Article Text

Download PDFPDF

COVID-19 ventilator rationing protocols: why we need to know more about the views of those with most to lose
  1. Whitney Kerr,
  2. Harald Schmidt
  1. Department of Medical Ethics and Health Policy, University of Pennsylvania Perelman School of Medicine, Philadelphia, Pennsylvania, USA
  1. Correspondence to Whitney Kerr, Department of Medical Ethics and Health Policy, University of Pennsylvania Perelman School of Medicine, Philadelphia, PA 19104, USA; wjea{at}


Withholding or withdrawing life-saving ventilators can become necessary when resources are insufficient. With rising cases in many countries, and likely further peaks in the coming colder seasons, ventilator triage guidance remains a central part of the COVID-19 policy response. The dominant model in ventilator triage guidelines prioritises the ethical principles of saving the most lives and saving the most life-years. We sought to ascertain to what extent this focus aligns, or conflicts, with the preferences of disadvantaged minority populations. We conducted a bibliographical search of PubMed and Google Scholar and reviewed all ventilator rationing guidelines included in major recent systematic reviews, yielding 589 studies before screening. Post screening, we found six studies comprising a total of 10 591 participants, with 1247 from disadvantaged populations. Three studies reported findings stratified by race and age, two of which stratified by income. Studies included two to seven principles; all included ‘save the most lives’. Involvement of disadvantaged minority populations in eliciting preferences is very limited; few studies capture race and income. This is concerning, as despite relatively small numbers and framing effects there is an observable and plausible trend suggesting that disadvantaged groups worry that dominant principles reduce their chances of receiving a ventilator. To avoid compounding prior historical and structural disadvantage, policy makers need to engage more fully with these populations in designing and justifying ventilator rationing guidance and review their adequacy. Likewise, clinicians need to be aware that their implementation of dominant triage guidelines is viewed with higher levels of concern by minority populations.

  • allocation of health care resources
  • distributive justice
  • ethics
  • resource allocation
  • COVID-19

This article is made freely available for use in accordance with BMJ’s website terms and conditions for the duration of the covid-19 pandemic or until otherwise determined by BMJ. You may use, download and print the article for any lawful, non-commercial purpose (including text and data mining) provided that all copyright notices and trade marks are retained.

Statistics from


Rationing of resources becomes necessary when demand outstrips supply. In the wake of the initial US surge of the COVID-19 pandemic, numerous hospitals and states developed plans for how to allocate ventilators.1–4 We seek to determine what is known about the views of historically and structurally disadvantaged populations on guiding rationing principles. This analysis matters practically for designing and reviewing triage models in the ongoing epidemic, and conceptually for assessing their underlying values.

The dominant model in ventilator triage guidelines prioritises the principles of saving the most lives and/or life-years to maximise benefits.1 5–10 The main aim is to avoid wasting ventilators on people likely dying while being ventilated or likely dying soon after hospital discharge. By selecting patients most likely to survive treatment, and to live sufficiently long afterwards, each ventilator generates more lives and life-years. These principles can be controversial due to their utilitarian underpinning, and because their ahistorical ‘colorblind’ perspective can compound further the disadvantage of already disadvantaged populations.10

In the US context, a particular concern is the impact of such principles on the chances of racial minority communities (black, brown, Indigenous) to receive life-saving treatment, who already experienced a disproportionate impact of COVID-19-related unemployment, hospitalisations, morbidity and mortality (with on average two and up to six times as many black, brown, and Indigenous residents dying from COVID-19 compared with whites).2 10–13 The effects of structural racism also surface in major discrepancies in US life expectancy, which matters for considering the remaining life expectancy after discharge.10 These discrepanies are largely due to circumstances of living, including limited access to public health goods and clinical services.10–13

Guideline authors frequently emphasise drawing on both expert and community input in establishing the guiding principles of ventilator rationing guidelines. However, recent systematic reviews highlight significant variations in the manner and transparency of engagement efforts.1 2 4 Further, two commonly cited community projects report extensive community engagement, but do not provide quantitative data around community feedback and ultimately recommend a save-the-most-lives approach take priority.5 6 Given the high stakes, an important question is to what extent dominant ventilator allocation guidelines align, or stand in conflict with, the preferences of the disadvantaged communities.


We conducted a bibliographical PubMed and Google Scholar search of studies on public attitudes towards rationing ventilators or other life-saving treatment in an influenza or respiratory illness pandemic, published between 1 January 2000 and 15 November 2020 (the search string is shown in figure 1). Using a snowball method, keyword searches were conducted in all journals where identified papers were published, complemented by a cited-works search of all identified studies. We reviewed all references to public engagement in recent systematic reviews of state-level guidance.1 2 4 We supplemented with a general internet search using the terms shown in figure 1. As a result, a total of 589 studies were identified and 6 met the inclusion criteria (figure 1). As a pragmatic proxy we operationalised the concept of ‘disadvantaged groups’ as racial minorities and low-income groups. For quantitative analyses, where studies provided data on both, we included the group with the largest number to avoid undercounting involvement of ‘disadvantaged population’.

Figure 1

Search schematic. MeSH, medical subject headings.


