Unit 9 Assignment
RESEARCH ARTICLE
How to Fairly Allocate Scarce Medical Resources: Ethical Argumentation under Scrutiny by Health Professionals and Lay People Pius Krütli1☯*, Thomas Rosemann2, Kjell Y. Törnblom1, Timo Smieszek3,4,5☯
1 Transdisciplinarity Lab (TdLab), Department of Environmental Systems Science, ETH Zurich, Switzerland, 2 Institute of Primary Care, University of Zurich, Switzerland, 3 Modelling and Economics Unit, Statistics, Modelling, and Economics Department, Public Health England, London, United Kingdom, 4 MRC Centre for Outbreak Analysis and Modelling, Department of Infectious Disease Epidemiology, Imperial College School of Public Health, London, United Kingdom, 5 Center for Infectious Disease Dynamics, The Pennsylvania State University, University Park, PA, United States of America
☯ These authors contributed equally to this work. * pius.kruetli@usys.ethz.ch
Abstract
Background
Societies are facing medical resource scarcities, inter alia due to increased life expectancy
and limited health budgets and also due to temporal or continuous physical shortages of
resources like donor organs. This makes it challenging to meet the medical needs of all.
Ethicists provide normative guidance for how to fairly allocate scarce medical resources,
but legitimate decisions require additionally information regarding what the general public
considers to be fair. The purpose of this study was to explore how lay people, general practi-
tioners, medical students and other health professionals evaluate the fairness of ten alloca-
tion principles for scarce medical resources: ‘sickest first’, ‘waiting list’, ‘prognosis’,
‘behaviour’ (i.e., those who engage in risky behaviour should not be prioritized), ‘instrumen-
tal value’ (e.g., health care workers should be favoured during epidemics), ‘combination of
criteria’ (i.e., a sequence of the ‘youngest first’, ‘prognosis’, and ‘lottery’ principles), ‘reci-
procity’ (i.e., those who provided services to the society in the past should be rewarded),
‘youngest first’, ‘lottery’, and ‘monetary contribution’.
Methods
1,267 respondents to an online questionnaire were confronted with hypothetical situations
of scarcity regarding (i) donor organs, (ii) hospital beds during an epidemic, and (iii) joint
replacements. Nine allocation principles were evaluated in terms of fairness for each type of
scarcity along 7-point Likert scales. The relationship between demographic factors (gender,
age, religiosity, political orientation, and health status) and fairness evaluations was mod-
elled with logistic regression.
PLOS ONE | DOI:10.1371/journal.pone.0159086 July 27, 2016 1 / 18
a11111
OPEN ACCESS
Citation: Krütli P, Rosemann T, Törnblom KY, Smieszek T (2016) How to Fairly Allocate Scarce Medical Resources: Ethical Argumentation under Scrutiny by Health Professionals and Lay People. PLoS ONE 11(7): e0159086. doi:10.1371/journal. pone.0159086
Editor: Yoel Lubell, Mahidol-Oxford Tropical Medicine Research Unit, THAILAND
Received: February 14, 2016
Accepted: June 27, 2016
Published: July 27, 2016
Copyright: © 2016 Krütli et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Data Availability Statement: Data are neither ethically or legally restricted, and are not third-party data. Data used in this manuscript are provided as S1 Dataset.
Funding: The Cogito Foundation funded this study (S-117/12). The Institute of Primary Care of the University of Zurich supported the data collection.
Competing Interests: The authors have declared that no competing interests exist.
Results
Medical background was a major predictor of fairness evaluations. While general practition-
ers showed different response patterns for all three allocation situations, the responses by
lay people were very similar. Lay people rated ‘sickest first’ and ‘waiting list’ on top of all allo-
cation principles—e.g., for donor organs 83.8% (95% CI: [81.2%–86.2%]) rated ‘sickest
first’ as fair (‘fair’ is represented by scale points 5–7), and 69.5% [66.2%–72.4%] rated ‘wait-
ing list’ as fair. The corresponding results for general practitioners: ‘prognosis’ 79.7%
[74.2%–84.9%], ‘combination of criteria’ 72.6% [66.4%–78.5%], and ‘sickest first’ 74.5%
[68.6%–80.1%); these were the highest-rated allocation principles for donor organs alloca-
tion. Interestingly, only 44.3% [37.7%–50.9%] of the general practitioners rated ‘instrumen-
tal value’ as fair for the allocation of hospital beds during a flu epidemic. The fairness
evaluations by general practitioners obtained for joint replacements: ‘sickest first’ 84.0%
[78.8%–88.6%], ‘combination of criteria’ 65.6% [59.2%–71.8%], and ‘prognosis’ 63.7%
[57.1%–70.0%]. ‘Lottery’, ‘reciprocity’, ‘instrumental value’, and ‘monetary contribution’
were considered very unfair allocation principles by both groups. Medical students’ ratings
were similar to those of general practitioners, and the ratings by other health professionals
resembled those of lay people.
Conclusions
Results are partly at odds with current conclusions proposed by some ethicists. A number
of ethicists reject ‘sickest first’ and ‘waiting list’ as morally unjustifiable allocation principles,
whereas those allocation principles received the highest fairness endorsements by lay peo-
ple and to some extent also by health professionals. Decision makers are advised to con-
sider whether or not to give ethicists, health professionals, and the general public an equal
voice when attempting to arrive at maximally endorsed allocations of scarce medical
resources.
Introduction The Universal Declaration of Human Rights [1] and its specifications in the International Cov- enant on Economic, Social, and Cultural Rights, Art. 12, adjudges everyone „the right [. . .] to the enjoyment of the highest attainable standard of physical and mental health” [2]. This provi- sion includes access to all the medical resources needed to live up to that standard [3]. How- ever, societies are facing situations when medical resources are scarce, and access to means of prevention, diagnosis, and treatment of those in need is not always guaranteed. Insufficient supply of medical resources is obvious in many developing countries where basic services are widely lacking [4–5]. Yet, also well-off countries are confronted by scarcities of medical resources such as donor organs, hospital beds during epidemics or after severe disasters, or unusually expensive drugs like Sofosbuvir to cure HCV infection [6] or Myozyme to manage Pompe disease [7].
Notwithstanding the need to reduce scarcities of critical medical resources worldwide, exist- ing shortages necessitate principles and rules prescribing how to allocate available medical ser- vices among the needy. Explicit rules addressing whom to prioritize are in place in many
Fair Allocation of Scarce Medical Resources
PLOS ONE | DOI:10.1371/journal.pone.0159086 July 27, 2016 2 / 18
countries regarding certain life-saving resources like donor organs and drugs against pandemic influenza (Box 1).
Ethicists offer moral guidance for how to fairly allocate scarce medical resources, e.g., [11– 13], and a number of allocation principles have been defined and balanced against each other. Persad et al. [14], for example, discuss eight principles: lottery, waiting list, sickest first, youn- gest first, number of lives saved, prognosis, instrumental value, and reciprocity (Table 1). In their view, most principles are fair, except for waiting lists and the sickest first principle, both of which were rejected as morally unjustifiable. The authors argue that waiting lists favour well-off people and are susceptible to corruption, while the sickest first principle ignores a patient’s prognosis and favours today’s sickest individuals over those who might be even worse-off in the future. However, medical allocation has been widely and controversially dis- cussed in bioethics and philosophy, e.g., [15–25], and Persad et al.’s positions are disputed in many ways, e.g., [26], yet it is out of the scope of this paper to substantially contribute to this ethical discussion. In this paper, we build mainly on the substantial set of ethical principles described by Persad and colleagues.
In contrast to the prescriptive approach in ethics, the social psychological focus is descrip- tive and explores people’s subjective perceptions of justice–“justice is in the eye of the beholder” [27]. Thus, opinions about what is a fair (in this article, the terms ‘fair’ and ‘just’ are used interchangeably) allocation of social resources vary with the context and may differ between individuals, groups and cultures [28]. The most commonly discussed principles are allocation according to needs, contributions (of effort, ability, and/or results), or equal amounts to all. See [29] for an overview of additional ‘generic’ allocation principles (into which most of the medical resource allocation principles focused by Persad et al. and other researches [14, 30], for example, can be translated).
Fairness judgments of resource allocation principles may be affected by a large number of factors [31] such as the allocated resource, per se [32], the social relationship [33], and the soci- etal context [34–35]. Even though studies within the empirical medical research tradition typi- cally focused on justice conceptions separately of patients, clinicians, lay people, and medical students, comparisons among these categories of people within the framework of one single study are lacking (like in the cases of [36, 30]). Previous studies have shown that different group identities appeared to affect moral judgments and behaviour differently, e.g., [37–38]. Existing empirical research has also focused on the impact of individual recipient characteris- tics like gender, age, life-style, health status, etc. [39–46], as well as on specific priority princi- ples and allocation situations [47–50].
