US Clinicians’ Experiences and Perspectives on Resource Limitation and Patient Care During the COVID-19 Pandemic

Key Points Question How have US clinicians planned for and responded to resource limitation during the coronavirus disease 2019 pandemic? Findings This qualitative study included interviews with 61 clinicians across the United States. While institutions planned for an explicit and systematic approach to resource allocation in crisis settings, this approach did not address many challenges encountered by frontline clinicians, leaving them to struggle with what constituted acceptable standards of care and to make difficult allocation decisions. Meaning The findings of this study suggest that expanding the scope of institutional planning to address a broader spectrum of resource limitation may help to support clinicians, promote equity, and optimize care during the pandemic.


Introduction
Since the first US case of coronavirus disease 2019  was diagnosed in mid-January 2020, 1 the pandemic has completely transformed health care delivery in this country. Early reports from frontline clinicians in global epicenters describing extreme shortages and bedside rationing of ventilators and intensive care unit (ICU) beds 2 prompted a national conversation about how to respond to similar challenges in the United States. 3 Hospitals and health care systems drew on frameworks developed by the Institute of Medicine (IOM) and other national organizations to guide care in resource-limited emergency settings. [4][5][6][7] Under the IOM's framework, resource allocation is intentionally siloed from other aspects of clinical care to ensure a fair process and spare frontline clinicians from the responsibility of having to ration scarce resources at the bedside. The IOM recommends a phased adaptation to resource limitation. Institutions first shift from conventional to contingency capacity, in which resources are adapted, optimized, and redistributed to maintain a standard of care that is functionally equivalent to usual care. If resources become so limited that a functionally equivalent standard of care can no longer be sustained, institutions then shift to crisis capacity, and care is redirected to provide the greatest aggregate benefit to the population. 8 Under crisis standards of care, a specialized triage team becomes responsible for rationing scarce resources and making decisions about which patients will and will not receive potentially life-saving treatments.
Although the IOM framework reflects lessons learned during earlier pandemics 9,10 and has been iteratively refined through ethical analysis 8 and community deliberation, 11,12 there have been few opportunities to test the framework in real-world clinical settings. Reports from past pandemics 13,14 and from early global epicenters of the COVID-19 pandemic, 15 press reports, 16,17 and perspectives published in the scientific literature 18,19 describe some of the challenges faced by frontline clinicians.
However, there is little empirical work describing the experiences and perspectives of US clinicians pertaining to resource limitation and clinical care during the COVID-19 pandemic. 20,21 To address this knowledge gap, we conducted a qualitative analysis of interviews with US clinicians during the pandemic.

Participants
We recruited clinicians from across the United States who had cared for patients during the COVID-19 pandemic and/or had been involved in planning institutional responses to resource limitation. We used purposive sampling to select a group of participants with a variety of roles who were practicing in a range of different settings. We began by recruiting clinicians with direct experience planning for and/or practicing in settings of resource limitation (intensivists, nephrologists, and triage team members) at our own institution (the University of Washington). We then expanded recruitment to include other groups of clinicians (eg, trainees, palliative care specialists, nurse care coordinators) and those practicing in other parts of the country. We used a snowballing recruitment strategy in which we asked participants to identify colleagues with relevant experience who might be interested in participating in the study. We did not exclude participants who had collegiate relationships with members of the research team. Interviews were conducted between April 9, 2020, and May 26, 2020. The University of Washington institutional review board approved this study and authorized verbal in lieu of written consent. Verbal consent was obtained from all participants. We report details of our methods using the Consolidated Criteria for Reporting Qualitative Research (COREQ) reporting guideline (eTable 1 in the Supplement). 22 bioethics). All but 1 interview, which included 2 participants at their request, were conducted 1-on-1, and 2 interviews were spread over 2 sittings. A semistructured interview guide (eTable 2 in the Supplement) was developed by 3 of us (C.R.B., A.M.O., and S.P.Y.W; A.M.O. and S.P.Y.W. are academic nephrologists and physician scientists with experience in qualitative methodology, geriatric nephrology, and palliative care) and included open-ended questions to elicit clinicians' perspectives and experiences pertaining to clinical care, institutional policies, and resource limitation during the pandemic. The interview guide was iteratively refined by 1 of us (C.R.B.) with input from 2 of us (A.M.O. and S.P.Y.W.) to allow for elaboration of emerging themes. Interviews were recorded and transcribed verbatim. To protect confidentiality, participants were offered the opportunity to review their written transcripts to confirm accuracy and identify passages that they did not want published, but they were not invited to review draft or final versions of the article. Participants also completed an online survey with questions about their demographic characteristics and practice experience.
The size of participants' home hospital (or the hospital at which they volunteered, if that was the focus of the interview) was ascertained by online search.

