As coronavirus disease 2019 (COVID-19) continues its global spread, we are faced with the very real problem of too many patients and not enough ventilators. In this issue of Critical Care Medicine, Sprung et al (1) consider the worst of these fears and attempt to provide guidance on critical care resource utilization. COVID-19 is a novel disease with an uncertain natural history (2). As a consequence of a lack of testing, the true case fatality rate cannot be known. In a worst-case scenario, mathematical modeling predicts illness and death of millions of people worldwide. Such an influx of patients over a relatively short period of time would behave like a mass casualty event and overwhelm healthcare and ICU resources. Sprung et al (1) describe the development of a triage model to manage expected ICU excess patient capacity under these circumstances. Their model draws from a review of scientific literature on the subject of epidemic ICU triage found within PubMed and Medline, relevant statements from other professional medical societies, and the opinions of the article’s authors.
The bottom line here is the creation of a rationing plan that will select some people for death who otherwise would not have died. Rationing as described will not save lives overall. Instead, it allows the rationing team to select who will live and who will die. In order to make this exceedingly troublesome idea palatable, a grading system is proposed that incorporates coexisting illnesses and age to generate a priority scale. The banality of the accounting is meant to engender confidence and to make acceptable a plan that, at its base, calls for the direct killing of patients without their consent.
Rationing scenarios long have been considered in bioethical thought through examples like the classical trolley problem (3). In the trolley problem, we consider killing more people or less by drawing on either a rule-based model or a utilitarian model. Sprung et al (1) make a utilitarian argument when they advocate for saving the maximum number of life years as opposed to the traditional first come first serve system that aims to save the maximum number of people. The article by Sprung et al (1) departs from classical utilitarian reasoning by suggesting that the benefits or pleasures of specific people are incommensurable because they differ in terms of the potential number of years they are likely to live.
Conversations about resource allocation are reasonable and currently necessary. Resource-poor ICU care is new for many western medical practitioners and citizens of western countries, and developing rational and ethical strategies before a worst-case scenario plays out has clear merit. Consider the controversy surrounding rationing choices that took place at Memorial Medical Center in New Orleans during Hurricane Katrina in 2005 (4). Protocol driven care in a variety of clinical situations may have the advantage of offsetting mistakes in reasoning that can occur under duress (5). Further, the article makes the fair point that certain forms of care allocation that might be construed as rationing have been in place prepandemic. Likewise, certain patients may not be eligible for ICU care as a consequence of the concern that the care itself would not reasonably result in the patient-desired outcome. Ethically, we accept when a patient refuses care, even if that refusal leads to death.
But while the article fairly identifies an important subject for our collective review, on the whole, the concept of the triage committee has disturbing flaws. The authors claim that a triage system will ensure fairness. The concept of fairness is problematic and is not necessarily applicable. Further, they argue that a triage system will provide “enhanced consistency.” Consistency is only a benefit if the triage system itself is proven to be workable, moral, and lawful.
The claim is made that clinicians may experience moral distress when making triage decisions. No data support is offered for this claim, nor for the claim that an independent triage team would reduce moral distress. This argument is problematic as disagreement exists as to the meaning of futility (6) and when we refuse to treat a patient on the basis of our claim of moral distress, we violate patient autonomy and our actions are arguably maleficent. The authors state that the triage algorithm will provide psychologic support without clarifying what sort of psychologic impairment is present and how that would be mitigated by this algorithm. Likewise, no data support the argument that only experienced critical care doctors should have the last word on triage. What sort of experience will be necessary? If the triage decision is ultimately up to the experienced critical care doctor, why create a separate triage team at all?
Fundamentally, the concept of greater benefit breaks down as the principle of promoting life-years lived is clearly cover for ageism. In the trolley problem, utilitarian analysis holds that five lives saved are better than one. The value of any individual life cannot be shown to be more valuable than any other life in a straightforward and agreed-upon fashion. Simply using potential life years or claiming that an older person is consuming life they no longer deserve (7) unreasonably places youthful vitality over gritty experience. Many older individuals are now involved in guiding a way forward from this pandemic catastrophe. Allowing older individuals to die would remove the wisdom that such a moment in history desperately requires. Although COVID-19 has been seen broadly in the population, the worst afflicted are generally older patients with coexisting health problems (8). In this sense, severe acute respiratory syndrome coronavirus 2 does not have an equal chance at sickening all members of society. The occurrence of COVID-19 in the population is not “fair” and therefore employing a fairness standard designed to maximize the population’s life-years is itself unfair. Indeed, since COVID-19 is survivable with good medical care, it might make more sense to focus more on the elderly rather than less. Younger patients without coexisting diseases may expect a better outcome if in the ICU and may need less care.
