Background

Concerns about the physical and psychological consequences for healthcare providers tasked with caring for patients with the novel-coronavirus SARS-CoV-2, the virus causing COVID-19, have been well documented [1,2,3,4,5]. From early in the pandemic, reports from countries around the world have detailed experiences of increased emotional distress, symptoms of depression, anxiety, insomnia and overall mental health disturbances among frontline staff [6,7,8,9,10,11]. Furthermore, retrospective studies following the severe acute respiratory syndrome (SARS) outbreak in 2003 have shown that some healthcare providers who worked in locations where contact with SARS patients was common continue to experience symptoms of posttraumatic stress disorder years later [12,13,14]. This is especially true in instances where psychological supports were minimal at the time [15]. Reduced confidence in training, fear of the unknown, limited hospital capacity, and insufficient support of personal protective equipment, are just some of the factors that have been shown to heighten healthcare worker experiences of negative psychological symptoms during the COVID-19 pandemic [16,17,18,19,20].

Prior to the start of the pandemic in 2020, national reports indicated that hospital visits via the emergency department [21], as well as staff burnout [22], were already on the rise in Canada [23] . In fact, Canadian ICUs were operating at close to full capacity in 2016, particularly in large urban academic centers, leaving little flexibility for surge coverage [24, 25]. Demand for ICU beds is expected to increase with an aging population [26, 27]. Notably, depleted human resources (e.g., nursing availability), affected both bed and ventilator use in the ICU at this time [28]. Existing levels of resource strain have been further exacerbated since the start of the COVID-19 pandemic. High patient volumes and overwhelmed supply chains have led to both national and global shortages of the resources required to care for critically ill patients (e.g. ICU beds, ventilators, staff, etc) [29,30,31] and to protect critical care providers from contagion (e.g. medical masks, N95 respirators) [32]. In extreme cases, insufficient resources have resulted in triaging life saving interventions (e.g. mechanical ventilation) [33] and unduly exposing providers to the virus [34,35,36]. Risk of viral transmission increases when performing aerosol generating medical procedures (AGMP) [37], particularly tracheal intubation, which is commonly performed by frontline critical care providers [38, 39]. Executed without adequate personal protective equipment (PPE)—gloves, gowns, eye protection, N95 respirators—may further compromise frontline provider safety [40]. It remains unclear how anticipated or actual resource shortages have affected critical care physicians and the delivery of patient care in ICUs in Canada during the pandemic. To address this gap in our knowledge, and to generate information that may be used to inform hospital and provider focused guidance, we sought physicians’ perceptions and experiences of caring for critically ill patients in the context of actual or anticipated resource strain across multiple institutions during the COVID-19 pandemic.

Methods

Study design

We used a qualitative description study design [41] conducted in accordance with the Consolidated Criteria for Reporting Qualitative Research. A qualitative research design is appropriate to explore topics where little information is previously known and to provide in depth information of participants’ own meanings and experiences. The use of qualitative inquiry offered the unique opportunity to develop rich insight into critical care physicians’ perceptions and experiences of resource strain in the ICU during the COVID-19 pandemic. The Research Ethics Boards at the University of Calgary (#REB20–0377) and Dalhousie University (# REB2020–5106) approved this study. Participants provided explicit oral consent in lieu of written consent.

Participants

We recruited critical care physicians using a recent and accessible sampling frame previously developed by our team [42] of all clinical and academic faculty from adult critical care training programs in 13 Canadian universities. We used non-probability purposive sampling [41] to invite a diverse sample of female and male physicians, currently employed, across years of practice, in provinces representing three (Atlantic, Central, Prairies) of Canada’s five [Atlantic, Central, Prairies, Pacific, Northern Territories] major regions to ensure variation in experiences. We aimed to recruit 5–8 (15–24 in total) participants per region. To this end, we consecutively emailed invitations until we achieved data saturation [43]. No additional eligibility criteria were applied.