Finally, and of special importance in the context of the study reported here, there is also a lack of comprehensive empirical studies that combine and compare the descriptive and pre- scriptive approaches [51]. While ethics provide the moral fundament, enforceability of rules in democratic societies require majority endorsement as well as consensus among stakeholder groups. Thus, the descriptive and prescriptive approaches may complement each other to bet- ter guide decisions about social resource allocation.
The major objective of the study reported here was to study how (a) four categories of peo- ple (lay people, general practitioners, medical students and other health professionals) evaluate the fairness of (b) ten allocation principles (see methods section) for (c) three scarce medical resources (donor organs, hospital beds, joint replacements). We (d) compared our empirically obtained fairness evaluations of the ten allocation principles with those derived ‘prescriptively/ ethically’ by Persad et al. [14] to reveal in/consistencies between the their version of the ethical (philosophical) approach and the empirical (descriptive/social psychological) approach. Finally, we (e) explored whether certain individual characteristics might account for observed variations in fairness perceptions.
Fair Allocation of Scarce Medical Resources
PLOS ONE | DOI:10.1371/journal.pone.0159086 July 27, 2016 3 / 18
Box 1. Allocation of scarce medical resources in Switzerland: current regulations and involvement of ethicists in the formulation of these regulations
Example 1: Prioritizing individuals during an influenza pandemic The Regulation on the Control of Communicable Diseases of Humans (Regulation 818.101.1; enacted 1 January 2016; replaced Regulation 818.101.23) sets the rules for pri- oritizing individuals in situations when vaccines effective against pandemic influenza (as well as other critical drugs to treat communicable diseases) might be scarce.
Any prioritization based on this regulation has to reflect generally accepted medical and ethical criteria as well as economic and societal concerns. The following groups are explicitly mentioned as eligible for prioritisation: 1. health care workers;
2. individuals, who have a higher risk of complications and adverse outcomes of the dis- ease than others;
3. individuals, who contribute to crucial services such as inner and national security, transport, communication as well as energy, water, and food supply.
Accordingly, individuals can be prioritized according to their instrumental value and according to their individual prognosis (cf. Table 1).
The Swiss Federal Office for Public Health (FOPH, which was in charge of formulat- ing the current and the preceding regulation) actively reflected and incorporated ethical arguments and positions coming from the Swiss National Advisory Commission on Bio- medical Ethics. Academic ethicist Urs Thurnheer (Karlsruhe University of Education) was actively involved in preparatory work which fed into the regulation on pandemic influenza preceding the current regulation (personal communication [8]).
Example 2: Allocating donor organs The Regulation on the Allocation of Organs for Transplantation (Regulation 810.212.4; enacted 16 March 2007; version of 1 May 2016) governs the allocation of the following donor organs: heart, lung, liver, kidney, pancreas, small intestine. Allocation rules are based on multiple criteria and differ (slightly) for different organs.
In summary, 1. medical urgency: first priority is given to patients whose life would be at immediate
risk if they would not receive the organ;
2. medical benefit: second priority is given to patients, for whom the greatest medical benefit is expected.
If two (or more) patients should have the same priority according to these criteria, the following criteria shall also be taken into consideration: medical urgency, waiting time, extended waiting time due to the patient’s blood group, fit of the tissue characteristics.
Regulation 810.212.4 also uses a combination of fairness criteria with the sickest first principle (cf. Table 1) being the most important criterion, prognosis being the second most important criterion. Other principles, including the waiting list (cf. Table 1), only play a role if at least two patients have the same priority according to the sickest first and prognosis principles.
Fair Allocation of Scarce Medical Resources
PLOS ONE | DOI:10.1371/journal.pone.0159086 July 27, 2016 4 / 18
Upon defining the Swiss Federal Office for Public Health commissioned an expert’s report by ethicist Beat Sitter-Liver (University of Fribourg) [9] which was used as a foun- dation for phrasing the Regulation on the Allocation of Organs for Transplantation (per- sonal communication [10]). Among more than 70 stakeholders, three institutions representing academic ethics (the Center for Ethics, University of Zurich; the Institute for Applied Ethics and Medical Ethics, University of Basel; the Swiss National Advisory Commission on Biomedical Ethics) were actively involved in the consultation process, which resulted in the enactment of the regulation.
Table 1. Allocation principles.
Allocation principle
Abbreviation Description Pros Cons Complete lives system
Sickest first SICK Prioritizes the sickest, i.e., those who have greatest need for treatment at a specific moment in time.
Intuitively obvious; sickest are also worst-off
Ignores post-treatment prognosis
Excluded
Waiting list ORDR Allocates services according to the individual’s position on the waiting list. Also known as ‘first-come, first- served’ principle.
Equality of opportunities; no discontinued interventions
Ignores relevant differences between individuals; favours the well-off; susceptible to corruption
Excluded
Prognosis SURV Prioritizes those with favourable prognosis, hence, those with the highest survival probability and duration.
Intuitively obvious; saves most life years
Does not consider distribution and number of lives saved
Included
Behaviour BHAV Prioritizes those who did not engage in risky behaviours that caused their condition or affected it negatively.
Promotes healthy life style; promotes individual responsibility
Reasons for individual behaviour ignored; conflict with liberty rights
Not considered
Instrumental value
IMPF Prioritizes those who’s function is essential to keep up fundamental services, e.g., health care professionals. Relevant, e.g., during pandemics.
Serves saving most lives Can encourage abuse of system
Included under certain conditions
Combination of criteria
COMB This allocation scheme includes a combination of criteria such as age (youngest first), prognosis and lottery.
Considers several morally relevant criteria; appropriate distributive justice
Discriminates older people Subset of Complete lives system approach
Youngest first YONG Prioritizes young over old individuals. Prioritizes worst-off; hard to corrupt
Ignores relevant other principles
Included
Lottery RAND Allocates medical services randomly among those who are in need of treatment.
Equal opportunities; little knowledge about recipients needed; easy to handle; resistant against corruption
Blind against other factors; treating people equally often fails to treat them as equals
Included
Reciprocity SERV Prioritizes those who have voluntarily provided societal services in the past.
Justice to people who have provided contributions in the past
Requires complex inquiries Included under certain conditions
Monetary contribution
MONY Prioritizes those who contribute to the costs of medical treatment.
Relieves public health system; reduces costs; reflects common societal principle that those who need more pay more
Favours wealthy people; undermines societal solidarity; makes allocation to worst-offs impossible
Not considered
Source: adapted from Persad et al. (2009, p. 424) [14]. Column ‘Abbreviation’ corresponds to Fig 1 and Table 2 in Results section.
doi:10.1371/journal.pone.0159086.t001
Fair Allocation of Scarce Medical Resources
PLOS ONE | DOI:10.1371/journal.pone.0159086 July 27, 2016 5 / 18
Methods An online survey containing 99 questions was conducted between December 2, 2013 and May 31, 2014. To avoid bias by row effects, questions were presented in random order, and the three types of scarce resources were randomized as well. A total of 1,790 respondents accessed the questionnaire, and 1,267 (71%) answered all questions. Incomplete datasets were excluded from the analysis. The questionnaire was pre-tested and discussed with peers for clarity of questions and logical, bias-reducing order among questions. For the current study only a subset of questionnaire items were used. The respective dataset can be found as S1 Dataset and S1 Text (see also S1 File).
Respondents Participants were recruited from three predefined pools: a market research panel (MRP), gen- eral practitioners (GP), and medical students (MS). The MRP consisted of a sample of the 25– 65 year old population from the German-speaking part of Switzerland (representative in terms of age- and gender-distribution). Most GPs were recruited from the cantons of Zurich and Lucerne (1) via invitation letters to all 1,415 GPs in the canton of Zurich who were enrolled with the Swiss Medical Association, (2) by asking medical networks to send invitation emails to their associated GPs (a pool of approx. 250 GPs), and (3) advertising the study in Swiss med- ical newsletters. MSs were approached by sending invitation emails to all students enrolled at the Faculty of Medicine at the University of Zurich. Incentives were offered to increase response rates (lotteries with 10 prices valued CHF 300 each for GPs and MSs and EUR 3 pay- ments for all online panellists (paid by the MRP provider)).
Respondents declared themselves as physicians (GP), medical students (MS), other health professionals (HP), or others (i.e., lay people, LP), see Fig 1. Socio-demographic profiles of these four groups are provided as S1 Table.
