The Yale Law Journal

VOLUME
130
2020-2021
Forum

Who Gets the Ventilator? Disability Discrimination in COVID-19 Medical-Rationing Protocols

27 May 2020

abstract. The coronavirus pandemic has forced us to reckon with the possibility of having to ration life-saving medical treatments. In response, many health systems have employed protocols that explicitly de-prioritize people for these treatments based on pre-existing disabilities. This Essay argues that such protocols violate the Americans with Disabilities Act, the Rehabilitation Act, and the Affordable Care Act. Such explicit discrimination on its face violates these statutes. Nor can medical providers simply define disabled patients as being “unqualified” because of disabilities that do not affect the ability to ameliorate the condition for which treatment is sought. A proper interpretation of the law may permit medical providers to use disability as a basis for a rationing decision where an individual’s underlying disability will kill the individual in the immediate term regardless of the treatment. However, as this Essay demonstrates, those circumstances will be narrow. Further, the law requires that such imminent-death determinations will be made based on the best available objective evidence, free from both bias against people with disabilities and devaluation of their lives.

Introduction

For years, conflicts over medical rationing have preoccupied participants in healthcare debates. But they have tended to take place in the abstract—as in discussions of what treatments will be covered under various healthcare reform proposals that may or may not pass. In many cases, discussions of these issues have tended to the demagogic—as in the trumped-up furor over the supposed Obamacare “death panels.”

Due to the coronavirus pandemic, the threat of medical rationing has become much more present. As the virus spreads throughout the United States, providers in various areas have found themselves lacking adequate medical facilities and equipment to handle the predicted number of cases. Hospitals faced with a crush of patients must confront questions of how to allocate scarce resources—notably, life-saving ventilators—at a time of severe shortage. Although our health systems seemed to weather the first wave of the COVID-19 outbreak without resort to overt rationing, it still seems far more possible than in the past that we will be in the triage situation that the law, bioethics, and policy literature typically merely hypothesizes.1 Should that come to pass, medical providers will truly be in a position to choose who shall live and who shall die.2

After the H1N1 outbreak in 2009, the Institute of Medicine urged states, in advance of the next pandemic, to adopt “crisis standards of care.”3 These standards were intended to guide how medical professionals would allocate their resources in times of “a substantial change in usual healthcare operations and the level of care it is possible to deliver, which is made necessary by a pervasive (e.g., pandemic influenza) or catastrophic (e.g., earthquake, hurricane) disaster.”4 Many states took up the call.

Unfortunately, the crisis standards of care adopted by hospitals and state agencies often employ explicit disability-based distinctions. Until the Office for Civil Rights (OCR) at the United States Department of Health and Human Services (HHS) intervened in April 2020, Alabama maintained crisis standards of care that “allowed for denying ventilator services to individuals based on the presence of intellectual disabilities, including ‘profound mental retardation’ and ‘moderate to severe dementia.’”5 Tennessee still lists “people with spinal muscular atrophy who need assistance with activities of daily living” as among those who will not receive critical care in a situation of scarcity.6 The University of Washington Medical Center’s guidelines provide that the goal in a crisis situation should be “[o]verall survival,” defined as “healthy, long-term survival, recognizing that this represents weighting the survival of young otherwise healthy patients more heavily than that of older, chronically debilitated patients.”7 Guidelines promulgated by Washington State provide that, in deciding whether to administer life-saving treatment to an individual, hospitals should look to the individual’s “baseline [pre-coronavirus]functional status” and should accordingly “consider [pre-existing] loss of reserves in energy, physical ability, cognition and general health.”8

Like the now-withdrawn Alabama policy, the Tennessee and Washington protocols, as well as similar ones in Kansas, Pennsylvania, Utah, and New York, have recently prompted state and national disability rights organizations, as well as individuals with disabilities, to file administrative complaints with OCR. The complaints allege that these and other provisions violate the federal disability discrimination laws.9 Although OCR resolved the Alabama complaint after the state withdrew its policy,10 and issued more general guidance addressing the issue,11 the other complaints remain pending. And the broader issue remains a live one.

In this Essay, written as this crisis unfolds, I argue that disability-based distinctions like these do indeed violate the law. The Americans with Disabilities Act (ADA), the Rehabilitation Act, and the Affordable Care Act (ACA) all prohibit healthcare providers from discriminating against qualified individuals with disabilities because of their disabilities. The explicit discrimination embodied in policies like those of Washington State, on its face, violates these prohibitions. Nor can medical providers simply define disabled patients as being “unqualified” because of disabilities that do not make them unable to benefit from treatment for the condition that they seek to ameliorate. Longstanding and authoritative interpretations of the law bar the use of such circular techniques to insulate disability discrimination from legal challenge.

