1 - COVID-19 and Clinical Ethics

Summary

Kierkegaard famously said life is lived forward and understood backwards. What is true for life is also true for pandemics and it is tempting to look at the dawning days of COVID-19 through the prism of reflection. But to do so would be to lose the lived experience of those who had to make fraught choices during the initial surge of the pandemic. What happened during March and April 2020 should not be obscured by a post hoc analysis informed by what we learned about the SARS-COV-2 virus or about ourselves since those early days. That early history, unsanitized as it was, is essential to framing subsequent discourse. As an academic physician charged with coordinating a clinical ethics service at a major academic medical center during the spring surge in New York City, I can attest that what happened during the spring of 2020 in New York City bears remembrance, not only to honor those who served – and died – but just as importantly, to inform clinical care and public policy. Those lessons were hard won and it would be unfortunate to lose them through the gauzy haze of memory.

I Living History

Kierkegaard famously said life is lived forwards and understood backwards.¹ What is true for life is also true for pandemics and it is tempting to look at the dawning days of COVID-19 through the prism of reflection. But to do so would be to lose the lived experience of those who had to make fraught choices during the initial surge of the pandemic. What happened during March and April of 2020 should not be obscured by a post hoc analysis informed by what we learned about the SARS-COV-2 virus or about ourselves since those early days. That early history, unsanitized as it was, is essential to framing subsequent discourse. As an academic physician charged with coordinating a clinical ethics service at a major academic medical center during the spring surge in New York City, I can attest that what happened during the spring of 2020 in the city bears remembrance, not only to honor those who served – and died – but, just as importantly, to inform clinical care and public policy. Those lessons were hard-won and it would be unfortunate to lose them through the gauzy haze of memory.

II Hometown Heroes

A daily ritual evolved that spring when New Yorkers came to the street each night to cheer, and revere, their health care heroes.² At 7:00 pm, people would gather on their fire escapes and street corners to clap, clang cow bells, and otherwise cheer for those toiling away in the hospitals. Nietzsche might have viewed it as the bread and circuses moment of the pandemic.³ New Yorkers wanted to express their gratitude. But they also needed to do so. They were scared and frightened and viewing health care workers as heroic made them feel better, more secure. Having said this, not a single one of my colleagues felt we were deserving of the adulation. Clinical failure, with which we were unaccustomed, had become the norm in our intensive care units (ICUs), both the regular kind and the pop-up variety that took care of the overflow of patients. During the weeks of March 22 through April 4, the weekly percentage of hospitalized New York City patients who subsequently died peaked (mean = 36.4 percent; range = 33.5 to 38.2 percent).⁴ With these numbers, we did not feel like heroes.

We did not feel deserving, but the public needed to believe in us and those we represented because they had to believe in something. Civil society was in a state of chaos. The stores were bare, the supply chain broken. People were suddenly dying. But all would be well because they put their faith in their superheroes, who would rise up, rescue, and save them. Except we could not back then, although we tried. And some died – trying.

It was a valiant effort. In addition to meeting a novel disease head on, which reduced the most skilled clinicians to novice practitioners, they were charged with making triage decisions because we did not have adequate supplies – of, for example, personal protective equipment (PPE), drugs, and ventilators – and operated without crisis standards of care, which New York State failed to invoke. This placed clinicians under untenable stress.

One needs a tragedy to have heroes, and the situation was tragic for those who were lost, their families and loved ones, as well as the clinicians who were placed in a position where they had to do more than act as professionals. Professionalism should have been enough of an expectation; heroism was a bridge too far.⁵ It was one that ultimately collapsed, both in how society came to view the doctors and nurses who stepped up, and in the mental health sequelae of the pandemic, what Victor Dzau, the president of the National Academy of Medicine, has described as its own epidemic.⁶

So much of the tragedy we were grappling with in New York could have been avoided. Others have written about the Trump Administration’s downsizing of pandemic surveillance as a national security issue and the pulling of observers from Wuhan, China, the presumptive ground zero of the pandemic.⁷ But on a much more local level, hospitals suffered from the failure of New York State to promulgate crisis standards of care in response to the pandemic. It became clear that this failure was an abdication of governmental responsibility and of the state’s obligation to protect its citizenry.