Questionnaire Participants were provided descriptions of three hypothetical situations in which the three types of scarce medical resources were to be allocated, i.e. (i) donor organs–an inelastic resource, (ii) hospital beds during a flu epidemic–an elastic resource, and (iii) joint replace- ments–an elastic resource (Table 2). The respondents were asked to give their advice on how they thought the three resources should be allocated (thus, generating descriptions of
Fig 1. Study sample. Top row: three different sources of respondents; bottom row: self-declared medical background.
doi:10.1371/journal.pone.0159086.g001
Fair Allocation of Scarce Medical Resources
PLOS ONE | DOI:10.1371/journal.pone.0159086 July 27, 2016 6 / 18
respondents’ prescriptive views) choosing from a list of nine allocation principles (i.e., selection of the most fair principle), each of which they also rated in terms of fairness along 7-point Likert scales ranging from 1 (totally unjust) to 7 (totally just) (Table 2). Not all allocation prin- ciples were included in every situation. Additional data included participants’ gender, age, reli- giosity, political orientation, and health status. Screen shots of the online survey are provided as S1 File.
Ethics Statement The research protocol was submitted to the ETH Ethics Commission for review and approval. The executive secretary reviewed the protocol and decided that, according to the law and
Table 2. Three hypothetical situations of scarce medical resource allocation, and their respective nine allocation principle alternatives.
Situation A: Donor organs Situation B: Hospital beds during epidemic Situation C: Joint replacements
One hundred organs (kidneys) are available yearly from voluntary and eligible donors. A team of consultants is responsible for the allocation of the 100 donated kidneys to some of the 500 individuals who are in need of a kidney transplant. For convenience, we assume that the kidneys are equally tolerable to all 500 individuals.
A very severe flu epidemic hits a mid-sized town (approx. 50,000 inhabitants) in Switzerland and, as a consequence, 2,500 individuals need hospital care. There are, however, only 500 hospital beds available. A team of consultants will allocate the 500 hospital beds to some of the 2,500 individuals in need.
In Switzerland, there are 5,000 individuals who are waiting for a life-quality enhancing treatment, e.g., hip-joint replacement. Thus, they don’t suffer from a life-threatening condition. This treatment is very expensive, and only 1,000 hip-joint replacements can be provided. A team of consultants will allocate the 1,000 replacements to some of the 5,000 individuals who are in need of it.
Problem (Question): In your opinion, how should the team of consultants proceed?
The 100 kidneys should be allocated: The 500 beds should be allocated: The 1,000 hip-joints should be allocated:
• to the sickesta individuals, i.e., those, who need the organ most urgently [SICK]
• to the sickest individuals [SICK] • to the sickest individuals (i.e., those whose hip problem results in the most severe reduction in life-quality) [SICK]
• according to the order of registration for a donor organ (i.e., those with the longest wait are prioritized) [ORDR]
• according to the order of falling sick (i.e., those who have been ill the longest are prioritized) [ORDR]
• according to the order of registration for surgery (i.e., those with the longest wait are prioritized) [ORDR]
• by prioritizing those who are likely to survive the longest because of the organ transplant [SURV]
• by prioritizing those who are most likely to survive the infection as a result of hospital care [SURV]
• by prioritizing those with the longest life expectancy [SURV]
• by favouring those, who have not become by own fault a medical emergency [BHAV]
• by prioritizing those, who have essential roles for keeping society operational (e.g., hospital staff) [IMPF]
• by prioritizing those, whose life-quality improvement needs are not self-inflicted [BHAV]
• according to a combination of criteria: age (youngest first), prognosis (longest survival with organ transplant), and by chance (drawing lots) [COMB]
• according to a combination of criteria: age (youngest first), prognosis (longest survival), and by chance (drawing lots) [COMB]
• according to a combination of criteria: age (youngest first), prognosis (longest life expectancy based on general state of health), and by chance (drawing lots) [COMB]
• according to age, prioritizing young individuals [YONG]
• according to age, prioritizing young individuals [YONG]
• according to age, prioritizing young individuals. [YONG]
• randomly, e.g., via a lottery [RAND] • randomly, e.g., via a lottery [RAND] • randomly, e.g., via a lottery [RAND]
• by prioritizing those who contributed in the past to the common good (e.g., by volunteering) [SERV]
• by prioritizing those who contributed in the past to the common good (e.g., by volunteering) [SERV]
• by prioritizing those who contributed in the past to the common good (e.g., by volunteering) [SERV]
• preferably to those who contribute substantially to the costs of the treatment [MONY]
• preferably to those who contribute substantially to the costs of the treatment [MONY]
• preferably to those who contribute substantially to the costs of the treatment [MONY]
Acronyms in brackets correspond to those in Table 1. Translations from German by the authors, see also S1 File. a Persad et al. define ‘sickest first’ as”prioritizing the person who will die soonest if she does not receive an organ”. Our definition is slightly different: ”those
who need the organ most urgently” should receive the organ. As ‘urgency’ may be defined in different ways, apart from helping those ‘who will die soonest’,
our data may not be unambiguously compared.
doi:10.1371/journal.pone.0159086.t002
Fair Allocation of Scarce Medical Resources
PLOS ONE | DOI:10.1371/journal.pone.0159086 July 27, 2016 7 / 18
pertinent regulations, a full evaluation was not required (decision 19 Sept 2013, EK 2013-N- 44), because pure survey research of this kind presents no risk of harm to participants, all par- ticipants were at least 18 years old, and also because there were no data protection concerns as no personally identifiable information (PII) were about to be collected.
Following our research protocol, all potential participants were informed about the purpose of this research and the expected duration of participation. They were informed about whom to contact for questions and concerns regarding the study. We explained that all information was collected in a fully anonymous manner. Finally, participation was voluntary and partici- pants had the opportunity to stop participation at all times before submitting the fully com- pleted online questionnaire. All survey questions relevant to this paper can be found as S1 File.
Participation implied consent and—in line with our research protocol submitted to the ETH Ethics Commission—we did not document informed consent because the research did not involve more than minimal risk and a signed consent document would have been the only record linking participant and data (PII), and, thus, making a fully anonymous research design impossible.
Statistical Analyses Fig 2 provides the following information for the allocation principles for each of the three situa- tions: (a) the frequency distribution of the Likert categories (totally just to totally unjust), (b) the proportion of participants favouring the respective principle over all other principles, (c) bootstrap 95% confidence intervals for that proportion as well as for the estimates of the bor- ders between categories ‘just’ (Likert scales 5–7) and ‘unjust’ (1–3). The figure was rendered using ggplot2 0.9.3.1, and the confidence intervals were simulated using Python code executed with Enthought Canopy 1.1.0. Group-wise means and standard deviations are provided as S2 Table.
Logistic regression models were estimated for all allocation principles as dependent vari- ables. The answer scales were merged into two remaining categories: (1) ‘just’ (Likert scales 5–7); (0) ‘other’ (1–4). For reasons of comparability, we estimated models with identical struc- ture for all questions, including medical background, gender, age, religiosity, political orienta- tion, and health state as independent variables (previous research by Skitka and Tetlock [39] suggested that these demographic variables may affect justice judgements). Regression models were estimated using IBM SPSS Version 22.
Results
Descriptive Statistics Substantial differences between GPs and LPs were obtained for all three allocation situations in terms of both fairness ratings and the fairest of all allocation rules (Fig 2). While GPs show dif- ferent response patterns for all the three allocation situations, the responses by LP are very sim- ilar in all three situations. The general response pattern of MSs is similar to that of GPs, while the responses by HPs resemble those by LPs (see S1 Fig).
LPs rated the sickest first principle and waiting list highest in all three situations. Lottery, monetary contribution, and reciprocity received the lowest ratings and are, hence, considered the most unjust allocation principles. All other principles were rated neither fair nor unfair— except in situation B. When LPs had to name the fairest principle, sickest first was chosen unequivocally for all three situations.
GP’s preferences differed for all three situations. Prognosis, sickest first, and combination of cri- teria were the highest-rated allocation principles for situation A. In contrast to LPs, waiting list was contested, while youngest first obtained solid support (majority of positive ratings).
Fair Allocation of Scarce Medical Resources
PLOS ONE | DOI:10.1371/journal.pone.0159086 July 27, 2016 8 / 18
Fig 2. Fairness ratings and forced choice of one single principle (percentage and 95% CI) by lay people versus general practitioners for three situations of scarce medical resource allocation. For the groups medical students and other health professionals see S1 Fig).
doi:10.1371/journal.pone.0159086.g002
Fair Allocation of Scarce Medical Resources
PLOS ONE | DOI:10.1371/journal.pone.0159086 July 27, 2016 9 / 18
Behaviour was a contested principle in situation A, and all other principles were considered unfair allocation principles in all situations. Sickest first, prognosis, and combination of criteria obtained clear majorities in situation B, whereas waiting list, youngest first and, interestingly, instrumental value were contested fairness principles during a flu epidemic. In situation C, we observe a high rating for the sickest first principle. In addition, prognosis and combination of crite- ria were considered fair principles, while waiting list, behaviour and youngest first were contested. When GPs had to choose the fairest of all principles, they again answered differently for all three situations. They clearly favoured combination of criteria in situation A. In situation B, sickest first, prognosis, and combination of criteria were chosen by about the same proportion of partici- pants. In situation C, sickest first and combination of criteria were the most favoured principles.
