Hospital Governance During the COVID-19 Pandemic: A Multiple-Country Case Study

ABSTRACT In response to the disruptions caused by COVID-19, hospitals around the world proactively or reactively developed and/or re-organized their governance structures to manage the COVID-19 response. Hospitals’ governance played a crucial role in their ability to reorganize and respond to the pressing needs of their staff. We discuss and compare six hospital cases from four countries on different continents: Brazil, Canada, France, and Japan. Our study examined how governance strategies (e.g., special task forces, communications management tools, etc.) were perceived by hospital staff. Key findings from a total of 177 qualitative interviews with diverse hospital stakeholders were analyzed using three categories drawn from the European Observatory on Health Systems and Policies framework on health systems resilience during the COVID-19 pandemic: 1) delivering a clear and timely COVID-19 response strategy; 2) coordinating effectively within (horizontally) and across (vertically) levels of decision-making; and 3) communicating clearly and transparently with the hospital’s diverse stakeholders. Our study gleaned rich accounts for these three categories, highlighting significant variations across settings. These variations were primarily determined by the hospitals’ environment prior to the COVID-19 crisis, namely whether there already existed a culture of managerial openness (including spaces for social interactions among hospital staff) and whether preparedness planning and training had been routinely integrated into their activities.


Introduction
Across the globe, the COVID-19 pandemic disrupted health systems in unprecedented ways, effectively providing a worldwide natural experiment testing health systems resilience. 1 Past experiences of health crises-SARS-CoV-1, influenza, Ebola virus disease, etc.pushed hospitals across the globe to develop new structures and instruments to better prepare and manage those crises.For instance, outbreaks of influenza prompted Canadian hospitals to incorporate crisis management simulation exercises for their staff, based on the national pandemic influenza preparedness plan published in 2006. 2 In these preparedness plans building from the past, there was however seldom mention of what to do in case of hospital human resource shortages-which was a major issue in the COVID-19 pandemic.Another missing piece was the strengthening of data management capacities, e.g., regularly updated and reliable information dashboards on hospital caseloads, and rapid systematic reviews to account for new evidence on emergent diseases. 3ithin health systems, health sector governance has been the key driver of effective pandemic responses and resilience. 1 For instance, Canada-based studies showed that during the COVID-19 pandemic, public health systems with autonomous decision-making structures displayed more agility in their response. 4,5Applied to the health sector, governance "involves the actions and means adopted by society to organize to promote and protect the health of its population. . . .p2) As the primary deliverers of health care services, hospitals and health care workers were the most directly affected components of health systems.Hospitals' capacity to modify their governance structures, including assigning new and changing roles and responsibilities, played a crucial role in their ability to respond to COVID-19 through the processes of shock absorption, adaptation, or transformation. 7o date, the majority of empirical studies investigating health systems resilience in the COVID-19 context were conducted at the macropolitical level, [8][9][10] rather than the organizational or hospital level.A recent European Observatory on Health Systems and Policies (EOHSP) report on health systems resilience during the COVID-19 pandemic highlighted diverse categories of strategies "for building back better." 9Among those, four are applicable to hospitals' governance structures: 1) delivering a clear and timely COVID-19 strategy; 2) coordinating effectively within (horizontally) and across (vertically) levels of government [i.e., decision-making]; 3) ensuring transparency, legitimacy, and accountability; and 4) communicating clearly and transparently with the population and stakeholders.Because our project is focused on governance, we omitted the third category-ensuring transparency, legitimacy, and accountability-as it represents the overarching goal of governance and risked producing redundancies across the reported findings.Indeed, we knew that the other analytical categories would lead us to report empirical findings relating to transparency and accountability, in particular).
The HoSPiCOVID research project is a qualitative multiple case study that examines the resilience of hospitals from multiple countries and continents, which offers both analytical granularity and comparability. 11or the present study, we attempted to provide answers to the following research question: how did hospitals' governance structures managed to transform and adapt in the face of COVID-19?Our multiple case study features the experience of six hospitals in four countries (Brazil, Canada, France, and Japan).Specifically, we reviewed the roles and responsibilities given to hospital governance bodies in each hospital and examined how they were perceived by hospital staff, using the three categories borrowed from the EOHSP report.