III Antecedents and the Surge

New York had a huge head start on disaster planning. As far back as 2007, the New York Task Force on Life and the Law, on which I serve, began deliberating the question of ventilator allocation in the context of an Avian flu pandemic.⁸ Ultimately, the Task Force issued a final report in 2015 delineating a ventilator allocation guideline. It was one we never hoped we would use, and it was written with a hefty dose of denial,⁹ but we did write it all the same.

At the heart of the plan was the Sequential Organ Failure Assessment (SOFA) score, a methodology originally designed to physiologically assess a patient’s need for ventilatory support if ill from influenza and to predict their short-term survival from that infection by tracking the functional status of several organ systems.¹⁰ The triage process would be put into place once public health authorities declared a public health emergency and invoked crisis standards of care.¹¹ In New York this would be declared by the governor. Crisis standards of care would replace the “usual” standard of care with one that was “sufficient” given the circumstances.¹² But these guidelines were never put in place and, as a result, in March 2020, individual hospitals had to make determinations about the allocation of scarce resources without government guidance.

This lack of government action compounded our worst fears about the depth of the pandemic. The most elite academic health care systems were overwhelmed by the flood of patients who were desperately ill and in need of intensive care. In those dark days, even the most skilled practitioner became a student. We all became novices, grappling with a new disease we did not understand. The practice of medicine, which is so dependent upon time, was for a spell, atemporal.¹³ It is worth recalling that even the most basic temporal dimensions of the virus, such as its period of incubation, duration of quarantine, and time course of treatments, were unknown. Collectively we were at sea. Practicing medicine without a clock is much like sailing without a compass. We were lost and striving to find our way.

To add to these challenges was the urgent redeployment of clinicians to meet staffing needs. Under normal circumstances, ICUs are run by pulmonology and critical care medicine attendings, their fellows and residents on rotation. To respond to the pandemic, doctors all over the hospital were reassigned to unfamiliar venues. Hospitalists, who provide in-patient care on medical services, worked in the ICUs as pediatricians backfilled on the medical services.

Psychiatrists handled phone consults for the sick and worried, and wondered if they had COVID-19 and should be tested. It was unfamiliar terrain for even the most experienced of practitioners.

The need was staggering. At New York Presbyterian Weill Cornell Medical Center between March 16 and May 10, 2020, we had 1,550 COVID-19 admissions and we increased our ICU capacity from 100 to 230 beds. Our colleagues at New York Presbyterian Columbia University Medical Center had 2,000 COVID-19 admissions and went from 117 to 300 ICU beds during the same period.¹⁴

But these numbers do not adequately convey the tragedy of individual narratives. In my role chairing our hospital’s ethics committee, I recall a case of a patient who was nearing death from respiratory failure.¹⁵ Her closest relative, a sibling, was approached about consenting to a do-not-resuscitate (DNR) order given the likely futility of resuscitation should she have a cardiac arrest. The ethics consult service was called when the clinical team encountered resistance. When our service became involved and elucidated the facts of the case, we found the reason for the sibling’s reluctance: another family member had passed away earlier in the day from COVID-19. It was just too much.

While this vignette was an outlier, it spoke to the burden of illness and tragedy experienced by clinical staff. Elsewhere, I analogized the onslaught of patients to a plane crash at LaGuardia Airport, except that the influx of patients continued for weeks on end.¹⁶ Hospitals across the city had to deal with a scarcity of medical personnel and material, notably PPE and ventilators. People wore single-use/disposable N95 masks for weeks on end. A nurse at Mt. Sinai Hospital was seen in a New York Post photograph wearing a garbage bag for protection.¹⁷ That was the practice in many institutions. Unprepared for the onslaught, we had to improvise.