Data collection

We iteratively developed a semi-structured interview guide and pilot tested it with two critical care physicians to ensure the core questions and probes adequately addressed targeted topics. In the pilot test of the guide, we removed one question that asked if the participants’ ICU had admitted any patients with a diagnosis of COVID-19. We also added a question as suggested by one pilot participant in the interview de-brief around the ethical and legal parameters of working under resource shortages in a pandemic. Interview questions asked physicians to reflect on three major topics: (1) current capacity of their primary ICU to meet the needs of critically ill patients with COVID-19, (2) existing or anticipated factors that would jeopardize physicians’ ability to manage the care of COVID-19 patients effectively, and (3) implications for the healthcare workforce and the delivery of patient care (Additional file 1). Interview questions were developed from topics of interest in the news media, clinical discussions and validated through discussion with ICU colleagues [44,45,46]. We included demographic questions at the end of the interview guide to capture participant age, sex, marital status, clinical base specialty, and ICU size.

Interviews and pilot tests were conducted in English via telephone in a private office by two senior research associates (LK, CdD), both female and with Masters degree training in qualitative and health services research. Participants provided informed verbal consent prior to participating. Interviews were collected between March 25 to June 25th, 2020. We digitally recorded the interviews and sent the audio files to a transcription company (www.rev.com/) to produce verbatim transcripts. The interviewers (LK, CdG) reviewed, cleaned, and de-identified the textual data prior to analysis. Participants were offered the opportunity to review their interview transcripts as a form of member checking to maximize validity.

Data management and analysis

Qualitative analysis was conducted between April 27 and July 25, 2020. We used NVivo 12 (https://www.qsrinternational.com/) to manage the data and facilitate thematic analysis. Two researchers (LK, CdG) first reviewed and coded a small sample of the transcripts (n = 3) independently and in duplicate using open coding [47]. Initial codes were compared and discussed with a senior qualitative researcher (JPL) to create a first draft of the codebook. The researchers then analyzed an additional three transcripts using both open and axial coding, iteratively refining the codebook until all relevant ideas were included. Deviant cases and exceptions within the data were sought and resulted in alterations to the codebook if the data were relevant to the research question, interview questions and emerging themes. The complete dataset (n = 15 transcripts) was then coded in duplicate with the finalized codebook. Fracturing of the data through axial coding was particularly useful in theme development as it enabled researchers to look across interviews for nuanced differences in participant perceptions. Both researchers used memos to document initial ideas during interviews as well as to document the relationships between codes during analysis. The researchers held weekly meetings during the 4-month period of analysis, wherein themes were developed, revised, and refined. Although our team was prepared to recruit further, saturation of overarching themes was achieved after the codebook was stabilized but prior to full analysis of the dataset (approximately halfway through the 4-month data analysis period).

Results

We contacted 44 potential participants of which we interviewed 15. Non-respondents were contacted once following the initial email invitation. We conducted all interviews between March 25 and June 25, 2020. Interviews lasted a median (IQR) of 21 (15.3, 26.1) minutes. Table 1 presents the participant characteristics. The participants worked across 14 different ICU’s. All but one participant reported that patients with a diagnosis of COVID-19 had been admitted to their ICU at the time of interview. One participant offered to participate in member checking of their transcript.

Table 1 Interview participant characteristics
Full size table

Analysis revealed three overarching themes and multiple associated subthemes. Overarching themes included: 1) Conditions contributing to resource strain (e.g., continuously evolving pandemic conditions), 2) Implications of resource strain on critical care physicians personally (e.g., safety concerns) and professionally (e.g., practice change), and 3) Enablers of resource sufficiency (e.g., adequate human resources) (Fig. 1).

Fig. 1
figure 1

Visual representation of themes and subthemes describing critical care physicians’ perceptions and experiences of caring for patients in the context of resource strain during the COVID-19 pandemic

Full size image

Overarching themes represent ideas that were largely echoed and emphasized by all participants in the study, while subthemes illustrate unique ideas related to a primary theme and provide a window into the diversity of participants’ perceptions and experiences. Subthemes were also shaped by the variability in pandemic preparations and COVID-19 case burden across ICUs and geographic regions. Quotations that exemplify themes and sub-themes are included in Tables 2, 3 and 4.