Logistic Regression Model Table 3 shows parameter estimates from the logistic regression models (reported are odds ratios, OR, and the corresponding 95% confidence intervals, CI). All subsequently reported group differences were statistically significant with p <0.05.
Medical background was the most influential independent variable. Major differences can be observed between reference groups LP and GP, LP and MS, and, partially, between the LP and HP groups.
GPs were between 1.71 and 4.68 times (OR) more likely than LPs to choose prognosis, com- bination of criteria, and youngest first. Further, GPs were less likely than LPs to choose waiting list and sickest first except in situation C, and reciprocity in situation B (OR 0.19–0.61). Simi- larly, like GPs, MSs (and HPs, but to a lesser degree) deviated from LP’s choices. However, MSs were almost six times as likely as LPs to choose prognosis in situation A. Further, they were twice as likely (situation B) and four times as likely (situation C) than LPs to choose sickest first. Most differences between HPs and LPs occurred in the context of situation A.
Men showed greater preferences than women for lottery in situation A (OR = 1.50), combi- nation of criteria (OR = 1.34) and instrumental value (OR = 1.65) in situation B, and youngest first (OR = 1.30) as well as monetary contribution (OR = 1.41) in situation C, while women considered sickest first principle to be more fair than men in all situations. They thought wait- ing list in situations B and C to be a more fair allocation principle than did men.
No significant age effect was observed in situation A, while in situations B and C increasing age correlated with an increased preference for sickest first (OR = 1.02), as well as for instru- mental value (OR = 1.02) in situation B, and behaviour (OR = 1.02) in situation C. The oppo- site was observed for combination of criteria in situations B (OR = 0.99) and C (OR = 0.98), and lottery (OR = 0.98) in situation and C.
Participants who declared themselves religious showed clearer preferences than non-reli- gious participants for youngest first in all three situations, for instrumental value in situation B (OR = 0.67), and for reciprocity in situation C (OR = 0.58). Those who were uncertain regard- ing their religiosity also deviated from the religious participants (ORDR, OR(A) = 0.55; MONY, OR(B) = 2.41; YONG, OR(C) = 0.54).
Political orientation had an effect on monetary contribution in all three situations and on behaviour in situations A and C: the more a participant was leaning towards the political right, the more likely s/he was to consider these principles to be fair. The opposite effect regarding the left-right spectrum (but to a lesser extent) was observed for combination of criteria in situa- tions A and B and for lottery in all situations: the more a participant was left-oriented, the more likely s/he was to consider these principles to be fair.
Finally, participants’ self-declared health state impacted their evaluation pattern in situation C: sickest first, waiting list, behaviour and combination of criteria were significantly more
Fair Allocation of Scarce Medical Resources
PLOS ONE | DOI:10.1371/journal.pone.0159086 July 27, 2016 10 / 18
T a b le
3 . O d d s ra ti o s (9 5 %
C Is ) o f a ll o c a ti o n p ri n c ip le s b y e x p la n a to ry
v a ri a b le s s u c h a s m e d ic a lb
a c k g ro u n d ,g
e n d e r, a g e ,r e li g io n ,p
o li ti c a lo
ri e n ta ti o n ,a
n d h e a lt h s ta te
in b in a ry
lo g is ti c re g re s s io n m o d e l.
A ll o c a ti o n p ri n c ip le
O d d s ra ti o ,O
R (9 5 %
C o n fi d e n c e in te rv a l, C I)
M e d ic a lb
a c k g ro u n d a
G e n d e r
A g e
R e li g io n b
P o l. O ri e n ta ti o n
S u b je c ti v e H e a lt h S ta te
M e d ic a lS
tu d e n ts
G e n e ra lP
ra c ti ti o n e rs
O th e r h e a lt h p ro fe s s io n a ls
M a le
Y e a rs
N o t re li g io u s
U n k n o w n
L e ft -r ig h t
V e ry
b a d -v e ry
g o o d
S it u a ti o n A : In e la st ic /f re q u e n te
ve n ts
(d o n o r o rg a n s)
S IC K
1 .0 7 (0 .6 2 – 1 .8 3 )
0 .4 7 (0 .3 1 – 0 .7 2 )
0 .5 2 (0 .2 8 – 0 .9 7 )
0 .7 0 (0 .5 1 – 0 .9 5 )
1 .0 1 (1 .0 0 – 1 .0 3 )
0 .7 5 (0 .5 5 – 1 .0 2 )
0 .7 4 (0 .4 2 – 1 .3 0 )
0 .9 7 (0 .9 1 – 1 .0 4 )
1 .0 3 (0 .9 0 – 1 .1 7 )
O R D R
0 .7 9 (0 .5 1 – 1 .2 2 )
0 .4 4 (0 .3 1 – 0 .6 3 )
0 .7 6 (0 .4 4 – 1 .3 3 )
0 .8 6 (0 .6 7 – 1 .1 0 )
1 .0 0 (0 .9 9 – 1 .0 2 )
0 .8 9 (0 .6 9 – 1 .1 5 )
0 .5 5 (0 .3 5 – 0 .8 7 )
0 .9 9 (0 .9 4 – 1 .0 5 )
1 .1 0 (0 .9 9 – 1 .2 2 )
S U R V
5 .7 7 (3 .4 5 – 9 .6 6 )
3 .3 5 (2 .2 6 – 4 .9 8 )
2 .8 1 (1 .5 5 – 5 .0 8 )
1 .1 4 (0 .8 9 – 1 .4 6 )
1 .0 0 (0 .9 9 – 1 .0 1 )
0 .9 0 (0 .7 0 – 1 .1 6 )
1 .0 4 (0 .6 5 – 1 .6 9 )
1 .0 3 (0 .9 7 – 1 .0 9 )
1 .0 4 (0 .9 4 – 1 .1 5 )
B H A V
1 .8 7 (1 .2 3 – 2 .8 6 )
0 .7 5 (0 .5 3 – 1 .0 6 )
1 .5 7 (0 .9 2 – 2 .6 7 )
0 .8 3 (0 .6 6 – 1 .0 5 )
1 .0 1 (1 .0 0 – 1 .0 2 )
1 .0 0 (0 .7 9 – 1 .2 7 )
0 .8 2 (0 .5 2 – 1 .2 9 )
1 .1 3 (1 .0 8 – 1 .2 0 )
1 .0 7 (0 .9 7 – 1 .1 8 )
C O M B
3 .7 1 (2 .3 4 – 5 .8 7 )
3 .7 9 (2 .6 2 – 5 .4 9 )
2 .1 3 (1 .2 5 – 3 .6 4 )
1 .1 1 (0 .8 7 – 1 .4 2 )
0 .9 9 (0 .9 8 – 1 .0 0 )
0 .9 5 (0 .7 4 – 1 .2 2 )
1 .1 1 (0 .6 9 – 1 .7 7 )
0 .9 4 (0 .8 9 – 1 .0 0 )
1 .1 2 (1 .0 1 – 1 .2 4 )
Y O N G
2 .1 3 (1 .3 9 – 3 .2 7 )
2 .9 0 (2 .0 4 – 4 .1 3 )
1 .9 0 (1 .1 2 – 3 .2 2 )
1 .1 2 (0 .8 8 – 1 .4 3 )
1 .0 0 (0 .9 8 – 1 .0 1 )
0 .7 8 (0 .6 1 – 1 .0 0 )
0 .8 0 (0 .5 0 – 1 .2 7 )
1 .0 3 (0 .9 8 – 1 .0 9 )
1 .1 0 (1 .0 0 – 1 .2 2 )
R A N D
1 .3 3 (0 .7 7 – 2 .3 0 )
1 .4 7 (0 .9 2 – 2 .3 4 )
0 .9 5 (0 .4 1 – 2 .1 6 )
1 .5 0 (1 .0 8 – 2 .0 8 )
0 .9 9 (0 .9 7 – 1 .0 1 )
1 .2 9 (0 .9 3 – 1 .7 9 )
0 .7 4 (0 .3 6 – 1 .5 0 )
0 .8 8 (0 .8 2 – 0 .9 5 )
1 .0 3 (0 .9 0 – 1 .1 9 )
S E R V
1 .3 6 (0 .6 9 – 2 .6 7 )
0 .7 5 (0 .4 2 – 1 .3 5 )
0 .9 9 (0 .4 1 – 2 .4 1 )
1 .3 7 (0 .9 4 – 2 .0 1 )
1 .0 0 (0 .9 9 – 1 .0 2 )
0 .8 5 (0 .5 7 – 1 .2 6 )
1 .2 3 (0 .6 3 – 2 .4 0 )
1 .0 5 (0 .9 6 – 1 .1 4 )
0 .9 0 (0 .7 8 – 1 .0 5 )
M O N Y
0 .7 0 (0 .3 1 – 1 .5 5 )
0 .6 9 (0 .3 8 – 1 .2 6 )
0 .8 9 (0 .3 4 – 2 .3 3 )