Six identified studies comprised a total of 10 591 participants, with 1247 from disadvantaged populations (figure 1 and table 1). Three studies reported findings stratified by race and age, with two of these three stratifying by income, or geographical locations serving as proxies for these demographics.8 14 15 Studies included as few as two principles to up to seven.8 14–18 All included save the most lives.8 14–18 Five included save the most life-years.8 14 16–18 Two studies directly included the principle of help the worst off (worded ‘save those most likely to die if not treated’ and ‘to patients who are most at need in the moment, which may mean fewer people overall survive, but doctors do not deny treatment based on age or health status’); four had no equivalent.8 14–18 Findings are summarised chronologically in the following sections.

Table 1

Summary of findings


In Biddison et al,8 83% of white participants selected save the most lives for ‘always or often use’, with the highest approval across six principles. By contrast, only a third of black participants shared this preference. In a follow-up study entailing the same six principles, Biddison et al 14 showed significantly lower odds among black participants to prefer ‘always or often use’ for saving the most life-years, compared with white participants (OR 0.34; 95% CI 0.21 to 0.58, odds vs first come, first served). In Pew Research Center,15 comparing two principles, 60% of black respondents favoured helping the worst off versus saving the most lives, while white participants were evenly split.


InBiddison et al 8 race tracked income, and economically advantaged groups clearly preferred save the most lives, compared with economically disadvantaged groups. In Pew Research Center,15 upper income respondents slightly favoured save the most lives (52%), while participants in the low income category favoured help the worst off (56%).


InBiddison et al 8 the majority of older participants (83%) preferred save the most lives for ‘always or often use’ compared with 33% of younger participants. Additionally, 18% of younger adults preferred the life cycle principle for ‘always or often use’ compared with 3% of older participants. Biddison et al 14 found slightly different results, stating that ‘those younger than 60 were significantly more likely to object to the use of life cycle criterion’. Pew Research Center15 also found differing results, with 58% of participants under 30 preferring save the most lives and 57% of participants over 50 preferring help the worst off.


We highlight here three key findings. First, overall, we found few studies that examined community preferences for rationing principles. Of the six studies identified only half captured preferences by race, income and age, which matters when considering who is represented in community samples and assessing involvement of disadvantaged groups. Second, in the studies that do examine between-group differences we found a trend, but also nuances, in preferences. Finally, framing effects in the number and wording of principles require attention.

The small number of studies, combined with the fact that only half reported on differences by race, income or age, limits understanding of group-specific community preferences for guiding rationing principles. Yet between-group differences are important as economically disadvantaged, racial minorities experience structural racism and far higher incidences of comorbid conditions and poor health.2 3 10 Only 11.8% of the 10 591 overall participants were from disadvantaged populations. Further, only one study presented a statistic quantifying the inclusion of Hispanic populations, and no studies included statistics on the inclusion of Indigenous communities.17

The existing data suggest that economically disadvantaged, minority populations more frequently prefer principles that offer them a better chance of receiving a ventilator, which stands in contrast to the common ‘colorblind’ emphasis on saving the most lives and life-years, which favours those with better baseline health.10 In addition to the data on preferences, Biddison et al 8 highlight in their discussion qualitative themes that emerged from sessions that black participants’ concerns were largely focused on social justice in ventilator rationing scenarios.

However, this point appears not to have been sufficiently salient to be explored universally further in subsequent work. Schoch-Spana et al 16 recreated the study design of Biddison et al 8 using a different geographical population in two areas of Texas. Yet no results were presented stratified based on location, and by implication race and income.16

Finally, studies included different ranges of ethical principles (from 2 to 7) that were worded and characterised differently. Numeric and conceptual framing of principles directly influences their interpretation and appraisal by respondents, and complicates cross-study conclusions, constituting the most important limitation of this study. For example, Biddison et al 14 framed the save the most lives principle as ‘Prioritize those most likely to survive the current illness’, while Pew Research Center15 stated ‘some patients don’t receive treatment because they are older or sicker’ (see table 1 for further variations).


There is little research on community preferences for rationing principles in the setting of ventilator rationing in a respiratory illness pandemic, particularly when it comes to surfacing views of historically and structurally disadvantaged populations. Different methodologies complicate drawing conclusions across studies, although we observed a trend that historically disadvantaged groups are concerned that dominant principles may compound the experience of prior disadvantage by reducing chances for a ventilator. Approximately half of available state-level guidance (13 of 25) refers to community engagement having informed the guidance, citing three of the six studies included in this review, although the integration of social science evidence is incomplete and untransparent (findings are not publicly available and it is not clear to what extent views of minorities have contributed to their formulation).1 Further study is needed to elucidate community preferences, especially preferences that may be influenced by systemic inequity. Citation of community support for a specific principle requires context on differing trends across groups, and policy makers should engage more fully with disadvantaged populations.


The authors would like to thank the anonymous reviewers for their thoughtful and thorough feedback which served to greatly strengthen this manuscript.



  • Contributors WK is a Master of Bioethics student at the University of Pennsylvania and conducted the empirical research. HS is an Assistant Professor of Medical Ethics and Health Policy at the University of Pennsylvania and has published more than 60 peer-reviewed empirical and conceptual papers in high impact journals on issues around rationing, resource allocation and priority setting in healthcare. HS conceptualised and mentored the project. Both authors contributed to the drafting of this manuscript. HS is the guarantor.

  • Funding The authors have not declared a specific grant for this research from any funding agency in the public, commercial or not-for-profit sectors.

  • Competing interests None declared.

  • Patient consent for publication Not required.

  • Provenance and peer review Not commissioned; externally peer reviewed.

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.