Methods
This was a multiple case study 12 using qualitative research methods.The cases were six hospitals: one in Brazil, two in Canada, one in France, and two in Japan.Their selection was guided by the imperative of ensuring a diverse pool of hospitals and specialties in our research project, as well as by research team members' prior engagement with certain hospitals. 11The six hospitals' main characteristics are outlined in Table 1.
In each hospital, research team members carried out qualitative interviews with a wide range of individual stakeholders: health care workers, senior and middle managers (i.e., respectively with high-level and midlevel responsibilities), clinical unit/department heads, supply unit heads, infection prevention advisors, and project managers.Interview guides are available upon request.Interviews lasted between 40 and 90 minutes.A total of 177 persons were interviewed (see Table 2).
Interviews were recorded, transcribed verbatim, de-nominalized, and transferred to qualitative data processing software programs for coding and subsequent thematic analysis.As a first step, we used the HoSPiCOVID overarching conceptual framework on health systems resilience in times of COVID-19 for coding. 11Our international research team met regularly via Zoom to compare findings from the different hospitals and settings; their agreement on research interpretations helped ensure the internal validity of our findings.Further, to confirm our findings, preliminary results were discussed and validated with key informants during results dissemination and lessonslearned workshops.As a second step, and in line with the research topic of hospital governance, the coded data were re-analyzed and charted using a shared Excel spreadsheet (available upon request).The spreadsheet charted the data according to the three EOHSP analytical categories focused on governance, 9 as mentioned above: 1) delivering a clear and timely COVID-19 response strategy; 2) coordinating effectively within (horizontally) and across (vertically) levels of decision-making; and 3) communicating clearly and transparently with the hospital's diverse stakeholders.Within each broad category, subcategories were developed through a deductive-inductive approach to capture context-specific themes reflected in our data (Supplementary File 1).Building on other works, 13 we also highlighted factors that facilitated or hindered the successful implementation of these categories of processes.Findings are presented in the Results section under these three broad categories.

Delivering a Clear and Timely COVID-19 Response Strategy
In all sites, the general perception among interviewees was one of timely and responsive decision-making.In mid-March 2020, every hospital management team rapidly set up a specific emergency committee (or crisis committee, or COVID-19 response task force) to manage the COVID-19 response within its hospital.These committees met regularly, i.e., every day at the beginning of the first COVID-19 wave (even twice daily in some cases, especially during the first few weeks), then weekly, then tapering to monthly.Their mandates covered the following tasks and responsibilities: reorganizing the physical infrastructure in line with current COVID-19 priorities (setting up red/ orange/green zones, etc.); increasing bed capacities; developing service continuity plans; closing or reducing activities in hospital units and services considered 'non-essential'; communicating on the major public health and hospital-specific decisions to all heads of hospital units and departments; collaborating with local government authorities and other health facilities (e.g., for sharing supplies of personal protective equipment [PPE]).
These decision-making bodies were considered the spearhead instrument for delivering timely COVID-19 response strategies.However, their roles and responsibilities were not always clearly communicated to hospital units, a limitation noted in the French hospital.In Canada, Site A relied on a preexisting emergency measures committee with a clear mandate, which had developed pandemic preparedness plans prior to 2020.Conversely, the French hospital's crisis unit, created to manage the pandemic response efforts, had no preexisting guides for managing exceptional health situations.In Japan, while no guidance was available, the two case hospitals had experience with preparing for disaster response, including exercises, which was useful in establishing and managing the task force.