Table 2 Subthemes theme 1, Conditions contributing to resource strain descriptions and quotes, Participant ID_SexAge_Region_InstitutionType
Full size table
Table 3 Subthemes for theme 2, Implications of resource strain descriptions and quotes, Participant ID_SexAge_Region_InstitutionType
Full size table
Table 4 Subthemes for theme 3, Enablers to resource sufficiency descriptions and quotes, Participant ID_SexAge_Region_InstitutionType
Full size table

Conditions contributing to resource strain

Several participants described their ICU as experiencing resource strain (e.g., limited essential supplies) at the time of interviewing, while others felt that strain would occur imminently if the number of patients requiring ICU services continued to rapidly increase in their region. Participants described several factors contributing to strain in their ICU, including: 1) Continuously evolving pandemic conditions (e.g., varying resource supply and demand, unpredictable patient surges), 2) Limited equipment (e.g., rationing of PPE, inadequate physical space in the unit) and supplies (e.g., medication), and 3) Staffing shortages (e.g. physician illness, increased patient demand). Of note, a small group of participants were becoming increasingly concerned about the resource availability in their ICU as requests to accept transfer patients from ICUs that had already reached capacity began to increase (Table 3). Other participants were particularly concerned about the possibility of having to admit multiple patients simultaneously (i.e., in the case of a long-term care home outbreak) and the demand this would place on their ICU and existing pool of limited resources (Table 2). As demonstrated in these examples, participants often framed their discussions of resource strain in the context of uncertainty about whether their ICU would be able to meet rapidly increasing demands as the pandemic progressed. Overall, participants were fairly confident in the pandemic planning occurring in their ICU yet remained unsure as to whether this planning would lead to actual preparedness for what was to come (Table 2).

Implications of resource strain

All participants in our study described both the personal and professional implications of working in the context of resource strain during the COVID-19 pandemic. In particular, two main subthemes related to the personal toll of their work emerged from interviews: 1) Concerns for personal and familial safety (e.g., transmitting the virus to family members), and 2) Psychological impacts (e.g., increased anxiety). Most striking was the degree of concern for the safety of family members (Table 3). For example, some participants made the difficult decision to self-isolate away from their immediate family members by residing in a hotel to ensure that they would not bring the virus home, while others took great pains to limit their risk of being a contagion by segmenting their home into “hot and cold zones” (i.e., hot zones being the spaces and surfaces that the physician in the family would inhabit or touch, cold zones being those designated to family members). Although the level of concern regarding transmitting the virus to family members did vary across participants from mild to serious, all participants indicated that if they had to work without appropriate PPE they would be very concerned about familial transmission. The unpredictable trajectory of the pandemic had a psychological impact on participants. Feelings of uncertainty about how the pandemic would develop and whether their unit’s planning and supplies would be sufficient, created increased anxiety in many participants as they “waited for the disaster to happen”. – Participant 4.

In addition to the noted personal implications of caring for critically ill patients in the context of pandemic resource strain, we identified three subthemes reflecting perceived implications to the healthcare workforce and delivery of patient care (i.e., professional implications):: 1) Changes to physician practice (e.g., less direct patient contact, more clinical work than normal, innovative solutions to PPE shortages), 2) Deteriorating health of colleagues (e.g., decline in colleagues’ physical and/or mental health), and 3) Legal and ethical considerations of working within resource scarcity (e.g., moral dilemma of acting versus not acting). In particular, participants from all regions described changes to their normal clinical practice, specifically emphasizing less direct patient contact as hospitals shifted protocols to minimize the number of people entering the rooms of patients with known or suspected COVID-19 (Table 3).

Several participants also described spending more time than usual on clinical service, while others expressed worrying more about colleagues who they believed were at a high risk due to pre-existing conditions. When asked about legal and ethical considerations of working during a pandemic without adequate PPE, some participants mentioned that their unit had not had any conversations regarding the legal requirements, while others had discussed the matter, yet remained unsure of their legal responsibilities (Table 3). Some participants further investigated their legal obligations by contacting their professional governing body.