A proper interpretation of the law may permit medical providers to use disability as a basis for a rationing decision where an individual’s underlying disability makes the individual unable to benefit from coronavirus treatment—either because that disability interferes with the treatment itself, or because the underlying disability will kill the individual in the very near term regardless of the treatment’s success in addressing the virus. But those circumstances will be narrow. And a proper interpretation of the law requires assurances that such imminent-death determinations will be made based on the best available objective evidence, free from bias against people with disabilities or devaluation of their lives.

Under this interpretation of the law, which I argue is the best interpretation of the existing legal materials, medical providers would retain substantial discretion to make resource-allocation decisions in a time of triage. They could make decisions based on non-disability factors. And where those factors do not dictate a decision, health systems would not be limited to a first-come, first-served approach. They could employ any other disability-neutral procedure that fits their professional judgment. Even a lottery would be fairer and more democratically legitimate than placing the burden on disabled individuals—individuals who already experience disadvantage as a result of societal discrimination, and who disproportionately lack access to the political and health-system processes that frame policies concerning medical rationing. But my point is not to defend a lottery or any other allocation procedure. Rather, my point is to argue that the law, best understood, rules out a protocol that puts disabled people at the back of the line because they have pre-existing medical conditions that do not make them unable to benefit from the treatment they seek.

In this Essay, I defend that reading of the law. Part I demonstrates that practices that expressly use disability as a factor in denying life-saving treatment discriminate because of a disability. Part II argues that it would be inconsistent with the law for a health system to treat a pre-existing disability as rendering an individual un-”qualified” for treatment for a new condition—at least where the pre-existing condition does not make the individual unable to benefit from the treatment. And Part III addresses the one significant body of legal precedent that might at first glance appear to stand in the way of these conclusions—the “Baby Doe” cases. As that Part shows, those cases conflict with more recent, authoritative developments in the law. And they do not apply to the COVID-19 situation even on their own terms.

I. discrimination because of disability

There are three major federal statutes that prohibit disability discrimination in the medical treatment context: the ADA; Section 504 of the Rehabilitation Act; and Section 1557 of the ACA. The ADA prohibits disability-based discrimination by state and local government agencies in its Title II,12 and it prohibits disability-based discrimination by private healthcare providers in its Title III.13 Section 504 prohibits disability-based discrimination by federal executive-branch agencies and recipients of federal financial assistance.14 And Section 1557 prohibits disability-based discrimination by health programs that receive federal financial assistance or are operated under a federal program or activity.15

Despite some slight variations in language, each of these statutes prohibits discrimination against individuals with disabilities because of their disabilities. And except for ADA Title III, all require that the disabled individual be “qualified” for the benefit or treatment they seek. ADA Title II provides that “no qualified individual with a disability shall, by reason of such disability,” be “subjected to discrimination” by a state or local government.16 Title III of the same statute says that “[n]o individual shall be discriminated against on the basis of disability in the full and equal enjoyment of the goods, services, facilities, privileges, advantages, or accommodations of any place of public accommodation.”17 Section 504 phrases the prohibition in this way: “No otherwise qualified individual with a disability” shall, “solely by reason of her or his disability,” be “subjected to discrimination under” a federally assisted program.18 And Section 1557 incorporates by reference the discrimination prohibition in Section 504.19

This Essay raises two interpretive issues relevant to medical-rationing decisions. First, are those decisions being made “by reason of” or “on the basis of” disability? Second, are the disabled individuals affected by those decisions “qualified” for the treatment they seek? I deal with the issue of qualification in Part II of this Essay. Here, I briefly discuss the issue of discrimination because of disability.

Many of the rationing protocols healthcare providers have been preparing to use in response to COVID-19 plainly inflict discrimination “by reason of” or “on the basis of” disability. Disability is an explicit factor used to deny treatment—a factor that will be decisive in many cases.20 The protocols raise none of the difficult issues that often stymie claims of disability discrimination: identifying discriminatory intent, determining appropriate reasonable accommodations, or delimiting the scope of the prohibition on disparate impact.21 Rather, they are straightforward examples of making treatment decisions because of disability. The harder question, to which I now turn, is that of qualification.