This citation from a Japanese hospital manager illustrates those decision-making processes and the anxiety associated with having to implement such processes: The biggest concern I had was the collapse of [general] medical care, and how to prevent it.I felt strongly that this was a very difficult task.After all, the Corona [SARS-CoV-2 virus] is considered a disaster.Well, it was a massive outbreak.So, we divided it into phases 1, 2, and 3, which is the BCP [business continuity planning] concept in the event of a disaster.We notified each department to prepare a BCP in each phase (Hospital manager, Japan, Site A).Decisions had to be made within short time frames, and the decision-making bodies were primarily set up to ensure fast, proactive, and adequate responses.In the French and Brazilian hospitals, anticipated bed capacity needs were based on the crisis unit's daily compilation of regional epidemiological forecasts.However, in some cases, such quick decisions were not well accepted by the staff on the front lines.In the French and Canadian hospitals, health care workers criticized the abruptness of some of the decisions to open or close services, characterized by their speed of execution.
A key facilitating factor that helped secure acceptance and build legitimacy for those decision-making bodies was that they involved clinicians and managers from diverse medical specialties and units across the hospital (e.g., emergency care, intensive care, infectious diseases, infection prevention and control, supply department, etc.), thereby ensuring the multidisciplinarity and representativeness of the decisions made.In three hospitalsthe French hospital, Site B in Japan, and Site A in Canada-management teams at the very top relied primarily on recommendations made by their own infection prevention and control (IPC) teams, which were led by infectious disease experts with credibility and leadership that extended beyond the hospitals' walls.Those teams were at the forefront of the response; they were tasked with constantly seeking updated scientific information about COVID-19 transmission risks and what to do about them, thereby strengthening the process of evidence-informed decision-making.A similar pattern occurred in the Brazilian hospital, where this work was carried out by the infectious diseases unit: Based on what we could gather from the scientific evidence coming out, we always tried to make decisions based on . . .perhaps not certainty, but less doubt.Because certainty . . .even the scientific studies had small samples, so we had to go on the experience that we had already had in . . .other epidemics (Nurse, Brazil).
In the Sites A of Canada and Japan, this feature occasionally led to hospitals' making decisions ahead of governmental policies, with the IPC teams gathering evidence faster than the ministries of health.In Canada's Site B, decisions affecting the intensive care unit were made based on the recommendations of a dedicated hospital team who met daily and weekly to gather the most up-to-date scientific evidence about COVID-19 and intensive care.
Not surprisingly, in all six hospitals the most significant hindering factors for fast and adequate response were the substantial scientific uncertainties about the SARS-CoV-2 virus and its transmission and transmissibility.These issues led to rapid and frequent updates in protocols-particularly PPE protocols-resulting in high levels of anxiety for hospital staff and ultimately in frustration and low adherence to protocols and rules.Such frustration was observed in all hospitals but one (Site A in Japan, see below), as illustrated in the French hospital, where this nursing assistant complained about decision-making bodies' lying about the equivalent effectiveness of surgical and FFP2 masks: At one point I was very disappointed when the masks were changed, when we were told, "no more FFP2, [now use] surgical masks, because actually that's enough."We know that we need FFP2.We would have preferred . . . the truth (Nursing assistant, France).
Site A in Japan was somehow an exception; thanks to minimal changes in infection prevention protocols and sustained application of the basic principles of infection prevention, the staff did not experience being overwhelmed by the latest information.

Coordinating Effectively Within (Horizontally) and Across (Vertically) Levels of Decision-Making
Horizontal coordination (i.e., across units and departments within a hospital) was considered quite effective between department heads (e.g., through the emergency measures committees), but less effective and successful between mid-level management and across the front lines.The vertical (top-down) configuration that was prioritized for faster decisions sometimes caused discrepancies in coordination processes and communication between units.In the French hospital, for instance, IPC team members served as information liaisons, transmitting decisions on unit reorganization (e.g., moving beds from one unit to another), without leaving much room for discussion between units.French respondents, particularly middle managers, complained about these "top-down" coordination processes that, in their view, did not promote horizontal coordination, as they did not allow middle managers and front-line workers to feel included in collective response efforts.Several interviewees felt they were treated as mere receptacles of information, rather than as active participants in its production and dissemination.Middle management professionals (e.g.health executive nurses) would have preferred to be more involved, as they considered themselves key actors in the hospital reorganization processes: The information was given without feedback.The fact that people ask questions can be disruptive, but from the outset, all the medical and paramedical staff needed to talk about messages that were misunderstood (Health executive nurse, France).