Enablers of resource sufficiency

All participants in this study identified factors that supported individual physicians and ICUs to effectively manage the care of critically ill patients with COVID-19. We clustered enablers into four interrelated subthemes: 1) Sufficient equipment and supplies (e.g., staff, ventilators, unit space, medication), 2) Comprehensive planning and communication (e.g., early pandemic preparation, clear and consistent PPE guidelines), 3) Adequate human resources (e.g., colleagues and personnel), and 4) Early pandemic preparation (e.g., surge planning) (Table 2).

Early pandemic preparation was a particularly salient factor to support resource sufficiency. Participants described several actions that their ICU leadership began executing early in the pandemic to plan for different surge and strain scenarios, such as, updated ICU plans (e.g., creating more ICU beds within and across hospitals), modified call schedules (e.g., creating backup call schedules), and redeployed staff from other units to the ICU to meet increased capacity demands. In addition, all participants mentioned that their units were considering solutions to mitigate and prepare for potential PPE shortages, including recycling, reusing and finding non-traditional PPE alternatives (e.g. painter’s masks, 3D printing face shields).

Similarly, participants across all regions indicated that the early implementation of clear and concise PPE guidelines was a crucial component of pandemic preparedness. Many also noted that while guidelines may need to shift in accordance with pandemic circumstances, keeping staff apprised of these changes in a clear and transparent manner (i.e., rapid knowledge mobilization and translation) was vital. At the same time, although there was consensus across our participant pool regarding the need for clear and concise COVID-19 guidelines, perspectives on what should be included in those guidelines to sustain high quality care and staff safety were more nuanced. Varied opinions were most noticeable in discussing what PPE should be used when entering the room of a patient with COVID-19. For example, participants generally agreed that full PPE (e.g., with N95 respirator) is needed in a closed-circuit intubated patient room because, as explained by Participant 11, “[there is] risk that at some point, the circuit becomes disconnected spontaneously, exposing everyone in the room [to the virus]”, but diverged on whether or not full PPE was necessary when entering the room of every patient confirmed to have COVID-19, regardless of medical procedure (Table 4).

Discussion

Caring for critically ill patients with a known or suspected novel infectious disease during a global pandemic is a complex task that requires multi-level (e.g., organizational, departmental, personal) planning and preparedness [48] to protect patients and health care workers [49]. The availability of resources is clearly an important mediator in successfully managing the care of critically ill patients in pandemic conditions [50]. Research during previous infectious disease outbreaks has shown that ample supplies of PPE significantly facilitated effective clinical care [51, 52], while insufficient or rapidly depleting PPE contributed to healthcare worker anxiety [53]. We conducted semi-structured interviews with critical care physicians in Canada to better understand their perceptions and experiences of caring for critically ill patients in the context of actual or anticipated resource strain during the COVID-19 pandemic. We identified three primary and interrelated themes: conditions contributing to resource strain, implications of resource strain on critical care physicians personally and professionally, and enablers of resource sufficiency.

When considering the management of essential resources required to care for patients with COVID-19, academic and mainstream media sources have been particularly concerned about adequate supplies of equipment (i.e., ventilators) and possible PPE shortages [29, 54, 55]. Our participants reflected similar concerns. In particular, the most salient and worrisome conditions that participants noted involved circumstances outside of their control (e.g., patient surges, supply chain shortages, etc.). At the same time, key enablers to support resource sufficiency in our study consisted of comprehensive planning and communication as well as early preparation. The issue of resource strain is likely further exacerbated by a healthcare system that already faced a growing demand for acute care with an aging population demographic [56]. Some participants described that their units had developed and enacted detailed plans to prepare for surge scenarios, including a tiered response of staffing and capacity depending on the number of patients with COVID-19 admitted to hospital. Other participants felt that their units were less prepared in that they were already experiencing shortages of PPE and were being presented with measures and mandates by their institutions to ration and conserve remaining supplies.

In addition to echoing the importance of adequate material resources, several of our participants also stressed the need for units to ensure that adequate numbers of properly trained personnel were available to care for incoming critically ill patients with COVID-19. In this respect, of particular concern to our participants were occurrences of large outbreaks at long term care facilities and outbreaks in neighbouring cities that would require hospitalization in their unit. Participants carried reports of global surges (where ICU bed need exceeded capacity leading to resource shortages and triage scenarios) [29] at the back of their minds, and largely did not know how their ICU’s pandemic planning would hold up in a similar scenario.