II. are disabled individuals “qualified” for life-saving treatment?

The major disability rights statutes generally limit their protection to “qualified” individuals with disabilities. ADA Title II defines a “qualified” individual as someone “who, with or without reasonable modifications to rules, policies, or practices,” meets “the essential eligibility requirements for the receipt of services or the participation in programs or activities provided by a public entity.”22 The Rehabilitation Act does not contain text defining “qualified” (or “otherwise qualified,” the language used by Section 504), but cases under the statute apply a similar rule: an individual is “otherwise qualified” if that individual is able, with “reasonable accommodation” if necessary, to perform the “essential functions” of the job or meet the eligibility requirements for a program.23

A. The General Rule: Disability as Such May Not Be a Basis for Disqualification

The most plausible legal defense of disability-based medical-rationing rules is that the affected individuals’ disabilities render them un-”qualified” for the life-saving treatment they seek. In times of triage—for example, when there are not enough ventilators and other resources to go around—a health system needs to deny these resources to some people who need them to survive. So long as medical professionals adopt reasonable criteria for making these sorts of tragic choices,24 the argument would go, those criteria should constitute the qualifications for the relevant benefit. And here, the argument would continue, it is reasonable to choose people without underlying disabilities in rationing these scarce resources, because they would benefit more from receiving life-saving treatment. They would (by hypothesis) live longer lives afterwards, they would enjoy their lives more, and they would contribute more economically to society.

Whatever one thinks as a philosophical or policy matter about rationing treatment based on quality-adjusted life years—an issue on which there is contentious debate25—the argument that medical professionals can simply define disability as disqualifying seems directly foreclosed by legal precedent. In Alexander v. Choate,one of its early cases under Section 504, the Supreme Court recognized that states could not permissibly evade the bar on disability discrimination simply by turning the absence of a disability into a qualification for a job or benefit: “Antidiscrimination legislation can obviously be emptied of meaning if every discriminatory policy is ‘collapsed’ into one’s definition of what is the relevant benefit.”26 Thus, the Court held that “[t]he benefit itself, of course, cannot be defined in a way that effectively denies otherwise qualified handicapped individuals the meaningful access to which they are entitled.”27

To adopt the absence of an underlying disability as a qualification for receiving life-saving treatment would, at least on its face, appear to conflict with that principle. The disability discrimination laws appear to erect a strong rule that disability as such may never be the basis for denying individuals opportunities.

To be sure, an individual’s disability may make it impossible or impracticable for them to satisfy certain eligibility criteria, even if the criteria are defined with no reference to disability. For example, a state may legitimately demand that bus drivers operate their vehicles safely; with current technology, blindness is simply inconsistent with that requirement. Safe operation of motor vehicles is an interest that can be defined without any reference to disability. The recognition that some disabilities, given the current state of technology, are simply incompatible with that interest reflects a simple and uncontroversial empirical judgment.

But disability in the medical-rationing context is different. Refusing to allocate scarce treatment resources to patients with pre-existing disabilities does not rest on a simple and uncontroversial empirical question. It rests on a series of value judgments: what should we be seeking to maximize when we allocate scarce healthcare resources? Lives saved, quality of life, prospective economic output of those we save? How do we determine quality of life? Do impairments to physical or mental functions necessarily limit an individual’s quality of life? What if the people who experience those impairments do not think so?28 It is practically difficult, if not conceptually impossible, to disentangle our answers to those questions from our judgments about disability and the value of life with a disability.29

Even when we answer what seem like the purely empirical questions here, our answers are likely to be inflected by our underlying views about disability and by overt or implicit biases against disabled people. Medical-rationing decisions are made in times of great uncertainty.30 They involve prognoses about the possibility of recovery and the length and quality of life following recovery that are themselves uncertain. It is natural for anyone resolving those uncertainties under such conditions to be guided by their underlying values and biases, even if only unconsciously. And there is ample evidence of widespread bias against people with disabilities among medical professionals31—a problem that is exacerbated by the underrepresentation of disabled people among their ranks.32

Philosophers could debate how we should answer these questions. Some, following the views of people like Peter Singer, might argue that many disabilities inherently limit the quality of life and that the quality of life ought to be central in determining how we should allocate scarce societal resources.33 Others might take the view that national productivity is the most important factor.34 Still others might disagree and argue that the quality of life is inherently subjective and that a purely productivity-focused view improperly reduces human beings to nothing more than “embodied net marginal product.”35

Wherever we might come down on the philosophical debate, the disability discrimination statutes resolve them for purposes of the law—at least insofar as the question is whether the absence of a disability can in and of itself be a qualification for needed medical treatment. Congress decided to bar the use of disability, standing alone, as a (dis)qualification for the receipt of needed benefits from the government or other medical providers. That decision may have reflected a fear that disability-based decisions will reflect irrational bias, or perhaps some broader commitment to disability equality. The legislative history is not specific. But the crucial point is that Congress made the decision.