1 .4 0 (0 .9 3 – 2 .1 1 )
1 .0 0 (0 .9 8 – 1 .0 2 )
0 .8 3 (0 .5 5 – 1 .2 7 )
1 .6 9 (0 .8 7 – 3 .2 6 )
1 .2 1 (1 .1 0 – 1 .3 3 )
1 .1 2 (0 .9 4 – 1 .3 3 )
S it u a ti o n B : E la st ic /r a re
e ve
n ts
(h o sp
ita lb
e d s)
S IC K
1 .9 9 (1 .1 2 – 3 .5 0 )
0 .5 7 (0 .3 8 – 0 .8 6 )
0 .5 7 (0 .3 2 – 1 .0 4 )
0 .7 3 (0 .5 4 – 0 .9 8 )
1 .0 2 (1 .0 0 – 1 .0 3 )
1 .0 3 (0 .7 7 – 1 .3 9 )
0 .9 9 (0 .5 7 – 1 .7 2 )
0 .9 4 (0 .8 9 – 1 .0 1 )
1 .0 8 (0 .9 6 – 1 .2 2 )
O R D R
0 .4 9 (0 .3 2 – 0 .7 5 )
0 .1 9 (0 .1 3 – 0 .2 7 )
0 .9 6 (0 .5 6 – 1 .6 6 )
0 .7 8 (0 .6 1 – 0 .9 9 )
1 .0 0 (0 .9 9 – 1 .0 2 )
0 .8 7 (0 .6 8 – 1 .1 1 )
0 .9 5 (0 .6 0 – 1 .5 0 )
1 .0 6 (1 .0 1 – 1 .1 2 )
1 .0 9 (0 .9 8 – 1 .2 0 )
S U R V
4 .0 3 (2 .5 2 – 6 .4 4 )
4 .6 8 (3 .1 6 – 6 .9 3 )
1 .4 3 (0 .8 4 – 2 .4 2 )
1 .2 0 (0 .9 4 – 1 .5 3 )
0 .9 9 (0 .9 8 – 1 .0 0 )
1 .0 0 (0 .7 8 – 1 .2 8 )
0 .7 3 (0 .4 6 – 1 .1 7 )
1 .0 0 (0 .9 5 – 1 .0 6 )
1 .0 3 (0 .9 3 – 1 .1 4 )
IM P F
2 .3 5 (1 .5 1 – 3 .6 6 )
1 .3 8 (0 .9 7 – 1 .9 5 )
1 .5 0 (0 .8 7 – 2 .5 9 )
1 .2 4 (0 .9 7 – 1 .5 9 )
1 .0 2 (1 .0 0 – 1 .0 3 )
0 .6 7 (0 .5 2 – 0 .8 6 )
0 .8 7 (0 .5 5 – 1 .3 9 )
1 .0 4 (0 .9 9 – 1 .1 0 )
1 .0 0 (0 .9 0 – 1 .1 1 )
C O M B
1 .9 3 (1 .2 6 – 2 .9 7 )
2 .6 0 (1 .8 3 – 3 .7 0 )
2 .2 4 (1 .3 1 – 3 .8 1 )
1 .3 4 (1 .0 5 – 1 .7 1 )
0 .9 9 (0 .9 8 – 1 .0 0 )
0 .8 2 (0 .6 4 – 1 .0 5 )
0 .9 4 (0 .5 9 – 1 .4 9 )
0 .9 3 (0 .8 8 – 0 .9 9 )
1 .1 6 (1 .0 4 – 1 .2 9 )
Y O N G
1 .4 0 (0 .8 6 – 2 .2 7 )
1 .7 1 (1 .1 7 – 2 .4 9 )
1 .4 4 (0 .7 9 – 2 .6 4 )
1 .3 0 (0 .9 9 – 1 .7 2 )
1 .0 0 (0 .9 9 – 1 .0 1 )
0 .7 5 (0 .5 6 – 0 .9 9 )
0 .8 6 (0 .5 1 – 1 .4 5 )
1 .0 1 (0 .9 5 – 1 .0 8 )
1 .0 7 (0 .9 5 – 1 .2 1 )
R A N D
1 .0 5 (0 .6 2 – 1 .8 0 )
0 .9 8 (0 .5 9 – 1 .6 0 )
1 .2 0 (0 .5 9 – 2 .4 7 )
1 .1 4 (0 .8 2 – 1 .5 7 )
0 .9 8 (0 .9 7 – 1 .0 0 )
1 .1 2 (0 .8 1 – 1 .5 6 )
0 .8 2 (0 .4 2 – 1 .5 8 )
0 .8 7 (0 .8 0 – 0 .9 4 )
1 .0 9 (0 .9 4 – 1 .2 6 )
S E R V
1 .2 8 (0 .6 5 – 2 .4 9 )
0 .4 8 (0 .2 6 – 0 .8 9 )
1 .3 5 (0 .6 1 – 2 .9 7 )
1 .6 5 (1 .1 3 – 2 .4 2 )
1 .0 1 (0 .9 9 – 1 .0 2 )
0 .7 5 (0 .5 1 – 1 .1 0 )
1 .1 3 (0 .5 8 – 2 .1 9 )
1 .0 2 (0 .9 4 – 1 .1 1 )
0 .9 9 (0 .8 5 – 1 .1 5 )
M O N Y
0 .3 2 (0 .1 3 – 0 .7 9 )
0 .5 0 (0 .2 6 – 0 .9 6 )
0 .6 7 (0 .2 6 – 1 .7 7 )
1 .0 3 (0 .6 9 – 1 .5 2 )
1 .0 0 (0 .9 8 – 1 .0 2 )
1 .1 2 (0 .7 4 – 1 .6 9 )
2 .4 1 (1 .2 9 – 4 .5 2 )
1 .2 7 (1 .1 6 – 1 .3 9 )
1 .2 4 (1 .0 4 – 1 .4 7 )
S it u a ti o n C : E la st ic /f re q u e n te
ve n ts
(j o in tr e p la ce
m e n t)
S IC K
4 .0 6 (2 .0 8 – 7 .9 3 )
1 .0 6 (0 .6 7 – 1 .6 8 )
0 .9 7 (0 .4 9 – 1 .9 4 )
0 .5 6 (0 .4 1 – 0 .7 6 )
1 .0 2 (1 .0 1 – 1 .0 4 )
1 .0 6 (0 .7 8 – 1 .4 6 )
0 .8 3 (0 .4 7 – 1 .4 5 )
1 .0 0 (0 .9 3 – 1 .0 7 )
1 .1 5 (1 .0 1 – 1 .3 0 )
O R D R
0 .8 6 (0 .5 6 – 1 .3 2 )
0 .6 1 (0 .4 4 – 0 .8 7 )
0 .8 6 (0 .5 0 – 1 .4 7 )
0 .6 8 (0 .5 4 – 0 .8 6 )
1 .0 0 (0 .9 9 – 1 .0 1 )
0 .9 1 (0 .7 1 – 1 .1 6 )
0 .6 7 (0 .4 3 1 .0 6 )
1 .0 1 (0 .9 5 – 1 .0 6 )
1 .1 5 (1 .0 4 – 1 .2 7 )
S U R V
2 .3 3 (1 .5 2 – 3 .5 7 )
2 .5 2 (1 .7 7 – 3 .5 8 )
1 .2 9 (0 .7 6 – 2 .1 8 )
1 .1 0 (0 .8 7 – 1 .3 9 )
1 .0 0 (0 .9 9 – 1 .0 1 )
0 .8 8 (0 .6 9 – 1 .1 2 )
0 .9 4 (0 .6 0 – 1 .4 7 )
1 .0 0 (0 .9 4 – 1 .0 5 )
1 .0 6 (0 .9 6 – 1 .1 7 )
B H A V
1 .9 5 (1 .2 8 – 2 .9 9 )
0 .6 9 (0 .4 9 – 0 .9 8 )
1 .3 8 (0 .8 2 – 2 .3 5 )
0 .9 9 (0 .7 8 – 1 .2 5 )
1 .0 2 (1 .0 1 – 1 .0 3 )
1 .0 6 (0 .8 4 – 1 .3 5 )
1 .1 9 (0 .7 6 – 1 .8 7 )
1 .1 2 (1 .0 7 – 1 .1 9 )
1 .1 2 (1 .0 1 – 1 .2 4 )
C O M B
3 .2 5 (2 .0 8 – 5 .0 8 )
3 .6 2 (2 .5 2 – 5 .1 8 )
2 .8 9 (1 .6 8 – 4 .9 5 )
1 .1 9 (0 .9 3 – 1 .5 2 )
0 .9 8 (0 .9 7 – 1 .0 0 )
0 .8 6 (0 .6 7 – 1 .1 0 )
1 .1 0 (0 .6 9 – 1 .7 6 )
0 .9 9 (0 .9 3 – 1 .0 4 )
1 .1 2 (1 .0 0 – 1 .2 4 )
Y O N G
2 .4 2 (1 .5 7 – 3 .7 5 )
2 .2 4 (1 .5 7 – 3 .1 9 )
2 .0 4 (1 .1 9 – 3 .4 9 )
1 .3 0 (1 .0 1 – 1 .6 6 )
1 .0 0 (0 .9 8 – 1 .0 1 )
0 .7 5 (0 .5 8 – 0 .9 6 )
0 .5 4 (0 .3 3 – 0 .9 0 )
1 .0 2 (0 .9 7 – 1 .0 8 )
1 .0 5 (0 .9 5 – 1 .1 7 )
R A N D
1 .2 3 (0 .7 4 – 2 .0 2 )
1 .2 3 (0 .7 7 – 1 .9 6 )
1 .3 4 (0 .6 8 – 2 .6 2 )
1 .1 4 (0 .8 4 – 1 .5 5 )
0 .9 8 (0 .9 6 – 0 .9 9 )
1 .1 4 (0 .8 4 – 1 .5 6 )
1 .1 2 (0 .6 3 – 1 .9 8 )
0 .9 1 (0 .8 4 – 0 .9 7 )
1 .0 6 (0 .9 3 – 1 .2 1 )
S E R V
0 .7 9 (0 .3 9 – 1 .6 1 )
0 .8 2 (0 .4 8 – 1 .4 1 )
0 .8 5 (0 .3 5 – 2 .0 5 )
1 .0 8 (0 .7 4 – 1 .5 7 )
1 .0 0 (0 .9 8 – 1 .0 2 )
0 .5 8 (0 .3 9 – 0 .8 6 )
0 .8 3 (0 .4 1 – 1 .6 8 )
1 .0 6 (0 .9 8 – 1 .1 6 )
1 .0 7 (0 .9 1 – 1 .2 6 )
M O N Y
1 .0 6 (0 .6 0 – 1 .8 5 )
1 .0 2 (0 .6 4 – 1 .6 4 )
0 .8 8 (0 .4 0 – 1 .9 3 )
1 .4 1 (1 .0 2 – 1 .9 4 )
0 .9 9 (0 .9 7 – 1 .0 0 )
0 .8 3 (0 .6 0 – 1 .1 6 )
1 .5 2 (0 .8 8 – 2 .6 1 )
1 .2 2 (1 .1 3 – 1 .3 1 )
1 .1 3 (0 .9 8 – 1 .3 0 )
a re fe re n ce
ca te g o ry
is la y p e o p le
b re fe re n ce
ca te g o ry
is re lig io u s
b o ld :s ig .p
< 0 .0 5 .
do i:1 0. 13 71 /jo ur na l.p on e. 01 59 08 6. t0 03
Fair Allocation of Scarce Medical Resources
PLOS ONE | DOI:10.1371/journal.pone.0159086 July 27, 2016 11 / 18
preferred by healthier participants. The same effect was observed for combination of criteria in situations A and B, and for monetary contribution in situation B (OR = 1.24).
Discussion Our data suggest that fairness ratings covary with (i) the rater’s medical background, (ii) the allocated resource, and with (iii) the individual factors gender, age, religiosity, political orienta- tion, and health status. The sickest first principle was clearly prioritized by LPs (see also [48]) in all three allocation situations and more so by females than by males. In theory, the sickest first principle favours the worst-off (i.e., those whose life may be at stake) and is equivalent to the need principle which is considered most fair when the recipient’s welfare is prioritized [33]. Sickest first, albeit to a slightly lesser extent, was also highly endorsed by HPs, MSs and GPs. Yet, it competed (especially in the donor organ situation) with prognosis and with combination of criteria.
