Hospital Resilience to the COVID-19 Pandemic in Five Countries: A Multiple Case Study

ABSTRACT Since the beginning of the pandemic, hospitals have been central to the COVID-19 response, often experiencing severe financial, material, and human constraints. In this special issue, we present some of the findings of the HoSPiCOVID research project. One of its main objectives was to compare hospital responses to the first and second waves of the COVID-19 pandemic in Brazil, Canada, France, Japan, and Mali. Studying and comparing how nine different hospitals coped with the pandemic in terms of preparedness and response allowed us to: 1) identify strengths and weaknesses of their responses, including challenges for hospital professionals; and 2) produce lessons learned, using a systematic approach to reflect and analyze their potential of resilience to the crisis. In the five countries, research teams conducted in-depth qualitative studies focused on nine large hospitals, using observation sessions, semistructured interviews with hospital professionals, and lessons learned workshops. The empirical work was supported by an original analytical framework on hospital resilience and a heuristic tool focused on configurations. The studies demonstrate that the hospitals were able to absorb and/or adapt to the crisis by deploying different coping mechanisms, which often required extensive involvement of hospital professionals. More extended study periods would be needed to assess the sustainability of these coping mechanisms and discern whether they have transformative potential. These international comparisons of hospital resilience, based on studies of contrasting contexts and epidemiological situations, allowed researchers to identify lessons learned to support hospital decision-makers in thinking more deeply about managing future health crises.


Introduction
As countries and health systems struggled to respond to the COVID19 pandemic, research teams were quickly mobilizing.Numerous calls were launched for pandemic research, and often for interdisciplinary research.Given the urgency of the situation and the availability of research resources, the gathering of a large multicountry interdisciplinary research team led to the creation of the HoSPiCOVID project.This research project, conducted from April 2020 to April 2022, aimed to study the resilience of public health systems, hospitals, and their staff to the COVID19 pandemic in Brazil, Canada, France, Japan, and Mali (https://uparis.fr/hospicovid/). 1 This special issue addresses one of the project's two main areas of study: the resilience of hospitals and their staff to the pandemic.The second area of the study, not addressed in this special issue, examined whether and how social inequalities in health were considered in the planning and design of contact tracing, 2 screening, 3 and vaccination. 4rough an interdisciplinary study, the objective of the HoSPiCOVID project was to conduct qualitative studies of hospital responses to make intercountry and/or thematic comparisons, and to learn from the experiences of staff in managing the COVID19 crisis.Our goal was also to disseminate the study findings to health care professionals, stakeholders, and researchers throughout the course of the research project.Based on our networks of previous collaborations 5,6 and our aim to compare cases with different epidemiological COVID19 contexts, research teams were formed in Brazil, Canada, France, Japan, and Mali.We chose these countries because they offered a variety of conditions useful for international comparison, as other research teams have done for countries in Europe, 7,8 North America, 9 and the Eastern Mediterranean region. 10We wanted to compare countries across different continents, with different health systems financing approaches and economies.In the context of emergency, when we planned this research and obtained its funding, we also wanted to be pragmatic.Thus, the countries and research teams were chosen for scientific reasons, by varying the contexts to better understand the resilience of case studies hospitals, but also for pragmatic reasons, to bring together interdisciplinary teams, and experts in domain with collaborations in the field.The hospital resilience component of this work aimed at understanding how, in these five countries, nine hospitals and their staff (both medical and nonmedical) responded to the pandemic.We choose these nine hospitals for pragmatic and heuristic reasons: first, for their importance in the pandemic response (most of them were COVID19 referral hospitals), and second, for their accessibility by external researchers in a time of urgency (structures were overwhelmed).
In many of these countries, strong coercive government measures were taken to contain the spread of the virus, protect the economy, and avoid overloading health systems, particularly hospitals. 7In most countries worldwide, hospital centrism is the norm, and hospitals often serve as points of entry into the health system. 8Referral hospitals have been particularly affected by the crisis, and those with specialty services in infectious diseases were often the first to begin caring for cases of COVID19. 10,11Many health systems worldwide have undergone multiple reforms over many years, 8 including neoliberal reforms that have reduced hospitals' operational and recruitment capacities, as happened in France. 12Thus, the pandemic occurred in difficult contexts, where the capacities of hospitals and their staff to prepare and cope were stretched beyond their organizational limits.To understand hospital resilience, it is essential to consider the relevant historical, organizational, political, and power dimensions. 13,14ome of these elements are addressed in the articles in this special issue and others are addressed in articles in other journals.
To introduce this special issue, we will first present the pandemic and hospital contexts of the cases analyzed.Then, we will discuss the conceptual and methodological issues encountered by the team, and the practical implications of the findings.