One good reason to treat the disability discrimination laws as resolving these debates relates to democratic legitimacy. People with disabilities have faced a long history of exclusion from democratic participation.36 Often, that exclusion has been overt, as in the denial of the right to vote to those under guardianship.37 Other times, the exclusion has resulted from an accumulation of decisions that made it impossible for many disabled people to have access to the political process.38 When government institutions make decisions that deny people with disabilities important benefits, we should worry that those decisions lack legitimacy. They were likely taken without the equal participation of those who are the most affected. We should especially worry when the immediate consequences of those decisions are life and death.

By contrast, the political process that led to the enactment of the disability discrimination laws did not exclude the nondisabled. As a group, the nondisabled are more likely to lose out from a move that denies medical providers the ability to make disability disqualifying for scarce life-saving treatments. When a group passes a law to put burdens on itself, there is little reason to worry that it is failing to take account of the full array of costs.39 Perhaps it is best to think of the enactment of the ADA and its sibling laws as a veil-of-ignorance moment. Members of Congress restricted the use of disability as a qualification for important benefits while accountable largely to members of the nondisabled public who did not know whether they would ever become disabled. That decision has more presumptive legitimacy than an ex post decision by those operating the healthcare system, who at that point know precisely who is and is not disabled, to place the burden of resource scarcity on disabled individuals—the very individuals who are most likely to have been excluded from such decisions.

B. Does Defining the Qualification as “[Comparative] Ability to Benefit” Solve the Problem?

1.The General Argument

What if we were to take the concept of qualification up a level of generality? Defenders of disability-based rationing might say that the relevant qualification for medical treatment is not the absence of disability but the ability to benefit from the treatment. And just like blindness and the bus driver, perhaps disability in many circumstances just makes it impossible (given current technology) to benefit from a treatment. If a person has a pre-existing condition such as advanced cancer that will inevitably kill them in the next week, and they then acquire coronavirus, is it sensible to say that they will benefit from ventilator treatment for their new disease if they are going to die so soon anyway? Such a person, one might say, is not qualified for treatment for the new condition.

And if we are in a triage situation, one might argue, the relevant qualification is not the ability to benefit in some absolute sense but the relative ability to benefit. If there simply are not enough life-saving treatments to go around, might a healthcare provider reasonably determine that the people who should get those treatments are the ones who are likely to live the longest, or live the fullest lives, as a result of the treatment? What makes a person “qualified” for treatment, on this view, is that the treatment will offer that person a higher marginal number of life years or quality-adjusted life years than the next person.

One might go further: perhaps one’s disability is relevant to that question even if one will not die immediately after receiving the treatment. If I have a pre-existing medical condition that will cause me to die ten years after receiving a successful treatment for my coronavirus, maybe we should give the treatment instead to someone whose post-treatment life expectancy is longer. There is a widespread (though hardly universal) intuition that we should give scarce life-saving treatments to younger people before we give them to older people.40 Perhaps the same principle should apply when a pre-existing disability will predictably shorten the post-treatment life of one patient vis-à-vis another. More controversially, if I have a pre-existing medical condition that means that my life will be of a lower quality than that of another patient after receiving a successful treatment for my coronavirus, maybe we should give the treatment to the other patient.

In times of extreme triage of specific treatments, these arguments of course deserve serious consideration. But there are substantial problems with them. One relates to democratic legitimacy. At any given moment, there is always scarcity of all sorts of healthcare inputs.41 Even if, as the economists insist, scarcity is itself a basic fact of life, the pattern of which products and resources are scarce in any given place and time is significantly the result of societal decisions. That is especially true in the richest society in human history—one whose health system has a well-documented problem of waste and inefficiency.42 As Frank Pasquale argues, “The current scarcity of care for the least well off is not a natural feature of the world; rather, it is epiphenomenal of repeated decisions not to impose certain tax burdens today even though they would have seemed perfectly fair 50 years ago.”43

Disabled people have a disproportionately small amount of input into decisions about the operation of the health system. They are underrepresented among the ranks of health professionals who constitute the system, and they experience significant barriers to participating in the democratic process that regulates the system.44 A significant point of the disability discrimination laws is to rectify these sorts of inequalities by forcing institutions to take equal account of the interests of disabled people when they make their decisions.

The perceived need to deny ventilators to coronavirus patients with pre-existing disabilities results not from scarcity as a natural fact, but from two societal decisions: first, the decision to fail to maintain an adequate stock of ventilators to serve all patients who would need them if a pandemic breaks out;45 second, the decision, once a pandemic breaks out, to use patients’ pre-existing disabilities as a basis for denying them the use of those devices. A process in which people with disabilities were equally represented vis-à-vis the nondisabled, and in which the interests of both groups were given equal concern, would not lead to both of these decisions. An equal process might instead have chosen to obviate the second decision by maintaining a sufficient stock of ventilators to serve every patient in the event of a pandemic. Or it might well have found the costs of such a policy to be extravagant when considered in light of the other possible uses of societal resources. But it would not then place all of the burden of that decision on individuals with pre-existing disabilities. The combination of both of these policies—allowing scarcity of ventilators, while imposing the life-or-death costs of that scarcity most heavily on disabled people—bespeaks a failure of democratic legitimacy.

