Become a Patreon!


Excerpted From: Govind Persad, Equal Protection and Scarce Therapies: The Role of Race, Sex, and Other Protected Classifications, 75 SMU Law Review Forum 226 (May, 2022) (104 Footnotes) (Full Document)

GovindPersadIn the third year of the COVID-19 pandemic, governments and hospitals have continued to struggle with the fair allocation of antibody and antiviral therapies that aim to prevent severe disease in patients who contract COVID-19. Spiking hospitalizations and the ineffectiveness of some older monoclonal antibody therapies against the Omicron variant exacerbated overall therapy scarcity, despite new oral antivirals. While scarcity has abated as cases have fallen, lack of federal funding risks renewed shortages.

Fairly distributing scarce treatments presents ethical questions. But it also presents legal ones--especially when characteristics shielded by equal protection law, such as race, sex, age, health, and even vaccination status, are considered. In this short Essay, I examine whether and when proposals for allocating scarce treatments may violate constitutional, federal, or state equal protection laws.

The legality of using some of these criteria, in particular race, was discussed during the vaccine rollout. But vaccine supply now exceeds demand in the United States, while many effective therapies are still in short supply or may become scarcer if funding is not renewed or viral drug resistance intensifies. Therapy allocation is therefore the most likely legal battleground for the near future of the COVID-19 pandemic--and is likely to arise for other pandemics and illnesses as well.


During the pandemic, COVID-19 deaths have “disproportionately affect[ed] Hispanic, non-Hispanic black, and non-Hispanic American Indian/Alaskan Native populations.” During the vaccine rollout, many states responded to this and other disparities by using social vulnerability indices to prioritize individuals from areas facing greater disadvantage. These indices do not base eligibility on an individual's race, but rather on their place of residence. Some indices incorporate racial demographics at the neighborhood level, but the use of race as a neighborhood-level variable is legally distinct from considering individual recipients' race. The use of indices directed vaccines where they could better prevent hospitalizations and deaths; it also served to remediate racial disparities.

A few states, such as Montana, Utah, and Vermont, as well as some vaccine sites, additionally or instead based eligibility on individual recipients' race. Many legal commentators and officials doubted such an approach satisfied current Supreme Court precedent, although some disagree. These prioritizations faced a few legal challenges or threats, but none have been adjudicated on their merits.

For the allocation of scarce COVID-19 therapies, some health systems have used social vulnerability indices similar to those used for vaccines. While some have grumbled about this, no lawsuits challenging these approaches have been reported. Particularly when the use of social vulnerability indices is framed in terms of preventing harm and promoting public health, rather than as a proxy for allocation by individual race, such lawsuits are unlikely to succeed. More recently, 15% of antiviral doses have been prioritized for Federally Qualified Health Centers that serve more disadvantaged populations.

In contrast, some states have instead or additionally proposed to use individual race to allocate scarce therapies. These proposals quickly faced criticism, legal challenges, and inquiries from elected officials; several have been withdrawn. In this Part, I examine the legal framework governing these prioritization proposals and conclude that the states were wise to withdraw their guidance given current precedent.

Since 1995, the Supreme Court has required that any governmental consideration of individuals' race to allocate benefits satisfy strict scrutiny. Strict scrutiny requires that such consideration serve a compelling government interest and be narrowly tailored to realize that interest. While strict scrutiny does not prohibit all consideration of race, therapy allocation based on individual race--even as one factor among many will struggle to satisfy these criteria, in particular narrow tailoring.

A. Compelling Interest

At least three different governmental interests might be advanced by using individual race to allocate scarce therapies. One is preventing hospitalization and death: if an individual's race predicts their likelihood of death or hospitalization if infected, allocating scarce treatments based on race could help reduce death and hospitalization. This is undoubtedly a compelling interest. But using it to justify race-based therapy allocation will be challenging. For instance, while prior to the advent of vaccines Black and Hispanic Americans faced disparately high risk of death and hospitalization, it is less clear that this reflected greater risk once infected, as opposed to greater risk of infection, or that it could not be accounted for by social vulnerability and high risk medical conditions (all factors that, of course, are shaped by structural racism). Most COVID-19 therapies are only prescribed to people who have tested positive. Additionally, the prevalence of immunity due to vaccination or prior infection and, as of 2022, the replacement of prior variants by the Omicron variant are likely to make pre-vaccine and pre-Omicron risk data less predictive of present risk. For instance, over the first four months of 2022, the CDC reports that the share of COVID-19 deaths among Black Americans has almost exactly equaled their share of the U.S. population; the share of deaths among Hispanic Americans has been a third less than their share of the population; and the share of deaths among non-Hispanic white Americans has been nearly one-fifth above their share of the population. (However, non-Hispanic white Americans nevertheless continue to have a lower than average risk of dying early in life from COVID-19: during 2022, they comprised less than half of deaths among people younger than 50. In contrast, during 2020, COVID-19 deaths among Black Americans were one and one-quarter times higher than their share of the proportion, while deaths among non-Hispanic white Americans and Hispanic Americans roughly equaled their share of the population. The share of Asian American deaths was less than their share of population during both years, and the reverse for Native Americans, but the share of total deaths occurring in each of these groups decreased between 2020 and 2022.