As a result, front-line health care workers felt inadequately informed and equipped, and that they were practicing care in suboptimal conditions.
An exception to the above was seen at the Japanese Site A. While decision-making processes there were highly centralized at the beginning, these top-down processes reportedly served as foundations for smooth coordination and collaboration between and within services.Thus, each operational committee of the COVID-19 special task force (e.g., intensive care committee, surgery committee) was able to coordinate front-line workers in different units while taking into account top management directives, thanks to close ties among diverse hospital actors: The person in charge of the fever consultation outpatient clinic was also a specialist in infectious diseases.She also had a very close relationship with her superiors.The chain of command was rather, well, let's call it a management team.The team members were able to work together smoothly and quickly (Pharmacist, Japan, Site A).
In the Japanese Site B, the heads of the most COVIDrelevant clinical departments served as special assistants to the hospital director even before the pandemic happened.In the COVID-19 special task force, horizontal coordination took into account "opinions from the field" brought to the task force by those special assistants.Horizontal communication among front-line physicians was routine because non-management physicians shared one large office.In both Japanese sites, unlike the coordination between physicians in different departments, coordination between nurses was done principally through the existing command and control structure.
Once the most pressing hospital decisions-notably IPC and PPE protocols-were communicated to department and unit heads, their implementation within those departments and units posed several challenges because they required substantial amounts of training and explanation.In every hospital, the availability of IPC professionals was a key asset for delivering training in all services and addressing the most pressing concerns and questions of front-line workers.In the French, Canadian, and Japanese sites, IPC professionals served as daily points of reference for resolving uncertainty on the front lines (e.g., in emergency rooms).In Brazil, the call for infectious disease student volunteers was instrumental in the study site, which is a university hospital: these students, together with professors and infectious disease professionals, were able to train medical and paramedical teams rapidly in the fundamentals of IPC and PPE protocols.However, in practice, IPC protocols were often difficult to implement in clinical routines due to the inexperience of newly hired professionals who had not received prompt or adequate training.
One key facilitating factor was a hospital culture favoring managerial openness (for instance, with fluid boundaries between hospital departments and units, and "open spaces" where staff can easily interact with the head physician).This was observed in Japanese Site B, Canadian Site A, and the Brazilian site: So, we ended up having to also do, sometimes . . .not only . . . the daily assessment, . . .but also support, so that the shift would be . . .less tiring, less heavy for those who were there in the front line on duty. . . .So . . .we started to spend more time together . . . .So, there was an increase . . .an intensification of interpersonal relationships . . .something like team spirit, helping each other . . .(Physician, Brazil).
This openness was also facilitated by prior experience with implementing preparedness plans (including crisis preparedness simulation training exercises), as seen in the Japanese and Canadian Sites A. In addition, in both Canadian sites, the presence (or absence) of a form of "horizontal trust" between front-line workers and team leads, department heads, or managers, was considered a key determinant of effective (or ineffective) response coordination within those hospitals.

Communicating Clearly and Transparently with the Hospital's Diverse Stakeholders
Another key ingredient of successful pandemic response was the transparency of communications within the hospitals.The six hospitals strove to make information available and accessible in a timely manner through multiple channels.Hospital-wide communication tools included: open conversations with hospital senior and middle managers (such as, in Japanese Site B, sharing the number of remaining N95 masks on any particular day); coaching and training teams for IPC and PPE protocols; regular information bulletins and newspapers; and posters on how to handle PPE.In addition, units and departments developed routine communications tools (e.g., mailing lists and WhatsApp groups for specific health care workers, in the French and Brazilian sites).In Canada's Site A, the combination of general and targeted communication strategies reportedly improved the hospital staff's feeling of security, especially because they were implemented rapidly: It was all about being up-to-date, making sure the communication was in place.[Our hospital] made big efforts in communications . . . in terms of the news bulletin coming every day . . . .So, really trying to make sure that . . . the information was circulating.That's what makes it . . .You know, I think it was well received . . .People felt safe overall (Social worker, Canada, Site A).