2. The Problem with Disability-Based Quality-of-Life Judgments

Even if they did not pose these problems of distributive justice and democratic legitimacy, disability-based criteria are likely to rest on erroneous—or at least highly controversial—judgments about the life prospects of those with disabilities. The point is most immediately obvious when considering quality-of-life judgments.

When medical professionals or others determine that a disability limits the quality of a person’s life, they are making a controversial normative judgment that often does not reflect the views of disabled persons themselves. “A massive body of research has demonstrated that people who acquire a range of disabilities typically do not experience much or any permanent reduction in the enjoyment of life.”46 Elizabeth Emens has extensively detailed the differences between what she calls the “inside” and “outside” views of disability:

From the outside, disability commonly looks like an unhappy place created by an individual medical problem for which the law sometimes provides special benefits to that individual. From the inside, disability often looks like a mundane feature of a no-less-happy life, rendered inconvenient or disabling largely by interactions with the surrounding environment, which legal accommodations alter in ways that sometimes provide benefits to many.47

To choose the “outside” view of disability held by most nondisabled people over the “inside” view held by many disabled people themselves reflects a normative choice.48 That choice might result from the devaluation of life with a disability—something that seems to fly in the face of the disability discrimination laws’ guarantee of equal treatment. Or it might result from a form of paternalism—an understanding that people without disabilities are better judges of the quality of a disabled life than are people with disabilities. “[P]aternalism has historically been one of the most significant contributors to the disadvantage people with disabilities experience.”49 And paternalism was a principal target of the disability discrimination laws.50 Given the well-documented evidence of bias by members of the medical profession against disability,51 we should be especially concerned that these judgments are being made in a way that suppresses disabled people’s interests and reinforces the problem of democratic legitimacy.

To the extent that disabilities do harm the quality of one’s life, that is often because of discrimination and societal decisions that have rendered significant opportunities inaccessible.52 To use those harms as a justification for denying life-saving treatment to disabled people imposes a form of “double jeopardy.” Societal discrimination against people with disabilities causes them to experience less full lives, which social institutions then offer as a reason for making the further decision to deny these already-disadvantaged individuals essential benefits.53

Considerations like these led the HHS, in the George H.W. Bush Administration, to reject the 1992 Oregon health-reform plan. Oregon’s plan implemented a form of rationing. HHS Secretary Louis Sullivan concluded that the rationing relied on quality-of-life surveys that “quantifie[d] stereotypic assumptions about persons with disabilities.”54 And when Oregon’s health commissioners conducted their own rankings of which conditions and treatments to prioritize, they relied on factors “including ‘quality of life’ and ‘ability to function’”—factors that, in Sullivan’s view, “expressly value a person without a disability more highly than a person with a disability in the allocation of medical treatment.”55 As a result, he concluded, the proposed rationing plan did not comply with the ADA.56 In line with Secretary Sullivan’s analysis, the ADA, Rehabilitation Act, and ACA should be understood to prohibit the rationing of life-saving treatment based on judgments about the effects of disability on a person’s quality of life.

3. The Problem with Disability-Based Quantity-of-Life Judgments

What about future life expectancy? A recent article in the New England Journal of Medicine co-authored by a number of influential bioethicists agrees that the rationing of life-saving treatment should not take account of assessments of a patient’s future quality of life, but that it should rest on assessments of future quantity of life.57 At least outside of the case in which it is clear that a person’s underlying disability will make them unable to benefit from coronavirus treatment—either because that disability interferes with the treatment itself, or because the underlying disability will kill the individual in the very near term regardless of the success of the treatment—consideration of the life-expectancy impact of the disability does not accord with the legal principles I have just discussed.

The intuition that age is an appropriate rationing criterion might lead us to think that the expected quantity of life should be treated differently than the expected quality of life. We should recall, however, that even if many people agree that age is an appropriate criterion, not everyone does.58 Among those who do not are the HHS Office for Civil Rights, which recently declared that the Age Discrimination Act of 1975 prohibits treatment decisions founded on “judgments about a person’s relative ‘worth’ based on,” among other things, “age.”59