Varied Pandemic and Hospital Contexts
The pandemic waves varied between countries regarding magnitude and timing (Figures 1 and 2).The hospitals are in major cities, and their catchment areas are local, with some also serving as referral centers for infectious diseases or pediatric patients (Table 1).Hospital bed capacity ranged from approximately 132 (Mali) to 850 (France) beds.Staff capacity ranged from just over 3,000 (Brazil) to over 10,000 (Canada, Site B).For COVID19 patients, five of the hospitals in Brazil, France, Canada (Site A), Japan (Site A), and Mali were quickly established as referral centers.The other hospitals, in Canada (Site B) and Japan (Site B), began receiving COVID19 patients once treatment protocols were established and the required infrastructure, including equipment, was put into place.Of the five study countries, Japan was the first to report a confirmed case of COVID-19, on January 12, 2020; France followed closely (Table 1).From the onset of the pandemic until December 31, 2020, France had the heaviest reported burden of COVID-19, with almost 40,000 cases per 1,000,000, followed by Brazil.Japan and Mali had the fewest reported cases during this period.

A Flexible and Heuristic Conceptual Framework for Resilience
Since the 2016 Fourth Global Symposium on Health Systems Research, numerous writings have been published on the concept of health system resilience. 15,16ndeed, the conference and the subsequent special issue of Health Policy and Planning allowed researchers interested in studying health systems and hospitals to explore the concept.The WHO recently (2022) published an 84-page tool identifying dozens of guides and normative texts on the resilience of health systems. 1715][16]18 The intent of this special issue is not to dive into the conceptual debate but rather to present our study results on hospital resilience, based on empirical data and our interdisciplinary approach. 19irst, we developed a common and broad definition of hospital resilience, as presented in the protocol: 1 . . . the capacities of a hospital facing shocks, stress or chronic destabilizing tensions (unexpected or anticipated, sudden or subtle, internal or external to the system), to absorb, adapt and/or transform in order to maintain and/or improve universal access to comprehensive, relevant and quality health care, without causing patients to fall into poverty.
The expected outcome of resilience is access to health care, particularly its various determinants, as proposed by Levesque et al. 20 We then created a conceptual framework of the dimensions and essential constructs for understanding the resilience of hospitals.The dimensions are: 1) the occurrence of events or shocks, such as the COVID-19 pandemic; 2) the effects on hospital organizational routines; 3) the strategies deployed to address them; 4) the impacts of these strategies (perceived as positive or negative); 5) impacts on access to care; and 6) resilience processes (absorption, adaptation, transformation) and their outcomes (destruction, deterioration, return to normal, improvement).Absorption represents the ability of the hospital to continue providing basic health services with the same level of resources and capacity.Adaptation represents the ability of the hospital to deliver the same level of healthcare services by relying on organizational adaptations to deal with a disruptive situation.Transformation represents the ability of the hospital to change its functions and structure in response to a shock.
Having a common conceptual framework proved indispensable for our research collaboration, as it was neither perceived nor used as a rigid framework but rather as a guide that could be adapted. 21,22evertheless, the challenges of using the framework within the team were numerous.These included achieving clarity around specific dimensions, given the novelty of the framework, grappling with the difficulties of reaching a common understanding of certain concepts, and ensuring consistency in meaning when translating the framework into several languages.Many discussions and readings were necessary to explain the conceptual framework, refine it, and improve its heuristic capacity when applied in the field for data collection and analyses.
Finally, we organized the analytical approach from a comparative perspective around a heuristic tool: configurations (and their recurrence in different contexts).We borrowed the use of configurations from the work of McKee and Healy, who perceived hospitals as a system and used configurations as part of their work in understanding complex systems. 8onfigurations were the preferred analytical tool to support approaches not only within each case (the hospitals of each country), but also between cases (comparative analyses).p. 41 In concrete terms, each research team produced summary tables to organize their empirical database around the dimensions of the hospital (adapted from Turenne et al. 16 ) in as many configurations as they thought relevant.Each configuration described the associations between: 1) the effects (positive or negative), caused by the pandemic; 2) the strategies implemented to deal with these effects; and 3) the impacts, perceived as positive or negative by the hospital's professionals to facilitate (or not) the crisis management, of these strategies on the hospital's organizational routines.
We drew inspiration from Ragin's methodological proposals suggesting that: "The analytic device that allows this is the truth table, which displays the data in a matrix of logically possible configurations of causal conditions.p71) Each research team identified between four and 19 different configurations, depending on the hospitals.Then, from the comparative analysis of these configurations, four common configurations (i.e., occurring in all countries) were identified: 1) the reorganization of infrastructures and spaces; 2) the reorganization of the work of hospital professionals; 3) the management of personal protective equipment (PPE), equipment, and supplies; and 4) the management of infectious risk.These common configurations facilitated the inter-country comparisons presented in this special issue and the related knowledge transfer activities (see below).

