observations provided us with a visual representation and map of the site, and provided insights into the layout and daily functioning of the ED for better context. We were able to observe the pace of flow at the ED, the location of physicians throughout shifts and communication exchanges. This helped to inform interviews and provided a cross-check for findings. For example, it was observed that while medical specialists had a dedicated room, EPs were often sat in a separate back room in their own discussions. This showed that while co-location put specialists on the ED, simply placing them there did not guarantee increased interaction. Additional archival data sources were utilized for triangulation and to inform our interview guide. Sources reviewed included news articles, reports concerning the pilot and intervention outcomes, professional association reports, EP curriculum, governmental reports and regulatory monitors.

Analysis

We employed an inductive analysis approach (Gioia et al., 2012), and the analysis was iterative throughout the research process. As data were collected, ideas and themes were identified and refined through group discussion in the research team and based on feedback from organizational insiders. In line with our interview approach, three distinct phases of analysis can be determined. During the first phase of interviews, analysis was largely focused on deepening our understanding of the dynamics and interfaces in the emergency care process, identifying types and forms of integration and collaboration, and gaining insights and reflections of professionals on their collaboration and its effects on care processes and outcomes (e.g. lead times). Questions driving this first phase of analysis (during the first ten interviews) included what are the barriers to integration? How do groups work together? How does the presence of medical specialists influence collaboration? What impact does collaboration have on outcomes such as patient flow and speed?

Codes began largely descriptive, as we mapped out the organizational context and gained insights into the dynamics and nuances of the care process itself. Summaries of each interview and of emergent themes were shared and evaluated between the first interviewer in Stage 1 and the first author. Ongoing discussions of emergent themes and codes continued throughout data collection to abstract codes from a descriptive level (barriers to integration, work processes, roles) to a second level of abstraction, identifying categories of barriers such as relational aspects (e.g. familiarity, trust) and cultural aspects (e.g. approach to work, educational background) and to begin connecting between the level of individual interactions and outcomes, such as patient flow. After we completed and analyzed the initial round of interviews (January 2018), a data summary was shared with the organization, and a meeting was arranged with the ED management, a medical manager and the first and fourth author to discuss emergent themes. The emergent themes were congruent with what organizational members had seen “on the ground,” and we agreed collectively to carry on with the research without modifying the interview protocol.

As we carried on interviewing in phase two, and participants had more experience with co-location, the boundaries between groups became clearer. For example, we recognized that what we had previously categorized as change resistance was actually the presence of more persistent boundaries that existed between groups. We began to trace and see patterns across our former codes that indicated several types of boundaries. We then revisited the literature to consider how our a priori constructs fit the emergent themes, revising our approach as needed. Ultimately, what we saw was that our findings were in line with the literature on the antecedents for integration, but that the emphasis on relational and cultural aspects was much greater and more nuanced than we found in the literature. It emerged that respondents perceived these aspects as exhibiting a strong influence on the operational performance, for example, by increasing time to treatment or stalling discharge decisions.

Building on our findings in the first phase, we saw that there was a clear divide between EPs and medical specialists, and therefore, we incorporated a constant comparison between these groups into our analysis. In this second stage, codes from Phase 1 were merged and grouped into higher levels to represent types of professional boundaries as we demonstrate here. To address our second research question, we then revisited how each boundary was perceived to influence integration efforts and patient flow (Table 4) and essentially “worked backwards” from demonstrating the existence of boundaries to exploring the effects. Following the completion of the second phase of interviews (March 2018) and analysis, we again fed back our findings to the organization via an informal report and management summary. Feedback from organizational insiders highlighted the accuracy and novelty of our findings from an organizational perspective.

