The debate so far

Several arguments have been put forward both in favour of and in opposition to taking individual responsibility into account within healthcare policy. Sharkey and Gillam recently surveyed the literature concerning this topic, summarising the existing arguments and responses, as well as pointing towards gaps where further development of the discussion is needed. Drawing on their analysis, I will briefly describe the debate as it stands in order to situate my contributions within the literature.3

In what has been called the medical argument, proponents of involving personal responsibility in healthcare policy have emphasised the decreased likelihood of treatment efficacy when treating self-inflicted illnesses (ie, an alcoholic who needs a liver transplant is expected to have a poor prognosis),4 ,5 while respondents point to the lack of evidence for this claim (ie, alcoholics appear to do equally as well as others who receive liver transplants—see refs. 6 and 7 for similar counterevidence with regard to smokers and heart surgery). Another argument for taking responsibility into account suggests that public support for healthcare resources that rely on individual donations like blood and organs will decline if self-caused illnesses are treated the same as other illnesses,8 but in response it has been noted that public opinion could be founded on prejudice and is therefore not a good basis for policy.9

Several theorists who argue against the inclusion of responsibility in healthcare policy have made the point that any attempt to implement policies that take personal responsibility into account would be impractical, for several reasons. Those given include the extensive time and resources that would be required to assess each individual's responsibility for a given condition,10 ,11 the difficultly of accurately determining responsibility in relation to confounding factors,12 ,13 the impossibility of assessing if the patient had the foresight to predict what their poor choices might result in10 ,13 and the inevitability of social biases playing a role in attributions of responsibility.14 Additionally, the universalisation argument claims that if responsibility were truly taken into account, the number of behaviours that would have to be accounted for within these policies—any dangerous jobs, hobbies, sports, etc—would be enormous and unfeasible.9 ,15 ,16

Two moral arguments have been put forward in support of taking personal responsibility into account, suggesting that obligations towards others who have a claim to medical resources4 ,16–18 and obligations towards one's own health justify treating individuals differently if they fail to fulfil such obligations.19 ,20 In response to these moral arguments, authors have pointed out that in many of these cases, individuals are not in fact responsible for their poor health outcomes. A lack of responsibility may be due to a lack of control as a result of addiction, mental illness, genes or advertising,12 ,21 ,22 or the myriad number of factors that contribute to a condition like lung cancer (eg, pollution, genes, radiation may play a part as well as smoking),16 ,23 or because the risk-taking behaviour that leads to poor health outcomes is negative, and helps to maintain psychological well-being.11 ,24 Others have responded to the moral arguments by reasoning that even if one is responsible for a negative health outcome, this should not result in restrictions of medical care, as such policies would violate the patient–physician relationship by reducing trust6 ,9 and asking the physician to play an inappropriate role in judging and blaming patients.12 ,25

Two new contributions

Sharkey and Gillam's3 analysis of the literature concludes with a worry that the debate has become stagnant and repetitive, and requires further development of existing arguments or the introduction of new arguments. In response to their plea, I hope here to expand and develop an existing argument against incorporating responsibility into healthcare policy as well as to offer a critical response to an argument that has been made in favour of such policies.

The first argument takes inspiration from the work of Peter Ubel, who has warned that we should be wary of any attempt to implement healthcare policies based on personal responsibility as it will inevitably be impacted by our social biases.14 ,26 I would like to expand this worry to apply not only to instances of implementing policy, but to the entire literature concerning personal responsibility and health. In order to ground this worry, I examine the debate through a wide lens, and ask broadly what ‘personal responsibility’ appears to be tracking within this literature. I consider three possibilities, including causal responsibility, moral responsibility and culpability, and conclude that none of these is able to fully account for the focus of the discussion concerning personal responsibility and health. I argue that the exclusive focus within this debate on a small subset of those who are morally responsible for their poor health outcomes is best explained by social biases that lead proponents of taking responsibility into account to only propose policies restricting care for individuals engaging in stigmatised behaviours, while failing to mention many socially acceptable behaviours that also contribute to poor health outcomes. I end with a challenge for proponents of assessing personal responsibility within healthcare to expand the debate to include such non-stigmatised behaviours or to provide an alternative explanation for the narrow focus of the debate.