What do these Articles Teach Us?
This issue contains both country-specific articles and intra-country and inter-country comparative articles.
The first set of country-specific articles is structured around the concept of configurations.The first three articles summarize the main configurations studied by country (Brazil, Canada, France, and Japan).Following a brief analysis of the hospital context and the effects of the COVID-19 pandemic on the organizational routines of the selected hospitals, the authors identify the strategies put in place by the hospitals and their staff and describe the impacts of these strategies as perceived by the hospital staff participating in the study.Finally, the authors discuss and analyze the resilience of hospitals and staff by identifying facilitating and constraining organizational and individual factors, and then point out areas for future research.In their study, de Araújo Oliveira et al. identified several factors that may facilitate hospital resilience in Recife (Brazil), including financial resources, new hospital protocols, support network, professional education, and proactive leadership. 25 Gabet et al. showed that Quebec's recent health reforms limited the Quebec integrated health and social services resilience, especially because of the highly centralized governance, and the administrative and institutional heaviness. 26Chabrol et al. highlighted the unprecedented referral hospital mobilization in the Parisian region (governance, leadership), the hospital relying heavily on the resilience and commitment of the hospital professionals. 27he second set of intra-country articles consists of comparative analyses led within countries.Paz de Sousa et al. studied and compared the resilience of three hospitals in the Recife region of Brazil, showing that adaptation and absorption processes were observed in all hospitals. 28Honda et al. underlined that strategies implemented at one level by two hospitals in Japan could be problematic at other levels. 29David et al. studied and compared the reorganization of hospital staff work and practices induced by the COVID-19 pandemic in two hospitals in Quebec, highlighting that adapting work through reassignment strategies could ensure the absorption of the influx of COVID-19 cases into hospitals but that these strategies also could have deleterious effects on hospital workers. 26he third set of inter-country articles consists of comparative analyses led between countries, focusing on specific themes, and comparing the experiences of health professionals in several countries.Hou et al. presented results from a comparative analysis of the communication and information strategies implemented by four hospitals and their staff in Brazil, Canada, and France to reduce the risk of COVID-19 infection, hospital staff often considering communication to help debunk fake news and in reassuring them. 30Gautier et al. examined and compared the governance strategies implemented in six hospitals in Brazil, Canada, France, and Japan and their perceptions by hospital staff, demonstrating that the transparency of communication within the hospitals was a key ingredient of a successful pandemic response. 31Coulibaly et al. compared the measures relating to family visits implemented during the COVID crisis in France and Mali, showing that measures taken to keep families away from where COVID patients were hospitalized were poorly received by families. 32response And Honda et al. identified and compared organizational factors that facilitated the creation and implementation of hospital innovative approaches in Brazil, Canada, and Japan, such as having adaptive and flexible organizational structures, building and maintaining functioning communication systems, having committed leadership, ensuring all staff have a common understanding of hospital missions and values, and establishing social networks that facilitate the creation and implementation of new ideas. 33n the last article of this special issue, Dagenais et al. described the lessons learned of the experiences lived by the hospital professionals in Brazil, Canada, France, Japan, and Mali to inform future decision-making within hospitals. 34

