I am a big fan of QALYs, and not just because you get to play exciting games of Things That Suck Worse Than Depression Scavenger Hunt. (So far the list includes literally being born without a brain.) I think they’re a great first attempt at actually quantifying utilitarianism and solving the interpersonal utility comparison problem.
A lot of my disabled friends really hate QALYs. They say: “QALYs are literally saying that my life is worth less than an abled person’s. This isn’t theoretical: people have advocated distributing organs on the basis of QALY maximization. Sure, I’d like a cure for my chronic pain issues, but I don’t want to die.”
Worse, some of my utilitarian friends have agreed with them.
(Side note: talking about organ allocation is fairly ridiculous, given that about four-fifths of people on the organ transplant waiting list are waiting for kidneys, and it is possible for living people to donate kidneys. Instead of encouraging people to donate kidneys to strangers, the US medical system inexplicably makes it difficult. Also, if you’re in good health, you should donate a kidney.)
QALYs take two things into account: first, the number of additional years you will live if you have a treatment; second, how much better your life will be if you have that treatment. That is perfectly sensible if you are trying to figure out whether it’s cost-effective to cover my mental health care, because the effects of my mental health care are that I live longer and have a better life. However, this is only a good method of figuring out who should get an organ if you believe that disability is the most important factor in someone’s happiness, to the point that it completely swamps every other consideration.
Ideally, we’d look at people’s happiness setpoints. Unfortunately, we don’t have a way to measure happiness setpoint objectively, and if saying they are extremely happy is what it takes to get an organ most people will say so. However, there are demographic factors correlated with happiness: we should not only favor abled people, but married people, people who attend religious services regularly, Republicans, and rich people. Furthermore, people who are well-rested after only six hours a night have an extra eighth of a life-year per year compared to those who require ten; how much you sleep is just as important as, say, mild intellectual disability.
I notice that no one has invented Sleep-And-Political-Party-Adjusted-Life-Years, much less advocated their use in organ allocation. I feel like this is probably because disability feels very mediciney, the sort of thing doctors should be concerned about, whereas religiosity is clearly out of their sphere.
I suspect my utilitarian readers are half going “okay, you’re right, sleep-and-political-party-adjusted-life-years are kind of absurd”, and half going “actually, that sounds GREAT, exactly how we should allocate kidneys.” However, QALYs do not pass the enemy control ray test.
Non-utilitarians mostly are interested in allocating organs “fairly”, so a proper analogy will involve some people acting quite out of character. So imagine that the Catholic Church, in a sudden fit of consequentialism, invented the Telos-Adjusted Life Year: in the future, you will be far less likely to get an organ from a Catholic hospital if you’re in a gay relationship, you’ve gotten divorced, or your medical history includes contraception use. Both utilitarians and AU Catholics are using the same rule: we’re giving organs to people whose lives more closely fit our idea of the good life. It’s just that utilitarians think the good life is being happy or having your preferences satisfied, whereas AU Catholics think that the good life involves using your passions in a way compatible with their final end. If we want a leg to stand on when condemning the use of Telos-Adjusted Life Years to distribute organs, we shouldn’t use QALYs to distribute them either.
I would like to propose the radical idea that philosophers should figure out the definition of the good life, and the medical system should cure sick people. We wouldn’t give Derek Parfit a scalpel and shove him into the operating theater, and we shouldn’t ask the medical bureaucracy to solve all of normative ethics for us.
Shiggity said:
If you’re limiting the discussion to cost effectiveness, what about time-discounted expected future medical expenses? Disablities are then neutral, except insofar as they cost the system money, which in allocation of scare resources is both fair and compassionate, even if non-consequentialistd get their fuzzies in a frazzle when you try to save the most lives in a way they don’t like.
But those people are already probably opposed to QALYs anyway.
LikeLiked by 1 person
Nisan said:
Is there any method of allocating health care that passes the enemy control ray test?
LikeLiked by 4 people
wireheadwannabe said:
I would imagine that going purely by expected years of life added would be better than other options even if it isn’t perfect.
LikeLike
Decius said:
“Distribute scarce resources so as to maximize overall goodness”.