The second half of this paper develops a critical response to an argument that has been put forward by several theorists who advocate for taking individual responsibility into account, and has not yet been challenged. Dubbed the behaviour change argument by Sharkey and Gillam, this line of reasoning suggests that holding individuals responsible for their illness will lead to positive outcomes in healthcare.16 ,17 ,19 ,27 In response, I provide reason to doubt the connection between holding individuals responsible for their health and motivating healthy behaviour change, or increasing overall positive health outcomes. I also examine the case of West Virginia's Medicaid reform, an instance of healthcare policy that explicitly restricts or expands care on the basis personal responsibility, and raise worries that such policies could lead to an increase in overall medical costs. Together, these points raise the possibility that involving attributions of responsibility in healthcare could result in a policy makers' worst nightmare: worse health outcomes and higher costs.

Argument 1: more than responsibility

In order to understand how personal responsibility might justify differential treatment within a healthcare context, it is important to ask what is meant by personal responsibility within these arguments. In the analysis that follows, I will attempt to answer this question by relying on both how responsibility is described by those making arguments in favour of taking personal responsibility into account within healthcare, and the examples they offer of what such policies might look like in practice. In hopes of discovering that which unites the various positive arguments offered within this literature, I will examine three possibilities as to what might underlie personal responsibility within these debates, and upon finding them each unsatisfactory, offer an alternative explanation.

Causal responsibility

As several authors who advocate for taking responsibility into account within healthcare policy refer simply to individual responsibility, personal responsibility or responsibility for ill health,2 ,16 ,28 ,29 it is possible that what is meant by personal responsibility is merely causal responsibility. An individual is thought to be causally responsible for an event if they committed an act that directly contributed to the event, where the event would not have happened without the act.ⁱ For example, you would be causally responsible for tripping someone if you stuck your leg out and caused them to fall. Similarly, you would be causally responsible for developing an ulcer if you took aspirin every day for an extended period of time. However, causal responsibility would also include cases where an individual tripped over your leg after you had been pushed to the ground, as well as cases where you developed an ulcer as a result of regularly taking aspirin but had no knowledge of the connection between anti-inflammatory drugs such as aspirin and the risk of developing an ulcer. From a brief glance at the literature on personal responsibility and health, it is clear that the focus of this debate is more narrower than mere causal responsibility. The primary cases considered within discussions of responsibility and health include examples where someone becomes ill as a result of smoking, drinking, drug use, eating poorly or attempting suicide; these cases involve not only causal responsibility, but also control over one's actions and knowledge of the risks involved.

Moral responsibility

These features suggest that the kind of responsibility at play within this debate might be moral responsibility. Moral responsibility can be thought of as a subset of causal responsibility, since it is usually thought that causal responsibility is required for moral responsibility (this is what grounds the famous principle of ‘ought implies can’), but not all cases of causal responsibility involve moral responsibility, since there are many cases in which an individual did not have enough control or knowledge to be held accountable for their actions.30 Most authors who write in favour of taking responsibility into account in healthcare explicitly emphasise a more narrow conception of responsibility than one that is merely causal, requiring both control over one's actions and knowledge of the risks. Feiring, for example, suggests that one is responsible only for outcomes that could have been avoided, while Glannon suggests that control must be present for one to be held responsible for unhealthy behaviours.4 ,31 Persson highlights the fact that these policies are meant for individuals who ‘knowingly contribute’ to their poor health outcomes, while Buyx emphasises that personal responsibility ‘requires knowledge’.2 ,16

However, if we take moral responsibility to involve both control over one's behaviour and knowledge of the risks entailed in such behaviours, there is reason to doubt that moral responsibility is all that guides discussion of personal responsibility in healthcare policy. As pointed out by proponents of the universalisation argument mentioned above, there are a seemingly limitless number of examples of cases in which individuals freely and knowingly take risks with regard to their health.9 ,15 ,16 ,21 The cases that are seen most often in the literature, those that involve smoking, drinking, drug use, poor eating habits and attempted suicide, certainly contain these features on many occasions, but so do a wealth of other cases that rarely or never arise in discussions of taking responsibility into account in healthcare.