The authors do not address the legal implications of such a triage system. In all countries that follow the common law, medical benefit to the patient governs the lawfulness of decisions to withhold treatment without consent or to withdraw it (9). We must consider how individual physicians will become legally responsible and vulnerable in this context. It is ethically allowable and also lawful to remove a ventilator from a patient as requested by that patient or his/her substitute decision maker (proxy) and assuming the patient/proxy has agency. It is ethically acceptable and also lawful to remove a ventilator from a moribund patient with consent of the proxy but the withdrawal of treatment is still treatment and as such, requires that consent or a court order where the treatment is futile (10). If a patient is removed from a ventilator without consent or court order and as a consequence that patient dies, the act of removal would constitute a homicide and be regarded as wrongful death, if viewed as a civil wrong, and manslaughter or murder if viewed as a criminal wrong. In America, the recent passing of the Coronavirus Aid, Relief, and Economic Security (CARES) Act (11) attempts to immunize volunteer healthcare workers from liability while treating COVID-19 patients. Section 3215 addresses limitation on liability for volunteer health professionals during COVID-19 emergency response. The Act raises the bar for a successful claim from proof of ordinary negligence to gross negligence. Protection against liability is not extended to all caregivers. The plain language of the Act confines it to volunteers, not paid staff, and that must occur rarely in American ICUs.
Malpractice insurance policies are based on historic risk of injury and compensable damages being awarded. Doctors cannot change the risk unilaterally and expect coverage. If the standard of practice would now include the possibility of the involuntary removal of a ventilator, insurance companies would very likely write a different policy at a different rate. Malpractice insurers cannot be expected to provide coverage for an uninsured risk. In the absence of the suspending of tort law that covers medical malpractice, physicians will find themselves in a potentially perilous legal position should they cause death by the removal of a ventilator under the proposed Sprung et al (1) resource utilization. Furthermore, one should expect a similar insurer reaction to a lawsuit under the old policy over a decision to withhold treatment from a patient who would have been offered the treatment before the Sprung et al (1) reforms.
Rationing teams are highly problematic. As imagined by Sprung et al (1), they draw little support from bioethics or the law and they support ageism. Although planning for a worst-case scenario has appeal, such plans inadvertently may invite the very worst outcome of what they intended to mitigate. Better, we struggle mightily when we think about ending a life without consent. Our patients deserve better than the coldness of an algorithm. This pandemic will pass and when it is done, our solution, final and otherwise, will be held up to the highest scrutiny.
1. Sprung CL, Joynt GM, Christian MD, et al. Adult ICU Triage
During the Coronavirus Disease 2019 Pandemic: Who Will Live and Who Will Die? Recommendations to Improve Survival. Crit Care Med 2020; 48:1196–1202
2. Ramphul K, Mejias SG. Coronavirus disease: A review of a new threat to public health. Cureus 2020; 12:e7276
3. Andrade G. Medical ethics and the trolley problem. J Med Ethics Hist Med 2019; 12:3
4. Charatan F. New Orleans doctor is charged with giving lethal injections during floods. BMJ 2006; 333:218
5. Wemm SE, Wulfert E. Effects of acute stress on decision making. Appl Psychophysiol Biofeedback 2017; 42:1–12
6. Bernat JL. Medical futility: Definition, determination, and disputes in critical care. Neurocrit Care 2005; 2:198–205
7. Emanuel EJ, Persad G, Upshar R, et al. Fair allocation of scarce medical resources in the time of Covid-19. N Eng J Med 2020 Mar 23. [online ahead of print]
8. Verity R, Okell LC, Dorigatti I, et al. Estimates of the severity of coronavirus disease 2019: A model based analysis. Lancet Infect Dis 2020 Mar 30. [online ahead of print]
9. Supreme Court of Canada: Cuthbertson v. Rasouli, 2013 SCC 53, at par. 191. Available at: https://scc-csc.lexum.com/scc-csc/scc-csc/en/13290/1/document.do
. Accessed March 23, 2020
10. Zivot JB. The withdrawal of treatment is still treatment. Can J Anaesth 2014; 61:895–898
11. Congress.gov: H.R. 748-CARES Act 116th Congress (2019-2020). Available at: https://www.congress.gov/bill/116th-congress/house-bill/748
. Accessed May 12, 2020