Abstract
With the limited initial availability of COVID-19 vaccines in the first months of 2021, decision-makers had to determine the order in which different groups were prioritized. Our aim was to find out what normative approaches to the allocation of scarce preventive resources were embedded in the national COVID-19 vaccination schedules. We systematically reviewed and compared prioritization regulations in 27 members of the European Union, the United Kingdom, and Israel. We differentiated between two types of priority categories: groups that have increased infection fatality rate (IFR) compared to the average for the general population and groups chosen because their members experience increased risk of being infected (ROI). Our findings show a clear trend: all researched schedules prioritized criteria referring to IFR (being over 65 years old and coexisting health conditions) over the ROI criteria (eg occupation and housing conditions). This is surprising since, in the context of treatment, it is common and justifiable to adopt different allocation principles (eg introducing a saving more life-year approach or prioritizing younger patients). We discuss how utilitarian, prioritarian, and egalitarian principles can be applied to interpret normative differences between the allocation of curative and preventive interventions.
Keywords: COVID-19, health care ethics, prevention, prioritarianism, vaccination, vulnerability
I. BACKGROUND
The efficient and fair allocation of scarce medical interventions encompasses a wide range of particular issues, including diagnostic tests, the availability of healthcare personnel, organ distribution, access to various therapies, or vaccines. In the early stage of the COVID-19 pandemic, many professional associations, healthcare institutions and governmental bodies published or updated prioritization guidelines regarding the allocation of scarce medical resources, eg beds or artificial ventilation in intensive care units.1 Later, in the second half of 2020, many governments published detailed prioritization schedules for the distribution of COVID-19 vaccines, which were scarce goods at the turn of 2020 and 2021. Unlike guidelines on medical treatment, official schedules on the distribution of medical prevention in the case of COVID-19 have not yet been analyzed or compared in scholarly journals. The only comparative analysis, which we are aware of is that of Schmidt, Weintraub et al., who analyzed the COVID-19 vaccine allocation schedules in the United States to check whether jurisdictions (ie 50 states, the District of Columbia, five cities and eight territories) adopted proposals to reduce inequity using disadvantage indices and related place-based measures.2 Thus, our main aim is to provide the first systematic international comparison of the official prioritization schedules for vaccinations in 29 countries and to analyze values and principles implicitly embedded in these documents. Although some suggest that prioritization during pandemic raises structurally similar dilemmas in the cases of diagnosis, treatment, and prevention,3 we will highlight and analyze the specific nature of allocation schemes in the case of prevention.
The main challenges regarding the interpretation of any prioritization guidelines or schedules stem from their valued-loadedness and the plurality of different principles implicitly embedded in them. The specific content of such schedules (eg the order of groups prioritized) may be interpreted as representing different and conflicting types of ethical principles.4 Four categories of such principles are commonly treated as generally relevant in healthcare contexts.5 The first requires treating people equally, eg by providing them with equal opportunities in the form of a lottery or the ‘first-come, first-served’ rule. Second, prioritarianism recommends favoring the worst-off, which are understood as either the sickest or the youngest. Third, utilitarian principles require the maximization of total benefits, that is, either saving the most lives or those with the best prognosis or those who have the most life years ahead of them. Fourth, rewarding social usefulness by either promoting other important values (a future-oriented aspect) or rewarding those who have implemented some important values (a past-oriented aspect). A guideline or schedule motivated by these last principles might prioritize front-line health care workers or research participants in vaccine trials. Of course, these four general principles do not exhaust all relevant values, and in such situations, there may exist a genuine plurality of intuitively plausible principles that give conflicting recommendations.6 There are also scholarly debates on the importance or weight of particular principles in a given context.
In the particular context of prioritizing access to pandemic vaccines, the most commonly articulated goals are preventing illness and saving lives from the virus. However, these two general aims are framed differently in the scholarly literature: as saving lives, benefitting the greatest number of individual people, minimizing years of life lost (YLL), maximizing quality of life years saved, maximizing quality-adjusted life-years (QALY), saving the worst off (ie those with the poorest prospects), saving those most likely to recover, saving younger lives, saving those most likely to contribute to a flourishing society (either economically or socially), saving those who can most usefully contribute to minimizing the impact of the pandemic.7
In the context of the recent pandemic, it has been claimed that the most important among the principles mentioned above is the utilitarian one, which requires the maximization of total benefits: ‘saving more lives and more years of life is a consensus value across expert reports’.8 However, with respect to the limited time and information in the COVID-19 pandemic, the authors admit that it is justifiable to concentrate solely on saving more lives (with a reasonable life expectancy) and treating saving more years of life as a subordinate principle used when the likelihood of survival of different individual patients is similar. They specify that in the case of COVID-19 vaccination distribution, this rule would justify giving older people priority for vaccines immediately after health care workers and first responders. In the case of patients above some threshold of risk (ie those over 60 years of age and with coexisting conditions), they propose using random selection instead of any fine-grained categorization of this large group. However, another article about ‘global vaccine allocation’ by the same main author was published in the journal Science a few months later (September 11, 2020). This proposed using Standard Expected Years of Life Lost (SEYLL) averted per dose of vaccine as the metric for premature death. SEYLL calculates life years lost compared to a standardized reference life table and is an explication of what we termed minimizing years of life lost (YLL) in the previous paragraph.9 Other authors suggested refining particular principles, eg utilitarian10 or prioritarian.
Discussing our results, we highlight several issues characteristic for the allocation of preventive interventions that correspond to the three main ethical principles mentioned above (in the following order: utilitarian, prioritarian, and egalitarian).
II. METHODS
In our research, we systematically examined what schedules regarding COVID-19 vaccines were adopted by national decision-makers. We compared official vaccination schedules, as presented on official government websites, in 29 countries with similar organizational and material resources: 27 European Union members, the United Kingdom, and Israel. In all research countries, public healthcare had a monopoly on the distribution of COVID-19 vaccines. Since the EU explicitly did not want to give precise guidance to country authorities on vaccine schedules because of different ‘country-specific epidemiological situations’ and ‘flexibility in terms of changing objectives’, we assumed that the various national schedules in the EU represent both different factual circumstances and varying value choices.11 As the schedules changed in time, we decided to analyze data available on August 15, 2021. We decided not to include the United States in our analysis, as its regulations differ in each of its 64 jurisdictions, especially in adopting policies towards vaccination of disadvantaged communities.12 In particular, we were interested in setting priorities among different groups within countries. Since an agreement between EU countries13 aimed at allocating access to vaccine doses according to the population distribution key, we are not discussing the problem of the international distribution of vaccines.
It should be noted that one may apply three main approaches to deriving values from normative regulations (including priority settings), namely justificatory, expressive, and consequential. These approaches differ in two interconnected respects: each points to a different dimension of a regulation and assumes a different basis on which the ascription of values is made.
According to the justificatory approach, values embedded in a particular schedule are understood as those that have guided an official’s decision to adopt this schedule. In this sense, values are assigned to a regulation on the basis of the intentions of the agent, that is, the goals and reasons of a decision-maker. However, since there is no direct access to the intentions of decision-makers, these goals and reasons must be identified based on available evidence, in particular the official justifications for legal regulations, public statements of government officials, and legislative history (documents produced during the process of decision-making). The justificatory approach tacitly assumes the credibility of such sources, namely, that publicly presented reasons mirror those that, in fact, governed a decision. In turn, the expressive approach refers to the values of a regulation as perceived from the external perspective, that is, by an observer who charitably reads the text of the regulation and strives to make sense of its provisions. In contrast with the previous account, the expressive approach does not assign priority to the viewpoint of the decision maker and focuses on values that provide the best rationale for a given priority setting (no matter how the decision maker sees this rationale). From this perspective, an investigation of official justifications as well as the legislative history is not necessary. The relevant values are to be determined on the basis of content of the setting and interpreted in light of the social context in which the setting was established. Finally, the consequential approach is focused on the actual results of the implementation of a legal regulation. On this account, neither official justifications nor social context of implementation are necessary to derive values from a particular schedule. All that is needed according to this approach is the knowledge of the social consequences that have occurred due to the establishment of a vaccine priority setting.
Our inquiry is based on the second approach, the expressive. In the context of our research, this means that relevant values are derived from particular schedules on the basis of the content of these schedules and the social circumstances of their establishment. In line with the approach, we adopted an external—observer—perspective, which in our case consists of the point of view of an expert equipped with knowledge regarding the theoretical discussions on priority setting and distributive justice.
There are several reasons for the selection of this approach. First, and most importantly, it provides the most suitable basis for comparing state policies on vaccine prioritization. It should be noted that different states adopted very different communication strategies—the general aims of the schedules and recommendations were not always explicit in revealing the value judgments, eg the German government published extensive justification of priority,14 the Polish one gave no explanation. The same problem applies to the analysis of social consequences of different priority settings. In contrast, all information required by an expressive approach (in particular, officially published vaccination schedules and data concerning the specificity of SARS-CoV-2 at the time of establishing these schedules) is publicly available with respect to all researched states. Second, it is widely believed that the expressive dimension of legal and political actions is crucial for understanding the meaning of these actions, as well as for the purposes of their moral assessment. This view is common among contemporary legal scholars.15 Third, considering the main objectives of COVID-19 vaccination programs, an expressive approach appears to be particularly suitable for the analysis of these programs. Since COVID-19 vaccine policies have attracted considerable social attention and were addressed to each and every citizen, their success was to a large extent dependent on how particular regulations concerning vaccine distribution (including prioritization schedules) were perceived by ordinary persons. Fourth, a methodology based on an expressive approach is accepted in the relevant research area. A similar research method was applied in the inquiry conducted by Schmidt et al.16 on vaccine prioritization in the context of the US.
We know that there are certain limitations to an expressive approach focusing on the perspective of the observer. One particularly important aspect is that it may yield different interpretations of the values embedded in the schedules. This phenomenon may be treated as analogous to the case of the underdetermination of scientific theories; ie, the situation in which different scientific theories may be observationally equivalent of the same evidence.17 Proponents of instrumentalism in the philosophy of science claim that this is not a problem because scientific theories are merely instrumentally useful representation devices. In contrast, scientific realists would not agree with the thesis, arguing that there is a fact of the matter as to which theory offers the right explanation of the observations. Analogically, if we substitute ‘principles’ for ‘scientific theories’, ‘justify’ for ‘be observationally equivalent’, and ‘schedules’ for ‘evidence’, we can conclude that many different normative principles may justify the same vaccination schedule (or some pattern visible in many schedules). The phenomenon has been noticed in ethics; for example, Parfit famously argued that some versions of consequentialism, kantianism, and contractualism essentially coincide in their recommendations and can be seen as attempts to climb the same mountain from different sides.18 However, in this paper, we will not analyze whether this observation has any metaethical implications, particularly whether this speaks for metaethical anti-realism or realism.
In the research, we used data (COVID-19 vaccination schedules and recommendations) accessible online on official government websites. The governments presented their policies in various ways. It is especially noticeable when it comes to terminology, eg with many phases describing corresponding groups (like different variants of ‘medical workers’ or ‘administration workers’). These differences are easily explained by the diversity of languages (only some of the plans were available in English at the time) and the varied legal systems of European countries. There are also notable disparities in the sizes of groups when it comes to vaccination phases: some countries opened the possibility of vaccination for smaller groups one by one, and some vaccinated larger cohorts at bigger intervals.