When we disagree about what is good, the mind-control ray stops working.
LikeLike
Vadim Kosoy said:
I suggest that each organ donor will be able to determine the criteria according to which their organs will be distributed. In addition, the system will automatically trade organs between donors in order to preserve efficiency.
LikeLike
Shea Levy said:
“I would like to propose the radical idea that philosophers should figure out the definition of the good life, and the medical system should cure sick people.” This completely ignores the problem? Resources are limited, as things are the medical system can’t cure all sick people, so which sick people should be cured?
LikeLiked by 8 people
Andrew said:
This is the comment I came here to write. Assuming we have 1,000 kidneys and 10,000 people who need kidneys who gets them? Also, why talking about just kidneys? Live liver donation is also an option and 40,000 people in the US die every year from a lack of donations. (http://www.mayoclinic.org/departments-centers/transplant-center/liver-transplant/choosing-mayo-clinic/expertise-innovation-research/living-donor)
LikeLike
Josh Morrison said:
40,000? That sounds high (there’s only 15K people waiting for livers).
Liver donation is way more burdensome than kidney donation (recovery times are much longer; complication rates are way higher).
LikeLiked by 1 person
pomeron said:
There are two different problems here, the moral problem and the technical problem. The moral problem is determining which utility function to maximize. This problem Ozy wants to delegate to philosopheres and I want to delegate to the organ donors. The technical problem is how to maximize the utility function (more generally how to arrive at the Nash bargaining solution). This problem ahould remain within the medical system.
LikeLike
Josh Morrison said:
Another problematic feature of allocating organs by QALYs is it would direct organs more to whiter, richer, better-educated people (who tend to be healthier and thus more “able to use” the organ). I find this to be a powerful thought pump towards some form of egalitarianism (since it seems especially fucked up that a worse-off group could be additionally punished on the grounds that “utility told me to”).
Also, thanks for the kidney donation shoutout. Not for everyone but I recommend it!
http://www.waitlistzero.org
LikeLiked by 1 person
Decius said:
It’s not obvious why the delta-QALYs would be higher for a healthier recipient. Surely the biggest factor in expected change in QALYs is rejection risk?
LikeLike
Josh Morrison said:
For any individual patient, the closeness of a match is important. That doesn’t mean baseline health doesn’t matter. For example, giving an organ to a 75 year old recipient will likely yield fewer QALYs than giving to someone who’s 25, because the 25 year old has fewer health risks besides the failed organ. Similarly, if rich people live longer than poor people, saving the life of the rich person will yield more QALYs than saving a poor person.
For me that’s a place where utilitarian principles break down (redistributing resources to the best off just because it yields more utility seems wrong to me).
LikeLike
Decius said:
How many QALYs per year does a poor person contribute?
LikeLike
blacktrance said:
If everyone’s utility is of equal value and giving kidneys to rich white people generates more of it, then that’s the egalitarian thing to do. It’s less egalitarian if some people matter disproportionately because of their membership in a group.
LikeLiked by 2 people
Toggle said:
My first instinct was to tut-tut your heuristic of “reject methods that benefit wealthy educated white people” (because social justice raaar), but on consideration I can think of some fairly strong arguments in your favor.
In particular, prioritizing aid by QALY is going to end up amplifying those inequalities in the same way that money tends to flow towards the rich. It’s a system that works on a power law, and therefore tends to artificially create utility monsters from populations that started with only minor differences in expected outcome. After all, now that this person has a working kidney, they can make even better use of their time…
As we’ve seen in many other domains, centralizing influence like that tends to have a destabilizing impact on the system as a whole, at least past a certain point. If we want to avoid a brittle society, it is probably worth our time to maintain a fairly gentle QALY gradient.
LikeLike
Zakharov said:
You can allocate organs to rich people without the obvious moral quandary by simply selling the organs, with the added benefit of greatly increasing the organ supply. Of course, this does create other issues.
LikeLike
Hedonic Treader said:
Richer people would also be willing and able to pay more for organs than poorer people. Surely there is some utilitarian potential in that fact. It’s just mostly blocked by law and society because it’s a taboo tradeoff.
