Skip to content

The efficiency of treating fat smokers

Recent changes to the NHS raise the potential for health care providers to deny treatment based on an individual’s characteristics, such as their weight or whether they smoke. I think this calls for a reminder of the implications of discriminating in this way, and why, I think, we would do best to avoid it.

Public preferences

The NHS is paid for by the public, so we generally accept that it should do what the public wants (unless the public gets it wrong, of course). A growing body of literature exists on public preferences for the distribution of health care, and the ways in which people might like to discriminate; at least one review of the literature has existed since 2005. When it comes to organ transplantation, some people show a preference for the recipient to have not caused their own illness. Elsewhere, polls have found that around 40% of people would give priority to such individuals. Despite this representing a large minority it does not indicate wide public support for discrimination based on this. Dolan et al (2008) reported that people give 8% less weight to individuals whose illness can be partially attributed to their own lifestyle.

So, it is not clear that the public wants the NHS to discriminate in this way. If it were clear it would presumably be included in NICE guidance. However, cynical readers will of course recognise that such policies are driven by financial incentives rather than a desire to act in the public’s interest.


8% said Dolan and co. This suggests that, if an intervention is close to NICE’s threshold QALY value, it may be justifiable for the fat smokers to miss out. Afterall, their health gain is of a reduced value to society. However, the real efficiency question is whether individuals responsible for their own health problem are actually likely to have a smaller health gain from treatment. This is a very real possibility, as fat smokers are more likely to experience complications in treatment and whatnot. More research should explicitly consider the impact of individual characteristics on the cost-effectiveness of treatment. But people’s capacity to benefit from treatment could be influenced by many things. If we are to discriminate based on efficiency arguments then we must also discriminate between people based on race, age, sex, sexuality, occupation, where they live, and who-knows-what else.

For me, none of this matters. Clinicians can make judgements in individual cases, and it may be efficient to do so, but I find it hard to imagine a situation where discriminating against individuals based on arbitrary distinctions would increase efficiency. If you don’t treat the fat smokers they’ll probably get more ill. They’ll probably die young. When the cancer spreads and the myocardial infarctions become frequent, it’ll cost a lot to save their lives  – a cost that just might have been avoided.

We now have a newsletter!

Sign up to receive updates about the blog and the wider health economics world.

0 0 votes
Article Rating
Notify of

Oldest Most Voted
Inline Feedbacks
View all comments
Matthew Franklin
11 years ago

That’s kind of the point mate, their capacity to benefit is exactly the same so on this basis you would not make a distinction between the two people, but given the process of getting to a particular state of being perhaps we should distinguish between the two. Move away from the objective goal of health maximisation and step back in to the objective goal of maximising societal welfare (aggregated utility) which is kind of the point of health maximisation but within the budget of the healthcare system. We have scare resources given the budget constraint, and therefore two people in the same health state should in theory get the same treatment as they should both have the same capacity to benefit, but this is only efficient if we assume that we have the resources to treat both. In reality we do not and therefore in the ideology of an efficient healthcare system we need to start making distinctions between people even in the same health state of which process should be our method of making allocative decisions.

Back to the capabilites approach, it is not the ability to do or be before the health state, it is the ability to do or be given the health state they are currently in. Back to me original example a persons ability to do or be is reduced by health state A, however person X has already accepted this reduced capability of health state A as something they are willing to accept (i previously referred to it as a negative externality), but person Y chose not to smoke in order to avoid the health state and the new (reduced) capability set associated with the health state. In this example person Y has a greater ‘right’ to healthcare over person X. And as previously mentioned if we chose to allocate healthcare on this basis it could promote self care as well as improve allocative efficiency to improve societal welfare.

Matthew Franklin
11 years ago

Your discussion intrigues me Chris, and although i do not suggest that “fat smokers” should be denied care, i do think there is a precedence to suggest that their right to healthcare should be weighted different to the general public. In fact using a twist on the capability approach, it is possible to discuss why we should weigh the benefit of those who choose to smoke, drink, and eat to excess less than the general public when it comes to healthcare. Plase note my following discussion has some limitations such as: weights should be allocated only in the case where a persons healthcare need is affected by the previous decisions (process) they have made; this benefit is based on aggregated utility rather than health maximisation; and the weights would have to be based on value judgements rather than any preference-based method.

So to start off, the capability approach suggests that it is a persons ability to do and be that should be the objective of policy makers. To this extent ‘process’ should be accounted for when making healthcare allocation decisions. If we have two people in health state A (for the purpose of discussion health state A is a smoking related disease, any will do) under the current methods for cost-effectiveness analysis focusing on health maximisation both persons are of equal ‘need’ of healthcare. However, lets say person X is a smoker, and person Y is a passive smoker; both have the same health state but using value judgements (which i think follows in line with your discussion) society would deem that person Y actually has a higher ‘need’ in relation to this person ought to receive healthcare more than person X. The reason for this is that person X has already had the choice to weigh off his own personally choice of suffering the health state and chosen that the health state is worth the process of smoking, where as person Y has had the ability to choose to smoke and chosen not to but has ended up in the same health state anyway. Under these conditions person Y has the greater need for healthcare because the healthcare system should offset the negative externality (the health state) by providing healthcare to person Y over person X (given that person X has already deemed that the negative externality of the health state was worth the process, and even paid for the privilege).

I have very much simplified my example, and arguably not said that much extra than what you have already mentioned. However, if policy decisions (and those tools that aid such decisions e.g. economic evaluations) were changed to focus on process as well as outcome then this could effectively represent a healthcare system that promotes self-care and uses revealed preference rather than stated preference for decision making (a persons choice to smoke in spite of the health risks are essentially revealed preference [ignoring risk adversity and intertemporal choice], which is better than stated preference as currently used). In these types of discussions you need to keep in mind that decisions that account for process suggest that if person X reaches health state A, but then chooses to quit would also be weighted greater than someone who still refuses to quit, and if they then choose to smoke again after treatment they are weighted less again. Such policy and evaluation could work as a deterrent as well as a method of resource allocation. In this case (and i think i have rushed my discussion to really be that convincing) allocation based on process (i.e. smoking and eating to excess) can improve efficiency and deal with issues of equity that are important to society.

Bren Collins
12 years ago

Most NHS areas in England ration bariatric surgery and have more restrictive criteria than those recommended by NICE, despite NICE’s modelling suggesting that it is a cost-effective intervention. But if individuals who are denied bariatric surgery get other diseases linked to being obese like diabetes, cancer or CVD then treatments for these diseases will seldom be rationed in the same way. I think this is because there is a clear relationship between individual behaviour, obesity and the need for bariatric surgery that makes it an unpalatable intervention even if it is cost-effective in the long term. And also because the NHS needs to make savings it is easier to make them where people can be blamed for their health problems, when in fact everyone lives on a continuum of behaviour where many people might eat a little too much, drink a bit too much alcohol or not exercise enough, so unless there can be some kind of ‘blame coefficient’ for how much we are all to blame for our own health then trying to make these kind of value judgements about other people is very dangerous.

12 years ago

An argument from the venerable Peter Singer on why we should (ethically) encourage policies that discourage weight gain:–pay-more

Join the conversation, add a commentx