The collected data are summarized and compared in two tables. Table A in the Supplementary Materials shows the prioritization adopted by 29 countries, where columns present subsequent groups from those vaccinated as the first ones (group 1) through the groups prioritized next, up until vaccinations are available to the general population. Out of necessity, the table presents a shortened description of vaccinated cohorts, omitting details such as, eg specializations of medical workers or very specific descriptions of occupational groups. Table B in the Supplementary Materials (with an excerpt, Table 1, below) first presents the position in vaccination schedules of two groups vaccinated in almost all of the researched countries: the frontline medical workers as well as personnel and residents of nursing homes (two first columns). We assume that the reasons why they were prioritized are mixed: both direct and indirect. And then, in the following four columns, we present our interpretation of why the subsequent groups have been prioritized. In these columns, we present only the groups we had reasons to believe were prioritized with the aim of protecting them directly but not, at least primarily not, because of someone else’s interests. In many cases, if someone was vaccinated with priority due to the interest of another person, she would be described in the schedule only in relation to others. For one, contacts of pregnant people (Germany) or household members of patients with certain chronic diseases (Latvia) were described as ‘people in contact with’ certain persons, not like other groups who are described as being vaccinated because of their own characteristics. On the other hand, home-based nursing care employees were prioritized in several countries, and there may be indirect and direct reasons for their prioritization. The reason being that they work often with seniors, as well as because they are completing tasks of healthcare workers. The first types of cases were omitted from our analysis, and the second ones were included in it.
Table 1.
COVID-19 vaccination prioritization in selected countries: increased risk of SARS-Cov-2 infection (ROI) vs. increased infection fatality rate (IFR); (an excerpt from Table B, included in Supplementary Materials)
|
Table 1 (an excerpt from Table B): Columns present the categories of vaccination groups. Rows present countries as well as prioritisation. On the top there a country that prioritised people almost entirely based on their increased infection fatality rate (IFR): The United Kingdom. At the bottom is the example of country that additionally also used many other factors that we interpreted as targeting people with increased risk of infection (ROI): Germany. In the middle, there are two countries that implemented a mixed approach: Poland and Slovakia. This first country did not take into consideration the housing condition of individuals, while the second did not consider housing or health condition in its prioritisation schedules. Both Poland and Slovakia also omitted care homes prioritisation.
We then differentiated between two main types of priority categories: groups that have an increased infection fatality rate (IFR) compared to the average for the general population and groups chosen because their members experience an increased risk of being infected (ROI). The reason for such an interpretation is that in the case of COVID-19, the individual risk of death (for simplicity, we assumed that decision makers focused on preventing deaths) depends on these two factors. Then we distinguished two subcategories in each category. Increased IFR stems from an individual’s physical state: suffering from certain health conditions or just being of an older age. Increased ROI is mainly determined by factors related to measurable social mobility—an increased number of social contacts compared to average in the population. In the analyzed schedules, we distinguished between two factors: working and housing conditions.
Such an interpretation excluded certain types of priority groups from Table B that are included in the prioritization schedules presented in Table A. The first were people vaccinated because of their occupation when their core work was not related to an increased ROI (mostly state and local authorities or small groups like Tokyo 2020 Olympic athletes or Eurovision Song Contest participants). Other than that, analyzing the ethical aspects of prioritizing some people only or primarily because of other people’s interests goes beyond the scope of this article, but it is a promising field for future research.
III. RESULTS
Our in-depth analysis of the results presented in Table A shows that all countries incorporated multiple variables to categorize populations, then grouped these subpopulations into a single schedule with three to nine phases in which subsequent subpopulations were allowed to obtain vaccines up until the point when vaccinations are available to the general population.
Two groups were vaccinated first in almost all of the researched countries: frontline medical workers as well as personnel and residents of nursing homes. However, there were a number of exceptions. In Denmark, healthcare professionals were not only vaccinated after the residents of nursing homes, but also after people 85 years and older, as well as after people over 64 years of age who received both personal care and practical assistance. Bulgaria, Poland, and Slovakia did not specify nursing home residents as a priority group, but since their policies included the rapid vaccination of the oldest residents, most patients at nursing homes would still have been vaccinated almost immediately.
Each analyzed country emphasized prioritizing senior members of society (either dividing them into a few fine-grained cohorts or treating those above some age threshold as one group, as in, eg Croatia, Belgium, Romania, Latvia, and Israel). The other most commonly prioritized groups were the following: people who are more vulnerable to severe symptoms of COVID-19 than the rest of the population due to a preexisting health condition, people with bad health but not necessarily more vulnerable to COVID-19, people vital to maintaining the state apparatus, people who are more exposed to infection than the general population for various reasons (especially occupation or housing conditions). There were also singular cases of the prioritization of more specific groups. Most of them are those who care for and/or come in contact with vulnerable individuals (eg Austria, Finland).
Our in-depth analysis of the results presented in Table B (and its abbreviated version) concentrates on increased IFR and increased risk of infection (ROI).
III.A. Increased IFR
First and foremost, old age was universally treated as a reason for priority in our interpretation, the only difference being the precise age that constitutes membership in the first prioritized group. In 5 cases, it is 85 years of age and more; in 11 cases, it is people over 80 years of age. 2 countries first vaccinated people over 75, and another 2 chose seniors over 70. Six countries vaccinated people over 65 years of age. Israel and Hungary vaccinated the over 60 cohort first and The Netherlands vaccinated by year of birth, starting from the oldest.
As a critical point from which we classify ‘seniors’ as a cohort in the context of SARS-Cov-2 infection, we adopted the age of 65. The IFR of COVID-19 only for people over this age was known to be higher than 1 per cent at the time when schedules were designed—and significantly higher than amongst all of the younger cohorts (see this meta-analysis published in December 202019). The newest meta-analysis mostly confirms earlier ones: the lowest IFR occurring at age 7 years (0.0023 per cent) and increasing exponentially through age 30 years (0.0573 per cent), 60 years (1.0035 per cent), and 90 years (20.3292 per cent).20 Most (20) of the countries researched put people over 65 among the initial four phases of vaccinations. The countries that left out some of the people over 65 in the first phases still gave priority to the oldest (70+ people) before those who were younger but still vulnerable. As a result, vulnerable groups with a higher IFR of COVID-19 than seniors were mainly vaccinated later in time, eg people with Down’s syndrome, those with cirrhosis, or people after transplantations.21 There were some countries that focused almost entirely on age groups in their vaccination policies at the expense of not including other possible prioritization claims. Among them, the most notable are Slovakia and the United Kingdom. Slovakia, after vaccinating frontline COVID-19 workers from many fields, employed only the age criterion, making no exceptions for any other groups. The UK based its policy on age groups, but in phases 4 and 6 also included groups characterized as suffering from underlying health conditions.
The second universally considered reason for prioritization, noticeable from the observer’s perspective, is health condition increasing the risk of the severe course of COVID-19 and even the death of an individual. There are numerous health problems that are reported to increase the fatality rate in the case of a SARS-Cov-2 infection,22 the most notable being hypertension,23 cardiovascular disease,24 diabetes,25 chronic obstructive lung disease,26 chronic kidney disease,27 and cancer,28 although the catalogue is much wider.29 All but one of the countries researched (Slovakia) included this factor in their policies, but to a very different extent. In some countries there was an extended list of diseases that qualified people for earlier vaccination. For example, policymakers in Luxembourg divided this category into four extensive groups, including many different kinds of conditions. Similarly, many other countries (eg Germany, Italy, or Latvia) covered a wide catalogue of health problems in their prioritization policies. There are also countries that included significantly fewer conditions, also putting them all in one category and without differentiating between different levels of risk. For example, Poland listed (in phase 4) only dialysis patients, oncological patients treated with chemotherapy or radiation therapy after December 31, 2019, patients after transplants undergoing immunosuppressive therapy, and patients undergoing chronic mechanical ventilation. It greatly narrowed the number of people in this category, as widely occurring conditions such as, eg hypertension,30 cardiovascular disease,31 or diabetes32 were not listed, and priority for cancer patients was also restricted.
Some state schedules mentioned mental health problems among the reasons for prioritization. Studies have found that preexisting mental health disorders correlate with both high ROI and high IFR in case of COVID-19.33 The prevalence of physical illnesses and risk factors connected to worse COVID-19 outcomes such as, eg obesity, cardiovascular diseases, diabetes, or HIV is higher amongst individuals with mental health problems.34 It also links to alterations in the immune system, which is a probable cause of increased COVID-19 ROI in this population.35 In addition, people with mental health disorders are often subjected to socioeconomic risk factors, such as poverty, unstable job situation and working in unsafe conditions, poor access to healthcare, homelessness or living in overcrowded settings, etc.36 All of the above make it impossible to place them in only one category of prioritization rationale, so we decided not to include mental health issues as either IFR or ROI prioritization.
Among the conditions listed in the prioritization policies, there are also those that we do not have reasons to believe are responsible for the direct and measurable increase in IFR for patients in the case of COVID-19: Some countries prioritize people who are simply in bad health, receive personal care and practical assistance (Denmark), or have a disability (eg Romania).
III.B. Increased ROI
Increased ROI connected with a given occupation is strongly supported by data in the case of healthcare workers, especially those dealing with patients.37 Other jobs at increased risk include, for example, such large groups as retail workers.38 Estimates also highlight protective services, including police officers and firefighters, personal care jobs such as childcare workers and domestic caregivers, and social services occupations as those being at risk.39There were a number of outbreaks in schools that affected teachers and other school staff,40 but these were limited due to lower susceptibility to COVID-19 among children and adolescents as well as the introduction of online teaching.41 Some models predicted that school and kindergarten teachers who worked in person would be at increased risk of infection.42 Studies conducted in Norway and Sweden showed that teachers were at medium risk, after the most exposed occupations from the services sector like bartenders, waiters, transportation workers, etc.43 After initial prioritization has been done, some findings emerged that measure the incidence of COVID-19 related to in-person teaching with masking requirements lower than its prevalence in general communities.44 It shows the significant uncertainty faced by policymakers when it comes to measuring ROI.
The vaccination of groups distinguished because of occupational reasons was carried out in most cases after the groups for which we assumed increased IFR, the exception mostly being frontline social workers (eg Finland) and sometimes also others engaged directly in the fight with COVID-19, like police forces (eg Slovakia), armed forces (eg Portugal), firefighters (France), administration (Hungary), etc. Home-based nursing care workers were sometimes also included (Austria, Finland, France, Germany, Italy, Latvia, The Netherlands, Romania, Spain, Sweden). Other groups distinguished by occupation fall behind the biologically vulnerable. In 10 cases (Austria, Czech Republic, Germany, Greece, Latvia, Lithuania, Malta, Poland, Romania, and Spain) the priority was given to teachers, probable cause being the need to end remote teaching. In some cases, people with ‘absolutely necessary cross-border travel activity’ or diplomats and their families were prioritized (Austria, Lithuania, Romania, Slovenia). Only in 6 countries (Austria, Germany, Ireland, Latvia, Romania, Slovenia) were grocery store workers prioritized, but only in 2 of them (Austria and Germany) they were prioritized not as a final group before the general population. It is also worth noting that there was no prioritization for most people working in the service sector, facing high ROI, like waiters, bartenders, caterers, taxi drivers, etc. As many as 6 countries (Belgium, Croatia, Cyprus, Luxembourg, Israel, the United Kingdom) did not prioritize any occupation, except for medical workers (and social workers in the case of the UK). In some of the countries, there were also categories containing employees who are critical for different sectors (Finland, Hungary, Latvia) and providers of vital services (Estonia). However, the inclusion of these groups might be explained by pragmatic considerations and not only by the increased ROI.