After being informed that co-location would become standard practice, follow-up interviews were conducted with a small subset of participants (N = 6) 18 months later, in October 2019, to see how dynamics had shifted over time. We used these interviews to probe respondents on how boundaries may have shifted and in regards to perceived process improvement. In this final phase of analysis, we first coded the six follow-up interviews independently, using the existing coding structure as well as open codes, and then paired them down and compared with findings from phase two. The follow-up interviews offered us the ability to see how boundaries shifted (or not) over time and helped to account for any potential “growing pains” that may have come through in the early phases of reconfiguration, providing further validation of our initial findings (see Table 3 for a coding tree). Archival sources were further integrated during this final stage of analysis to further validate and prompt critical reflection of our coding structure.

Emergency care in The Netherlands

ED is a hospital department designed to treat patients with acute and emergent illnesses. ED is open 24 h a day for patients with acute complaints and works to stabilize, treat and discharge patients in a rapid manner. Upon arrival, patients are triaged by a nurse, assigned to a specialty, then seen by a designated specialty resident and/or an EP who will consult a specialist as needed (Figure 1). Unlike other wards, which are typically organized by specialty, the ED is a shared space in which patients for all specialties enter the hospital, and for which, all specialties maintain [partial] responsibility. While ED staff triage, assess and sometimes treat patients, all patients ultimately “belong” to a specialty, and assessment, treatment and discharge decisions require close collaboration between the ED and specialty department. Once patients are assessed and treated, they are either discharged home, referred to another care facility or admitted as an inpatient into the hospital.

Organization of emergency care delivery

In our study, a reconfiguration was happening, which placed specialists on site at the ED during peak hours. Therefore, in our study, we are able to assess integration challenges based on two alternative configurations of care: (1) “the distance model” and (2) “co-location model.” In the distance model, EPs and specialty residents staff the ED, and medical specialists (senior doctors) are on call for consults. In the co-location model, medical specialists physically staff the ED during peak hours as standard practice; however, outside of peak hours, the distance model resumes, and remains the standard across the country. Therefore, it is important to assess how collaboration can function in both situations and to uncover inherent integration challenges across both configurations to highlight suggestions for best practice. Theoretically, it is interesting to observe how the removal of structural boundaries [via co-location] may alleviate, or create new challenges to integration.

The distance model: structural boundaries

The distance model: interpersonal boundaries

Trust and mutual understanding

Interviewees named trust as a core element in improving collaboration and emergency care. In the relationship between EPs and specialists, some respondents cited a lack of trust as a reason for delays in ED processes. This lack of trust also extended to the realities of the daily work and demands on the ED. One EP highlights how specialists question if they are actually needed to support at the ED when they are asked to come by Eps:

Yes, the availability was a barrier. We experienced a number of times that we call a specialist to say “well it is busy right now for your specialism, you should come and help”. The first thing they would do was to call their own resident to ask if it was busy. Come on man, he is busy. He barely knows that he is alive, so what would he know about his [patient] population? He does not. So, those are things we really hope are changing by this project. – EP

Because EPs have a shorter educational program and are seen as “non-specialists,” they occupy a lower status than specialists. Their expertise is valued by specialists for having a general overview of the patient and the performance of certain tasks such as stabilization, but specialists continued to defend knowledge boundaries to avoid any role blurring and restrict expansion of the EP role into their work domains. This restricts relational integration and inhibits joint working as specialists reinforce hierarchical control. In addition, the introduction of EPs was framed as a form of delegation rather than the creation of a separate domain of practice. Specialists expected EPs to carry out the work they were unable, unwilling or too busy to perform, but specialists maintain end [and overall] responsibility and jurisdiction.