For example, if one chooses to live in a city with hazardous levels of air pollution or a high crime rate, or to take a job that involves working in dangerous conditions or leads to severe levels of stress, one is freely and knowingly putting oneself at risk for poor health outcomes. Similarly, individuals who are injured in a motorcycle accident or while engaging in an extreme sport, people who develop skin cancer after spending too much time in the sun or in a tanning bed or women who have surgery-specific complications during a scheduled C-section, all fit within this conception of moral responsibility, and yet are not part of the discussion of how personal responsibility might be taken into account within healthcare. There is certainly no discussion of denying care for teenagers who have been in car accidents, despite it being the number one predictor of their mortality.32 This suggests that something more than moral responsibility is at play in discussions of personal responsibility in healthcare.

For clarity, I will refer to those cases that arise repeatedly within the literature (smoking, drinking, drug use, poor eating habits and attempted suicides) as the ‘considered cases’, and those that do not arise within the literature but involve moral responsibility for one's health (extreme sports, elective surgeries, motorcyclists, etc) as the ‘unconsidered cases’. My aim here is to determine what sets the considered and unconsidered cases apart, for that criterion is likely to shed light on what is meant by personal responsibility within debates about the inclusion of responsibility in healthcare policy.

Culpability

It is possible that the cases cited most frequently within this literature are the ones in which individuals are more culpable for their actions than the unconsidered cases. While many have warned about the risks of allowing judgements of blame to enter into healthcare settings,22 ,25 ,33 and most proponents of taking responsibility into account in healthcare avoid any discussion of blame or retribution, some have explicitly acknowledged the importance of culpability within this debate.17 Perhaps, different degrees of culpability are driving authors to focus on the considered cases and ignore the unconsidered cases. So, what might contribute to different degrees of blame in the context of taking care of one's health? In a criminal justice setting, what typically contributes to an increase in blameworthiness are differences in intent and severity, for example, we punish murder more harshly than manslaughter because of the difference in intent, while we punish manslaughter more severely than stealing because of the difference in severity. In cases of personal responsibility for poor health outcomes, it is worth exploring whether or not there are differences with regard to either intent or severity between the considered and unconsidered cases. A difference in intent quickly presents itself as an unlikely candidate. Surely, an individual who smokes does so without the intention of developing lung cancer, just as an individual who plays professional football does so without the intention of ending up with a brain injury. The intentions, in both these cases, stand apart from any poor health outcomes, although the risks are known and the actions are freely chosen.

A difference in severity seems more plausible, which might be understood in this context as a difference in the amount of harm caused. Several arguments have been put forward in favour of assessing personal responsibility within healthcare that rely on harm as the foundation for differential treatment of individuals seeking medical care. We can understand harm within the context of healthcare as occurring when one knowingly makes choices that lead them to place an undue burden on the healthcare system, thereby restricting other's options for care. Differential treatment can then be defended on the basis of solidarity or fairness.16 ,31 As Grant et al27 have argued, smokers take up more resources than non-smokers and should be denied angiography or coronary bypass surgery as a result. Similarly, Feiring has suggested that an obese patient suffering from a medical condition who is unwilling to sign a contract ‘of frequent medical follow-ups to help her lose weight’ should be given lower priority with regard to care because ‘when resources are limited we owe it each other to do what we can to make medical treatment efficacious’.4 These arguments suggest that what sets the considered cases apart from the unconsidered cases is a difference in blame, based on a greater harm to others through a greater use of medical resources.

However, this is only the case if we take a short-term perspective on an individual's use of resources. If we look at the resources that smokers require over an entire lifespan, they take up far less resources than non-smokers because they are less likely to live into old age where healthcare costs are highest.34 ⁱⁱ We can expect similar mortality rates and lower lifetime costs for those who drink alcohol, use drugs, attempt suicide or have poor eating habits.35–37 While some of the unconsidered cases also have high mortality rates, especially those involving dangerous driving, the considered cases clearly do not involve a higher degree of harm on the basis of using too many resources; so, this cannot be what drives theorists to focus on the considered cases.