Crowded housing situations are another instance where the risk of SARS-Cov-2 infection increases, as best proved in the case of care homes,45 homeless shelters46 and prisons.47 Although WHO48 recommended taking into account prison populations and people living in precarious living conditions at the end of the second phase of vaccination (ie when 20 per cent of the most vulnerable part of the population is vaccinated) this was typically not included in the researched schedules—20 of them have not mentioned the housing situation (except for care homes). When it has been mentioned, it applied to residents in precarious living conditions (Austria, Germany, Greece, Ireland) or people living in closed structures like prisons, centers for migrants and refugees, or shelters (Cyprus, Germany, Greece, Ireland, Latvia, Romania). Additionally, Bulgaria prioritized ‘vulnerable groups from the population due to high epidemiological risk of infection attributable to their way of life’ as the last group before the general population. Figure 2 summarizes the results described in the last two sections.
Figure 2.
The figure presents the number and percentage of countries researched that include different categories in their COVID-19 vaccination policies.
IV. DISCUSSION
The allocation of healthcare resources occurs on three levels of abstraction: macro, meso, and micro. First, on the most general level (eg national) when decision makers establish general rules of prioritization between individuals, for example, those with different levels of COVID-19 related risk of death, when not everyone’s healthcare needs can be met at the same moment. Prioritization at this level usually concerns as yet unidentified or statistical individuals.49 Second, the meso level concerns regional or institutional allocation and may concern both identified and statistical individuals. And third, the micro-level concerns situations where frontline practitioners allocate resources between two or more identified people when not everyone’s needs can be met. This last case is discussed by philosophers and bioethicists in many highly idealized examples, eg ‘should I save the life of a 70-year-old who can expect to live 20 years more; or a 40-year-old who can expect to live 15 years more?’
The allocation of therapeutic resources or interventions may operate on all three levels, while the allocation of preventive resources, in particular the COVID-19 vaccine schedules we analyze, is established only on the macro level. Of course, some decisions at the mesolevel may in practice influence the way vaccinations were distributed, but this is not the focus of our paper. Therefore, unlike many therapeutic interventions, in the case of prevention of COVID-19, at the time of intervention, individuals are often classified as eligible for the vaccine not based on their individual current or past conditions, but on the characteristics of the group or cohort to which they belong. In contrast with therapeutic interventions, there are no identified individuals whom we want to help, but only individuals with different levels of risk of COVID-19 related death. These observations are particularly important for the interpretation of the allocation of preventive interventions. After deciding on some prioritization schedule, we may be able to count effectiveness of this intervention for some group or cohort, that is, for example, evaluate how many people would have died from this group within some period if we had not vaccinated them. However, it is often impossible to determine which particular people are beneficiaries of the vaccination schedule, that is, which particular individuals would have died if a different prioritization schedule had been chosen instead. In other words, a decision maker may be statistically certain that, let us say, 1000 people out of some particular group of 1 million would have died if a different prioritization schedule had been chosen instead, but it may be impossible for them to establish the identify of beneficiaries of this schedule.
In this part, we discuss how the comparison of COVID-19 vaccine schedules may be helpful in interpreting the different value choices regarding the prioritization. In particular, we are interested in how three groups of principles (utilitarian, prioritarian, egalitarian) commonly treated as relevant in the healthcare contexts50 were embedded in the vaccine schedules in 29 countries.
IV.A. A Utilitarian Perspective: Currency and Epistemic Limitations
Let us start with utilitarian principles that require the maximization of total benefits. We understand this approach as giving recommendations that identify relevant values (conceptual analysis), assign a numerical scale to the identified values (measurement), assume full interpersonal comparability, and finally estimate the size of the overall value in different expected scenarios (aggregation). Therefore, putting aside the problems with measurement, comparability, and aggregation, the crucial question is what values vaccination schedules promote under utilitarian interpretation.
Some authors have postulated that saving the most lives should be the main ‘currency’ of the utilitarian approach in the context of the COVID-19 pandemic.51 However, other proponents of utilitarian approaches criticize the life-saving view as too narrow. Instead, the quality-adjusted life-years (QALYs) approach is commonly treated as the ‘most’ consistent with utilitarian approaches,52 and some authors defended the years-of-life saved approach as the proper expression of utilitarianism.53 In the case of COVID-19 Giubilini et al., contrasting the UK approach to some other selected recommendations and schedules (WHO, US, German), noticed that ‘It is often taken for granted that the criterion for prioritizing access to COVID-19 vaccines is vulnerability to COVID-19: the most vulnerable should get the vaccine first. The term “most vulnerable” is often taken to mean those with the highest probability of dying if infected’.54 The authors proposed widening the spectrum of utilitarian values relevant in this situation and taking into account different types of intrinsic (numbers of lives saved, years of life saved, quality of the lives saved, quality-adjusted life-years (QALYs)) and instrumental benefits (protecting healthcare systems and broader societal interests, eg prioritizing critical workers and having dependents). One of Giubilini et al.55 arguments favoring the diversification of the currency used in the schedules referred to the consistency of the health allocation schedules and guidelines. They suggested that not only saving more life years but also the quality of life and QALYs should be included in the vaccine distribution schedules because, as they argue, it would guarantee consistency within different policies regarding the allocation of scarce healthcare resources in the cases of treatment and prevention.
In fact, in most guidelines regarding treatment (eg ventilators in the case of COVID-19), it is not only saving lives that is considered and prioritized, but also a variety of other factors, particularly the probability of short-term survival as well as long-term considerations such as life expectancy and the quality of future life. For example, many guidelines regarding allocating scarce treatment resources take into account the very probability of survival (of some medical procedure), which in the US was very often based on the SOFA ‘Sequential Organ Failure Assessment’56 and in the UK on the frailty scale.57 In March 2020, the Italian Society of Anesthesia, Analgesia, Resuscitation and Intensive Care (SIAARTI) issued a series of recommendations that prioritized the allocation of resources to patients that have higher probability of survival and life expectancy.58 On the contrary, the vast majority of groups with prioritized access to vaccination was included in the schedules not on the basis of their past or current state of their health, but mainly based on their uncertain and narrowly understood prospects related to COVID-19 infection: as one may interpret, the worse their prospects were in this matter and the more probable that they may die because of COVID-19, the higher on the vaccination priority list they found themselves. The ‘worseness’ of the prospect was established mostly (although not entirely) on the basis of the age and comorbidities of the individuals (see blue areas in Figure 1).
Figure 1.
Compares Germany and the UK as countries representing contrasting approaches when it comes to their policies concerning COVID-19 vaccination prioritisation. The figure presents vaccination’s queues from top to bottom - from highest priority to not prioritised. The estimated sizes of the groups were retrieved from governments materials, with the exception of the groups marked with an asterisk (*) that are estimated by the authors. The corresponding groups are connected with lines, for example, since caregivers have no priority in the UK, caregivers fall into the wide group of vaccinated at the end, so this group is linked with ‘not prioritised’.
However, it would be a mistake to treat different utilitarian values embedded in vaccination schedules as necessarily conflicting with each other, that is, we should not assume that maximizing one value in some vaccination schedule (eg lives saved) must always prevent maximizing other values (eg years of life saved or QALYs). Indeed, some models of the effects of different vaccination strategies demonstrated that vaccine prioritization based on age dominated both in terms of reducing total lives saved and in terms of maximizing QALY59 or years of remaining life expectancy.60 Therefore, under some assumptions about vaccine availability, effectiveness, and safety, different values accepted by different types of utilitarian approaches can lead to acceptance of the same schedule. In an extreme case, which is not far from the case of COVID-19 pandemic, when the risk of dying increases dramatically with increasing age, utilitarian approaches that implement years of life saved might favor the old-first schemes.
One might ask whether these schedules that relied on age to a broader extent (eg UK) really represent a pure version of the ‘saving the most lives’ approach. This is far from obvious. First, the dependence of vaccination schedules mostly on age groups can be controversial even within this framework. It certainly has certain pragmatic advantages, as it eases communication and is simple to verify. However, it leaves out some groups vulnerable to COVID-19 in terms of high IFR, eg people with Down’s syndrome or those after transplants. Our analysis shows that these populations in most countries faced longer time without access to the vaccine than people over 65 years of age, despite many of them having higher IFRs than many people from the oldest cohorts.61 What is more, it is relatively easy to verify if someone belongs to those groups.62 Despite that, only in five countries (Austria, Denmark, Italy, Latvia, and Malta)63 people with comorbidities associated with extremely high risk of severe COVID-19 were prioritized in earlier phases than those between 65 and 70 years of age. In the next section, we will analyze an additional prioritarian argument stating that some groups vulnerable to COVID-19 in terms of high IFR are worse off since they are less likely to even reach senior age because of their health status.
Furthermore, there is another reason why relying mainly on IFR (and mostly ignoring ROI as in the UK) does not represent a pure version of the ‘saving the most lives’ approach. It is easy to imagine situations in which members of some group have a lower IFR than members of some other group (on average). However, still, members of this first group have a higher total risk of death from COVID-19 (on average). As an example, imagine two groups: A (a cohort 55–64) and B (a cohort 65–74), for which real-life estimations of IFRs are the following: 0.75 and 2.5 per cent, respectively.64 For simplicity, we assume that there is no other relevant information about risk factors for any member of these two groups. It is easy to calculate that if only ROI (within some period) is at least about 3.3 times larger for members A than for the members of B, an average member of A still has the greater relative risk of dying for COVID-19 (within some period), even though her IFR is much lower than for an average member of B’s. Of course, assuming such a difference in ROI may seem arbitrary and slightly artificial. Still, there are some reasons to believe that a somewhat younger cohort may be more exposed to the virus in real life (eg because many of them still have to go to work).
However, one could argue that the concentration on IFR, which is visible in many schedules, particularly in age cohorts, has a pragmatic justification based on the asymmetry of evidence. In the case of COVID-19, including someone in a high-risk group may have a different meaning. On the one hand, it may be based purely on medical premises (IFV); on the other hand, it may be primarily social-based (ROI). We hypothesize that this aspect is particularly interesting while analyzing utilitarian principles: schedules implement principles that depend on such social factors to a lesser extend because it is much more difficult to predict the results of their implementation. Moreover, there are also practical reasons underlying this asymmetry, since identifying individuals with higher ROI is much more complicated than identifying those with higher IFR. This may explain why many jobs typically viewed as precarian, such as waiter or Uber driver, were not prioritized, as those jobs can sometimes be additional activities for people generally focused on other careers. In addition, there are many migrants working in those jobs, not necessarily legally, so the authorities may not have enough resources to even begin the process of prioritization of those groups.
Considering the context of the situation, one may assume that at the time of establishing schedules there was more evidence of better quality about IFRs for different cohorts and groups than about ROIs, which may be estimated only on very scarce and random data. IFRs for various groups may be established precisely and reliably based on large meta-analyses (eg see this meta-analysis of 130 large studies),65 whereas ROI is usually established mainly on small observational studies.66 We can assume that the confidence of rational agents should adequately reflect the strength of evidence (or lack of it). In situations in which we have incomplete or sparse evidence for a proposition, philosophers like to say that it is rational to have imprecise levels of confidence in the proposition, and such confidence can be represented by an interval.