There were concerns expressed by specialists about the competencies of EPs to treat patients beyond basic care and determine an appropriate path for the patient, e.g. making discharge or admission decisions. This also relates to end-responsibility of specialists, who take on liability for patient safety. For EPs, this lack of discretion was a source of tension and frustration, especially when coupled with low availability or reachability of specialists. EPs expressed that, in some cases, they should be able to make admission decisions and their judgment should be trusted. Eps’ lack of discretionary power can lead to long delays, particularly for groups with limited availability:

I'm all done with my work and I just need to get the “amen” from the gastroenterologists and that can take hours. – EP

This lack of discretion is a result of the formal system that does not grant discretion to EPs to manage their own patients. When specialists placed trust in EPs, there was more flexibility, and informal systems were created to enable quicker decision-making and improve outflow. Informal agreements prove to be an efficient workaround of rigidity in the system, and allow increased discretion of EPs. Such agreements were fostered by relational work, and in spite of structural constraints and cultural boundaries at a more macro level. Yet, workarounds also introduce variability across specialists and specialty groups, creating discrepancies in shared understandings of roles and responsibilities more globally.

The distance model: cultural boundaries

The co-location model: exploring the impact of care organization

It was clear that by putting specialists directly on the ED, capacity increased and the potential for significant consultation delays decreased. By having specialists physically present on the ED, the sequential chain is reduced, improving information and operational integration. Additionally, members of both groups expressed that interpersonal relationships grew stronger, and face-to-face contact enabled more trust and understanding, particularly between specialists and EPs. However, in the new model of care, roles and responsibilities were not always immediately clear. In particular, the role of the EP became more ambiguous and debated, unearthing tensions between medical specialists and EPs. In particular, cultural issues were heightened due to misaligned goals and working culture. Continued investment in devising new protocols and establishing new ways of working are needed to overcome current boundaries.

The co-location model: structural boundaries

The co-location model: interpersonal boundaries

The co-location model: cultural boundaries

Reinforcing roles and status differences

With co-location, tensions regarding the EP role and its future development became unearthed, leading to uncertainty on both sides about the future. Specialists highlighted the expansion of EPs to supervisory roles, taking on resident education and “moving to the back office” (increased administrative and supervisory roles). They viewed this trend as undermining the EP's primary role to see patients for specialists. This tension was further emphasized in the move toward co-location:

Also, for them [EPs} it is a culture change. They wanted, their ultimate [goal] is their own department. The big chief from the ED and that they can say “that patient is for the surgeon, that patient is for internal medicine” That is the way it works in the US and other countries, but not in Holland at the moment and not in the coming 10–20 years. That is always a question of disturbance. – Intensivist

In interviews, EPs denied any claims that they wish to take over the ED. Some EPs even spoke directly of a “closed ED” format (used in the USA and Australia) as an especially negative development that would restrict the necessary collaboration and specialist view needed for complex patients. Yet, such tensions and statements highlight that there are unclear and potentially mismatched expectations about the role and future of the EP.

A focus on professional integration: unveiling unique barriers and antecedents

With this research, we have shown that professional integration necessitates an increased focus on improving attitudes (see Van der Vaart and van Donk, 2008) and breaking down interpersonal barriers between professionals (Shaw et al., 2017) to achieve high-quality collaboration among professionals that are interdependent but reside across boundaries. A focus on professional interactions, therefore, requires us to expand our current conception of integration (Leuschner et al., 2013), e.g. expanding our current understanding of relational integration to more seriously consider the antecedents to trust development in particular supply contexts. Such antecedents were identified in our study as the need for familiarity between professionals and understanding of each other's skills, roles and work processes. Additionally, considerations of power, status and other interpersonal dynamics are needed to open up our theorizing (Harvey, 2016) and better understand how integration can be fostered or inhibited [also at higher levels] within these interactions. Considering the central role and influence of professional interactions in service operations contexts and healthcare in particular (Lewis and Brown, 2012), it is essential that we begin to consider antecedents of all aspects of integration at lower levels of analysis.

