Article Text

Download PDFPDF

Phase-dependent justification: the role of personal responsibility in fair healthcare
  1. Kristine Bærøe1,
  2. Cornelius Cappelen2
  1. 1Department of Public Health and Primary Health Care, University of Bergen, Bergen, Norway
  2. 2Department of Comparative Politics, University of Bergen, Bergen, Norway
  1. Correspondence to Dr Kristine Bærøe, Department of Public Health and Primary Health Care, University of Bergen, Kalfarveien 31, P.O. Box 7804, 5020 Bergen 5020, Norway; Kristine.Baroe{at}


The main aim of this paper is to examine the fairness of different ways of holding people responsible for healthcare-related choices. Our focus is on conceptualisations of responsibility that involve blame and sanctions, and our analytical approach is to provide a systematic discussion based on interrelated and successive health-related, lifestyle choices of an individual. We assess the already established risk-sharing, backward-looking and forward-looking views on responsibility according to a variety of standard objections. In conclusion, all of the proposed views on holding people responsible for their lifestyle choices are subjected to reasonable critiques, although the risk-sharing view fare considerably better than the others overall considered. With our analytical approach, we are able to identify how basic conditions for responsibility ascription alter along a time axis. Repeated relapses with respect to healthcare associated with persistent, unhealthy lifestyle choices, call for distinct attention. In such situations, contextualised reasoning and transparent policy-making, rather than opaque clinical judgements, are required as steps towards fair allocation of healthcare resources.

  • Distributive Justice
  • Ethics
  • Political Philosophy
  • Public Policy

Statistics from

Request Permissions

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


The INTERHEART study of 30 000 men and women in 52 countries showed that at least 90% of heart disease is lifestyle related.1 And according to a new report from the WHO, 63%, or 36 million, of the 57 million deaths worldwide in 2008 were caused by chronic, non-communicable diseases.2 The WHO claims that this has much to do with lifestyle. Smoking, drinking, lack of exercise and unhealthy eating habits all contribute to chronic disease. A highly relevant question arises: Is it fair to hold people responsible—in one way or another—for lifestyle choices with a potential adverse health impact?

The overall goal of this paper is to examine different ways of holding people responsible for healthcare-related choices. We limit our discussion to recent and specific conceptualisations of responsibility, namely those involving blame and sanctions. This is because we are interested in types of responsibility that may have real and profound consequences for people. Furthermore, we focus strictly on fairness considerations and leave discussions about incentives aside, and we concentrate on typical lifestyle choices that are known to increase the risk of illness, such as smoking and consumption of alcoholic beverages. Thus, we exclude from the analysis risky sport activities that may cause injuries.

We add to the debate by showing how some of the recent conceptualisations of responsibility can be coherently structured according to the interrelated and successive choices of an individual. This analytical approach helps us (1) provide a systematic, overall discussion of responsibility ascription in health (partly by modifying one of the conceptualisations we discuss), (2) identify how basic conditions for discussions about healthcare, fairness and responsibility ascription alter along a time axis and (3) conclude by offering preliminary policy recommendations.

Healthcare and responsibility

In the discussion about healthcare provision and responsibility ascription, the concept of time becomes vital in several respects. Different conceptualisations of responsibility can be differentiated according to (1) when in a lifecycle—from initial risk behaviour to post-treatment behaviour—responsibility is ascribed; and (2) whether the responsibility is ascribed for past behaviour, for present behaviour or for future behaviour. Figure 1 summarises prominent responsibility mechanisms that are differentiated on the basis of these characteristics.

Figure 1

A chronological overview of individuals’ potential needs for healthcare interventions related to their unhealthy lifestyle choices. We argue that adequate, moral–political justification for holding people responsible for their lifestyle choices in healthcare contexts calls for specified, phase-related considerations of the different conceptualisations of responsibility.

The phase of initial risk

The phase of initial risk is unique in that it encompasses an individuals’ entire life (including when she develops high-risk markers or risk-related diseases). In this sense, it is ‘timeless’, which separates it from the other phases that are defined by health-related events occurring. Even before developing a disease people can be held responsible for risky lifestyles. More precisely, voluntary risk takers (eg, smokers and heavy drinkers) can be held collectively responsible for the aggregate increase in treatment costs associated with their behaviours.