Even if we focus on those who do agree that age is an appropriate criterion, it is far from clear that they hold that position because they believe health systems should maximize the number of life-years saved. Many people defend the use of age based on grounds that are backward rather than forward looking—“the feeling that everyone is entitled to some ‘normal’ span of health (usually expressed in terms of life years, e.g. ‘three score years and ten’).”60 The implication is that “anyone failing to achieve this has in some sense been cheated, whilst anyone getting more than this is ‘living on borrowed time.’”61 On this view, we may legitimately deny treatment to people who have already had their “fair innings” if that is necessary to protect others from being cheated out of the same opportunity.62 But that does not mean that we believe that “each life year is equally valuable,”63 or even that we should always prefer saving the younger person to saving the older person. For example, one bioethicist who purports to “support age-related rationing for COVID-19 patients” argues that “when the demand outstrips the supply” of ventilators, those devices should not be given to patients who are eighty or older.64 He contends that “people who have reached that milestone have enjoyed an opportunity to live a complete life.”65 But that reasoning does not rest on the principle that the health system should maximize the expected-life years saved. For example, a “complete life” or “fair innings” principle would treat a twenty-five-year-old the same as a thirty-five-year-old, even though saving the younger patient would, all else equal, maximize the number of expected-life years.

Moreover, reliance on expected number of years of survival as a justification for explicitly disability-based rationing raises many of the same concerns about democratic legitimacy, medical bias, and double jeopardy that reliance on quality-of-life measures does. Once again, we would be forcing disabled individuals to face deadly consequences because of societal decisions not to invest in sufficient treatments—decisions from which people with disabilities disproportionately lack access. Once again, we would be relying on medical judgments that are likely to be inflected by bias.66 And once again, we would be denying life-saving treatment at least in part because of societal discrimination—for discrimination against disabled individuals plays a key role in the poor health outcomes some of them experience.67

Society does not, of course, universally endorse expected number of years of survival as a basis for discrimination in the provision of life-saving treatment. Nor should it. For example, there remains a gap in life expectancy between African Americans and whites, despite real progress in recent years.68 There is also a substantial and growing gap between richer and poorer people.69 Yet few would defend rationing ventilators based on a patient’s race or wealth—even if it would be a good proxy for number of years of expected survival following treatment. We recognize that race and wealth are morally arbitrary for these purposes, that minorities and poor people were disproportionately excluded from the decisions about how to invest in healthcare, and that requiring them to forgo life-saving treatment because of their groups’ poor life expectancy inflicts a kind of double jeopardy.70 The disability discrimination laws place disability on a similar plane.

In a forthcoming piece in this Journal, Govind Persad defends the use of disability-based criteria that may result in exclusion from life-saving care when those criteria are based on quantity-of-life, but not quality-of-life judgments.71 He begins with a factual premise: people whose disabilities lead to deprioritization under such exclusions are likely to make up a small minority of the overall population of people with disabilities who need treatment. Based on that premise, Persad argues that denying life-saving treatment to those whose disabilities will predictably shorten their post-recovery lives will be better for people with disabilities “as a group” than the alternatives. That is, he argues, because it will enable limited treatments to save more lives overall. Such triage practices “save[] more people with disabilities than random selection, even if people with disabilities might—though this is debatable—comprise a greater proportion of the (smaller) number saved under random selection.”72

But there are three problems with that argument. First, his factual premises, while in some respects plausible, are neither obvious nor proven. Whether or not more people with disabilities will benefit from a policy containing disability-based exclusions depends on a comparison of the distribution of disabled people in two different groups. The first group is the population to whom the policy denies life-saving treatment. The second is the population to whom the policy provides life-saving treatment. Whether more people with disabilities will benefit from the disability-based exclusion also depends on the resources (such as time occupying a scarce ventilator) that would be consumed by members of each population if they were to be provided treatment. The ADA’s definition of disability is broad, and Persad is surely correct that “the overwhelming majority of disabilities do not affect” either “patients’ prospect of benefit or the quantity of resources they need.”73 But the population that needs scarce life-saving treatment is unlikely to be drawn randomly from the underlying population. We thus cannot know that a policy explicitly denying treatment to some individuals based on their disabilities is going to benefit more people with disabilities overall.

Second, antidiscrimination laws protect individuals. An employer who refuses to hire a woman for a particular job because she is a woman has illegally discriminated on the basis of sex even if it hires other women.74 A fair “bottom-line” distribution of benefits and burdens across groups does not provide a defense to proven explicit discrimination against an individual based on a protected category.75 That is a particularly important point in the context of the federal disability discrimination laws, where the definition of disability is so broad that it covers people with an immense variety of conditions.76 Maybe denying life-saving treatment to individuals with cystic fibrosis could save more individuals who take daily medication for high blood pressure. Both of those conditions count as disabilities under the ADA and associated statutes. But it would be bizarre to defend such a decision on the ground that it is better for disabled people “as a group.” The law provides no basis for such a defense.