For example, returning to our earlier example, one may be almost completely confident that IFR for cohort A (55–64) is 0.75 per cent, because this belief is based on a large and reliable meta-analysis. However, one cannot usually be as confident that ROI is at least about 3.3 times larger for members A than for the members of B (65–74), because in this case (let us assume for the sake of this example), your belief is based only on one study, which gave such result, but the study itself was small, observational, and as yet has not been replicated. Therefore, it could be argued that if one wants to minimize the risk of dying for large cohorts of people in such cases, it is rational to prioritize an older cohort (B), since it is rational not to ground your actions on the results of this second, more ambiguous study.67 One possible explanation for this may refer to the ambiguity aversion: a decision maker can be more certain how many people will be saved if they prioritize the older cohort (B), in comparison with prioritizing the younger cohort (A). In this second case, the number of saved people may be either higher (if ROI is, in fact, larger than 3.3 times for members A than for the members of B), or lower (if ROI is, in fact, smaller than 3.3 times for members A than for the members of B). The other explanation may refer to avoidance of the worst possible scenario, which would be realized when a decision maker would prioritize the younger cohort A believing that its ROI is at least about 3.3 times larger for its members than for the members of B, but their belief would turn out to be false and ROI would be much lower.
This evidential asymmetry may explain why categories concerning social factors that may lead to increased ROI are much less often and less consistently used in the policies of the analyzed countries than those justified by increased IFR. For example, no occupation appears on every or almost every list (except for healthcare workers), and housing conditions were typically overlooked in the analyzed recommendations (except for nursing homes). Many schedules prioritize public employees when they are engaged in combatting the pandemic but leave out other, often low-paid essential employees working in conditions that may have increased risk of infection, eg grocery store workers or delivery persons (eg France, Hungary, Slovakia). In many countries, a significant number of those workers are migrants68 living in more crowded spaces and with worse access to healthcare.69 The most comprehensive policies in that matter are probably the German and Romanian ones (although the prison population was left out in Germany and precarious living conditions were omitted in Romania), but it comes at the price of intricacy.
However, the asymmetry of reliable information does not explain why vaccination schedules have not apparently incorporated other utilitarian ‘currencies’ for which it is generally possible to have reliable statistical data, particularly years of life saved. In some cases, it would be indeed possible to count how different vaccination schedules would produce different outcomes in terms of, for example, years of life saved and many modelling studies incorporated such statistics.70 Imagine a hypothetical case where we vaccinate 100 people as early as possible, each 90 years old, preventing 10 people from the earlier death on Covid. However, it is expected that they will live, on average, until 92, so we would save 20 additional years of life in total. Instead, if we prioritize the vaccination of 100 people, each 70 years old, preventing only 2 persons from the earlier death on Covid. But in this case, it is expected that they will live, on average, until 85, so we would save 30 years of life in total. In the first case, of course, we are saving many more lives (10) than in the second case (2). In this second case, we are saving more life years (30) than in the first case (20).
This type of case was mentioned by Giubilini et al.,71 who complained that vaccine schedules in the UK concentrate on ‘those with less expected time left to live—say, a 90-year-old man in a care home—are prioritized over those who are still relatively vulnerable to COVID-19 but are likely to live longer—say, an otherwise healthy 70-year-old woman’. In fact, actual priority schemes cannot be justified by utilitarian calculus in the case of care homes. Giving highest priority to nursing home residents, what was commonly implemented in the schedules (see Table A), could only have been justified on a preventing most deaths basis. In fact, their residents had the highest mortality risk from COVID-19, but many of them had the shortest predicted life expectancy and quality of life. For example, Giubilini et al.72 refer to the case of dementia arguing that since dementia reduces both length and quality of life and significantly affects an individual’s expected QALYs, patients with severe dementia should, in their opinion, have a lower priority. Unfortunately, they do not provide practical guidance on how to incorporate such factors into existing schedules in practice, for example, how to measure the reduction in quality of life due to dementia and how to weigh this reduction with other factors. Instead, they refer only to extreme and catchy examples like this one: ‘it is a low priority to spend limited resources on somebody who is permanently unconscious, compared to a person who is in full possession of their mental faculties’.
To summarize, it seems on the one hand that the utilitarian approach promises to provide a straightforward solution to vaccine allocations by calculating and weighing the benefits of particular schemes in a quantifiable manner: ‘[T]his mathematical certainty that the utilitarian approach provides is appealing, especially in times of uncertainty such as the present one’.73 On the other hand, the COVID-19 vaccination schedules clearly show the epistemic limits in applying any utilitarian principles, particularly any currency more sophisticated than saving the most lives. This problem is not typical for our case because any principle that includes expected consequences to the moral evaluation of a schedule must take into account the problem of ‘cluelessness’, which states that in many cases, a decision maker has not the faintest idea whether or not a schedule A maximizes benefits to a higher extent than B.74 However, it is particularly important in the case of preventive interventions in general because of many more uncertainties, for example, regarding the social dimension of the recent pandemics, the pattern of social contacts and its influence on the pandemic dynamics, the efficacy of different nonpharmaceutical interventions for the reproduction number, etc. Moreover, any rationing scheme that would incorporate a more subtle criterion for prevention prioritization, eg quality of life or QALYs, is even more susceptible to critique (although the reasons are similar for both treatment and prevention): it may rely on inaccurate stereotypes about the quality of life; lack the perspective of disabled people; neglect the fact that poorer quality of life often results from social injustice.
IV.B. Prioritarianism: Who Counts as the Worst off?
Derek Parfit’s seminal presentation of prioritarianism states that ‘[B]enefiting people matters more the worse off these people are’.75 This view has been thoroughly discussed in philosophy and healthcare ethics, but even among those who agree that the worst off should be given some priority, eg within the healthcare domain, there is no agreement on who counts as ‘the worst off’.76 The dimension on which we concentrate in this section is how one can conceptualize someone as the worst off in the context of vaccination prioritization schedules. In the discussions about prioritarianism in healthcare, some argue that a decision maker should categorize the worst off by referring only (or primarily) to their entire lifespan (like a life-time prioritarianism concerned with distributions over entire lives77 or only (or mostly) to some part of their lifespan (like a time-slice prioritarianism78). This distinction is also visible in an article by Persad, Wertheimer et al.79 who distinguished ‘youngest first’ from ‘sickest first’ prioritarianism (the first may suggest a lifetime, the second time-slice prioritarianism). However, as we shall see, neither of these understandings of the worst off can be used to interpret the COVID-19 vaccine distribution schedules.
The main motivation for this first view stems from the assumption that individuals are the units of ultimate moral significance for public policies, and the standard account of personal identity asserts that individuals typically extend through time, from birth to death, as single persons. This approach is supported by views that assume that we should categorize the worse off by referring to their age: younger people are relatively worse off than older people because they have lived fewer life years.80 One possible extension of this form of prioritarianism is ‘the complete lives system’, which concentrates on the question of how long a patient has been alive, but also takes into consideration prognosis, saving the most lives, equal chances, and instrumental values.81 In the context of rationing ventilators and critical care beds during the COVID-19 pandemic, supporters of this approach highlighted that younger people should receive priority ‘not because of any claims about social worth or utility, but because they are the worst off, in the sense that they have had the least opportunity to live through the stages of life’.82 This is also the common way in which prioritarianism is interpreted in the health economy: ‘prioritarianism is more likely than utilitarianism to prioritize the young in the allocation of vaccine doses, despite the fact that they might have lower overall benefits from being vaccinated’.83 In this case, ‘benefits’ are understood as decreasing chances of dying from COVID-19.
One argument in favor of preferring younger refers to the pairwise comparisons of people’s claims.84 The argument can be visualized in the example we used in the previous section when we introduced two cohorts (A: 55–64 and B: 65–74) with different IFR: 0.75 and 2.5 per cent, respectively. Even if we assume that the individual risk of death for everyone in the older cohort is higher than for everyone in the younger cohort, one can still argue in favor of prioritizing people from the younger cohort. This is because there is no one among the older cohort who would be harmed by an earlier death as much (or nearly as much) as a person from the younger cohort who would die of COVID-19 as they would have lived much longer otherwise. This argument assumes that what counts is not the number of people saved by a vaccine schedule, but the fulfillment of the strongest individual claims. If it is not possible to vaccinate all, it would be better to vaccinate this group in which there are individuals with the strongest claims (that is, individuals who would lose the most if they died prematurely). This argument corresponds to the claims that prioritizing older patients may exacerbate inequalities because early death is strongly correlated with other forms of social disadvantages.85
In fact, some triage policies during COVID-19 pandemic may be interpreted as implementing ‘the complete lives system’, eg about 50 per cent of US ventilator triage policies included an age criterion, which prioritized younger (of course, this is not the only possible theoretical justification of using age criterion in such situation, the other is for example survival rates when connecting to a ventilator).86 However, including age has been criticized in the US context as ‘likely to constitute illegal age discrimination under existing federal law’.87 In Europe, although the ethical guidelines declaratively were very sensitive to the risk of discrimination arising from strict triage criteria, and particularly regarding age, some of them used the age limit as an exclusion criterion in certain situations.88 In contrast to treatment guidelines, our analysis clearly shows that this principle has not been implemented in the case of COVID-19 vaccine schedules. On the contrary, we observe blanket exclusions of younger people from early vaccine access.
According to the second understanding of prioritarianism, what matters is only how well-off individuals are in specific periods (where ‘period’ may mean a moment, a day, a week, a year, a decade, etc.). More precisely, what is essential is the level of well-being that accrues to the individuals in the period in which they receive (or could receive) a benefit. This approach, which seems to dominate in healthcare discussions about prioritarianism, treats current severe suffering as something that requires special concern and assumes that the initial state of the patients from which the distributive intervention may occur is the most crucial for evaluation. Shlomi Segall claims that the bedrock prioritarian intuition ‘lends itself to, and only to, the shortest temporal unit for which we can speak of a person’s welfare’.89 Under this interpretation, prioritarianism is a principle close to the Rule of Rescue, and the view distinguishes itself from utilitarianism by treating current severe suffering as something that requires special concern. For example, Rulli and Millum90 claim that the institutional duty of easy rescue ‘could be grounded in the same considerations that ground prioritarianism’. Similarly, Torbjörn Tännsjö: ‘[a]ccording to prioritarianism… severe suffering (at a moment) calls for special concern’.91
In the theoretical discussions about vaccination schedules there were also voices that postulated a need to pay special attention to groups who are disadvantaged either directly by having been disproportionately impacted by COVID-19 due to persistent structural and systemic disadvantage, or even indirectly, eg they are generally marginalized and disenfranchized. Here is an example of such an approach: ‘we believe that such plans (for COVID-19 and for other future diseases) should have more targeted and inclusive outreach and be implemented as systemic efforts capable of adequately and effectively protecting marginalized populations and disadvantaged/disenfranchised groups’.92 One way to implement this type of prioritarian concern is a categorized priority system, which divides resources into multiple categories, enabling the use of different criteria for the allocation of resources within each category.93 Some percentage of vaccines may be reserved in every stage of the vaccination schedule to realize a particular principle or principles (eg prioritizing the worst off, however defined) while the remainder could be allocated in line with the main principle, eg saving the most lives. In fact, such a system was introduced in many US states where by March 30, 2021 51 jurisdictions (out of 64) had prioritized specific zip codes in combination with metrics such as COVID-19 incidence or adopted disadvantage indices. For example, Massachusetts officially declared that it would allocate 20 per cent of vaccines to communities with a disproportionate burden of COVID-19 and high social vulnerability.94 In contrast, no schedule we analyzed has used such a system, which may stem from the fact that health inequalities in many analyzed countries are not as enormous as in the US. On a more general level, analyzed schedules show little sign of prioritization of people because of their current suffering or any other form of economic or social deprivation. Rare instances of prioritizing in those situations may be spotted only in a few countries that prioritize people with disabilities (eg Romania, Sweden), even when the specific chronic illness or health problem they suffer from is not a direct risk factor in the case of contracting SARS-Cov-2.