Then, as critical situations arose, tools were also created.You know, reminder posters, signs, new tools like that.So that also reassured [staff on the front lines] . . .Signs on how to put on protective equipment that were posted in the rooms most likely to have COVID patients, and especially instructions on how to remove it.You know, the order of removal, what are the steps to remove it without getting contaminated, what do you do if you get contaminated . . .These tools were put in place in the first week [in March] (Nurse, Canada, Site A).
According to respondents in Japanese Site B, being transparent meant sharing all information, including inconvenient information (e.g., hospital-acquired COVID-19 clusters), as soon as possible.Conversely, transmitting all information may result in information overload, as happened in the Brazilian hospital, the French hospital, and Canada's Site B. In France, where health care workers felt overwhelmed by the profusion of e-mails, the hospital and the crisis unit struggled to find an efficient way to regulate the flow of information coming from multiple sources.To resolve the e-mail overload issue, WhatsApp groups were implementedone for medical staff and one for the entire infectious diseases department, particularly to help with organizing night shifts for the medical staff of the unit.However, these solutions were developed spontaneously and only in certain units.
A key facilitator for the success of targeted communication strategies (i.e., within teams/units) was the prior existence of trust relationships between team leaders and team members.A case in point was the IPC unit of the Canadian Site A-particularly when it came to precautionary measures: [Our IPC team supervisor] would tell us . . .what to expect . . .Because she has a direct connection to the Ministry [of health], and she's the one who helps a lot with the recommendations . . .So . . .she really doesn't hesitate to tell us, "Okay, this is what the ministry recommends."And then, she'll tell us whether she agrees or disagrees, and then she gives all the evidence and everything to prepare us.And I find that this, again, adds to that trust in the team (IPC advisor, Canada, Site A).
In contrast, in the French site, such trust between hospital executives and staff seemed to have been lacking prior to the pandemic.This acted as a major hindering factor.
Supplementary File 2 offers an overview of facilitators and barriers for each analytical category.

Discussion
Our study provides key insights from multiple hospital settings across four continents on the determinants of successful (or unsuccessful) COVID-19 response governance tools and strategies.These unique findings are based on the accounts of two crucial categories of hospital staff-health care workers and hospital managers.To our knowledge, this is one of the first studies reporting findings on hospital governance tools and strategies featuring the views and perceptions of these two participant categories.We identified key facilitating and hindering factors, most of which had to do with to the clarity of roles and responsibilities assigned to key decision-makers, and with conducive experiences.The latter refer to hospitals' prior experience of preparedness planning and training, as well as to the existence of a culture of managerial openness that supports and sustains trust relationships between senior managers, middle managers, and front-line workers, which in turn facilitates communication and horizontal coordination.These findings were particularly salient in the Japanese and Canadian cases. 14These two broad determinants were the only governance determinants that appeared to produce variations across hospital settings.
A culture of managerial openness is conducive to health care workers' feeling included in collective response efforts.It also has positive health outcomes for patients.For instance, in the United Kingdom, instilling a culture of openness, which included implementing an independent system in which staff members could report errors or concerns without being brought into conflict with their employers, led to lower mortality rates in 137 National Health Service hospital trusts. 15In the hospital context, openness can be defined as an environment in which "staff will freely speak up if they see something that may negatively affect patient care" and "feel free to question the decision or actions of those with more authority." 16Even though the emergency context might jeopardize their ability to maintain such an approach, hospital decision-making bodies should at least safeguard some features of the openness culture, such as providing spaces for casual social interaction and debriefing for staff.In fact, in our own lessons-learned workshops, study participants noted that such open discussion spaces-which they were experiencing in the workshop itself-were much needed in such a stressful context.