These shortages were the product of a just-in-time approach in which the presumption is that when there is a surge in need, resources will be readily available from the supply chain. This is cost-effective during normal times and avoids having to expend resources on supplies that sit unused in inventory. Some of these items may have a time-limited shelf-life, so having excess inventory represents a potential fiscal loss. A just-in-time approach works when there are isolated pockets of need and resources can be obtained expeditiously. It fails miserably when there is a sustained and systemic stressor or when the supply chain breaks. In response to the inadequacies of a just-in-time supply chain, hospitals are migrating to a just-in-case approach, which stockpiles resources and enhances staff preparedness training and readiness.¹⁸

But that would be a lesson from the pandemic. During the spring of 2020, we had to innovate to survive. To that end, hospitals created new ICUs out of thin air. Elective surgeries were canceled. Operating rooms were redeployed to provide ICU care and ventilators were reconfigured to provide support to two patients at a time.¹⁹ Pop-up ICUs were built in hospital lobbies and football fields.²⁰ Parks were converted to field hospitals as the USS Comfort,²¹ the Navy’s hospital ship, made port in New York Harbor.²²

All of this was done to provide care to an explosive volume of patients who had a novel disease that we did not yet fully understand. In those early days, we had no effective therapies. We were treating patients empirically with pharmaceutical agents that would not be found to be efficacious (hydroxychloroquine and convalescent serum). In the beginning, we eschewed high-dose steroids, which were later found to be effective and a true game changer with respect to mortality.²³ We were months away from monoclonal antibodies, much less the miraculous mRNA vaccines that would hold so much promise.²⁴

At one point in our institution, we were down to three days of dialysis fluid because so many patients with COVID-19 developed renal failure when critically ill. This was wholly unexpected since COVID-19 was initially believed to be a respiratory disease. We soon learned that COVID-19 also caused vasculitis and renal failure. This epidemic of kidney disease within the broader pandemic quickly led to a shortage of available supplies of dialysate, a problem that first manifested itself institutionally when the Renal Service asked for an ethics consult. They wanted guidance about who should get dialyzed for kidney failure and the quality of that intervention given the short supplies.

This type of question was new for ethics consultation. Unlike a traditional Clinical Ethics Consultation, which centers around the care of individual patients, this question required us to think about groups of patients who would receive care on a particular unit or clinical service. Here, the group was comprised of all those patients who might need dialysis and make a claim on scarce resources. In this scenario, the choice was to use scarce resources selectively and fully dialyze those patients we thought most likely to survive. This would provide the usual standard of care to a select few. The alternative was performing sub-optimal dialysis for a larger number of patients in the hope of temporizing until more dialysate could be secured. We recommended the second course of action.

In our published analysis of ethics consultation performed during the spring surge,²⁵ the first papers published on ethics consultation during the pandemic in the United States, we dubbed this collective consult as a Service Practice Communications/Intervention (SPCI).²⁶ This was a second level of ethical analysis pertaining to groups of patients rather than individuals. As such, it was a new epistemic category of consultation prompted by the exigencies of the pandemic and the need to think in utilitarian terms rather than the deontological ones that inform care under normal circumstances. In addition to SPCIs, we also provided normative advice to the institution under the guise of what we described as an Organizational Ethics Advisement (OEA).²⁷ OEAs were advice to leadership at the hospital, university, and system levels, mediating disputes between units, and real-time education about developments in New York State law that might have a bearing on patient care.

Based on an historical review of our case notes, phone logs, and emails, we estimated conservatively that we performed 2,500 SPCIs and OEAs during the six-week period we studied.²⁸ This was an unprecedented amount of activity, prompted in part by the clinical and regulatory contingency of the situation during the spring surge. While the biology of the coronavirus was a force of nature, the lack of state guidance with respect to crisis standards of care was a consequence of human nature. It would turn out that human nature was more difficult to control than the virus: New York’s inability to provide direction to clinicians and institutions during the crisis constituted an abdication of leadership.²⁹

IV Abdication

In failing to promulgate crisis standards of care, the state stepped away from its responsibilities and failed to fulfill its duty during a crisis. This was especially disheartening to clinicians. In addition to grappling with an unfamiliar disease, they were forced to make ethical choices that turned usual presumptions of care on their head amidst the utilitarian demands of a public health emergency. They felt vulnerable to retrospective critique and legal liability because the regulatory context did not adequately acknowledge this need.³⁰ In the back of our minds was the experience of doctors following Hurricane Katrina and the question of whether physicians would be vulnerable to prosecution after the crisis had passed.³¹

While the governor did insert the Emergency or Disaster Treatment Protection Act of 2020 into the state budget, which mitigated some concerns about professional liability, the provisions of the Act were ambiguous, at least as viewed by those of us in the clinical community, insofar as it related to questions of resuscitation and the provision of critical care.³² It did nothing to coordinate services across the state and bring resources to underserved areas, an area of omission that would become more glaring as the pandemic wore on.