We call this the risk-sharing view. It can be implemented by, for example, levying a health tax on alcohol, tobacco, sugary beverages—and other risky products—so as to make risk-takers collectively pay the additional treatment costs associated with their consumption.

Ideally, the health tax should equal the increase in aggregate treatment costs for any given (voluntary) health risk, for example, the health tax on cigarettes should be set at a level where total revenues equal the cost of treatment associated with smoking. Obviously, things are more complicated if matters unrelated to healthcare are accounted for. To illustrate, those with healthy lifestyles have higher life expectancy than those with unhealthy lifestyles, and thus receive more in, for example, pensions, while those with unhealthy lifestyle arguably have lower work productivity. If such aspects are taken into equation, the precise level of the fair health tax is not straightforward.

Versions of the risk-sharing view have been discussed by among others Schmidt, Cappelen and Norheim, Roemer and Le Grand.3–7 The healthcare system itself does not give patients differential treatment on the basis of personal responsibility according to this view; rather, personal responsibility is captured by the health tax itself. Since risk-takers have paid for the expected and aggregate treatment costs related to their behaviours, they will not be denied or given suboptimal treatment by the healthcare system later on.

First phase of intervention

In the first phase of intervention, the individual has entered the healthcare system and is (a) likely to develop a lifestyle-related disease due to identified high-risk markers calling for preventive measures, or (b) has actually developed a lifestyle-related disease calling for healthcare (and health policy) interventions. At least two types of responsibility ascription are possible in this phase, often referred to as the backward-looking view and the forward-looking view.

According to the backward-looking view, personal responsibility is ascribed at the bedside for past behaviour. In its strictest version, defended by Rakowski, a publicly-financed healthcare system has no moral obligation to treat a disease traceable solely to an avoidable risk.8 A similar view is proposed by Segall (though he recognised that we may have other moral reasons than those rooted in justice to treat people regardless of responsibility).9 On a more lenient interpretation, personal responsibility is used as a prioritising criterion; for example, if liver disease is caused by long-term heavy drinking (alcoholic liver disease), the patient should be given lower priority on the waiting list than people whose need for orthotropic liver transplant cannot be traced to lifestyle.3 ,10 ,11 Yet another alternative is to introduce (increase in) copayment.3 ,12

Forward-looking responsibility is also an option in the first phase of intervention. Forward-looking responsibility means that a person is held responsible for the consequences of her future choices rather than for what she did in the past. There are different conceptualisations of forward-looking responsibility, such as Schmidt's conceptualisation of health responsibility as co-responsibility;3 and furthermore Richardson's view that future (moral) responsibility should be dependent upon people's situations and their roles, and thus institutionally differentiated.13 We focus here on a conceptualisation proposed by Feiring that may involve blame and punishment.14i According to Feiring, when patients are diagnosed with a disease x—which might be linked to lifestyle—it is too harsh to give them suboptimal treatment. Rather, they are given the following option. If they are willing to sign a contract committing to a lifestyle change supported by medical follow-ups, they are given equal treatment to those who have x but not the associated lifestyle. However, if patients are not willing to do so they will be held responsible by being given lower priority.

Nth phase of intervention

In this phase, a new medical need arises related to the same lifestyle as in the first phase of intervention. Both the backward-looking and forward-looking views are attributable in this phase. According to the latter, if patients in the prior phase have failed to meet contractual commitments, they should be considered to have a weaker claim on resources than others who either did comply or never had a high-risk lifestyle in the first place. However, if contractual commitments are met, they are given treatment on equal terms with all the others. The Nth phase of intervention can serve as a proxy for all consecutive phases where a new medical need arises that is related to the same lifestyle that caused the medical need in the previous phase (further relapses).

Objections to responsibility

In this section, we briefly present some of the most prominent objections against responsibility, before discussing them more carefully in relation to the various types of responsibility ascriptions made possible in the different phases. The focus is mostly on the extent to which the responsibility views prima facie fall victim to the objections we present, rather than on how they—arguably—can escape them.