As we have shown, structural, interpersonal and cultural boundaries inhibit collaboration and integration in various ways, including restricting the trust and commitment (relational integration) needed for successful joint working (operational integration) and the availability of expertise and sharing of knowledge (information integration). Boundaries result in different clinical views and approaches to patient care (restricting relational integration) but also lead to particular views about teamwork itself “that must be negotiated through the course of work” (Finn, 2008, p. 108). To work together and achieve integration, medical professionals must constantly negotiate boundaries in their everyday work. Ongoing and historical functional differentiation forms chasms across specialties or departments in terms of their mindset, functions, approach to time (Lawrence and Lorsch, 1967; Jansen et al., 2009) and create “pragmatic boundaries” (Carlile, 2002). This creates clear barriers to cross-functional work, joint working and goal alignment, as each group comes to operate independently within their own cultural and clinical framework over time. However, as we saw in our case, once professionals begin to intermingle and work together outside their subunits, interpersonal boundaries are reduced and integration becomes more likely and performance can improve. Co-location, therefore, provides a good basis for fostering the relational and cultural change needed to truly improve integration.

Through first removing structural constraints and reducing thresholds for interaction, professionals gain exposure to different workflows and processes, and observe others' work and skill. This exposure and interaction, as result of the co-location, may encourage professionals to begin to shift their mindsets, e.g. toward other groups (specialists viewing EPs work as valuable) and processes (gaining awareness of the external issues facing the ED and the importance of urgency). This can help to build both goodwill (EPs gaining confidence in specialists' commitment to ED processes and improvement and working together) and competence-based trust from specialists toward EPs (Sako, 1992) and even potentially shift role boundaries (Niezen and Mathijssen, 2014), increasing the likelihood of integration at all dimensions. However, as our research demonstrates, creating opportunities for connection and building more direct (versus indirect) interfaces simply lays the groundwork for the more in-depth and long-term relational and cultural work that is needed to enable true integration at the individual level (Fawcett et al., 2008; Zhao et al., 2011). It became clear in follow-up interviews that such shifts require long-term, continued investment in cultural change at all levels of the organization, and change agents who champion a clear vision of the future. Redesigns should be planned with this in mind and may use structural change as a basis to incorporate additional strategies to overcome interpersonal and cultural boundaries.

Why integrative efforts fall short: the importance of relational and cultural dynamics

Our findings are in line with previous work that showcases the power of building cross-functional interfaces and connectedness to foster more understanding (Daft and Lengel, 1986), knowledge exchange and integration (Jansen et al., 2009). However, as we have shown here, implementing integrative roles (e.g. EPs) and creating new linkages between individuals (Vickery et al., 2003) is not enough to overcome the more deeply entrenched interpersonal and cultural boundaries that exist and are continually fostered through specialization and differentiation. While organizational reconfiguration (i.e. co-location) helped to alleviate structural boundaries, it did not help to alleviate the underlying problems of resolving other boundaries between interdependent groups, and proves an expensive [and only partial] solution.

Other scholars have emphasized the fact that structural reforms or system redesign alone does not lead to performance improvement (see also Hyer et al., 2009), and our study shows that despite its ability to potentially improve processes of care (Hyer et al., 2009), co-location is not enough in and of itself to overcome integration challenges. As we have shown, entrenched interpersonal and cultural boundaries continued to induce significant delays in the care process, which have been shown to have potentially negative consequences for both patients, staff and system performance (Morley et al., 2018). As our findings show, creating direct links between professionals (e.g. via co-location) may also work to reinforce boundaries between groups, incidentally undermining integration and performance. This demonstrates the need to go beyond “cross functional and cross firm business processes” (Leuschner et al., 2013) and more seriously consider the importance of interpersonal relationships and their influence on operational performance at a higher level. When we recognize professionals as key linkages in operational processes, we reveal a need for more consideration of interpersonal, psychological and cultural dimensions. We reveal with this study, and encourage other scholars to take up this call, that these dimensions also filter up, and significantly shape outcomes and care processes.