Stigma

A more likely possibility is that what grounds emphasis on the considered cases in this debate are biases towards particular groups and lifestyles. Peter Ubel14 has warned of the role of stigma in policies that consider personal responsibility and health, noting that ‘any attempt to base allocation on personal responsibility is impossibly entangled with our social judgments about the desirability of [a behaviour]’. In a clever demonstration of this bias, Ubel et al asked subjects to distribute hypothetical transplantable hearts to individuals who both were and were not part of three groups: intravenous drug users, cigarette smokers and people who ate high fat diets against the doctor's advice. Two other conditions were introduced, which varied whether or not these unhealthy behaviours were in fact the cause of the organ failure that led them to require a transplant and how good their prognosis was. Ubel et al found that across all conditions, subjects were reluctant to give organs to any hypothetical individuals with a history of unhealthy behaviour, but were significantly less likely to give the transplantable hearts to intravenous drug users, the most stigmatised behaviour of the three. Interestingly, the subject's willingness to give the organs to unhealthy recipients was influenced by relative prognosis, but not by whether or not the behaviour caused the organ failure.ⁱⁱⁱ The most significant finding, however, is that the factor that primarily determined subject's decisions regarding organ allocation was unhealthy behaviour, regardless of their projected prognosis after receiving an organ or whether or not that behaviour was causally responsible for their illness.26

Corroborating Ubel's worry, evidence from moral psychology suggests that attributions of both causal and moral responsibility shift as a result of the social desirability of an action.30 ,38–40 For example, Alicke found that attributions of responsibility for a driver in a car crash were significantly lower when participants believed the driver was rushing home to hide a gift for his parents' anniversary than when they believed he was rushing home to hide cocaine from his parents.39 It has also been shown that individuals have a tendency to see actions as active rather than passive if they disapprove of them. For example, previously established ethical views on abortion predicted whether subjects were likely to see a case of a pregnant woman who did not adopt a special diet to save a vitamin-deficient fetus as either killing the fetus or letting it die.41

Taken together, this research suggests that personal responsibility in healthcare is likely to appear more often to us if behaviours that are not socially approved are present as well. When the considered cases are examined in relation to the unconsidered cases, it seems very likely that bias, not merely responsibility, is guiding the focus of this discussion; each of the considered cases (smoking, drinking, drug use, obesity, past suicide attempt) is a highly stigmatised behaviour,42–45 while none of the unconsidered cases are. This exclusive focus on stigmatised behaviours within the debate is especially worrisome, given the way in which socially undesirable behaviours have been treated within healthcare contexts in the past, including the chemical castration of homosexuals and the forced sterilisation of people with mental disabilities. Furthermore, there is ample evidence that suggests that stigma and bias still play a significant role in healthcare settings today.46–50

This suggests that what underlies personal responsibility within the debate over taking responsibility into account in healthcare is not only the way in which an individual knowingly contributed to a negative health outcome, but also whether or not the individual engaged in a socially undesirable behaviour. There is no reason to make more space for such biases in healthcare, as policies that are based on attributions of personal responsibility are sure to do. As Arthur Caplan15 has warned, ‘as increasing emphasis is placed on the role that personal responsibility plays in health policy and resource allocation, great care will be required lest sin become one of the tests increasingly applied at the bedside to determine who will live and who will die’. In light of this, I challenge the proponents of incorporating responsibility into healthcare policies to either expand the scope of their arguments to include unconsidered cases as well, or to provide an alternative explanation for the exclusive focus on highly stigmatised behaviours within the literature.

Argument 2: potential consequences

My second argument concerns the potential consequences that could fall out of policies that aim to take responsibility into account within the delivery and distribution of healthcare. In what Sharkey and Gillam have called ‘the behaviour change argument’, several justifications have been put forward in favour of such policies on the basis of positive effects that can be anticipated to occur as a result. In particular, it has been suggested these policies will motivate healthy changes in behaviours,17 ,19 ,27 improve overall health outcomes16 and decrease overall healthcare costs.20 I will address each of these proposals in turn.