The harshness objection

An often cited criticism against responsibility ascription in a healthcare context is that it is too harsh to deny people public healthcare simply because their disease is caused by voluntary risk exposure.4 ,6 ,9 ,14–16 We have a public obligation to help people who are in need of medical intervention, independent of why intervention is needed, assuming that helping would not impose unacceptable sacrifices on those providing the help.17

The objection of intrusion

A second argument against responsibility is that the process of establishing whether some health disadvantage can be attributed to choice or circumstance may be intrusive, demeaning and shameful for the person under assessment.3 ,4 ,10 ,14 In the same manner, Wolff argues that assessment of responsibility shows distrust and a lack of equal respect.17

The objection of avoidability

Prima facie, it only seems fair to hold a person responsible for choosing a health risk if they could have avoided it. However, it is not obvious what it means to be able to avoid risk.3 ,4 ,7 ,12 ,14 ,15 ,18 ,19 First, it is not clear when and how social and cultural influences may constitute unavoidable barriers to healthy lifestyle choices. Second, a particularly hard case involves addiction (eg, alcoholism). Perhaps it was voluntary triggered, but to what extent should patients be held responsible for continued addiction? It is, for example, unlikely that people voluntary engage in harmful, excessive drinking.20 Third—and this relates to Segalls condition of ‘reasonable avoidability’9—it seems fair to hold people responsible for a risk only if it could have reasonably been avoided. However, some risks are so good (the odds involved) that they are hard, if not impossible, to resist.ii In such situations, it might be considered unfair to hold people responsible for choosing the risk. But when is a risk so good (relative to the safe or safer alternative) that it should be considered ‘reasonably unavoidable’? Fourth, even though a lack of safe alternatives forces a person to choose a risk, it could still seem reasonable to hold her responsible, for example, if she would have chosen the risk even in a counterfactual situation with safe alternatives.21 Thus, in order to fairly ascribe responsibility, we must investigate what she would have done in a hypothetical situation where her opportunity set involved both safe and risky options.

The objection of causality

It can often be problematic to obtain information concerning the extent to which a patient's disease is caused by lifestyle or by uncontrollable genetic conditions. Establishing a causal relationship between behaviour and outcomes is impossible in many cases. And if causality cannot be established, responsibility cannot be ascribed.4 ,22


The phase of initial risk: objections

Both the back-looking and the forward-looking views are related to healthcare interventions and are thus irrelevant to consider in this phase. Rather, in this initial phase, responsibility can be ascribed by holding risk-takers collectively responsible for the aggregate increase in treatment costs associated with the risk they choose; we referred to this as the risk-sharing view. It avoids the harshness objection since it does not allocate health treatment on the basis of personal responsibility. However, it can still create burdensome situations. To illustrate, a possible result of a health tax on cigarettes is not necessarily less smoking among low-income people, but rather fewer dollars spent on nutritional food and other essentials—conceivably leading to more illness.

The risk-sharing view avoids the objection of intrusion since, on this view, healthcare treatment is not conditional on past behaviour. Responsibility is not ascribed by means of backward-looking policies at the bedside, but rather through increased taxation/premiums prior to the need for treatment.

The risk-sharing view does not fully escape the objection of avoidability. Recent research showing a higher incidence of risky behaviours among low-income groups suggests that socioeconomic factors, such as inadequate income or lack of education, influence lifestyle.23 It could therefore be argued that some people (eg, the low-educated) are less responsible for risk exposure—such as smoking—than others (eg, the high-educated).7 In that case, it is unfair that all risk exposures pay the same amount of health tax (or premium increase). Walker has argued against tax (or minimum price) on, for example, alcohol precisely because it unfairly places considerable burdens on those already most disadvantaged in society.20

Albertsen and Knight have argued that the objection of causality applies to the risk-sharing view:Both the tax and mandatory insurance options do, however, involve placing unavoidable costs on people whose risky behavior in the end does not result in bad health, a controversial outcome in the luck egalitarian literature.10