Generally, under the model that people respond to incentives more than we would ideally want them to, some inequality is justified under consequentialism.
LikeLike
Decius said:
Isn’t saying “My life is less valuable because I have depression” isomorphic to “My depression is part of my life and has negative value”?
LikeLike
tailcalled said:
Only if you’re consequentialist.
LikeLike
Decius said:
Can non-consequentialists talk about QALYs?
LikeLike
tailcalled said:
Yes, but I doubt they care about them as much.
LikeLike
Vamair said:
I wonder if QALY is a powerful anti-abortion argument.
LikeLike
Lambert said:
A pro-choice argument is that abortions should take place when the negative effect on the mother’s quality of life due to the child is greater than the number of qalys of the child.
Also, pro-life arguments from this standpoint need to address why abortions are bad but not conceiving whenever possible is not.
LikeLiked by 1 person
tailcalled said:
I think Vamair’s point is that under naive QALY measurement, abortion is literally murder. This, of course, depends on you caring about the QALYs of persons that do not yet exist, which would imply the repugnant conclusion.
LikeLike
Vamair said:
Right. And unless we’re dealing with an unusual case, under naive QALY, that’s much worse than almost any other murder. The fetus has a whole life before them, years of childhood have nonzero quality, so any other person’s life is worth less. Even if we don’t care about the QALY of nonexistant people, it’s still not clear we shoudn’t care about a fetus, as the fetus already exists as a human organism (even if not as a person).
I believe there probably should be something like an additional “productivity quotient”, where productivity is thought of as an ability to improve the lives of the other people. But I doubt we can calculate it, especially without guilt-tripping people in process.
LikeLike
Creutzer said:
If a toddler, let alone a fetus, is more valuable than an adult, this looks like a reductio ad absurdum of QALYs.
LikeLike
Lambert said:
But then a sperm & an egg are more valuable than a fœtus at 30 weeks. This makes not having a child literally worse than murder. Celibacy is as bad as killing a baby someone every 9 months.
LikeLike
Vamair said:
Creuzer, this was intended as reductio ad absurdum to point out some of the inadequacies inherent in the concept. Or our moral intuitions. Probably the concept.
Lambert, sperm and egg – probably not. At least we can pull out a trick and say that before conception they’re genetically not a human organism, so there’s nothing to calculate QALYs for. Zygote, on the other hand…
LikeLike
Hedonic Treader said:
Tailcat said:
“This, of course, depends on you caring about the QALYs of persons that do not yet exist, which would imply the repugnant conclusion.”
There are some misconceptions about the Repugnant Conclusion that are worth addressing. The Repugnant Conclusion doesn’t just follow from “potential people matter”, it requires very specific assumptions:
1) Potential people matter and
2) the marginal utility of additional lives + their externalities will be net-positive and
3) this will remain true until the average quality of life is “barely worth living”
Note that, even though it is often claimed, the Repugnant Conclusion does *not* imply maximizing population, or that the resulting population should be at or near subsistence. For example, “barely worth living” could be significantly above subsistence because certain inevitable sufferings are bad enough to subtract the utility.
The Repugnant Conclusion also doesn’t imply real-world natalism. For example, one can accept it but also believe that funding technological innovation (e.g. invent hedonium) is more important right now than having more children. Or one can technically accept it but be a pessimist about human quality of live altogether.
For those who are interested, there is a good paper online called “In Defence of Repugnance” by Michael Huemer on the Repugant Conclusion, including a chapter about caveats and how it does not necessarily map onto real world population policy.
LikeLike
Decius said:
It is an equally powerful anit-birth control argument…
LikeLike
J. Goard said:
And an argument against doing almost anything else whenever reproductive sex could be had. Unless one of the three position is accepted:
(a) consequentialism is invalid or not absolutely correct;
(b) lives can have negative value (and, in fact, it must be fairly likely that this can be the case, under predictable circumstances)
(c) value is not additive.
LikeLike
Decius said:
(b1) a potential human can cause more harm than good.