Sickest first was age-dependent in situations B (hospital beds) and C (joint replacement), i.e., considered increasingly fair with advancing age of the rater—most likely because of older people’s increased mortality risk from influenza and their higher prevalence of suffering from worn-out (hip) joints.
Our empirical data do not support the normative claims by ethicists Persad et al. [14] that the sickest first principle is not morally justifiable. As our respondents were asked how they thought the three resources should be allocated (i.e., how fair the respective allocation principle is), we may assume that we have tapped their moral standpoint. If so, this may pose a challenge for ethicists (of Persad et al.’s orientation) as well as for health care administrators. However, ethicists may argue that normative requirements cannot be deduced from empirical data [52]. Whether or not this is true, it may be unwise to ignore the discrepancy between empirically tapped normative standpoints and ethicists’ moral conclusions derived on the basis of non- empirical deductions. Furthermore, moral standards may shift over time, and decisions in a democracy will not be sustainable in the long run unless legitimized by a majority.
The waiting list principle is also in contrast to what ethicists suggest. It is considered very fair by LP (and more so by females than by males in situations B and C) and to a lesser extent by MSs and HPs. However, this principle is contested by GPs. Waiting list may seem attractive at first glance, as it implies equality of opportunity [53], but the GPs may recognize its inherent shortcoming (Table 1).
Prognosis is top ranked in situations A and B by GPs, MSs (and HPs in situation A only). To estimate a patient’s chances of survival given a particular treatment requires the knowledge and experience possessed by GPs—and to some extent HPs and MSs as well. This may best explain the discrepancy between these three groups and LP.
Whether one should take into consideration if a person’s behaviour was harmful to her/his health or not is contested by respondents. Political orientation varies markedly with fairness conceptions, i.e., the more a respondent was leaning to the political right, the more likely s/he considered this principle to be fair. Behaviour has to do with responsibility, and it is well known that those on the right side of the political spectrum stress individual responsibility. Earlier studies also suggest that people tend to negatively sanction those who are deemed responsible for their predicament, e.g., abuse of alcohol and corresponding need for liver trans- plant [39–40, 44–45, 54].
Instrumental value is a principle that is included in several pandemic preparedness plans [55]. It prioritizes those (e.g., health care workers) who are pivotal in keeping essential services functioning. However, it is a questionable principle from a moral standpoint, because its focus is on efficiency that does not consider individual needs [17]. Likely results from the application
Fair Allocation of Scarce Medical Resources
PLOS ONE | DOI:10.1371/journal.pone.0159086 July 27, 2016 12 / 18
of this principle highlight the dilemma between favouring the worst-offs (individual perspec- tive) and saving the most lives (utilitarianism, societal perspective).
Combination of criteria is a principle favoured by GPs and MSs, particularly in situations A and C. Practitioners are exposed to patients on a daily basis and are well aware that ‘one-size- fits-all’ approaches may not be feasible when allocating scarce medical resources. The need for differentiation is obvious, as neither sickest first, waiting list, or prognosis may be adequate allo- cation principles for all specific situations of medical scarcity. The combination principle may also reflect current practice (see Box 1), particularly when donor organs are allocated. Both political orientation and health status affect the response pattern: the more left-leaning and the healthier respondents are, the more likely they are to consider a combination of criteria fair.
Youngest first is another controversial principle, as illustrated in a recent German study [49]. Only GPs in situation A consider this principle fair to a certain extent. This may reflect the controversy of giving a young person a chance to live a full life-span and at the same time not excluding old people [56–57]. Furthermore, religious (more than non-religious) respon- dents favour youngest first. Most religious people in Switzerland belong to a Abrahamic religion which implies that “there is a time for everything [. . .] time to give birth and a time to die; a time to plant and a time to harvest what is planted” [58]. In other words, younger persons have to be granted time to “harvest”.
Counter-intuitively, neither group considers lottery to be fair. This principle is frequently rejected, although it is a very fair principle from a moral standpoint, as it gives everybody an equal chance/opportunity [59]. The major disadvantage, as Persad et al. [14] point out, is that lottery is insufficient, blind to other relevant factors. The more left-oriented respondents are, the more likely they are to consider this principle fair. This can perhaps be explained by the fact that lottery is a type of equality which is a major value in left-oriented groups. On the other hand, monetary contribution is often opted by right-oriented persons who favour individual responsibility and less government involvement. Still, in view of the fact that health care regula- tors may come under huge pressure to balance increasing health costs and patients’ needs (e.g., by expensive new therapies), this clear rejection of monetary contribution cannot be ignored.
Limitations The samples of participants in our study may not be representative of the populations from which they were drawn. Although we made sure that the participants from the MRP matched the age and gender distributions of the German-speaking part of Switzerland, and the fact that other demographic variables (level of education, political orientation) also seem to be balanced and in line with our expectations, we cannot rule out selection and self-selection biases.