Our results show that the COVID-19 schedules may be interpreted as implementing the prioritarian principle in a different sense from the two discussed above. The fact that in each country, individuals above 65 years were given special access to vaccines very early on, most schedules included people with characteristic comorbidities that increase their expected IFR but ROI-based criteria were not systematically and consistently applied may be interpreted as implementing the very specific prioritarian principle: the worst off are those who have the highest risk of death if infected, that is, the highest IFR. This index, and more generally, the future health prospects of an individual in the case of COVID-19, can be established either by just categorizing one into some age cohort (what does not take into account individual characteristics) or by also taking into account characteristic comorbidities that increase expected IFR (takes into account current or past conditions of particular individuals). These two methods coexisted, although to a very different extent, in all analyzed schedules. Only a few schedules recognized that the total risk of death for COVID-19 depends not only on IFR, but also on the very probability of being infected—these schedules (eg German) included criteria that we characterized as reflecting ROI.
The difference between these three dimensions may be again visualized in the example of two cohorts (A: 55–64 and B: 65–74) with different IFRs: 0.75 and 2.5 per cent, respectively. These countries that prioritized people almost exclusively on the basis of age would order members of these two cohorts in a ranking close to lexical order (ie almost every individual belonging to B would be higher on the vaccination schedule than every individual from A, with some minor exceptions concerning, for example, healthcare workers). In contrast, these countries that included more high-risk patients would allow many individuals from A (or even from younger cohorts) with higher IFR than the average for their own cohort to be on the same level (or even higher, since some countries put high-risk patients on the very top of the prioritization schedule) than individuals from B. Finally, these schedules that used ROI took into account the fact that some subgroups of A may have a similar or even higher risk of death related to COVID-19 because their members have a relatively high probability of being infected.
Of course, since no policy is a pure representation of one particular normative theory, no schedule represents a pure version of these approaches. Nevertheless, these schedules that prioritized people almost exclusively on the basis of age and comorbidities (eg UK, see Table 1 and Table B in Supplementary Materials) and took into account no (or relatively very few) additional factors related to ROI may be interpreted as the closest to this understanding of prioritarianism that defines the worst off as those who have the highest IFR. On the other side of the spectrum, these approaches that included not only cohorts and groups with the highest IFR, but also groups with higher ROI (eg Germany, see Table 1 and Table B in the Supplementary Materials) can be interpreted as closest to this understanding of prioritarianism that defines the worst off as those who have the highest risk of COVID-19 death, which is a different criterion from pure IFR. Interestingly, this last approach overlaps with these utilitarian approaches that concentrate on saving most lives.
However, even in this case there would be an interesting difference between utilitarian and some prioritarian approaches because utilitarianism (and some versions of prioritarianism) does not distinguish the way some consequences appear. The literature on prioritarianism distinguishes two ways to generalize prioritarianism from cases of certainty to those of uncertainty, the so-called ‘ex ante’ and ‘ex post’ prioritarianism.95 The first states that a decision maker should maximize the sum of priority-weighted expected well-being (‘ex ante’). Under this interpretation, the assessment concentrates on effects on individuals’ chances of death (their ‘prospects’) and the worst off are those who have the worst prospects. The second states that a decision maker should maximize the expected sum of priority-weighted well-being (‘ex post’). Under this interpretation assessment, the assessment focuses on the intervention’s expected effects on the overall pattern of outcomes, and the worst off are those who in fact will die on COVID-19 (even if their identity is not known ex ante).
It is easy to imagine cases where two different vaccination schedules may result in the same number of avoided deaths, but the saved individuals may belong to different groups. Let us go back to our standard case of two cohorts (A: 55–64 and B: 65–74) with different IFRs: 0.75 and 2.5 per cent, respectively. Moreover, for the sake of this example, let us assume that both groups are equally numerous (eg 1 million people), and ROI (within period t) for A is 33, (3) per cent whereas for B it is 10 per cent. Under such assumptions we can predict that the number of expected deaths related to COVID-19 will be more or less the same in both groups within period t (about 2500). If we use the individual risk of Covid-19 death as our measure of how well-off people are (and so for fixing prioritarian weights), every member of the younger cohort is equally well off in comparison with members of the older cohort because everyone has the same total individual risk of Covid-19 death (0.25 per cent). From a utilitarian or ex post-prioritarian perspective, it does not matter, which group will be vaccinated first (assuming that we cannot vaccinate both groups simultaneously). However, in A many more people will have COVID-19 than in B (333,333 vs. 100,000), and in B those who will have COVID-19 will die much more often. Therefore, from some ex ante prioritarian perspective, this may be a reason to vaccinate the older cohort B first. Followers of this approach may argue that being infected with SARS-CoV-2 is what makes an individual worse off in the relevant sense (not just belonging to some age cohort). Thus, a vaccination schedule under this interpretation focuses all ‘moral attention’ only on those people who will be infected with SARS-CoV-2 in these two cohorts, even if their identity is not currently known (or even impossible to learn, since the unpredictability of the pattern of social contacts and spread of the virus). The rationale for this is simple: people in these cohorts who will not be infected with SARS-CoV-2 bear no risk of death because of COVID-19, so we do not have to care about them in this regard. Under this interpretation, the previous case incorrectly generalized risk to every member of these two cohorts. To sum up, from an ex ante prioritarianism two policies might have the same expected outcomes—so, be identical from an ex post perspective (eg in terms of number of people who die)—but differ ex ante. The lesson we can learn is that the focus of most analyzed schedules on IFR may represent an ex ante prioritarian approach.
IV.C. Equality: Choosing without Preferring
The value of equality was almost commonly recognized in international documents as a guiding principle for the allocation of the COVID-19 vaccine. WHO SAGE framework includes equal respect (consisting of recognition and treatment of all human beings as having equal moral status and particular interests as deserving of equal moral consideration) in the list of six principles relevant for the vaccine distribution, while the EU Strategy for COVID-19 vaccines mentions equitable access for all in the EU among three main objectives for the EU policy. Furthermore, some theorists96 place a special emphasis on equality in this context. In all these cases the principle of equality was distinguished from the prioritarian standard and refers to ex ante equality, that is, to an equal chance of some vaccine-related benefit (what does not necessary imply an ex post equality, that is, equality in terms of outcome, eg vaccination uptake or COVID mortality).
In many papers about the allocation of rare healthcare resources, the principle of equality is understood formally as the requirement to distribute some goods on the basis of the ‘first-come, first-served’ or by chance (ie by lottery).97 Harris98 mentions an amendment to these two rules that would allow people the option of giving away their equal priority to others. Most authors reject the ‘first come, first serve’ approach, considering it inferior to lotteries. The main reason for this is that the ‘first come, first serve’ approach may unacceptably exacerbate inequalities, ie may benefit people who are resourceful and socially privileged to be the first in lines. Schmidt99 even calls the rule ironically let-me-use-my-connections-and-pointy-middle-class-elbows approach.100 Moreover, some authors argue that waiting time is not intrinsically morally significant.101
It has been argued that the random distribution of deficit healthcare resources has many practical advantages: it supports an equal claim to scarce resources, prevents small differences from drastically affecting outcomes, is simple, and does not require any knowledge about recipients from a decision maker. In the context of treatment during the recent pandemic, Harris,102 who treats equality as the main relevant value in many healthcare contexts, suggested that ‘choosing without preferring’ is the only morally justified way of distributing scarce healthcare resources, that is, choosing between lives without doing so in a way that shows a preference for the life or person chosen. Developing his previous ideas on this topic, he argued that this is the only method that counts every person for one and none for more than one: ‘so long as we each wish to live out the rest of our lives, however long that turns out to be, then if we do not deserve to die, we each suffer the same injustice if our wishes are deliberately frustrated, and we are cut off prematurely’.
Harris assumes that every other distribution except the ‘choosing without preferring’ approach would be ‘insulting’ because it would treat people’s lives as ‘worthless’ (in the sense that they would be ‘worth less’ than those of others). According to this reasoning, any allocation rule other than equalitarian must implicitly assume judgments of comparative valuing between lives, which, in fact, is morally impermissible. Although he speaks about treatment, it seems, at first sight, that an analogical argument could be applied to prevention, that is, vaccinations should be randomly assigned among individuals, no matter how vulnerable they are, that is, how high IFR or ROI some of them may have. However, surprisingly, in the context of vaccination he resigns from this categorical stance and allows to prioritize such distribution of vaccines that prioritize ‘the most vulnerable to COVID-19 until there is sufficient supply for all’.103 This proposal resembles earlier approaches that tried to mix an egalitarian approach with some utilitarian components, for example, a weighted lottery approach where, every candidate for life-saving treatment would have some chance in competition with every other candidate, but those who stood to gain more life years would have greater chances.104
In the debate on COVID-19 vaccines, a weighted lottery allocation approach was conditionally defended by Schmidt.105 However, he grounded his argument not in equal chances, but in priority for the underprivileged that should reflect different levels of underlying disadvantages, both health and not-health related. In practice, Schmidt106 postulates the use of weighing based on measures such as the Area Deprivation Index (ADI), which combines income, education, employment, and housing quality data to rank neighborhoods by socioeconomic status disadvantage.107 Alternatively, a categorized priority system mentioned in the previous section108 could reserve some percentage of vaccines in every stage of the vaccination schedule to be distributed randomly (either in equal chance or weighted lottery).
Considering the unique nature of the egalitarian principle, our analysis shows that this principle was only adopted in the analyzed vaccination schemes in a limited form. The egalitarian approach serves there exclusively as a second-order principle, namely, as a pattern of distribution within already prioritized groups (that is, groups which are distinguished on the basis of some other criteria). In particular, the ‘first come, first serve’ approach was to distribute vaccines within subsequent groups. Furthermore, and particularly noteworthy, no researched priority setting adopted any chancy mechanism to distribute COVID-19 vaccines—either in the version of identical chance lottery or a weighted lottery. This may suggest that random distribution, which is often discussed by philosophers, has in fact limited practical applications in the prioritization of healthcare prevention.
V. CONCLUSIONS
The comparison of COVID-19 vaccine prioritization schedules from 29 countries shows relevant similarities among state policies. Our crucial finding is that all of the settings researched rely largely on priority criteria referring to individual IFR. In particular, being over 65 years old was commonly regarded as a reason for prioritization and coexisting health conditions were almost universally recognized as such a reason. Although particular states introduced IFR criteria in different forms, the common thing was the noticeable priority of IFR criteria over the ROI ones (where the latter includes, eg occupation and housing conditions). The IFR criteria were generally placed higher on the priority-settings than the ROI criteria.
This dominance of IFR criteria is certainly not trivial and surprising, since in the context of treatment, it is rather common and justifiable to adopt very different distribution schemes (eg introducing saving more life-years approach and prioritizing younger patients). Moreover, many authors noticed not only that far more resources are devoted to treating disease than to preventing it, but also that preventive services are usually subject to more scrupulous cost-effectiveness assessment than treatment.109 As we have argued in the discussion, this justification of this pattern of priority criteria in the case of COVID-19 vaccination is far from clear, with neither utilitarian, prioritarian nor egalitarian principles providing a straightforward basis for this distributive arrangement. From the utilitarian perspective, the dominance of IFR criteria seems to get support from the ‘saving most lives’ principle; however, this support does not come without reservations. According to the account of prioritarianism, this recognized approach might be seen as a specific version of ex ante prioritarianism. Finally, one may also discern some egalitarian traits in the schedules, but only to a very limited extent with the ‘first come, first served’ version of egalitarianism operating as a second-order principle (that is, a principle that provides a pattern of distribution within first-order groups).