At the start of the pandemic, the concerns were more localized to hospital care. Many clinicians wanted more specificity about triage and decisions about unilateral resuscitation. Might it be possible to set limits in the face of three pressing issues: (1) overwhelming scarcity; (2) the futility of resuscitation given advanced disease; and (3) the risk of contagion to providers who sought to revive patients. Let us take these three issues in turn.

With respect to scarcity, at the peak of the surge it was not uncommon for multiple patients to be in imminent need of intubation at the same time. One anesthesiologist reported that it was not uncommon for him to intubate more than ten patients on a single shift.³³ Under normal conditions this was unheard of, and it placed a strain on the ability to provide timely care as would have been the case pre-pandemic. Some of the patients who were being intubated were certain to die and yet, absent any change in New York State law regarding resuscitation, there was no way to unilaterally write DNR orders if consent could not be obtained from surrogates.³⁴

Turning to the question of futility, at that early juncture during the pandemic, patients who had a cardio-pulmonary arrest invariably died. The initial survivor data from Wuhan reported that only 2.9 percent survived.³⁵ An audit of ICU care in the United Kingdom reported in early April 2020 was similarly dire. The thirty-day mortality of patients who were ventilated in intensive care was 49.9 percent for all patients and 68.1 percent for those 70 and older.³⁶ So, the issue was one of futility and the utility of these interventions, an inter-relationship that is often overlooked when we consider medical interventions.³⁷

Finally, there was the question of proportionality and the burdens-to-benefits ratio associated with resuscitation.³⁸ As just noted, the benefit at that juncture was fleetingly low, approaching zero. In contrast, the risks were exceedingly high for practitioners, who would be exposed to aerosolized secretions during emergent intubations and cardio-pulmonary resuscitation (CPR). This would expose clinicians to the risk of contagion, a danger compounded by the inadequate availability of PPE at the time and the fact that those who were performing intubations might not have been as skilled as those normally called upon to perform that task prior to the pandemic. The risk of contagion was quite real. Spain, which was a few weeks ahead of New York with respect to the course of the pandemic, reported that 18.5 percent of health care workers contracted the coronavirus.³⁹ This combination of factors, very low benefit coupled with high risk, made the procedure ethically unbalanced given risks outweighed benefits: low benefit to the patient against the risk of contagion to staff. This combination of factors made resuscitation disproportionate.

V For Want of Crisis Standards of Care

Scarcity, futility, and contagion would seem to argue for a change in resuscitation policy in New York State and the promulgation of some sort of triage mechanism along the lines of the 2015 Ventilator Report utilizing the SOFA mechanism. This seemed to be where the New York State Department of Health (DOH) was headed in late March. Despite the granularity of these conversations, the Department neither promulgated guidelines nor explained their decision not to act – either then or since.

The possible reasons why guidelines were not offered by the DOH are complex and multivariate, and one can only speculate. Politically, it was easier to build up capacity, as they did, rather than admit that resources were limited, as later documented in a report from the US Department of Health and Human Services Office of the Inspector General.⁴⁰

The most objective reason that guidelines were not offered was likely scientific and a question of the prognostic utility of the SOFA methodology with respect to COVID-19 respiratory failure. The 2015 report and the SOFA methodology were designed to respond to an Avian flu pandemic, not COVID-19. Even weeks into the coronavirus pandemic, it became clear that the SOFA methodology was not a perfect fit: The Avian flu had a much quicker time course than did patients with COVID-19. For example, in the context of the Avian flu, patients triaged to one of the SOFA color categories were to be reevaluated at two- to three-day intervals. This was far too soon to evaluate patients with respiratory failure from COVID-19, who could take a month to recover and get off a ventilator.