LikeLike
Zykrom said:
The Catholics Are Right
LikeLike
Hedonic Treader said:
J. Goard and Decius both make important points, which show that total utilitarianism does not imply natalism, unless very specific assumptions are added:
1) the additional lives + their externalities have net-positive utility and
2) it just so happens that this will be reached at the maximum possible human population (to imply strong natalism)
Weaker natalism only requires that the marginal utility of additional humans is net-positive at the moment. I think the most likely realistic candidate is selective weak natalism: The marginal utility of *some identifiable subset* of additional people is net-positive (i.e. some people should have more children, but not others).
But consider: no utilitarian discussion of natalism is complete without the opportunity cost of kids (http://reducing-suffering.org/the-cost-of-kids/).
LikeLiked by 1 person
Vamair said:
I would like to provide a toy example. Let’s imagine a person’s utility U is the person’s resources R minus 1000 (the quality needed for a life worth living, is the same as 1, as we don’t use any units, but looks bigger). For any person with U = 5000 there may be 5001 people with U = 1, so if we’re deciding which population (along with their resources) to let live in a case of a catastrophe, the 5001 one is going to be the survivors. But the most efficient allocation of the resources R is to create only a single immortal person (or, more closely to the real world, a smallest population big enough to survive indefinitely) and to give them all the resources. I.e. antinatalism. The real person’s utility grows not as fast as a linear function, so the situation is different, but to illustrate the fact that the repugnant conclusion doesn’t imply natalism, this toy example is enough.
LikeLike
Henry Gorman said:
An elegant way to deal with some of the utilitarian arguments against contraception here– assume that only people who actually will exist in the future matter. If your course of action forecloses their existence (ie: if you get an abortion or use birth control), then their utility does not matter. If future people’s existence is uncertain, we could roughly weight it by the probability that they will come into existence.
LikeLiked by 1 person
Lambert said:
The question then is: does a fœtus exist, in the sense above?
LikeLike
Vamair said:
A utility oscillator
Let’s say there is a scientist that wants (utility 1) to create a monster and a future monster-person that would prefer not to exist (utility of existence -2). Now the scientist is a utilitarian and is calculating the utility. Utility of a 0%-chance monster-person is 0, so they’re free to act on their desire to create the monster. But then as the probability rises, the total utility drops, so the scientist starts to have doubts. Until the probability is 50% and rising, when they start trying to stop the creation of the monster. But when the probability of monster is zero again, the scientist can start their project anew.
LikeLiked by 1 person
tailcalled said:
tailcalled said:
Whoops, apparently the comment system removes things between lt/gt signs. The comment above should read:
Obligatory comment pointing out that before you start optimizing the distribution of transplants using consequentialism, there are a lot of other things about the medical system that can be optimized for a much, much greater increase in utility.
LikeLike
Merve said:
I know we’re largely speaking in hypotheticals here, but nonetheless, before anyone gets too alarmed, I should point out that in general, physicians don’t make decisions on the basis of QALYs (or DALYs or what-have-you). It’s hospital administrators and health economists who use QALYs, usually to allocate resources across units or types of treatments. To my knowledge, nobody actually makes organ allocation decisions between two patients on the basis of QALYs – or at least, they don’t do so explicitly. Whether or not they should is another matter entirely.
LikeLiked by 1 person
Ghatanathoah said:
I wonder if you could beat the enemy control ray test by changing the wording to “Allocate organs to people whose lives most close match their own personal idea of a good life,” rather than our idea of a good life. That is what separates most forms of utilitarianism from other ideologies, they take what other people want in consideration, instead of considering there to be an “objective good life.” (I’ve read some Catholic thinkers actually consider this to be moral relativism, but I don’t see it as such since it still has objective standards for interpersonal morality). Hopefully building that into the directive, rather than leaving the definition of “good life” nebulous, would beat the ray.
Or would the enemy control ray filter that directive into something else?