Positive behaviour change

In defence of a healthcare system that allocates resources on the basis of personal responsibility, Grant et al have argued that the best way to motivate smokers to give up smoking is to refuse to perform angiography or bypass surgeries on them, while Kass has made the case that refusing or reducing benefits for smokers will contribute to better health outcomes.19 ,27 Similarly, Smart has suggested that giving equal access to healthcare for all, without taking responsibility into account, may actually encourage people to adopt irresponsible and unhealthy lifestyles.17

While none of these authors offered evidence for these claims, there is reason to doubt that policies based on responsibility for ill health would lead to positive changes in health behaviour. First of all, most of the considered cases involve some level of addiction or compulsion (ie, smoking, substance use, unhealthy eating), a hallmark feature of which is that the behaviour continues in the face of adverse consequences. Recent evidence has linked this tendency to what has been called a reduced sensitivity to punishment in individuals engaging in addictive behaviours.51 ,52 It seems even more unlikely that individuals who are planning a suicide attempt will be deterred by a policy that might jeopardise their transplant candidacy in the future. Second, behaviour change is not primarily motivated through fear and punishment. While debates over the particulars of how to motivate positive behaviour change still remain, there is widespread agreement that changing individual attitudes, building experiences of self-efficacy and altering societal norms all play an important role, while punishment is given little credit.53 It does seem hard to imagine that the distant threat of being denied care in the case that one might develop lung cancer will make a significant difference to an individual's current smoking habits. If it did, we would expect to see reduced smoking rates in the large percentage of the US population that is not covered by health insurance, and yet this is the opposite of what is found.54 As Lockwood has expressed, ‘someone who is undeterred by the prospect of seriously damaging his health is hardly likely, in my opinion, to be deterred by the prospect of less than ideal health care thereafter’.

Improved health outcomes

In a more sweeping version of the behaviour change argument, Buyx16 has offered as justification for health policies based on responsibility, ‘the very positive effect that personal responsibility has in practice: it leads to better health’. However, it is not obviously true that improved health outcomes would result from restricting or reducing care on the basis of responsibility. In order to explore the likelihood of Buyx's prediction coming true, it is crucial to consider who would be most affected if we were to begin to redesign the healthcare system according to assessments of personal responsibility. The majority of stigmatised behaviours included in the considered cases—cigarette use, various kinds of substance abuse, poor eating habits—are taken up most frequently by those at the bottom of the socioeconomic hierarchy.55–57 Interestingly, alcoholism does appear to spread itself fairly evenly across tax brackets,58 but unremitting alcoholism is associated with disadvantaged neighbourhoods.59 Many of these behaviours co-occur with psychiatric illnesses, and especially so in the least well off.60 Regardless of the form of policy change that might be implemented—higher taxes, rewards for good behaviour, restrictions for the most unhealthy—those that will be most affected by such measures are already the most disadvantaged. Mary Simmerling has observed that poverty is already a contraindication for organ transplant because no one on Medicaid can actually afford the costly immunosuppressant medications required to survive after a transplant.61 What would be the consequences of making health resources even harder to access for this population?

The relationship between inequality and poor health outcomes is a fairly robust one. Research has shown that both health and longevity increase as a society becomes more egalitarian, while socioeconomic inequality leads to poorer health outcomes within a population.62 It has been observed that the greater the income distribution within a country, the lower the life expectancy. This shows that absolute wealth predicts health outcomes to a lesser degree than relative wealth, suggesting that being at the bottom of a tall pyramid is more harmful than being at the bottom of a short one. A contrast between Britain and Japan shows this effect quite clearly. While comparable in terms of income distribution and life expectancy in the 1970s, the two countries diverged in the following decades, with Japan's income distribution narrowing dramatically and Britain's growing significantly. Along with these changes, Japan's life expectancy rose to be one of the highest in the world, while Britain's sunk considerably.63 This evidence demonstrates a definite link between inequality and health outcomes, and should certainly give us pause before implementing policies that primarily serve to reward those at the top of the hierarchy while restricting care for those at the bottom. If widening this gap will in fact lead to worse health outcomes, it may be that taking personal accountability into consideration within health policies will have the unintended effect of making people unhealthier.