In our opinion, however, this would only be controversial if people generally prefer not to pool the risk, such as is the case in Las Vegas style gambling (eg, when playing the roulette and slot machines). However, it is highly likely that people prefer to pool healthcare risks, such as smoking (prior to any knowledge of whether the risk result in bad health) rather than to bear the consequences of them alone.16 Finally, we would like to emphasise that not all risky activities are taxable (eg, sitting on the couch all day) since they are not administratively controllable. Furthermore, some risks are non-linearly associated with bad outcomes, for example, it could be healthy to drink two glasses of wine per day, but unhealthy to drink a bottle of wine per day. The first objection is pragmatic and shows that the risk-sharing view cannot be perfectly implemented—it cannot cover all risks. The second one is philosophical, and illustrates that the risk-sharing view can sometimes produce unfair outcomes in the sense of holding people responsible for too much (for a risk they did not take).

First phase of intervention: objections

At this stage, preventive and curative intervention is called for, and thus both backward-looking and forward-looking responsibility can be ascribed. Even a ‘lenient’ version of the former view, is clearly subject to the harshness objection (eg, if waiting in line for organ transplant, lower priority can lead to deaths).

The backward-looking view falls victim fully to the objection of avoidability (how do we know whether or not the patient freely chose risk exposure?), the objection of causality (how do we know that a patient's disease is causally related to lifestyle?) and the objection of intrusion (eg, the patient may be asked to submit to genetic testing). We would like to emphasise though, that if the healthcare system were to facilitate lifestyle changes by offering easily accessible opportunities to change lifestyle free of charge, the backward-looking view would be less vulnerable to the avoidability objection. We elaborate this point more carefully later.

The forward-looking view is less of a victim to the harshness objection, since every patient whose treatment is conditional on signing a contract would have the opportunity to do so. Treatment is not conditional on past behaviour, but rather on committing to a lifestyle change. It is the capacity to form an intention about trying to change lifestyle that is called for at this stage—and not the (future) capacity to successfully comply with it. What then with those who refuse to sign the contract? It could be argued that they knowingly choose lower priority than they might otherwise get. To the extent that their refusal is voluntary, the harshness involved in providing lower priority is decreased.

Assuming that only those with unhealthy lifestyles must sign contracts, the forward-looking view falls victim to the objection of intrusion since in deciding who must sign a contract, information about past behaviours must be gathered. One could, however, argue that if treatment is conditional on signing a contract for everyone whose treatment can be improved by lifestyle change, then the bite of the intrusiveness objection is narrowed since we do not have to consider whether people are responsible for their lifestyles. Still, some might find being questioned about their bad habits intrusive.

Furthermore, to the extent that the forward-looking view requires information about patients’ past (or current) ‘irresponsible’ behaviour, it does not fully escape neither the objection of avoidability nor the objection of causality.

Nth phase of intervention: objections

In this phase, a contractually committed patient has a relapse (a new medical need associated with the same lifestyle as in the previous phase). According to the forward-looking view, the patients’ post-treatment behaviour is morally relevant, and suboptimal treatment should be given to all those who did not live up to their contractual commitment. According to Albertsen, it is therefore subject to all the objections attributed to the backward-looking view; in particular, he emphasises that, in the process of deciding which patients have failed to live up to their contractual commitment, ‘it is very difficult to assess what counts as choice and what counts as circumstance, and, furthermore, that a process of clarifying it could potentially be a demeaning and intrusive process’.12

However, we will argue that Feiring has not sufficiently discussed the relevant contractual requirements of the healthcare system for the contract to be fair, and we will therefore suggest a modified version of her forward-looking view. Assume that the healthcare system facilitates lifestyle changes by offering easily accessible opportunities free of charge, like enrolment into programmes along the lines that Feiring suggests (ie, stop smoking, stop drinking and lose weight programmes) with individually adjusted follow-ups. This would make responsibility less of a victim to the objection of avoidability, since conditions for responsible choice making, that is, easily accessible opportunities to change lifestyle, are present and equally so for all. Assume furthermore that, in evaluating whether or not a patient has lived up to the contractual commitment, it is the effort that should be evaluated and not the result.12 As long as contractually committed patients have tried to change their lifestyle by attending required programmes, they should be given priority on equal terms with the others. Then, simply checking for presence could be a feasible, non-intrusive way of ascribing responsibility.