Finally, Persad’s argument depends on the existence of “evidence-based” assessments of how a pre-existing disability affects life expectancy—and it depends on the evidence being good. But the biases I discussed above should lead to great skepticism about the quality of the “evidence” supporting express disability-based exclusions or deprioritizations in a rationing plan. A key goal of the antidiscrimination laws is to counteract those sorts of biases.

If it is clear that a person’s underlying disability will in fact interfere with the efficacy of the life-saving treatment, or that the person will die in the immediate term from that disability with or without that treatment, perhaps it is appropriate to say that the individual is not qualified. Here we would be applying an absolute criterion of benefit rather than the relative one that I explored and rejected above.77 But even here, humility should rule the hour. Medical professionals’ biases often lead them to make unduly negative prognoses regarding their disabled patients.78 Because disability rights law is designed significantly as a check on these biases, the law should not be read to permit health providers to make futility-of-treatment decisions for their disabled patients absent truly objective, persuasive, and reliable medical evidence.

The “direct threat” defense that applies when an entity engages in disability discrimination to prevent a safety risk is instructive here. The law allows such discrimination based on a determination that the individual poses “a significant risk” that “cannot be eliminated by reasonable accommodation.”79 Such a determination must rest on “the most current medical knowledge and/or the best available objective evidence” and “an expressly ‘individualized assessment’” of the disabled person.80 The law imposes these constraints to provide a check on biases that lead to exaggerated fear of safety risks stemming from disability.81

The widespread nature of medical bias against people with disabilities justifies a similarly restrictive approach here. Absent a clear, objective basis for concluding that a person will die in the immediate term as the result of a pre-existing disability—a determination that should be made through procedures that protect against bias—the law ought to be read to prohibit using such a disability as a basis for denying life-saving treatment that the person seeks.82

This does not mean that medical systems must take a first-come, first-served approach to providing life-saving treatment in triage situations. As Secretary Sullivan explained in his response to the proposed Oregon plan, a system may consider “a wide range of factors” that are “consistent with the ADA.”83 “These factors include, but are not limited to, the cost of medical procedures, the length of hospital stays, prevention of death, and prevention of contagious diseases”—and indeed “any content neutral factor that does not take disability into account or that does not have a particular exclusionary effect on persons with disabilities.”84 If a system does not wish to follow that path, it may allocate treatments on a lottery basis—a process that many people believe would be fairer.85 But under the best reading of the disability discrimination laws, they may not deny life-saving treatment simply because a patient has a pre-existing disability.

III. the “baby doe” cases

My arguments here might seem inconsistent with an older line of lower-court cases that rejected challenges to the withholding of medical treatment from newborns with developmental disabilities.86 Those cases, known as the “Baby Doe” cases, appear to hold that the disability discrimination laws cannot be employed to challenge medical treatment decisions. Although they have not been the subject of much academic discussion in recent years, they do occasionally pop up in litigation.87 Accordingly, it would not be surprising if a health system sought to defend a disability-based rationing decision by pointing to those cases.

In the leading “Baby Doe” case, United States v. University Hospital, the Second Circuit said that the Rehabilitation Act’s “otherwise qualified” language “cannot be applied in the comparatively fluid context of medical treatment decisions without distorting its plain meaning.”88 The court asserted that “[i]n common parlance, one would not ordinarily think of a newborn infant suffering from multiple birth defects as being ‘otherwise qualified’ to have corrective surgery performed.”89 It determined that challenges to nontreatment decisions would “invariably require lengthy litigation primarily involving conflicting expert testimony to determine whether a decision to treat, or not to treat, or to litigate or not to litigate, was based on a ‘bona fide medical judgment.’”90 And it concluded, based on a review of the legislative history, that “[C]ongress never contemplated that section 504 would apply to treatment decisions of this nature.”91

But there are two significant problems with relying on the “Baby Doe” cases here: the cases are inconsistent with more recent Supreme Court decisions, and they do not apply here even on their own terms.

As we have seen, in the year after University Hospital, the Court concluded in Choate that entities may not evade the bar on disability discrimination simply by turning the absence of a disability into a qualification for a benefit.92 By allowing hospitals to employ an infant’s disability as the basis for determining that the child was not “qualified” for medical treatment, that is precisely what the “Baby Doe” cases did.

The “Baby Doe” cases relied on the premise that the disability discrimination laws apply only in those contexts Congress specifically envisioned.93 But the Supreme Court decisively rejected that premise in Pennsylvania Department of Corrections v. Yeskey.94 In Yeskey, the Court held that Title II of the ADA applies to discrimination against disabled individuals incarcerated in a state’s prison system. Although the state argued that “Congress did not envisio[n] that the ADA would be applied to state prisoners,” the Court found that argument “irrelevant.”95 Even “assuming [the state’s argument] to be true,” the Court held, “the fact that a statute can be ‘applied in situations not expressly anticipated by Congress does not demonstrate ambiguity. It demonstrates breadth.’”96 Because Title II applies to any disability-based discrimination by a “public entity,” without making a specific exception for prisons, the Court refused to read such an exception into the statute. That conclusion fatally undermines the premise of the “Baby Doe” cases.