How to interpret the observed patterns of prioritization in COVID-19 vaccination schedules? Do they stem from some systematic differences between curative and preventive medical interventions that may influence the prioritization rules? For example, from the fact that prioritization in the case of preventive interventions always concerns merely statistical individuals? Or from the intricacy of ascribing causal claims to the case of preventive medical interventions, which may be understood as ‘a matter of causing the nonoccurrence of an event’.110 Surprisingly, the problem of priority settings in the cases of medical preventive interventions has not been systematically discussed in the literature to such an extent as in the case of treatment.111 Furthermore, in contrast with many medical treatments (eg the allocation of organs for transplantation), there is no well-established expert consensus on this matter.
The better interpretation may come from the analysis of the situation of the urgent healthcare crisis and the peculiar context that influenced the prioritization schemes for COVID-19 vaccination in a specific way. The theoretical ambiguity of vaccine distribution patterns might paradoxically be regarded as an advantage in political practice. The reason for this is that the legitimization criteria applied by bioethical experts and the general public typically differ, whereas social legitimacy, which is crucial for the effectiveness of vaccine policies, is mainly dependent on the latter. The tension between bioethics experts and public opinion was clearly visible in the case the allocation of respirators in the first phase of the COVID-19 in the US, where the decision not to give the respirator to disabled people or people suffering from certain diseases, albeit motivated on the basis of a well-considered bioethical reasoning, aroused protests and in some cases led to changes in the guidelines.112
The main challenge for social legitimacy in most cases of political decision-making seems to be that, on the one hand, citizens deeply differ in their moral views, while on the other hand, political regulations (including priority-schedules) should ideally be acceptable to the overwhelming majority of citizens.113 Since priority schedules are inevitably value-laden, social acceptability becomes difficult to attain. In such circumstances, sticking to particular normative theory (in particular, in the situation where there is no well-established consensus among stakeholders) may be seen by the public as controversial. Knowing this, one may conclude that the fact that established schedules could be interpreted and defended on different grounds may increase their legitimacy in the eyes of the public. Such schedules could be supported by an overlapping consensus of different moral views in John Rawls’ terms.114 Since the effectiveness of vaccination programs depends (at least partly) on their social legitimacy, this openness to various interpretations appears as a crucial factor.
Nevertheless, it should be clear that we do not argue for a more general claim that such vagueness in policy concerning highly disputable issues should be regarded as beneficial in all (or even most) cases. There are many threats involved in policies that seek to gain widespread social legitimacy. First, such policies often lead to decisions, which are questionable in the light of well-established democratic standards (such as the principle of equal respect). Furthermore, the growing distrust of experts and the popularity of ‘alternative sources of knowledge’ seems to push democratic societies into establishing less informed policies.115 At the same time, in situations where expert input is indispensable, political leaders tend to use scientific expertise as a way to lessen the responsibility placed on them and reduce their accountability.116 We are aware of these threats, and our conclusion is much more modest for this reason. We claim that the ambiguity of the researched vaccine policies could be regarded as an advantage considering the very special circumstances in which these policies were adopted. In particular, one should note that the pandemic has pushed public attention to questions about priority settings that are typically not in the spotlight. As it could have been anticipated during early days of the pandemic, the focus on health issues provided grounds for the rise of vaccine hesitancy and anti-vaxx movements. Last but not least, all of this happened on an unprecedented scale and under conditions of high uncertainty about the future dynamics of the pandemic.
Supplementary Material
ACKNOWLEDGEMENTS
The authors would like to thank Angelika Mus-Nowak for the help regarding all the graphics and tables used in the article. Special thanks also go to Piotr Bystranowski for the consult on graphic content and help with preparation of the Figure 1 and to Izabela Zwiech for helpful notes on preparing Table B (Supplementary Materials). The authors also thank Vilius Dranseika, Aeddan Shaw and Łukasz Woiński for their help with translation of the schedules from European languages not known to authors and valuable linguistic consult. Members of the team of the Interdisciplinary Centre for Ethics at the Jagiellonian University in Kraków, and the audience at the Oxford Global Health and Bioethics International Conference (June 27–28, 2022) provided useful feedback on earlier drafts. Tomasz Żuradzki is grateful to the Brocher Foundation (www.brocher.ch) for hosting him as a visiting fellow in March 2022.
FUNDING
This research has received funding from the European Research Council (ERC) under the European Union’s Horizon 2020 research and innovation program [grant agreement 805498].
Footnotes
For review of these schedules see: Hans-Jörg Erni, Urban Wiesing, and Robert Ranisch, Saving the Most Lives—A Comparison of European Triage Guidelines in the Context of the COVID-19 Pandemic, 35.2 Bioethics 125–134 (2021); Kristina Orfali, Getting to the Truth: Ethics, Trust, and Triage in the United States versus Europe during the Covid-19 Pandemic, 51.1 Hastings Center Report 15–22 (2021); Marta Perin and Ludovica De Panfilis, Among Equity and Dignity: An Argument-based Review of European Ethical Guidelines Under COVID-19, 22.1 BMC medical ethics 1–29 (2021); Alex Rajczi, Judith Daar, Aaron Kheriaty, and Cyrus Dastur, The University of California Crisis Standards of Care: Public Reasoning for Socially Responsible Medicine, 51.5 Hastings Center Report 30–41, 2021; Emily Score, Kyeong Yun Jeong, Mary Katherine Gaurke, Bernard Prusak, and Daniel P. Sulmasy, Rationing With Respect to Age During a Pandemic: A Comparative Analysis of State Pandemic Preparedness Plans, 161.2 Chest 504–513 (2022).
Harald Schmidt, Rebecca Weintraub, Michelle A. Williams, Kate Miller, Alison Buttenheim, Emily Sadecki, Helen Wu, Aditi Doiphode, Neha Nagpal, Lawrence O. Gostin & Angela A. Shen, Equitable Allocation of COVID-19 Vaccines in the United States, 27 Nature Medicine 298–307 (2021).
Govind Persad, Alan Wertheimer, and Ezekiel J. Emanuel, Principles for Allocation of Scarce Medical Interventions, 373.9661 The Lancet 423–431 (2009).
Douglas B. White and Bernard Lo, A Framework for Rationing Ventilators and Critical Care Beds During the COVID-19 Pandemic, 323.18 Jama 1773–1774 (2020).
Persad et al. supra note 3.
Saul Smilansky A Hostage Situation, 116.8 The Journal of Philosophy, 447–466 (2019).
For an overview of the literature on prioritizing access to pandemic infuenza vaccines, see: Jane H. Williams and Angus Dawson, Prioritising Access to Pandemic Influenza Vaccine: A Review of the Ethics Literature, 21.1 BMC Medical Ethics 1–8 (2020).
Ezekiel J. Emanuel, Govind Persad, Ross Upshur, Beatriz Thome, Michael Parker, Aaron Glickman, Cathy Zhang, Connor Boyle, Maxwell Smith, and James P. Phillips, Fair Allocation of Scarce Medical Resources in the Time of Covid-19, 382.21 New England Journal of Medicine 2049–2055 (2020).
Ezekiel J. Emanuel, Govind Persad, Adam Kern, Allen Buchanan, Cécile Fabre, Daniel Halliday, Joseph Heath et al., An Ethical Framework for Global Vaccine Allocation, 369.6509 Science 1309–1312 (2020).
Alberto Giubilini, Julian Savulescu, and Dominic Wilkinson, Queue Questions: Ethics of COVID-19 Vaccine Prioritization, 35.4 Bioethics 348–355 (2021).
European Union, European Union’s Vaccines Strategy,https://ec.europa.eu/info/live-work-travel-eu/coronavirus-response/public-health/eu-vaccines-strategy_en#possible-priority-groups (accessed July 21, 2021).
Schmidt et al. supra note 2.
Advance Purchase Agreements between the European Union and Pfizer Inc./BioNTech Manufacturing GmbH, Moderna Switzerland GmbH, Astrazeneca AB as well as Janssen Pharmaceutica NV, https://ec.europa.eu/info/sites/default/files/redacted_advance_purchase_agreement_biontech-pfizer_0.pdf, https://ec.europa.eu/info/sites/default/files/redacted_advance_purchase_agreement_moderna_0.pdf, https://ec.europa.eu/info/sites/default/files/apa_astrazeneca.pdf, https://ec.europa.eu/info/sites/default/files/jj_apa_202005071550.pdf (accessed Jul. 21, 2021).
Robert Koch Institute, STIKO-Empfehlung zur COVID-19-Impfung. Beschluss der STIKO zur 2. Aktualisierung der COVID-19-Impfempfehlung und die dazugehörige wissenschaftliche Begründung, Epidemiologisches Bulletin (2021).
Eg. Cass R. Sunstein, On the Expressive Function of Law, 144.5 University of Pennsylvania Law Review 2021–2053 (1996); Elizabeth S. Anderson and Richard H. Pildes, Expressive Theories of Law: A General Restatement, 148.5 University of Pennsylvania Law Review 1503–1575 (2000); Deborah Hellman, Judging by Appearances: Professional Ethics, Expressive Government, and the Moral Significance of How Things Seem, 60 Md. L. Rev. 653 (2001).
Schmidt et al. supra note 2.
Willard V. Quine, On Empirically Equivalent Systems of the World, 9.3 Erkenntnis 313–328 (1975).
Derek Parfit, On what matters. Vol. 1 2011.
Andrew T. Levin, William P. Hanage, Nana Owusu-Boaitey, Kensington B. Cochran, Seamus P. Walsh, and Gideon Meyerowitz-Katz, Assessing the Age Specificity of Infection Fatality Rates for COVID-19: Systematic Review, Meta-analysis, and Public Policy Implications, 35.12 European Journal of Epidemiology 1123–1138 (2020).
Reed JD Serensen, Ryan M. Barber, David M. Pigott, Austin Carter, Cory N. Spencer, Samuel M. Ostroff, Robert C. Reiner Jr et al., Variation in the COVID-19 Infection-Fatality Ratio by Age, Time, and Geography During the Pre-Vaccine Era: A Systematic Analysis, 399.10334 Lancet 1469–1488 (2022).
Robert Koch Institute, supra note 14.
Kalpana Thapa Bajgain, Sujan Badal, Bishnu B. Bajgain, and Maria J. Santana, Prevalence of Comorbidities Among Individuals with COVID-19: A Rapid Review of Current Literature, 49.2 American Journal of Infection Control 238–246 (2021).
Yanbin Du, Nan Zhou, Wenting Zha, and Yuan Lv, Hypertension Is a Clinically Important Risk Factor for Critical Illness and Mortality in COVID-19: A Meta-Analysis, 31.3 Nutrition, Metabolism and Cardiovascular Diseases 745–755 (2021).
Sung A Bae, So Ree Kim, Mi-Na Kim, Wan Joo Shim, and Seong-Mi Park, Impact of Cardiovascular Disease and Risk Factors on Fatal Outcomes in Patients with COVID-19 According to Age: A Systematic Review and Meta-Analysis, 107.5 Heart 373–380 (2021).
Zeng-hong Wu, Yun Tang, and Qing Cheng, Diabetes Increases the Mortality of Patients with COVID-19: A Meta-Analysis, 58.2 Acta Diabetologica 139–144 (2021).