This modified version of the forward-looking view does not fully escape the avoidability objection, namely that it is difficult to assess whether efforts to comply with contractual commitments are within or beyond personal control. However, we claim modestly that it seems more reasonable to hold patients responsible for attending required programmes if they are free of charge and easily accessible; and furthermore that it is less intrusive to simply check whether a person did attend, for example, a stop-smoking programme, compared with investigating why he/she did not quit smoking.

Finally, let us emphasise a further challenge associated with relapses. A comprehensive exploration of personal responsibility in healthcare must address whether or not there should be limits to how many interventions associated with unaltered, unhealthy lifestyle that can be carried out. (For a case study of a discussion of first-time relapse and claim on intervention, see Miljeteig et al.24) Ethically, limit setting requires reasonable justification toward those being adversely affected.25 In case of relapse, biomedical ineffectiveness (eg, inoperable tissue) would be a reasonable justification for limit setting to new interventions. However, when repeated relapses indicate that future interventions would be futile because of continued unhealthy lifestyle alone, the conditions for making claims about fairness (in a setting of resource scarcity) is generically different from within the first phase of interventions. When, if ever, is it fair to ration interventions to patients who repeatedly fail to alter their lifestyle? This question calls for transparent decision-making and contextualised reasoning and should not be left to the opaque, discretionary field of clinical judgement and informal practices. Clearly stated policies on when, if ever, people could be held responsibility for their unhealthy lifestyle in these cases could be educational and have the spin-off effect of discouraging certain lifestyle choices already in the phase of initial risk.

As summarised in table 1, the risk-sharing view is (relatively) immune to all the fairness objections we have presented except for the avoidability objection. The backward-looking view did not stand up to the test of any of the various lines of objections. Overall, Feiring’s forward-looking view did better than the backward-looking view in the first phase of intervention, but it could transmute into the backward-looking view in subsequent phases. Our modified forward-looking view fared a bit better; in particular, it is not subjected to the objection of intrusion, but the harshness, causality and avoidability objections still stand. We need to emphasise, however, that there are other versions of forward-looking responsibility not examined here that could be immune to the above objections.3iii However, these versions focus less on blame and punishment.

Table 1

An overview of justified objections to various fairness perspectives on personal responsibility in health according to different phases in a patient's history of lifestyle choices


Unless the backward-looking and forward-looking views presented here can be shown to convincingly circumvent the objections to them, holding people responsible in either of these ways in the sense discussed in this paper cannot be recommended as fair policies. The fairness of ascribing responsibility according to the risk-sharing view occurs, on the other hand, more acceptable. In the case of repeated relapses due to unaltered unhealthy lifestyle, fairness calls for a distinct approach, that is, transparent, political decision-making on when and how to limit interventions.


We wish to thank Jørgen Pedersen, Alexander Cappelen and colleagues in the research group Global Health Priorities for commenting on an earlier draft of this manuscript. We are also very grateful for invaluable comments from the reviewers and the editor of this journal.



  • Contributors The authors have contributed equally to the content and the preparation of this manuscript.

  • Competing interests None declared.

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

  • i Her view could easily be seen as inspired by Richardson's view in that responsibility depends on the situation a person finds herself in. However, Richardson's view does not involve the explicit focus on punishment found in Feiring's view.

  • ii The ‘goodness’ of a risk could be evaluated by, for example, its expected value (the expected value is calculated by multiplying each of the possible outcomes by the likelihood that each outcome will occur and summing all of those values). Thus, if the expected value of a risk is much higher than the alternative and safe (or safer) alternative, then it could be considered unavoidable.

  • iii On Schmidt's conceptualisation (2009), for example, personal responsibility is a matter of degree, and people are held responsible for future choices without blame and punishment.