Statistical Analysis
We conducted an inductive thematic analysis 23 of interview transcripts with the goal of discovering emergent themes describing clinicians' perspectives, experiences, and practices pertaining to resource limitation during the COVID-19 pandemic. Two of us (C.R.B. and A.M.O.) independently reviewed and openly coded transcripts until reaching thematic saturation (ie, the point at which no new concepts were identified), 24,25 which occurred after reviewing 30 interview transcripts. One of us (C.R.B.) coded all remaining transcripts for concurrence. Throughout the analysis, the 2 investigators iteratively reviewed codes, collapsed codes into groups with related meanings and relationships, and developed larger thematic categories, returning frequently to the transcripts to ensure that emergent themes were well-grounded in the data. 23,25 All coauthors (including A.G.W., a pediatric nephrologist, physician scientist, and bioethicist) reviewed example quotations and themes and together developed the final thematic schema. We used Atlas.ti version 8 (Scientific Software Development GmbH) to organize and store text and codes.

Results
We approached a total of 97 clinicians by email, of whom 75 (77%) agreed to participate and 22 (23%) declined or did not respond to our inquiry. Among those who agreed to participate, we purposively sampled 61 clinicians to participate in interviews from April 9, 2020, to May 26, 2020 ( Exemplar quotations are referenced in parenthesis and listed in Table 2, Table 3, and Table 4.

Planning for Crisis Capacity
Institutional leaders who participated in planning for crisis capacity described the challenges of adapting and operationalizing existing guidelines as well as the substantial moral weight of the task.
They were relieved when it became clear that these processes would likely not be needed at their institutions ( Table 2).

S Pacific
Some of the protocols have already been developed from the regional disaster planning[,] … but they're pretty broad, and how those are going to be actualized are part of the discussion.

V Pacific
They had a real problem with including any sort of age-based criteria in any guidelines because they thought it was using age as a social worth determinant …

Developing Allocation Algorithms
Clinicians who were involved in institutional planning described strong institutional support for their work to develop protocols to guide care should their region reach crisis capacity (quotation 1). They believed that establishing protocols in advance would allow for a more carefully considered approach (quotation 2) and would be reassuring to staff (quotation 3).
Their work involved developing actionable triage algorithms based on existing frameworks

Triage Team Members
Clinicians who had been appointed to triage teams were usually respected leaders in intensive care, palliative care, or bioethics who were recognized for their ability to collaborate and communicate (quotation 16). Some of the intensivists and ethicists with whom we spoke saw the work of the triage team as an extension of their usual work (quotations 17 and 18), but many clinicians saw this experience as entirely new (quotation 19). Clinicians described being motivated to participate in the triage team out of a sense of duty and desire to contribute (quotation 20) but were also cognizant of the moral weight and emotional burden of the task before them (quotation 21).

Relief that Crisis Capacity Had Been Averted
Clinicians involved in triage planning understood the processes they were developing to be intended exclusively for crisis settings (quotations 22 and 23) and saw the importance of optimizing resources under contingency capacity to avoid having to resort to crisis standards of care (quotation 24). While several clinicians described a period of intense planning early in the pandemic, by the time of our interviews, many were relieved to report that crisis standards of care were unlikely to be invoked at their institutions, and some had paused or disengaged from triage planning (quotations 25 and 26).

Adapting to Resource Limitation
Clinicians working during the pandemic were forced to grapple with multiple expected and unexpected forms of resource limitation that did not rise to the level of triggering crisis capacity. There was real urgency when we first started; we wanted to have a tool within 2 or 3 weeks. But then as the surge kept being delayed, and with social distancing, our surge never really materialized to any great extent. … As things went along, we decided we would probably never use this.

M South
We were able to get dialysis to everyone who needed it, but I didn't feel like we were necessarily able to provide enough of a dose of dialysis to make a meaningful contribution to their medical care. We were basically just keeping them hanging on by a thread over the course of the weekend.

M South
We went through the entire list at the beginning of the week and [said], this person has to dialyze these days, this person would probably benefit from a dialysis session, a third group person we could probably just string along and medically manage if we needed to.

R Northeast
No one was not getting dialysis, but there were a lot of people getting minimal dialysis. … Even though people were getting treated, resources were very stretched, and we delayed starting until our hand was forced. … Should you really wait for the potassium to get threateningly high? Probably not.

I Northeast
Two-hour treatments for people with a BUN of 250, you don't bat an eye at that stuff. It's like that's fine, the other person needs it too, or whatever. It's just because they're so many. Everybody gets a little bit of bad care.

A Pacific
Severe ARDS and prone and on pressers. They're all critically ill. There's no "can we make space?" That wasn't going to be a possibility. You can't take what under normal circumstances would receive 1-on-1 nursing care on a ventilator and say, "No, let's space it out to 2-to-1, or 3-to-1, and also give them a travel vent [ilator]." That's not a thing and not something we were willing to do.