Another factor to consider before we go down the road of favoring Republicans and churchgoers is that most people recognize a difference between happiness and goodness, and that some forms of happiness are of zero or negative value (such as happiness derived from believing in falsehoods). I think we can safely say that not all religions can be true at once, so that means the majority of people deriving happiness from religion are deriving happiness from belief in a falsehood, so their lives are considerably worse than they think they are. The same thing is probably true about happiness derived from having some political beliefs.
Needless to say, attempting to take those factors into account would really, really, really fail the enemy control ray test.
LikeLike
Ghatanathoah said:
QALYs are saying that abled and disabled peoples lives are worth exactly the same: Zero. The whole point of QALYs is that it’s not your life that’s valuable. It’s the things you want to do and achieve with your life that’s valuable. And therefore, we should allocate resources in such a way as to maximize people’s ability to do and achieve things with their lives. And that in turn means that if there’s some additional factor preventing your from doing things with your life besides lack of medical resources, it’s not an efficient use of resources to give them to you.
Of course, this doesn’t always imply we should favor the abled. A wheelchair using person whose greatest ambition is to be a mathematician is probably just as able to achieve their goal, if not more so, than an able person who wants to be a pro athlete.
LikeLike
Lambert said:
The problem lies not with QALYs, but that the quality adjustments are inaccurate, taking disability in to account too much.
LikeLike
Joe said:
The enemy control ray argument seems like a bad one to me. I’m going to attempt a reductio ad absurdum.
Here on Earth, doctors agree that health interventions should prevent diseases instead of cause them. So we think it’s acceptable for doctors to stick needles in people in order to vaccinate them against disease. However, on Threa, all the doctors agree that health interventions should *cause* diseases. In Threa, when doctors stick needles in people, it’s to make them sick. So if Earth doctors say “It’s morally OK to stick needles in people in order to prevent disease”, Threa doctors hear “It’s morally OK to stick needles in people in order to cause disease”. Therefore, it’s not morally OK to stick needles in people in order to prevent disease, by the Enemy Control Ray.
The enemy control ray argument might be a good way to reason about the *rhetorical* implications of what you say. Yes, in the real world, if you spread the idea that non-egalitarian distributions of health care are OK, Catholics are liable to seize on to that and promote some other agenda (they probably wouldn’t). But if we’re *actually trying to figure out what we want*, I don’t see why Catholic misinterpretations of our conversation should come in to play. And I’d guess that most readers of this blog would e.g. agree that e.g. saving young peoples’ lives should be prioritized over saving old peoples’ lives if there’s a tradeoff.
So the Enemy Control Ray is a good way to talk to your enemies but not a good way to figure out what you want. It might be a good way to craft laws, but not morality.
(FWIW, as an ex-disabled-person, if you asked disabled-me about allocating him less health care due to disability, I think he would have assented to that.)
LikeLiked by 3 people
Henry Gorman said:
In some ways, distributing more health care to the rich and educated because their lives will have more quality would be very naive utilitarianism. It might maximize the outcomes of health care, but there’s more to the world’s utility function than healthcare outcomes. Massively amplifying existing social equalities through healthcare seems like something that would have really important disutilities, since feeling like a member of an underclass makes people miserable.
LikeLike
Autolykos said:
I think the best internal argument against allocating more resources to the happiness of the rich is diminishing returns. Improving the lives of the poor is usually way cheaper, because there is more space above them.
If someone is already 95% happy, there are only 5% to go, and most of that may be very hard to reach (say, a pet tiger, a giant yacht and a trip to the moon), while someone who is only 40% happy may be a lot easier to bring up to 45% or more (e.g. by a hot meal every day). Utilitarianism based on QALYs may lead to counter-intuitive results in some cases, but they do not have to lead to utility monsters and other failure conditions. It also depends a lot on whether the effects of reduced QALY from multiple sources are additive, multiplicative or even more complex (I bet on the third).
LikeLiked by 4 people
Lambert said:
Only vaguely on topic, but I can’t seem to find any tables of actual QALY values anywhere. Does anyone know a good source for quality adjustments, as used by the medical community?
LikeLike
Paperclip Minimizer said:
Hey Ozy, did you changed your mind on this ? Because this does not seem to be very coherent with the thesis of Deontologist Envy.
LikeLike