The Rigor of Our Qualitative Research and the Challenges of Causality
This special issue reports the main results of research in hospitals based on a qualitative approach.The credibility of participants, rigor in data collection and analyses, triangulation, and confirmation from participants were integral components of our research approach.The reflexivity of our approach is also essential for qualitative research.
There is still a long way to go regarding biomedical and positivist experts understanding the rigor of qualitative analyses. 35They still often ask for evidence of the internal and external validity of the studies.The classic qualitative sampling procedures were therefore used in this project, to achieve a targeted and relevant selection of the participants, 36,37 while being careful to avoid "enclosed systems" 35 that could confine data collection to restricted circles without providing a holistic view of the phenomenon being studied.This sampling challenge was central to most of our data collection activities, complicated by power issues specific to the organizational structures that are hospitals. 8,38However, we took the time needed to investigate the hospitals and used methods to ensure we had original and rigorous data (some of our data collection spanned more than 18 months).In contrast to the recently published studies from the European Observatory on Health Systems and Policies in high-income countries, which were mainly based on the analysis of secondary data, 9 our project relied on primary and original empirical data collected over an extended period of time in hospital facilities in a mixture of economic levels of countries or through remote interviews with people working there.Our special issue also complements the multiple articles published at the beginning of the pandemic by healthcare professionals involved in hospitals who wished to share their experiences in a more reflexive manner.
In our research, the use of configurations was heuristic, both to help researchers account for complex phenomena and to support an approach aimed at international comparison.The analyses are not causal assessments that would allow us to assert that a particular strategy in the context of an organizational routine produced a specific effect(s). 39uccessional causality (i.e., A produces B, which produces C, with the context variables being controlled) does not make sense in our approach, where the hospital is understood as a complex system. 40Reflections on the generative causality of critical realism therefore inspire our work. 24,41he aim is not to discover an immutable law that would work in all contexts and for all hospitals but rather to identify regularities in configurations from one hospital to another whose occurrence would be frequent but not necessarily constant.In assessments grounded in critical realism, these regularities would be described in the form of "context-mechanism-effect" configurations, but in our research, we organized them around "contexts-effectsstrategies-impacts" configurations.Our analytical approach meant that we were not guided by variables, but rather by the cases (hospitals) in a holistic approach, to use Ragin's terms, 23 through configurations.Configurations make it possible to identify sufficient conditions associated with particular changes. 42oncerning the ability to extrapolate our results to a larger scale, caution is advised.It is important to note that we do not claim that the results of these analyses represent the diversity of what has occurred in all hospitals or health systems in each country.The resilience discussed in these articles was focused on specific hospitals and not on all the components of any health system.Our study reports "focused configurations" based on empirical case studies rather than "abstract configurations" based on a more general theory to use the distinction proposed by Pawson and Tilley. 24Our analytical approach is, therefore, more in line with Yin's 43 and Ragin's 23 proposals, which inspired our case study approach.Rather than referring to external validity or applying study results to a different context, Yin suggests, in the context of case studies, invoking analytical generalization, which "depends on using a study's theoretical framework to establish a logic that might be applicable to other situations."We have done this by applying the same health systems resilience conceptual framework 1 to a variety of specialized hospitals in different contexts, using configurations as an "analytic device." 23As Yin suggests, this analytical generalization approach proceeds in two steps, which we followed by means of these configurations.p18)

Between Researchers, Decision-makers, Professionals, and Lessons Learned
The teams involved in this research program are all convinced of the importance of producing knowledge that is useful not only to scientists but also, above all, to those responsible for health systems and hospitals.Like the European Observatory on Health Systems and Policies 9,44 and 18 countries in the Eastern Mediterranean region, 10 we undertook knowledge transfer activities.We produced lessons learned 45 specific to the contexts of our studied hospitals.The last article in this special issue reports on the lessons learned in each hospital through a process that included a series of deliberative activities (focus groups, workshops).We aimed to follow a robust and rigorous method of producing lessons learned from empirical data collection.A rapid review of the literature was conducted on this method to support the researchers. 46Based on this method, the article addresses the implications of the findings of the empirical articles for public policy and clinical practice in future crises.

Conclusion and Future Directions
This special issue provides a platform of knowledge about hospital resilience during the first and second waves of the pandemic in nine hospitals in five different countries.This work broadens our understanding of the coping strategies used in hospitals during critical periods of the pandemic and the perceived effects of these strategies.While the approach to studying health system resilience needs further methodological and conceptual consensus, this special issue highlights the value of considering the use of configurations to identify conditions (and drivers) within complex systems.Future work should include further refinement of our conceptual framework to test its robustness, replicability, reflexibility, and utility.We recommend longer study periods to understand whether coping strategies have transformative potential.We identified lessons learned to help deepen the reflections of hospital decisionsmakers, and we hope that our approach in producing lessons learned will inspire others.

Figure 1 .
Figure 1.Daily new confirmed cases of COVID-19 per million people.7-day rolling average.Source: Our World in Data

Figure 2 .
Figure 2. Daily new confirmed COVID-19 deaths per million people.7-day rolling average.Source: Our World in Data

Table 1 .
Descriptive characteristics of each hospital and epidemiological context (provided by the hospitals and by Our World in Data for the country-level estimates).