And contrary to the “Baby Doe” cases, the Supreme Court has now twice recognized that the disability discrimination laws apply in the “context of medical treatment decisions.”97 In Bragdon v. Abbott,98 the Court applied the ADA to a dentist’s refusal to treat a patient because she had HIV.99 And in Olmstead v. L.C.,100 the Court held that the ADA imposes significant obligations on states regarding the provision of mental health treatment. The Olmstead Court disavowed any holding “that the ADA imposes on the States a ‘standard of care’ for whatever medical services they render, or that the ADA requires States to ‘provide a certain level of benefits to individuals with disabilities.’”101 But it specifically held “that States must adhere to the ADA’s nondiscrimination requirement with regard to the services they in fact provide.”102

Notably, when the Supreme Court addressed the “Baby Doe” issue, in Bowen v. American Hospital Association,103 it did not embrace University Hospital’s position that the Rehabilitation Act does not cover medical treatment decisions. Justice Stevens’s plurality opinion rested on a very narrow ground. He concluded that a hospital’s withholding of treatment to disabled infants did not violate the statute where the parents of those infants refused to consent to the treatment: “[W]ithout the consent of the parents or a surrogate decisionmaker the infant is neither ‘otherwise qualified’ for treatment nor has he been denied care ‘solely by reason of his handicap.’”104 Chief Justice Burger concurred in the judgment to make five votes for Justice Stevens’s bottom line, but he issued no opinion to explain his reasoning.105

In Bowen, not a single Supreme Court Justice endorsed the Second Circuit’s University Hospital holding. And the Court’s subsequent cases stand in conflict with the Second Circuit’s decision. Those decisions do not provide a basis for exempting medical rationing from the disability discrimination laws.

There is a second reason why the lower-court “Baby Doe” cases cannot save the discriminatory medical-rationing decisions that are being adopted in response to the coronavirus pandemic. Those cases embraced a rule that the disability discrimination laws do not apply to a decision to deny treatment based on the disability itself being treated. The Second Circuit explained its position in University Hospital: “Where the handicapping condition is related to the condition(s) to be treated, it will rarely, if ever, be possible to say with certainty that a particular decision was ‘discriminatory.’”106 The Second Circuit’s argument was intuitive. If a disability is the reason why an individual needs the medical treatment in the first place, how can the refusal to provide that treatment discriminate on the basis of disability? There is no comparator who lacks that disability who was being treated better. For this reason, the Tenth Circuit, in its case relying on University Hospital, specifically distinguished “hypothesized situations in which the handicap that forms the basis of the section 504 discrimination bears no relation to the medical treatment sought but denied.”107

The refusal to provide coronavirus treatment to patients with pre-existing disabilities is precisely the case that the Tenth Circuit distinguished. A pre-existing condition like a developmental disability “bears no relation to the medical treatment”—a ventilator—that the challenged protocols will deny the disabled patient. Here, unlike the situation in which the courts believed themselves to be in the “Baby Doe” cases, the medical system’s action is the refusal to treat one condition based on an individual’s having a different disability.108 Even on their own terms, the “Baby Doe” cases do not bar application of the disability discrimination laws here.

Conclusion

Under the best reading of the ADA, the Rehabilitation Act, and the Affordable Care Act, the denial of life-saving treatments to individuals who have a distinct pre-existing disability violates the law. In a true triage situation—one the coronavirus pandemic may soon present—it may be necessary for the medical system to make tragic choices about how to allocate scarce resources. But denying those treatments to individuals because of their disabilities, when those individuals can benefit from them, is not one of the choices the law permits.

Samuel R. Bagenstos is the Frank G. Millard Professor of Law, University of Michigan Law School. Member, University of Michigan Institute for Healthcare Policy and Innovation. I thank Harold Pollack for a conversation that helped me significantly in framing this piece. I also appreciate comments I received from Don Herzog and participants at a Fawley workshop at the University of Michigan Law School. Thanks as well to the staff of the Yale Law Journal for their extremely helpful editorial suggestions on a highly compressed schedule. I am among the counsel for a coalition of disability rights organizations, who have filed several recent complaints with the Office for Civil Rights of the United States Department of Health and Human Services to challenge state plans for rationing medical services in response to COVID-19. This Essay represents only my personal views, however.