In our discussions, this prompted the perennial adage of not letting perfect be the enemy of good. Could the SOFA methodology be modified to respond to this emerging disease or would such a modification make the use of SOFA a non-evidence-based approach to ventilator allocation? Or would the use of a modified SOFA methodology be better than an ad hoc approach to rationing ventilators? When the DOH convened in March to discuss this question, the trend was toward modification of intervals with an eye toward data collection and further iteration in order to retrofit the methodology for the current pandemic. Whether or not this would be possible is still a question for debate, with the literature suggesting both the ability and inability of SOFA scores to risk stratify and predict mortality from acute respiratory failure from COVID-19.⁴¹

Another salient objection was that triage policies could be discriminatory. This is a serious objection but my view at the time was that the greater threat was unregulated triage, rationing, or priority setting, in which bias would sneak in without the proper oversight and accountability afforded by the law. Explicit guidelines could better protect the civil rights of people with disabilities, something that was in fact at risk during the pandemic, which saw the thirtieth anniversary of the Americans with Disabilities Act. Crisis standards of care, if properly conceptualized and implemented, could (and should) incorporate disability rights as part of any normative and legal framework. People with disabilities would be better off with transparent crisis standards of care that are properly designed and regulated than ad hoc decision-making, which could be discriminatory.

In a more recent analysis, I have argued that the SOFA methodology was indeed flawed when it came to the assessment of people with severe brain injury, but that is getting ahead of what I knew at the time.⁴² I have also worried that altering hard-won patient prerogatives about decision-making at the end of life could result in resurgent paternalism and an erosion of norms that we would come to regret.⁴³ Finally, there is an emerging literature on racism and SOFA scores. As one example, Tolchin and colleagues subsequently analyzed SOFA scores in non-Hispanic Black and Hispanic patients hospitalized in Yale New Haven Health System from March 29 to August 1, 2020. They found that non-Hispanic Black patients had greater odds of having a SOFA score greater than or equal to 6 when compared with non-Hispanic White patients.⁴⁴ But all of this was later, after the surge and in moments of quiet reflection.

VI Social Justice and the Clinic

Whatever the reason for the failure to promulgate crisis standards of care, the consequences were significant. Without this guidance, clinicians and institutions were left to their own devices to make judgments unilaterally. This placed practitioners under tremendous stress. The stress, however, was not equally distributed. It disproportionately burdened clinicians and communities in underserved areas.

By way of an example is correspondence with a physician-ethicist practicing in a hospital in the Bronx.⁴⁵ In a brave post on a national bioethics listserv at the height of the pandemic in New York, he wrote of the desperate situation on the ground. Overwhelmed by patients and by an inability to meet their needs, he wrote of unilateral decisions to withhold and withdraw life-sustaining therapy made at his city hospital.

He reported that, “we have mostly stopped performing CPR (notwithstanding absence of a DNR order) in cases where there was no chance of survival even with CPR.” He invoked the 2015 Task Force Report for guidance to withdraw life-sustaining therapy, guidance that – absent DOH crisis standards of care – remained an advisory document without the force of law. He concluded his email with an ethical justification: “Arguably, these withdrawals were acknowledgment of reality, not a true triage.”⁴⁶ It was unfortunate that a lone clinician had to be placed in this predicament, with its normative burden and associated liability risk.

His professional challenge as a practitioner was reflective of the broader community he served. At that juncture, society was just beginning to comprehend the disproportionate burden that communities of color experienced from COVID-19. A remarkable research letter in JAMA in late April 2020 would report that the Bronx had the city’s highest COVID-19 morbidity and mortality due to long-standing health inequities, poverty, dense housing, and a disproportionate number of essential workers.⁴⁷

Here, the gritty experience of the clinic, what Foucault called the “medical gaze,”⁴⁸ is illustrative of broader social forces: the lack of preparation for the coming plague and the endemic health disparities that compounded its consequences. It should not have taken COVID-19 for us to have been prepared, or to recognize and respond to, health inequity. If this lived experience fails to inspire a concerted response, it will only compound the tragedy of the pandemic. There are many lessons to be learned from this history. We must heed these lessons lest history repeat itself when the next pandemic hits, as it surely will.