Recommendations for practice/strategies to improve operational performance

In light of our findings, we suggest organizational changes that can be considered by organizations in their efforts to improve lead times and service delivery processes and outcomes. First, hospitals should focus on redesigning processes (see also Hyer et al., 2009) and conduct thorough analyses of where bottlenecks exist within care processes and patient journeys. As we have shown, there are several key decision-making moments within the ED care process. These decisions create interfaces between professionals across disciplines, departments and with diverse levels of experience. While focus is often on other steps in the process where delays are known to occur (e.g. triage, discharge decisions, bed placement), our study shows several key points throughout the process (e.g. determining patient ownership, interdisciplinary consultations) where delays can and do occur. Overlooking these moments leaves organizations unable to appropriately direct resources and improve care processes.

Second, hospitals may consider reworking the process by reconsidering the role of professionals involved in the process. For example, expanding the role of EPs may support a more efficient flow by granting more discretion to individuals closest to and overseeing the patient journey. While granting EPs full autonomy may work to reinforce boundaries, small changes such as providing EPs the autonomy to take more ownership, e.g. by allowing them to admit undifferentiated patients, can help avoid the lengthy delays that result from sequential consults. As other studies have shown, providing ED staff with admit rights can reduce lead times, leading to extra bed availability (Morley et al., 2018). However, structural role expansion will still require attention to relational and cultural dynamics to improve processes on the longer term as the ED remains a shared space that demands constant collaboration.

Third, as Lu et al. (2021) highlight in their recent work on sourcing collaboration, informal exchanges and relational dynamics are important and have strong effects on the more formal exchanges that take place. While our focus is different, the notion that we need more attention to informal and “softer” mechanisms is growing across scholars of operations management. Although formal integration mechanisms are still needed, our findings showcase the need for organizations to focus on more informal and relational mechanisms (see also Jansen et al., 2009) to achieve integration and reduce instances of crowding, such as the role of boundaries at the interfaces of care delivery (Nugus et al., 2010). As Huckman and Staats (2011) have shown in their work on fluid teams, team familiarity (members having prior experience working together) may support performance, particularly when members have different levels of experience (in our case, older and younger residents, medical specialists and EPs). Our study indicates that boundaries between EPs and specialists may be lessened by encouraging continued interactions outside of the direct working environment (see also Huckman and Staats, 2011), e.g. by creating opportunities and tasking both groups to collaborate on solutions for emergency department challenges. Such an effort reinforces the idea of shared responsibility and creates opportunity for meaningful feedback between the groups that may help break down cultural barriers in the workspace (e.g. specialists gaining understanding of the need to work more quickly and beginning to work parallel rather than sequentially) and can support productive interactions and integration between individuals.

Limitations and future research

In this study, we expand current notions of internal integration and bring focus to a lower level of analysis. However, we do not explore how professional integration may come to impact other levels of integration. Taking a multilevel approach to integration would be a particularly fruitful area to explore in future studies and would allow for further consideration of how the more social and cultural aspects of operations actually come to influence integration at multiple levels. Additionally, because we focus on one organization, we did not explore the presence of inter-organizational boundaries and how these may uniquely impact integration and performance. This adds an additional layer of complexity and warrants further attention, also to consider how boundaries transcend organizational borders.

Additionally, while we demonstrate the importance of focusing on the social and cultural aspects of operations, and call for a more dynamic view of integration, our findings are limited to one sector. Healthcare represents a unique professional service context (Senot et al., 2016) and the emergency care process has unique characteristics that may affect the level of professional integration, e.g. ad hoc teams and high variability. While this induces particularities, it may not be entirely dissimilar to the notion of fluid teams (Arrow and McGrath, 1995; Huckman and Staats, 2011), and therefore, we suggest that other studies work to tease out how particular team characteristics may effect integration efforts and outcomes. Additionally, studies that focus on this level of integration should be conducted across different professional and knowledge-intensive operational contexts (e.g. law firms, engineering made to order, universities) (Balthu and Clegg, 2021) to further examine the antecedents to integration at the level of professionals more generally.