Ethical Issues and the Allocation of Scarce Resources During a Public Health Emergency

IN RESPONSE:

We appreciate our colleagues' comments and are heartened to see that our analysis has stimulated further thought among clinicians and policymakers. We agree with Dr. Glass that government has an ethical obligation to carefully plan for foreseeable public health emergencies, such as pandemic influenza. However, because society's resources are limited, it is impossible to fully meet all emergency needs while still allocating adequate resources to competing routine societal needs, such as primary care, education, infrastructure, and defense. Public health emergencies remind us of the inevitable need to balance competing considerations when shaping public policy.

Dr. Hansen-Flaschen uses the 1952 poliomyelitis outbreak in Sweden as a test case for our multiprinciple allocation strategy. He raises the concern that even with a multiprinciple allocation system, clusters of patients may be indistinguishable on the basis of age, prognosis for survival, and life-years saved, requiring a tie-breaking mechanism. This is possible, but not necessarily problematic. If patients are indistinguishable on the basis of allocation principles set forth as morally relevant, then there is no compelling reason to prioritize any one over the others in the group. We advocate that if such a situation occurs, random allocation should be used to break ties.

Dr. Siegel eloquently argues that individuals with severe functional impairment should receive relatively less prioritization for life support. He also correctly points out that there is deep disagreement in society about whether functional status and social worth are material considerations when allocating scarce resources. This is an emotionally charged social issue. Because public trust and cooperation with restrictive measures will be crucial to a successful public health response, we think including such a controversial criterion that affects a very small, vulnerable patient group may ultimately be more detrimental than beneficial to an effective response. We agree that this is a key issue for robust public engagement.

Dr. Powell and colleagues assert that the allocation strategy we propose is similar to theirs because both propose to exclude patients from life support on the basis of certain criteria. We disagree. We argued against categorical exclusion of patients and instead favor assigning a priority score to all who would be eligible to receive life support in routine conditions, then having the cut-point for receiving life support determined by the availability of resources. Dr. Powell and colleagues also claim that we have gone against our own recommendation for genuine public engagement by proposing an alternative allocation strategy without first engaging the public. We had no intention to set policy; that is a task for policymaking bodies and officials. Instead, our goal was to inform policymakers and the public about alternative allocation strategies and their ethical implications. Our hope is that the ideas we present will be vigorously discussed during the public engagement process—and ultimately will be accepted or rejected by informed citizens and policymakers.

Most of the criticisms from Dr. Christian and colleagues seem to come from their misunderstanding of the purpose of our proposal. They fault it for failing to comprehensively address all of the organizational and logistic challenges of a public health emergency. However, our goal was not to create an organizational protocol. Their group did this admirably (1, 2). We sought to complement their work by providing a clear analysis of the ethical issues at stake when the demand for life support exceeds supply during a public health emergency. We also sought to provide policymakers and the public with an alternative set of allocation criteria that are ethically robust and accompanied by clear justifying arguments, a feature lacking from the efforts of past groups. The assertion by Dr. Christian and colleagues that our proposed multiprinciple allocation strategy is not feasible seems premature. The continuous allocation score that our approach yields could be categorized into the familiar 4 color-coded triage categories.

We are not persuaded by Dr. Christian and colleagues' defense of categorical exclusions of certain patient groups from access to life support. They claim that these exclusions are justified because these patients would have an exceedingly poor prognosis for 1-year survival or would require a disproportionate amount of scarce resources in order to survive. They exclude, for example, very elderly persons, patients with severe cognitive impairment, and patients with certain severe comorbid diseases. However, advanced age alone is not a reliable predictor of poor intensive care unit outcomes or disproportionate resource use (3, 4). Moreover, we are aware of no solid data to support the claim that individuals with severe cognitive disabilities require disproportionately more life support than matched individuals without cognitive disabilities. Of note, some of the excluded chronic diseases do not reliably predict poor 1-year outcomes or disproportionate resource use, as illustrated by our case example in Table 1 of our article.

Dr. Christian and colleagues also criticize our proposal for allowing a degree of clinical judgment in determining whether a patient has a life-limiting comorbid condition. We acknowledge that allowing clinical judgment also opens the door to physician bias. However, ample empirical and actuarial data could inform the development of lists of diseases known to be associated with various degrees of shortened life expectancy. These data could guide the assessment of life-limiting diseases in our allocation strategy.

Space constraints prevent us from providing a detailed account of whether life support should be withdrawn from some patients in order to provide it to others with a more favorable allocation score. Several groups have addressed this issue (1, 2), and we agree with their assessment that it would probably be necessary to withdraw life support from patients whose conditions worsen considerably, or from those who demonstrate via a time-limited trial that their need for resources will be disproportionate to the needs of others.