Our findings shed light on the importance of establishing ethical and legal parameters for healthcare workers as part of pandemic planning and preparedness [57]. In Canada, when a physician agrees to treat a patient, they have a legal duty to provide a certain standard of care [58]. In a state of emergency, individual provincial legislation may permit the provincial government to mandate physicians to perform certain services [59]. Yet, Canadian labor boards have outlined criteria to justify refusal to work which may be applicable during a pandemic [60]. Our participants highlighted the legal uncertainty surrounding physicians’ rights and obligations to continue to practice during a pandemic [59]. Some participants had conversations within their unit while others reached out to their governing bodies for guidance regarding the expectations of working as a physician during a pandemic and potentially without adequate PPE. This variability in awareness of the legal and ethical parameters for working during a pandemic highlights national discrepancies and suggests there are inadequate guidelines in Canada for what is expected of physicians during an infectious disease outbreak [59].

It is clear that the impacts of working during an infectious disease outbreak—particularly under conditions of real or anticipated resource strain—extend beyond the workplace setting. A recent systematic review highlighted the burden of mental health symptoms including anxiety, acute stress, depression and burnout among frontline healthcare workers during and following a disease outbreak [61]. Participants in our study experienced varying degrees of anxiety, particularly with respect to concerns for their family [62]. A small number of participants responded similarly to frontline workers around the world by finding alternate living accommodations (e.g., living in a hotel) to minimize exposure and risk to their families [63]. Other qualitative work indicates frontline healthcare workers in the China and the United States also dealt with anxiety about becoming infected themselves or family members becoming sick with COVID-19 [64, 65]. Other participants were much less concerned for the safety of their families, particularly their young children, potentially due to early reports indicating that the virus may not impact children as severely as adults [66, 67]. A general sense of anxiety and uncertainty was prominent throughout the interviews in regard to the pandemic placing the participants and their organizations in uncharted territory. National and provincial predictive models indicated first wave peaks that would induce resources shortages across the country [68,69,70]. These models may have positively impacted and motivated early pandemic preparations, yet they may have also contributed to the stress and anxiety of healthcare providers and the public alike. While Canadian ICUs have a similar number of beds to Western European countries [71] a comparison in terms of conditions contributing to and implications of resource strain in ICU during a pandemic warrant further exploration.

There are some limitations that should be considered when interpreting our findings. First, the perspectives shared by our participants may not be transferable across Canada, as the majority of our participants practiced at urban academic institutions. Smaller community, regional or rural centres may have had varying levels of preparation and numbers of patients with COVID-19 leading to differences in perceived or actual resource strain. Second, physicians who felt particularly well prepared or conversely, highly strained, in terms of resource shortages may have been more motivated to volunteer for an interview. We attempted to mitigate this by purposively sampling a diverse group of critical care physicians in provinces that had variable numbers of people affected with COVID-19. This was further supported by using telephone interviews which enabled participation outside of the researchers’ home locations. Historically, interviews have been conducted face-to-face as the frame of social interaction and cues (e.g., body language) fits better within this method [72]. However, there is a growing acceptability [73] to conduct telephone interviews and evidence suggests strengths over face-to-face interviews such as ability to concentrate on voice instead of face, the feeling of not being judged and easier rapport over the phone [74]. The use of multiple interviewers has the risk of generating different data due to different interactions and participant perceptions [75], to mitigate this each interviewer followed the same introductory script and wrote memos following each interview that informed analysis. We also kept the interviewers and analysts consistent between data collection and analysis. Non-probability purposive sampling is often critiqued for being ambiguous and subjective.

Conclusions

The COVID-19 pandemic has placed many of hospitals and healthcare providers in uncharted territory. This study investigated physicians’ perceptions and experiences of caring for critically ill patients in the context of resource strain, and identified contributing conditions, implications, and perceived enablers to resource sufficiency. Continuously evolving pandemic circumstances and a sense of uncertainty expressed by our participants highlight the importance of an organized national pandemic response plan for subsequent waves of COVID-19 and future pandemics.