Implementing managerial openness, however, may not be easy for hospitals that have traditionally employed a vertical decision-making approach. 17hile the horizontal approach is not yet dominant, several leadership styles such as inclusive leadership may support a stronger sense of collective ownership from all hospital staff. 18Inclusive leaders invite and appreciate members' contribution while enhancing their feeling of being part of the group (i.e., belongingness). 19In fact, "shared decision-making" involving senior and middle managers and staff from the front lines was identified as a key capacity for hospital resilience in a recent review. 20everal COVID-specific innovative governance strategies were implemented from the very beginning of the crisis, in particular the creation of multidisciplinary emergency measures committees or task forces and rapid scientific literature reviews done by IPC or infectious disease units to gather the most up-to-date information on COVID transmission routes and ways to protect patients and staff. 21Those units swiftly reorganized and acquired more human resources (including, in some cases, students and volunteers), who guided the formulation of measures adopted by the emergency committees (e.g., updating PPE protocols) and then oversaw their implementation throughout the hospital, from front-line staff to senior management.At the same time, this constant updating-including updates to PPE and IPC protocols-created significant anxiety for health care staff who had to deal constantly with COVID transmission on the front lines.Similar to findings from the United Kingdom, 22 honest communications about the remaining PPE stocks and why protocols were being changed were considered effective strategies by respondents in the Japanese and Canadian sites.
Questions remain as to the sustainability of such innovations, which belonged to a "governance of exception" within specific time frames.How might some of the key "wins" of such innovations be maintained?Several hospitals recruited additional human resources to ensure swift delivery and application of IPC protocols; will these new hires be retained?Will funding for continued training in IPC be sustained over time?Some authors believe that, at a minimum, the prominent position of IPC professionals, which the pandemic context brought to the forefront of the response and which had often been neglected prior to 2020, will continue in most hospitals. 23,24Sustaining those innovations would also require renewed leadership with visionary, longterm planning to "build back better" hospitals and health systems. 25mplementing hospital-wide communication channels along with tailored messaging for specific teams and units, while maintaining a clear and transparent communication style, was identified, in our study cases, as a key driver of success in hospital governance throughout the pandemic.Such a communication mix effectively minimized information asymmetry.Targeted communication tools such as WhatsApp groups helped prevent information overload for managers of clinical units and supported their daily coordination operations, while broad messaging (via mail lists) ensured that all hospital staff could have access to key information in a timely and adequate fashion.These observations in our cases concur with accounts of health systems resilience in geographical regions not covered in this study. 25p6) Our study adds to the body of empirical literature assessing determinants of health systems resilience by focusing on tools and strategies used to govern the response in hospitals.The present multiple case study provided rich accounts from six hospital settings across the globe, highlighting some significant variations.These variations were primarily determined by the hospital's environment prior to the COVID-19 crisis, namely whether there already existed a culture of managerial openness (including spaces for social interactions among hospital staff) and whether preparedness planning and training had been routinely integrated into their activities.When such openness was lacking, front-line staff appeared to be more vocal in their criticism of the hospital's pandemic governance.
Our study also has limitations.We were unable to systematically assess the hindering and facilitating factors for each strategy/tool reported in each hospital, due to time and feasibility constraints on the data collection process (e.g., recurrent COVID waves that often prevented in-person interviewing) that did not allow for documenting each of these strategies and tools in-depth.

Conclusions
While being qualitative accounts, and therefore not generalizable, our findings do reflect a variety of geographical contexts and hospital characteristics, from which a few hypotheses can be drawn.Additional empirical research, particularly quantitative research, could explore what components of managerial openness and prior crisis preparedness experiences are conducive to a successful COVID-19 pandemic response.More specifically, further quantitative studies may be needed in order to confirm that the factors highlighted in our paper-which are only based on health care workers' and managers' accounts, can actually explain (and to what extent) the success or failure of these governance strategies and tools.Building upon our findings on the determinants of effective communication of COVID-19 decisions in hospitals, other studies could also explore what combination of targeted and hospital-wide communication tools would be most effective to improve hospital staff's access to key information.Finally, future research should also investigate the long-term financial and/or implementation sustainability of new governance mechanisms put in place in the context of the COVID-19 health crisis.

Table 1 .
Characteristics of the six hospital cases.

Table 2 .
Respondents' position and level of responsibility.