The other two interests that could support the use of individual race are, first, reducing racial disparities, and second, rectifying historical racial discrimination. These interests, however, may not always reach a legally compelling level. Rectifying historical discrimination has only been accepted as legally compelling when a past discriminator is rectifying its own missteps. Meanwhile, courts may question whether reducing racial disparities in medical outcomes constitutes a compelling interest.

B. Narrow Tailoring

Even if a compelling interest such as reducing death and hospitalization can be established, using individual race to allocate novel, scarce therapies is likely to fail a narrow tailoring analysis. Narrow tailoring can be understood as requiring that the consideration of individual race be the only, or best, way of achieving the compelling interest at issue. Narrow tailoring disallows uses of individual race that are overinclusive (prioritizing people whose prioritization would not serve the compelling interest) or underinclusive (failing to prioritize some people whose prioritization would serve that interest).

Post-Adarand, uses of individual race often fail on the narrow tailoring analysis. This is likely to be especially true for definitions that do not make distinctions within racial groups or lump racial groups together. Assigning the same number of priority points to all treatment candidates who are not white is not narrowly tailored because it lumps together candidates whose race-associated risks vary substantially. Prioritizing specific racial groups over others would likely still be both over and underinclusive. For instance, even though Asian-American race may now be associated with lower risk, specific Asian-American groups have been at high risk in the pandemic. Additionally, associations between race and risk are likely to be confounded by other factors, such as vaccination status: while Asian-Americans have recently died at lower rates than some other racial groups, they are also highly vaccinated.

More fundamentally, the use of individual race will likely fail narrow tailoring because, in the COVID-19 pandemic, racial identity is an (imperfect) proxy for exposure to societal factors that increase risk, rather than itself a cause of risk. This makes the use of individual race to allocate therapies legally different from its use to diversify workplace or educational settings, where appropriate representation of people with specific racial identities is the governmental interest at issue, or to counterbalance historical identity-based racism. Using race as a proxy for societal factors that elevate medical risk is also different from using race as a proxy for biological factors such as genetics or ancestry--a medical practice whose legality remains unsettled. Even a clear association between race and risk is not enough: narrow tailoring will require approaches that address the social factors that link race with medical risk, not the shortcut of using race as a proxy. Using individual race as a proxy for these factors may appear administratively convenient, but is insufficient to satisfy narrow tailoring. Analogously, Asian-Americans on average score highest on the mathematics section of some college entrance examinations, but this correlation would not license dispensing with examinations and prioritizing students who identify as Asian-American for admission, even as one factor among many, given the better tailored alternative of using examination results themselves. Strict scrutiny disallows the use of individual race as a proxy for medical risk if the compelling interest could instead be realized either by directly assessing the risk factors for which race is proxying or by using alternative proxies for risk, such as social vulnerability indices that may include race as a neighborhood variable.

Additionally, courts have also required “serious, good faith consideration of workable race-neutral alternatives” prior to the implementation of policies that classified individuals by their race. Indeed, prior failed trials of race-neutral alternatives were relevant even under earlier precedents that applied only intermediate scrutiny to classification by individual race. This presents further problems for an initial allocation rule for a novel therapy that begins by using individual race.

Is a pandemic different? Current litigation over therapy allocation has already been dismissed for want of standing as scarcity abates. But other decisions during the pandemic further suggest that, if courts were to reach the merits, use of individual race--particularly before individually race-neutral alternatives for distribution have been tried-will be highly unlikely to pass muster under current precedent. And the Supreme Court of 2022 is substantially more skeptical of racial classifications than the courts that decided Adarand and Parents Involved.

While the law applying to allocation policies voluntarily implemented by private hospitals or clinics is slightly different, allocation decisions based on individual race face similar obstacles. This is particularly true where the institution allocating therapies receives governmental funds and so comes under the umbrella of federal law. Standards appear more lenient for a private organization that receives no federal funds, but the legal situation is highly uncertain. In addition, some states further limit the use of race in decision-making, which could support state-law legal challenges to prioritization rules.

Last, prioritization based on Native American status would likely fare better legally than prioritization based on individual race. But even this is in flux in the courts, with the Supreme Court set to hear arguments that classifications based on Native American status should be treated more like racial classifications. Ultimately, the use of social vulnerability indices or other population-level classifications is clearly the legally soundest approach to mitigating harm as well as addressing disparities. And, under current precedent, eschewing the consideration of individual treatment candidates' race makes the biggest legal difference--not describing individual race as just one factor among many (which is still likely to fail) or avoiding the use of even neighborhood-level racial data (which is likely unnecessary).

[. . .]

In this Essay, I have discussed the legality of policies that allocate scarce therapies on the basis of legally protected identities. I have argued that race-based allocation is on the most tenuous legal ground, followed by sex-based allocation. Allocation based on age or health status will typically be legal, and allocation based on vaccination status presents novel issues. While I have framed my assessments in the context of the COVID-19 pandemic, scarcity is sadly not confined to pandemics, nor will COVID-19 likely be the last pandemic we face. These issues will remain relevant for other scarcities and future pandemics.

Assistant Professor, University of Denver Sturm College of Law JD, PhD, Stanford University.

Become a Patreon!