Reduced healthcare costs

In his discussion of the appropriate role for blame in healthcare, Martin remarked that holding people responsible for their dangerous health behaviours ‘might be the most cost-effective’, while also acknowledging that this is an empirical question at heart.20 Since there are few policies in place that incorporate assessments of individual responsibility, the empirical evidence is limited, but the case of West Virginia's implementation of a Medicaid reform that took place between 2006 and 2010 is one source. This reform explicitly embraced personal responsibility as a central factor in determining whether or not individuals would receive what was called the ‘basic plan’ or an ‘enhanced plan’ that offered unlimited prescriptions (as opposed to the usual four per month), and otherwise unavailable diabetes care, tobacco-cessation programmes, substance use and mental health services, among other resources. In order to receive the enhanced plan, members had to opt in as well as sign and adhere to a ‘Medicaid Member Agreement’, which included a list of responsibilities such as attending programmes prescribed by physicians, showing up for appointments on time and not using the emergency room (ER) for non-emergencies. Violations of this agreement could result in being deferred to the basic plan.64 As noted by Steinbrook, many of the patients that would benefit most from these additional services are those that would have the hardest time complying with the Member Agreement requirements. He suggests several reasons including language barriers, psychiatric illness, medication side effects and transportation that might inhibit patients’ abilities to fulfil these requirements and result in them being denied enhanced Medicaid services.64 Schwartz65 draws attention to the potential psychological effects of such a policy, in that ‘holding patients to standards of responsibility that are unreasonably difficult for them to meet can be disempowering’, while Bishop and Brodkey66 observe that ‘this plan asks the most vulnerable population to do more with less ability to accomplish what we ask of them’.

While these points emphasise the worries raised above in relation to both changes in behaviour and improvements in health outcomes as a result of policies that take responsibility into account, the economic aspects of West Virginia's Medicaid reform are also disconcerting. When ER visits were analysed after the fact by Gurley-Calvez et al, it was found that individuals on the enhanced plan had a lower likelihood of visiting the ER in a non-emergency, but that adults on the basic plan increased their primary-care treatable ER visits by 7%. Overall, visits to the ER increased after the reform was implemented, because by 2009, only 14% of members had opted in for the enhanced plan.67 These results suggest that, at least in this particular case, policy designed to restrict services for those who are unable or unwilling to comply with the assigned health protocols may lead to increases in healthcare costs.

Taken together, the above discussion warns that basing decisions regarding healthcare allocation on assessments of personal responsibility is unlikely to contribute to positive behaviour change and has the potential to lead to overall worse health outcomes and higher spending on healthcare services. While the evidence I have considered is insufficient to determine with certainty what the effects of policies that take personal responsibility into account might be, I hope to have offered at least the beginning of a rebuttal to arguments that suggest policies informed by personal responsibility will lead to positive outcomes.

Summary

In an investigation of the meaning of ‘personal responsibility’ within discussions of policies that consider individual accountability for negative health outcomes, I have made the case that causal responsibility, moral responsibility and culpability do not track the narrow focus of this debate on highly stigmatised behaviours. Instead, I have argued that biases towards the lifestyles that take centre stage within the literature play a significant role, and have challenged proponents of such policies to either expand the scope of the debate or provide an alternative explanation for the neglect of poor health outcomes that arise through socially approved behaviours. Additionally, I have offered a response to the behaviour change argument, which anticipates positive outcomes in response to policies that take responsibility into account. I have provided reason to doubt that individuals would refrain from engaging in these unhealthy behaviours as a result of these policies, and warned that by restricting care for the least well off, these types of measures may only serve to increase inequality and thereby lead to a probable increase in negative health outcomes. Finally, I have considered the case of West Virginia's Medicaid reform and cautioned that higher costs in healthcare could potentially be another unwanted result of health policy founded on the notion of personal responsibility.