Q Northeast
We were happy to be able to offer something. That was a positive. As I said, acute PD, we weren't sure how successful it was going to be, it did allow us to offer something for a period of time.
Distinguishing clinical care from rationing 44 F Northeast When you cross that line and say that you're rationing care, you have done something that is potentially taboo. And that is going to be in the newspapers in a completely different way.

AA South
Under normal times we would've been a little more aggressive with saying we're not going to try to keep doing this because it's not working, and they're not getting better. But I think that because of the sensitivity, the concern that people are going to be withholding care and this institution doesn't want to be seen like that, as a whole we were less likely to have those conversations. Nevertheless, these limitations could compromise care, require that they make difficult allocation decisions, and engender moral distress (Table 3).

Limited Institutional Response to Resource Limitation
Although none of the clinicians with whom we spoke reported a shift to crisis capacity at their institutions, they nevertheless described being faced with a range of expected and unexpected forms of resource limitation (eg, dialysis machines, staff, routine supplies) (quotation 27) that could arise in a haphazard manner with little warning (quotation 28). Some expressed frustration that resource shortages they were seeing in practice were not acknowledged as such by hospital or regional leadership (quotations 29 and 30) or felt unsupported by colleagues at neighboring medical centers (quotation 31). When not available from their institutions, some clinicians resorted to obtaining health care equipment through personal contacts or even fabricated it themselves (quotations 32 and 33).

Redefining Standards of Care
Clinicians were strongly motivated to avoid situations in which they would have to categorically deny needed treatment to any patient (quotation 34) and went to great lengths to develop alternative treatment options (quotation 35). This might involve using unorthodox therapies or nontraditional approaches to care delivery that could be suboptimal or potentially harmful (quotations 36 and 37).

D Pacific
People had been shamed for wearing masks a few weeks ago, and then I wondered if it was some kind of, "I'm not going to use PPE," like, it was just for weak people. I'm not sure. But I was really shocked. … They were all sitting around talking, and I walked by with a mask, and it almost seemed like they kind of looked at me funny.

EE
Pacific I call her [a patient's wife] and say, "Unfortunately, he's on 100%, I would have to put him on the ventilator." And she says, "Absolutely not, he's not in distress." So, she can't see him. … I said to her, "But I'm an expert! I can tell you he needs to be intubated." And she says, "No, he's not in distress." … Normally having families there, they see how often you're in that room trying to take care. I think I was in that room for hours and hours that day, and so they build that trust. Rapid pace of change and uncertainty 67 L Northeast I was getting multiple emails, multiple times a day regarding best practices and various new articles that were coming out.
… This is what we're going to be doing for high-flow nasal canula, and noninvasive positive pressure, and now we're going to reuse our N-95s. So just lots of things rapidly evolving.

C Pacific
There's just constant stuff, right? There's news from medical journals, news from reports from other cities and what their experiences have been. There's projections upon projections upon projections. And I think all of that amounts to this incredible torrent.

L Northeast
[The intensivists] were happy with any kind of unproven therapy, even if there were risks. I think they were desperate to also give families good news, even though things looked grim. There was just a lot of desperation to keep patients alive, probably because they'd been so traumatized by patients unexpectedly dying.

74
AA South I was actually expecting having to do some deescalation and some heated discussions. … [I] explained to the families also that it wasn't just to protect the patient's comfort and to not do something for them that wouldn't be beneficial, but also for the medical providers who would have to be in the high risk situation like a code. … The families were quite receptive to that and felt that they didn't want to be putting health care workers at risk either.

75
M South I didn't bring up resource limitation on the phone; that's not appropriate with a patient's family, but I think they sensed that was going on. Somehow they picked up on that, and they got very upset with my suggestion that maybe we forego dialysis knowing that his mortality was very high. Like, it just sucked, because in general I feel like I do these conversations pretty well, and partially, looking back on it, I think in part it was maybe my own anxiety around the situation, the conversation. I remember the lady saying I sounded "rushed[,]" … sounded "detached." … I remember being very ashamed about the conversation.

76
O Northeast I kind of had a deal with telling him, "Listen, you're too stable for dialysis right now," even though if not in a pandemic he would've gotten dialyzed 2 days ago and today, but that we're pushing him again. He was just extremely frustrated. He said, "This is crazy, what is this?" He was like mad at me personally. All I could do was try to explain our perspective here, that we're overwhelmed. This is the situation we're in. This is an international, this is a pandemic. … I don't think it means much to someone who's supposed to get dialysis and they're short of breath, being denied the treatment they need.