Jia Song, Ming Zeng, Hai Wang, Chuan Qin, Hong-Yan Hou, Zi-Yong Sun, San-Peng Xu et al., Distinct Effects of Asthma and COPD Comorbidity on Disease Expression and Outcome in Patients with COVID-19, 76.2 Allergy 483–496 (2021).
Ruyi Cai, Jinshi Zhang, Yifan Zhu, Lin Liu, Yueming Liu, and Qiang He, Mortality in Chronic Kidney Disease Patients with COVID-19: A Systematic Review and Meta-Analysis, 53.8 International Urology and Nephrology 1623–1629 (2021).
QuanQiu Wang, Nathan A. Berger, and Rong Xu, Analyses of Risk, Racial Disparity, and Outcomes of Cancers and Coronavirus Disease 2019 from Electronic Health 4 Records in the US, 10 JAMA Oncol (2020).
For a general estimation of preexisting risk of mortality from COVID-19 see an umbrella review by Robert Koch Institute, supra note 14.
Patricia M. Kearney, Megan Whelton, Kristi Reynolds, Paul K. Whelton, and Jiang He, Worldwide Prevalence of Hypertension: A Systematic Review, 22.1 Journal of Hypertension 11–19 (2004).
Prachi Bhatnagar, Kremlin Wickramasinghe, Julianne Williams, Mike Rayner, Nick Townsend, The Epidemiology of Cardiovascular Disease in the UK, 101.15 Heart (2015).
Nita Gandhi Forouhi, and Nicholas J. Wareham, Epidemiology of Diabetes, 38.11 Medicine 602–606 (2010).
Eg. QuanQiu Wang, Rong Xu, and Nora D. Volkow, Increased Risk of COVID-19 Infection and Mortality in People with Mental Disorders: Analysis from Electronic Health Records in the United States, 20.1 World Psychiatry 124–130 (2021); Huazhen Yang, Wenwen Chen, Yao Hu, Yilong Chen, Yu Zeng, Yajing Sun, Zhiye Ying et al., Pre-pandemic Psychiatric Disorders and Risk of COVID-19: A UK Biobank Cohort Analysis, 1.2 The Lancet Healthy Longevity e69-e79 (2020); Maxime Taquet, Sierra Luciano, John R. Geddes, and Paul J. Harrison, Bidirectional Associations Between COVID-19 and Psychiatric Disorder: Retrospective Cohort Studies of 62 354 COVID-19 Cases in the USA, 8.2 The Lancet Psychiatry 130–140 (2021); Martin Maripuu, Marie Bendix, Louise Öhlund, Micael Widerström, and Ursula Werneke, Death Associated with Coronavirus (COVID-19) Infection in Individuals with Severe Mental Disorders in Sweden During the Early Months of the Outbreak—An Exploratory Cross-Sectional Analysis of a Population-based Register Study, 11 Frontiers in Psychiatry 609579 (2021); Luming Li, Fangyong Li, Frank Fortunati, and John H. Krystal, Association of a Prior Psychiatric Diagnosis with Mortality Among Hospitalized Patients with Coronavirus Disease 2019 (COVID-19) Infection, 3.9 JAMA Network Open e2023282-e2023282 (2020); Katlyn Nemani, Chenxiang Li, Mark Olfson, Esther M. Blessing, Narges Razavian, Ji Chen, Eva Petkova, and Donald C. Goff, Association of Psychiatric Disorders with Mortality Among Patients with COVID-19, 78.4 JAMA Psychiatry 380–386 (2021).
Marc De Hert, Christoph U. Correll, Julio Bobes, Marcelo Cetkovich-Bakmas, D. A. N. Cohen, Itsuo Asai, Johan Detraux et al., Physical Illness in Patients with Severe Mental Disorders. I. Prevalence, Impact of Medications and Disparities in Health Care, 10.1 World Psychiatry 52 (2011).
Victor Mazereel, Kristof Van Assche, Johan Detraux, and Marc De Hert, COVID-19 Vaccination for People with Severe Mental Illness: Why, What, and How?, 8.5 The Lancet Psychiatry 444–450 (2021).
Eg. Id; Marc De Hert, Victor Mazereel, Johan Detraux, and Kristof Van Assche, Prioritizing COVID-19 Vaccination for People with Severe Mental Illness, 20.1 World Psychiatry 54 (2021).
Eg. Anoop SV Shah, Rachael Wood, Ciara Gribben, David Caldwell, Jennifer Bishop, Amanda Weir, Sharon Kennedy et al., Risk of Hospital Admission with Coronavirus Disease 2019 in Healthcare Workers and Their Households: Nationwide Linkage Cohort Study, 371 BMJ (2020); Güven Çelebi, Nihal Pişkin, Arzum Çelik Bekleviç, Yurdagül Altunay, Ayşegül Salcı Keleş, Mehmet Ali Tüz, Bülent Altınsoy, and Demet Hacıseyitoğlu, Specific Risk Factors for SARS-CoV-2 Transmission Among Health Care Workers in a University Hospital, 48.10 American Journal of Infection Control 48, no. 10 1225–1230 (2020).
Fan-Yun Lan, Christian Suharlim, Stefanos N. Kales, and Justin Yang, Association between SARS-CoV-2 Infection, Exposure Risk and Mental Health Among a Cohort of Essential Retail Workers in the USA, 78.4 Occupational and Environmental Medicine 237–243 (2021).
Marissa G. Baker, Trevor K. Peckham, and Noah S. Seixas, Estimating the Burden of United States Workers Exposed to Infection or Disease: A Key Factor in Containing Risk of COVID-19 infection, 15.4 PloS one e0232452 (2020).
Eg. Arnaud Fontanet, Rebecca Grant, Margrethe Greve-Isdahl, and Devi Sridhar, Covid-19: Keeping Schools as Safe as Possible, 371 BMJ (2021); Jay K. Varma, Jeff Thamkittikasem, Katherine Whittemore, Mariana Alexander, Daniel H. Stephens, Kayla Arslanian, Jackie Bray, and Theodore G. Long, COVID-19 Infections Among Students and Staff in New York City Public Schools, 147.5 Pediatrics (2021).
Eveline Otte Im Kampe, Ann-Sophie Lehfeld, Silke Buda, Udo Buchholz, and Walter Haas, Surveillance of COVID-19 School Outbreaks, Germany, March to August 2020, 25.38 Eurosurveillance 2001645 (2020).
Michael Zhang, Estimation of Differential Occupational Risk of COVID-19 by Comparing Risk Factors with Case Data by Occupational Group, 64.1 American Journal of Industrial Medicine 39–47 (2021).
Karin Magnusson, Karin Nygård, Fredrik Methi, Line Vold, and Kjetil Telle, Occupational Risk of COVID-19 in the First Versus Second Epidemic Wave in Norway, 2020., 26.40 Eurosurveillance 2001875 (2021); Folkhälsomyndigheten, Förekomst av COVID-19 i olika yrkesgrupper. Bekräftade COVID-19 fall i Sverige 13 mars—27 maj 2020. [Prevalence of COVID-19 in Different Occupational Groups. Confirmed COVID-19 Cases in Sweden March 13–May 27, 2020.] Stockholm: Folkhälsomyndigheten; (2020), https://www.folkhalsomyndigheten.se/contentassets/5e248b82cc284971a1c5fd922e7770f8/forekomst-covid-19-olika-yrkesgrupper.pdf (accessed Jul. 21, 2021).
Amy Falk, Alison Benda, Peter Falk, Sarah Steffen, Zachary Wallace, and Tracy Beth Høeg, COVID-19 Cases and Transmission in 17K–12 Schools—Wood County, Wisconsin, August 31–November 29, 2020., 70.4 Morbidity and Mortality Weekly Report 136 (2021); Jay K. Varma, Jeff Thamkittikasem, Katherine Whittemore, Mariana Alexander, Daniel H. Stephens, Kayla Arslanian, Jackie Bray, and Theodore G. Long, COVID-19 Infections Among Students and Staff in New York City Public Schools, 147.5 Pediatrics (2021).
Nathan M. Stall, Kevin A. Brown, Antonina Maltsev, Aaron Jones, Andrew P. Costa, Vanessa Allen, and M. P. Hillmer, COVID-19 and Ontario’s Long-Term Care Homes,Science Briefs of the Ontario COVID19 Science Advisory Table (2021); Michela D’ascanio, Marta Innammorato, Lara Pasquariello, Dario Pizzirusso, Giulio Guerrieri, Silvia Castelli, Aldo Pezzuto et al., Age Is Not the Only Risk Factor in COVID-19: The Role of Comorbidities and of Long Staying in Residential Care Homes, 21.1 BMC Geriatrics 1–10 (2021); Shamez N. Ladhani, Anna Jeffery-Smith, Monika Patel, Roshni Janarthanan, Jonathan Fok, Emma Crawley-Boevey, Amoolya Vusirikala et al., High Prevalence of SARS-CoV-2 Antibodies in Care Homes Affected by COVID-19: Prospective Cohort Study, England, 28 EClinicalMedicine 100597 (2020).
Emily Mosites, Erin M. Parker, Kristie EN Clarke, Jessie M. Gaeta, Travis P. Baggett, Elizabeth Imbert, Madeline Sankaran et al., Assessment of SARS-CoV-2 Infection Prevalence in Homeless Shelters—Four US Cities, March 27–April 15, 2020, 69.17 Morbidity and Mortality Weekly Report 521 (2020); Elizabeth Imbert, Patrick M Kinley, Ashley Scarborough, Caroline Cawley, Madeline Sankaran, Sarah N Cox, Margot Kushel, Juliet Stoltey, Stephanie Cohen, Jonathan D Fuchs, Coronavirus Disease 2019 (COVID-19) Outbreak in a San Francisco Homeless Shelter, Clinical Infectious Diseases (2020).
Monik C. Jiménez, Tori L. Cowger, Lisa E. Simon, Maya Behn, Nicole Cassarino, and Mary T. Bassett, Epidemiology of COVID-19 Among Incarcerated Individuals and Staff in Massachusetts Jails and Prisons, 3.8 JAMA Network Open e2018851–e2018851 (2020).
World Health Organization. WHO SAGE Roadmap For Prioritizing Uses of COVID-19 Vaccines In The Context of Limited Supply (2020), https://www.who.int/publications/m/item/who-sage-roadmap-for-prioritizing-uses-of-covid-19-vaccines-in-the-context-of-limited-supply (accessed Jul. 21, 2021).
For the distinction between different levels of identifiability, see: Tomasz Żuradzki, The Normative Significance of Identifiability, 21.4 Ethics and Information Technology 295–305 (2019).
Persad et al. supra note 3.
Emanuel et al. supra note 9.
Steven Schwartz, Jeffrey Richardson, and Paul P. Glasziou, Quality-Adjusted Life Years: Origins, Measurements, Applications, Objections, 17.3 Australian Journal of Public Health 272–278 (1993).
Mark S. Stein, The Distribution of Life-saving Medical Resources: Equality, Life Expectancy, and Choice Behind the Veil, 19.2 Social Philosophy and Policy 212–245 (2002).
Giubilini et al. supra note 10.
Id.
Deborah Hellman and Kate M. Nicholson, Rationing and Disability: The Civil Rights and Wrongs of State Triage Protocols, 78 Wash. & Lee L. Rev. 1207 (2021).
Dominic J. Wilkinson, Frailty Triage: Is Rationing Intensive Medical Treatment on the Grounds of Frailty Ethical?, 21.11 The American Journal of Bioethics 48–63 (2021).