Further, with regard to our selection of allocation principles following Persad et al. [14] (there is a slight difference in how we defined sickest first), we may have overlooked principles that might have been relevant and perhaps preferred by our respondents (see, e.g., [48]). Further, our participants might have responded differently given additional time to reflect more thoroughly on the presented allocation situations and on the implications of the various allocation principles
The generalizability of our findings to other populations is limited. Switzerland is among the ten wealthiest countries (GDP capita, at purchasing power parity [60]) and, hence, scarcity problems exist on a very different level and affect less people than in many other countries. As a consequence, our findings may not apply in its entirety to societies in poor parts of the world, where scarcities of basic medical resources are widespread or, for that matter, even to other wealthy countries. Diverging perceptions of what is fair are also likely to exist due to individual experiences with different healthcare systems. Further, there is ample evidence in the literature on justice, that cultural differences in fairness judgements are common [54].
Fair Allocation of Scarce Medical Resources
PLOS ONE | DOI:10.1371/journal.pone.0159086 July 27, 2016 13 / 18
Finally, comparisons between studies on principles for social resource allocation, in this case, between those with a focus on medical resources, may not be entirely valid. The reason is that respondents are not always asked to evaluate or rank order the principles in terms of the same criterion. Our study tapped prescriptive justice evaluations, while others may concern evaluations in terms of preferences, importance, endorsement, support, or stated behavioural intentions.
Conclusions Ethical reasoning is prescriptive and asks ‘what ought to be’. The present study complements this perspective via a social psychological, empirical description of respondents’ prescriptions. They were asked according to which principle (from a list of nine per resource) they think three resources should be allocated, and how just and fair they consider each one of the nine allocation principles to be. The social psychological approach involves ‘what is’, in this case what is a fair allocation of scarce medical resources according to the respondents’ subjective perceptions?
However, it may be unwise to derive normative (moral) principles from empirical results [52]. A majority opinion is no guarantee against moral wrong-doings, as the history of human- kind has repeatedly shown. Nevertheless, and due to the possibility that the prescriptive prefer- ences of the general public and the ethicists’ theoretical moral derivations may not necessarily be in agreement, a generally accepted foundation is crucial, on the basis of which allocation principles for scarce medical resources are morally justified and democratically accepted. Empirical insights cannot be ignored in the context of normative justice research, and vice versa. Ethicists as well as health care regulators need to take into consideration what people per- ceive as just allocation of medical resources. Lest we risk the two justice perspectives to become completely detached from each other. From an academic perspective this may not be a problem, but from a clinical and societal perspective it would clearly matter: normative claims may be considered unrealistic and even outright unfair, and may—as a result—be ignored in daily life.
We identified such a gap regarding the popular principles sickest first (endorsed by all groups) and waiting list (not favoured by GPs), both of which conflict with the ethical perspec- tive of Persad et al. [14]. We assume that most GPs are aware of the inherent disadvantages of the waiting list principle (Table 1), and that the other groups would agree on less favourable fairness judgements of this principle, if they were equally well informed. On the other hand, we expect that most groups would reject Persad et al.’s [14] arguments against the sickest first prin- ciple as practically inapplicable, for instance and in particular their criticism of the principle’s inherent trade-off between the neediest today versus those of the future. If ethicists believe their arguments to be the most ethically beneficial, attempts to influence people’s fairness per- ceptions are needed in order to narrow the gap between both justice perspectives.
We think societal consensus among respondent groups is possible, even though their justice evaluations may diverge. Existing differences in opinions among different categories of stake- holders prompts us to face the problem of how to select a (or a combination of) legitimate deci- sion maker(s) to make possible fair allocations of scarce medical resources.
In summary, our study
• widens the scope of the discussion about how to fairly allocate scarce medical resources by combining ethical reasoning and empirical data,
• is comprehensive and takes us beyond earlier empirical studies of scarce medical resource allocation, in that we include ten allocation principles, three medical resources, and four eval- uation groups,
Fair Allocation of Scarce Medical Resources
PLOS ONE | DOI:10.1371/journal.pone.0159086 July 27, 2016 14 / 18
• enlightens academic as well as societal discussions, as our empirical data contradict certain ethical standpoints about major allocation principles, and
• may challenge current practice and health regulators, as it reveals major differences in fair- ness conceptions between different categories of people.
Considering the nature of our results we recommend that (1) a dialogue is initiated concern- ing the prescriptive ethicist’s and the general public’s views of justice concerning medical resource allocation, and that (2) health administrators, decision-makers, and allocators of scarce medical resources be alerted to the necessity of finding fair, efficient, and legitimate solu- tions. To cite Richardson “[. . .] ethical principles should be broadly consistent with commu- nity values but [. . .] community values should, in turn, be subjected to ethical scrutiny, debate and revision.”, p.5 [61].
Supporting Information S1 Dataset. Dataset used for the current study. (CSV)
S1 Fig. Fairness ratings and forced choice responses (percentage and 95% CI) by medical students versus other health professionals regarding three situations of scarce medical resource allocation. (EPS)
S1 File. Screenshot of the online questionnaire relevant to this paper. The file includes trans- lated text from German. (PDF)
S1 Table. Socio-demographic profile of respondent groups. a Percentage; b 11-point scales ranging from 1 = most left to 11 = most right. (DOC)
S2 Table. Mean (M), standard deviation (SD), F-statistic, and p-values of fairness ratings of nine allocation principles by medical students, general practitioners, other health profes- sionals, and lay people for three situations of scarce medical resource allocations. 7-point Likert scales ranging from 1 = totally unjust to 7 = totally just. (DOC)
S1 Text. Codebook of dataset used for the current study. (DOCX)
Acknowledgments We acknowledge and express our gratitude to the Cogito Foundation for funding this research and to the Institute of Primary Care at the University of Zurich for supporting the data collec- tion. We also thank Sandro Bösch for assembling and laying out figures, Maria Rey for admin- istrative support, the project’s advisory board (Drs Daniel Koch, Christian Studer, and Jakob Zinsstag) for giving generously of their time and for their helpful suggestions, Drs Hans C. Matter and Elvira Del Prete for their advice on the legal history of donor organ and pandemic influenza vaccine allocation in Switzerland, several ETH faculty members for participating in a pilot study, and the respondents for their time to complete the questionnaire. We gratefully acknowledge Dr Govind Persad’s comments that helped to sharpen up the quality of this man- uscript, as well as the helpful comments of two reviewers and the journal’s editor.
Fair Allocation of Scarce Medical Resources
PLOS ONE | DOI:10.1371/journal.pone.0159086 July 27, 2016 15 / 18
Author Contributions Conceived and designed the experiments: PK TS TR KT. Performed the experiments: PK TS TR. Analyzed the data: PK TS TR KT. Wrote the paper: PK TS TR KT.
References 1. Universal declaration of human rights [Internet]. 1948. Available: http://www.ohchr.org/EN/UDHR/
Pages/Introduction.aspx.
2. International Covenant on Economic, Social and Cultural Rights [Internet]. 1966. Available: http://www. ohchr.org/EN/ProfessionalInterest/Pages/CESCR.aspx.
3. Wolff J. The Human Right to Health (Norton Global Ethics Series): WW Norton & Company; 2012.
4. Backman G, Hunt P, Khosla R, Jaramillo-Strouss C, Fikre BM, Rumble C, et al. Health systems and the right to health: an assessment of 194 countries. The Lancet. 2008; 372(9655):2047–85.
5. Maitland K. Management of severe paediatric malaria in resource-limited settings. BMC Medicine. 2015; 13:42. doi: 10.1186/s12916-014-0263-6 PMID: 25858094
6. Brennan T, Shrank W. New expensive treatments for hepatitis C infection. Journal of the American Medical Association. 2014; 312(6):593–4. doi: 10.1001/jama.2014.8897 PMID: 25038617
7. Fleck LM. Fair funding of extraordinarily expensive medication. Bioethica Forum 2011; 4(3):96–8.
8. Matter HC. Bern: Federal Office of Public Health; 2016.
9. Sitter-Liver B. Gerechte Organallokation—Ethisch-philosophische Überlegungen zur Verteilung knap- per medizinischer Guüter in der Transplantationsmedizin. Bern: Universität Freiburg, Philosophie Dd; 2003.
10. Del Prete E. Bern: Federal Office of Public Health; 2016.
11. Cookson R, Dolan P. Principles of justice in health care rationing. Journal of Medical Ethics. 2000; 26 (5):323–9. PMID: 11055033
12. Beauchamp TL, Childress JF. Principles of biomedical ethics: Oxford University Press, USA; 2001.
13. Daniels N. Just health: meeting health needs fairly: Cambridge University Press; 2008.
14. Persad G, Wertheimer A, Emanuel EJ. Principles for allocation of scarce medical interventions. The Lancet. 2009; 373(9661):423–31.
15. Liss P-E. Hard choices in public health: the allocation of scarce resources. Scandinavian Journal of Public Health. 2003; 31(2):156–7. PMID: 12745767
16. McConnell T. Allocating Scarce Medical Resources. The International Encyclopedia of Ethics. Oxford UK: Blackwell Publishing Ltd; 2013.