G Northeast
There was so much that was unknown about the trajectory, and also, … we didn't know, are we going to run out of dialysis fluid at some point? … We left it as, their loved one was very sick, and it needed to be a day-by-day conversation with the ICU team. Just to make sure that we weren't causing more harm or difficulties than benefit with them. I think that's more how we left it.
For example, some nephrologists described triaging patients for hemodialysis based on immediate need (quotations 38 and 39), delaying dialysis until there was an emergent indication (quotation 40), and/or prescribing shorter treatment times. As 1 clinician explained, "everyone gets a little bit of bad care" (quotation 41). Rarely, clinicians were able to draw a clear line between acceptable and unacceptable care (quotation 42), but most focused on doing the best they could under the circumstances (quotation 43).

Distinguishing Clinical Care From Rationing
The notion of rationing generally had negative connotations (quotation 44), and some clinicians even

Moral Distress
Many clinicians were fearful of having to ration resources (quotation 55). Even in lower acuity and outpatient settings, some struggled with whether it was acceptable to provide suboptimal care (quotation 56) and worried about the potential harms of disrupted care practices (quotation 57). A sense of responsibility for poor outcomes could take a substantial emotional toll (quotation 58).
Some clinicians felt that explicit guidelines would be helpful in limiting this moral distress (quotation 59), while others felt personally responsible for poor outcomes regardless of whether they were adhering to institutional recommendations or requirements (quotation 60).

Multiple Unprecedented Barriers to Care Delivery
Clinicians described multiple barriers to care delivery during the pandemic. These challenges compounded and were difficult to disentangle from the effects of resource limitation (Table 4).

Contact Limitation
Policies and practices were modified to limit physical interaction between staff, patients, and family members with the dual goals of reducing viral spread and conserving personal protective equipment.
While necessary, policies to limit contact with patients were seen as being detrimental to care and to

Discussions With Families About Disrupted Care
Many clinicians commented that families and patients were often quite understanding of care disruption (quotation 72), and that they were surprised at how accepting some families could be when care was compromised by the need to conserve resources and/or protect clinicians (quotations 73 and 74). However, some clinicians did describe contentious interactions with family members (quotation 75) and patients (quotation 76) who felt that care was being inappropriately withheld.
Other clinicians described deliberately avoiding mention of resource limitation when talking with families (quotation 77).

Discussion
Our thematic analysis of interviews with US clinicians who were directly involved in patient care and/or strategic planning during the COVID-19 pandemic highlights the real-world complexity of adaptation to resource limitation. Clinicians described patterns of institutional planning that mirrored the IOM's phased approach, which assumes a common understanding about what constitutes usual standards of care with a plan for a coordinated regional response when these become untenable.
However, consistent with prior anecdotal reports, 27 other factors that compounded the association of resource limitation with quality of care during the pandemic included the need to limit contact between clinicians and restrict visitors, the rapid pace of change, and the lack of scientific evidence.
Existing frameworks guiding institutional approaches to resource allocation in crisis settings [4][5][6][7] represent an important step toward promoting fairness through transparent allocation processes. 36 However, our results suggest that a narrow focus on crisis capacity may fail to address the full spectrum and complexity of challenges to providing high-quality care encountered by frontline clinicians during the COVID-19 pandemic (Figure). Insights gained from our study suggest several strategies to better support clinicians and guide clinical care as the pandemic evolves. First, regional authorities and institutions could prepare guidelines, protocols, and defined standards of care in advance to address the myriad types of resource limitation and challenges to providing high-quality care that arise long before declaration of crisis capacity. Second, in addition to their narrow role in allocating scarce resources under crisis capacity, triage team members could collaborate and/or consult with frontline clinicians to address challenges related to more nuanced forms of resource limitation. 37,38 Finally, as the community moves beyond the current crisis and plans for the future, training in bioethics may help to support clinical teams in navigating the value conflicts that can arise when resources are limited in both pandemic and usual care settings. 39,40 Limitations This study has limitations. Our results may not capture the experiences and perspectives of clinicians practicing in other parts of the world, in regions of the United States not represented in our study, or in specialties not represented by our participants (eg, pediatrics). 41 We also did not collect information about the COVID-19 caseload at each participant's institution at the time of the interview, which may have shaped participants' experiences and perspectives. Because resource allocation can be a sensitive topic, participants may have felt limited in discussing some aspects of their experiences. Furthermore, the dynamic nature of the pandemic makes it likely that new challenges not identified in our study will arise over time and practices may evolve in response to early experience.

Conclusions
In this qualitative study, many clinicians described institutional planning for crisis capacity, but this did not always address real-world challenges to providing care when resources were limited.
Expanding the scope of institutional planning beyond crisis capacity may be helpful in supporting clinicians and addressing moral distress, promoting equity, and optimizing care as the pandemic evolves.