Marco Vergano, Guido Bertolini, Alberto Giannini, Giuseppe R. Gristina, Sergio Livigni, Giovanni Mistraletti, Luigi Riccioni & Flavia Petrini, Clinical Ethics Recommendations for the Allocation of Intensive Care Treatments in Exceptional, Resource-Limited Circumstances: The Italian Perspective During the COVID-19 Epidemic, 24 Critical Care 165 (2020).
Sam Moore, Edward M. Hill, Louise Dyson, Michael J. Tildesley, and Matt J. Keeling, Modelling Optimal Vaccination Strategy for SARS-CoV-2 in the UK, 17.5 PLoS Computational Biology e1008849 (2021).
Joshua R. Goldstein, Thomas Cassidy, and Kenneth W. Wachter, Vaccinating the Oldest Against COVID-19 Saves Both the Most Lives and Most Years of Life, 118.11 Proceedings of the National Academy of Sciences e2026322118 (2021). Because of our expressive approach we do not analyze how far, if at all, official vaccination schedules followed such modelling studies, and which values have been implemented in these models. For an analysis of value judgments in a COVID-19 vaccination model, see: Stephanie Harvard, Eric Winsberg, John Symons, and Amin Adibi, Value Judgments in a COVID-19 Vaccination Model: A Case Study in the Need for Public Involvement in Health-Oriented Modelling, 286 Social Science & Medicine 114323 (2021).
Robert Koch Institute, supra note 14.
L. Devlin and P. J. Morrison, Accuracy of the Clinical Diagnosis of Down Syndrome, 73.1 The Ulster Medical Journal, 4 (2004).
Although the UK put ‘clinically extremely vulnerable individuals’ before people aged 65–69 (in phase 4), UK policymakers generally do not consider people with Down’s syndrome or people after transplants as part of this group. Instead, they are part of the phase 6 ‘persons with underlying health conditions’ group (The United Kingdom’s Department of Health and Social Care’s Joint Committee on Vaccination and Immunization, Advice on Priority Groups for COVID-19 Vaccination, https://www.gov.uk/government/publications/priority-groups-for-coronavirus-Covid-19-vaccination-advice-from-the-jcvi-30-december-2020/joint-committee-on-vaccination-and-immunisation-advice-on-priority-groups-for-Covid-19-vaccination-30-december-2020, 5–8 (2020) (accessed Jul. 21, 2021)). At the same time, the description of ‘clinically extremely vulnerable individuals’ in government documents is very vague and it is explicitly stated that some members of the groups included in ‘persons with underlying health conditions’ can be classified as ‘clinically extremely vulnerable’ and vaccinated earlier, ie, in phase 4, before people aged 65 to 69.
Levin et al. supra note 19.
Id.
Eg., see this observational study in one supermarket on 104 workers: Lan et al. supra note 38.
Caspar Hare, Risk and Radical Uncertainty in HIV Research, 43.2 J Med Ethics 87–89 (2017).
Lukas Kleine-Rueschkamp, and Cem Ozguzel, COVID-19 and Key Workers: What Role Do Migrants Play in Your Region? OECD Policy Responses to Coronavirus (COVID-19) (2020), https://www.oecd.org/coronavirus/policy-responses/Covid-19-and-key-workers-what-role-do-migrants-play-in-your-region-42847cb9/ (accessed Nov. 14, 2021).
Ruxandra Paul, Europe’s Essential Workers: Migration and Pandemic Politics in Central and Eastern Europe during COVID-19, 6.2 European Policy Analysis 238–263 (2020).
Eg. Joshua R. Goldstein, Thomas Cassidy, and Kenneth W. Wachter, Vaccinating the Oldest Against COVID-19 Saves Both the Most Lives and Most Years of Life, 118.11 Proceedings of the National Academy of Sciences e2026322118 (2021).
Giubilini et al. supra note 10.
Id.
Thana Cristina de Campos-Rudinsky, and Eduardo Undurraga, Public Health Decisions in the COVID-19 Pandemic Require More Than ‘Follow the Science’, 47.5 Journal of medical ethics 296–299 (2021).
Hilary Greaves, Cluelessness, 116.3 Proceedings of the Aristotelian Society 311–339 (2016).
Derek Parfit, Equality and Priority, 3 Ratio 202–221 (1997).
Dan W. Brock, Priority to the Worse Off in Health-Care Resource Prioritization, Medicine and Social Justice 362–372 (2002); Daniel Hausman, The Significance of ‘Severity’, 45.8 Journal of Medical Ethics 545–551 (2019); Lasse Nielsen, Pandemic Prioritarianism, 48.4 Journal of Medical Ethics 236–239 (2022).
Trygve Ottersen, Lifetime QALY Prioritarianism in Priority Setting, 39.3 Journal of Medical Ethics 175–180 (2013).
Shlomi Segall, Why Inequality Matters (2016).
Persad et al. supra note 3.
Greg Bognar, Fair Innings, 29.4 Bioethics 251–261 (2015). It is important to distinguish here a different motivation for prioritising the younger. In the early days of pandemic (June 2020) some suggested that under specific assumptions (which, as it later turned out, were not accurate) it would be more efficient to prioritize younger, even if our aim was to protect the most vulnerable (eg those with the highest IVF). For example, Giubilini et al. (Alberto Giubilini, Julian Savulescu, and Dominic Wilkinson, COVID-19 Vaccine: Vaccinate the Young to Protect the Old?, 7.1 Journal of Law and the Biosciences (2020)) argued that protecting the most vulnerable might require prioritizing vaccinating children in order to maximize the benefits of indirect immunity for the elderly and the other vulnerable groups (see also Rosamond Rhodes, Justice in COVID-19 vaccine prioritization: rethinking the approach, 47.9 Journal of Medical Ethics 623–631 (2021)).
Persad et al. supra note 3.
Persad and Lo, supra note 4.
Maddalena Ferranna, J. P. Sevilla, and David E. Bloom, Addressing the COVID-19 Pandemic: Comparing Alternative Value Frameworks, in: Prioritarianism in Practice. (Matthew D. Adler and Ole F. Norheim ed., 2022)
Thomas Nagel, Equality, in: Ibidem, Mortal Questions (1979). For using this method in the context of COVID-19, see: Frances Kamm, Moral Reasoning in a Pandemic, Boston Review (2020).
Govind Persad, Monica E. Peek, and Ezekiel J. Emanuel, Fairly Prioritizing Groups for Access to COVID-19 Vaccines, 324.16 Jama 1601–1602 (2020).
Armand H. Matheny Antommaria, Tyler S. Gibb, Amy L. McGuire, Paul Root Wolpe, Matthew K. Wynia, Megan K. Applewhite, Arthur Caplan et al., Ventilator Triage Policies During the COVID-19 Pandemic at US Hospitals Associated with Members of the Association of Bioethics Program Directors, 173.3 Annals of Internal Medicine 188–194 (2020).
Alex Rajczi, Judith Daar, Aaron Kheriaty, and Cyrus Dastur, The University of California Crisis Standards of Care: Public Reasoning for Socially Responsible Medicine, 51.5 Hastings Center Report 30–41, 2021; see also: Mary Katherine Gaurke, Bernard Prusak, Kyeong Yun Jeong, Emily Scire, and Daniel P. Sulmasy, Life-Years & Rationing in the Covid-19 Pandemic: A Critical Analysis, 51.5 Hastings Center Report 18–29 (2021).
Marta Perin and Ludovica De Panfilis, Among Equity and Dignity: An Argument-Based Review of European Ethical Guidelines Under COVID-19, 22.1 BMC Medical Ethics 1–29 (2021).
Shlomi Segall, Why Inequality Matters (2016).
Tina Rulli and Joseph Millum, Rescuing the Duty to Rescue, 42.4 Journal of Medical Ethics 260–264 (2016).
Torbjörn Tännsjö, Setting Health-care Priorities: What Ethical Theories Tell Us (2019).
Mirko Farina, and Andrea Lavazza, Advocating for Greater Inclusion of Marginalized and Forgotten Populations in COVID19 Vaccine Rollouts, 66 International Journal of Public Health 1604036 (2021).
Tayfun Sönmez, Parag A. Pathak, M. Utku Ünver, Govind Persad, Robert D. Truog, and Douglas B. White, Categorized Priority Systems: A New Tool for Fairly Allocating Scarce Medical Resources in the Face of Profound Social Inequities, 159.3 Chest 1294–1299 (2021).
Schmidt et al. supra note 2.
Michael Otsuka and Alex Voorhoeve, Why It Matters that Some Are Worse off than Others: An Argument Against the Priority View, 37.2 Philosophy & Public Affairs 171–199 (2009); Nils Holtug, Prioritarianism, in: Oxford Research Encyclopedia of Politics (2017); Matthew D. Adler, Measuring Social Welfare: An Introduction (2019).
Persad et al. supra note 85.
Persad et al. supra note 3.
John Harris, Why Kill the Cabin Boy?, 30.1 Cambridge Quarterly of Healthcare Ethics 4–9 (2021).
Harald Schmidt, Vaccine Rationing and the Urgency of Social Justice in the COVID-19 Response, 50.3 Hastings Center Report 46–49 (2020).
See also: James F. Childress, Putting Patients First in Organ Allocation: An Ethical Analysis of the US Debate, 10.4 Cambridge Quarterly of Healthcare Ethics 365–376 (2001); Norman Daniels, Fair Process in Patient Selection for Antiretroviral Treatment in WHO’s Goal of 3 by 5, 366.9480 The Lancet 169–171 (2005); Persad et al. supra note 3.
Tyler M. John, and Joseph Millum, First Come, First Served?, 130.2 Ethics 179–207 (2020).
John Harris, Combatting COVID-19. Or, “All Persons Are Equal but Some Persons Are More Equal than Others?”, 30.3 Cambridge Quarterly of Healthcare Ethics 406–414 (2021).
Id.
Dan W. Brock, Justice and the ADA: Does Prioritizing and Rationing Health Care Discriminate Against the Disabled?, 12.2 Social Philosophy and Policy 159–185 (1995).
Schmidt, supra note 99.
Id.
Schmidt et al. supra note 2.
Sönmez et al. supra note 93.
Halley S. Faust, and Paul T. Menzel, eds., Prevention vs. Treatment: What’s the Right Balance? (2011).
John Collins, Preemptive Prevention, 97 Journal of Philosophy 223–234 (2000).
But see: Faust and Menzel, supra note 109.
Orfali, supra note 1.
Andrea Lavazza and Mirko Farina, The Role of Experts in the Covid-19 Pandemic and the Limits of Their Epistemic Authority in Democracy, 8 Frontiers in Public Health 356 (2020).
John Rawls, Political Liberalism. (1993).
Tom Nichols, The death of Expertise: The Campaign Against Established Knowledge and Why It Matters (2017).
Christopher M. Weible, Daniel Nohrstedt, Paul Cairney, David P. Carter, Deserai A. Crow, Anna P. Durnová, Tanya Heikkila, Karin Ingold, Allan McConnell, and Diane Stone, COVID-19 and the Policy Sciences: Initial Reactions and Perspectives, 53.2 Policy Sciences 225–241 (2020).
Contributor Information
Karolina Wiśniowska, Interdisciplinary Centre for Ethics, Jagiellonian University, Grodzka Street 52, Kraków, Poland.
Tomasz Żuradzki, Institute of Philosophy & Interdisciplinary Centre for Ethics, Jagiellonian University, Grodzka Street 52, Kraków, Poland.
Wojciech Ciszewski, Law and Administration Department, Jagiellonian University, Bracka Street 12, Kraków, Poland.
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.