17. Emanuel EJ, Wertheimer A. Public health—Who should get influenza vaccine when not all can? Sci- ence. 2006; 312(5775):854–5. PMID: 16690847
18. Gibson JL, Martin DK, Singer PA. Setting priorities in health care organizations: criteria, processes, and parameters of success. BMC Health Services Research. 2004; 4:1.
19. Hirose I. Should we select people randomly? Bioethics. 2010; 24(1):45–6. doi: 10.1111/j.1467-8519. 2008.01706.x PMID: 19508614
20. Huesch MD. One and done? Equality of opportunity and repeated access to scarce, indivisible medical resources. BMC Medical Ethics. 2012; 13:11. doi: 10.1186/1472-6939-13-11 PMID: 22624597
21. Nelson RM, Drought T. Justice and the moral acceptability of rationing medical care: the Oregon experi- ment. Journal of Medicine and Philosophy. 1992; 17(1):97–117. PMID: 1545187
22. Moss AH, Siegler M. Should alcoholics compete equally for liver transplantation? Journal of the Ameri- can Medical Association. 1991; 265(10):1295–8. PMID: 1995977
23. Peterson M. The moral importance of selecting people randomly. Bioethics. 2008; 22(6):321–7. doi: 10. 1111/j.1467-8519.2008.00636.x PMID: 18445094
24. Rhodes R. Justice in medicine and public health. Cambridge Quarterly of Healthcare Ethics. 2005; 14 (01):13–26.
25. Thompson AK, Faith K, Gibson JL, Upshur RE. Pandemic influenza preparedness: an ethical frame- work to guide decision-making. BMC Medical Ethics. 2006; 7:1.
26. Kerstein SJ, Bognar G. Complete Lives in the Balance. American Journal of Bioethics. 2010 2010; 10 (4):37–45. doi: 10.1080/15265160903581718 PMID: 20379920
27. Walster E, Berscheid E, Walster GW. New directions in equity research. In: Berkowitz L, Walster E, edi- tors. Equity theory: Toward a general theory of social interaction. New York: Academic Press; 1976.
Fair Allocation of Scarce Medical Resources
PLOS ONE | DOI:10.1371/journal.pone.0159086 July 27, 2016 16 / 18
28. Deutsch M. Distributive justice: a social-psychological perspective. New Haven: Yale University Press; 1985.
29. Törnblom KY, Kazemi A. Distributive justice: Revisiting past statements and reflecting on future pros- pects. In: Cropanzano RS, Ambrose ML, editors. The Oxford handbook of justice in the workplace. New York: Oxford University Press; 2015.
30. Schwappach DL. Does it matter who you are or what you gain? An experimental study of preferences for resource allocation. Health Economics. 2003; 12(4):255–67. PMID: 12652513
31. Törnblom KY. The social psychology of distributive justice. In: Scherer KR, editor. Justice: Interdisci- plinary perspectives. Cambridge: University Press; 1992. p. 177–236.
32. Foa UG, Converse J, Törnblom KY, Foa EB. Resource theory. Explorations and applications. New York: Academic Press, Inc.; 1993.
33. Deutsch M. Equity, equality, and need: What determines which value will be used as the basis of distrib- utive justice? Journal of Social Issues. 1975; 31(3):137–49.
34. Bauman CW, Skitka LJ. Moral disagreement and procedural justice: Moral mandates as constraints to voice effects. Australian Journal of Psychology. 2009; 61(1):40–9.
35. Krütli P, Stauffacher M, Pedolin D, Moser C, Scholz RW. The process matters: Fairness in repository siting for nuclear waste. Social Justice Research. 2012; 25(1):79–101.
36. Matania E, Yaniv I. Resource priority, fairness, and equality-efficiency compromises. Social Justice Research. 2007; 20(4):497–510.
37. Leavitt K, Reynolds SJ, Barnes CM, Schilpzand P, Hannah ST. Different hats, different obligations: Plu- ral occupational identities and situated moral judgments. Academy of Management Journal. 2012; 55 (6):1316–33.
38. Johnson RE, Lord RG. Implicit effects of justice on self-identity. Journal of Applied Psychology. 2010; 95(4):681. doi: 10.1037/a0019298 PMID: 20604588
39. Skitka LJ, Tetlock PE. Allocating scarce resources—a contingency-model of distributive justice. Journal of Experimental social Psychology. 1992; 28(6):491–522.
40. Furnham A, Thomson K, McClelland A. The allocation of scarce medical resources across medical con- ditions. Psychology and Psychotherapy-Theory Research and Practice. 2002; 75:189–203.
41. Furnham A, Petrides K, Callahan I. Prioritizing patients for surgery: Factors affecting allocation of medi- cal resources for kidney transplantation, IVF, and rhinoplasty. Journal of Applied Social Psychology. 2011; 41(3):588–608.
42. Furnham A, Ariffin A, McClelland A. Factors affecting allocation of scarce medical resources across life-threatening medical conditions. Journal of Applied Social Psychology. 2007; 37(12):2903–21.
43. Furnham A. Factors relating to the allocation of medical resources. Journal of Social Behavior and Per- sonality. 1996; 11(3):615–24. PMID: 11660693
44. Fortes PA, Zoboli EL. A study on the ethics of microallocation of scarce resources in health care. Jour- nal of Medical Ethics. 2002; 28(4):266–9. PMID: 12161584
45. Lenton AP, Blair IV, Hastie R. The influence of social categories and patient responsibility on health care allocation decisions: Bias or fairness? Basic and Applied Social Psychology. 2006; 28(1):27–36.
46. Wiseman D. Medical resource allocation as a function of selected patient characteristics. Journal of Applied Social Psychology. 2006; 36(3):683–9.
47. Neuberger J, Adams D, MacMaster P, Maidment A, Speed M. Assessing priorities for allocation of donor liver grafts: survey of public and clinicians. BMJ. 1998; 317(7152):172–5. PMID: 9665895
48. Cicognani E, Mancini T, Nicoli MA. Criteria for the allocation of medical resources: Citizens' perspec- tives. Journal of Applied Biobehavioral Research. 2007; 12(1):13–34.
49. Diederich A, Winkelhage J, Wirsik N. Age as a criterion for setting priorities in health care? A survey of the German public view. PLOS ONE. 2011; 6(8).
50. Ubel PA, Loewenstein G. Distributing scarce livers: the moral reasoning of the general public. Social Science & Medicine. 1996; 42(7):1049–55.
51. Hope T. Empirical medical ethics. Journal of Medical Ethics. 1999; 25(3):219–20. PMID: 10390674
52. Hume D, Norton DF. A treatise of human nature. Oxford: Oxford University Press; 2000.
53. Childress JF. Putting patients first in organ allocation: An ethical analysis of the US debate. Cambridge Quarterly of Healthcare Ethics. 2001; 10(4):365–76. PMID: 14533403
54. Luyten J, Kessels R, Goos P, Beutels P. Public preferences for prioritizing preventive and curative health care interventions: A discrete choice experiment. Value in Health. 2015; 18(2):224–33. doi: 10. 1016/j.jval.2014.12.007 PMID: 25773558
Fair Allocation of Scarce Medical Resources
PLOS ONE | DOI:10.1371/journal.pone.0159086 July 27, 2016 17 / 18
55. Swiss influenza pandemic plan: Strategies and measures to prepare for an influenza pandemic [Inter- net]. Federal Office of Public Health. 2013. Available: http://www.bag.admin.ch/influenza/01120/01132/ 10097/10104/index.html?lang=en.
56. Daniels N. Am I my parents' keeper?: an essay on justice between the young and the old: Oxford Uni- versity Press New York; 1988.
57. McKerlie D. Justice Between the Young and the Old. Philosophy & Public Affairs. 2001; 30(2):152–77.
58. Bible The. Ecclesiastes 3.2 [Internet]. Available from: http://biblehub.com/ecclesiastes/3-2.htm.
59. Broome J. Selecting people randomly. Ethics. 1984:38–55. PMID: 11651785
60. GDP per capita PPP (current international $) [Internet]. 2015 [cited August 10, 2015]. Available: http:// data.worldbank.org/indicator/NY.GDP.PCAP.PP.CD?cid=DEC_SS_WBGDataEmail_EXT.
61. Richardson J. Empirical Ethics: Or the Poverty of Ethical Analyses in Economics and the Unwarranted Disregard of Evidence in Ethics. Working Paper 120. Monash University, Australia: Centre for Health Program Evaluation, 2001.
Fair Allocation of Scarce Medical Resources
PLOS ONE | DOI:10.1371/journal.pone.0159086 July 27, 2016 18 / 18
Copyright of PLoS ONE is the property of Public Library of Science and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use.