I think the key point of contention is contained here:
"The marginal unit of health care doesn’t come clearly marked. If we want to cut the marginal unit of health care (for example, following Robin’s recommendation to cut health care in half) we need to cut specific things. If you would otherwise get ten treatments in a year, you need to cut out five if you want to halve health care like Robin suggests. Which five? You could make the decision centrally (the medical establishment decides some interventions are less valuable than others, and insurance stops covering those) or in a decentralized free-market way (customers get less insurance, increasing the cost of medical care and causing them to make harder trade-offs about when to get it), but somebody has to make this decision at some point. On what basis do they make it?"
Scott is essentially saying "we can't identify which 5 treatments fail the cost/benefit analysis" which is true at the aggregate, nationwide level, but I think you are saying "but patients CAN sufficiently identify, for themselves, at least on average, which 5 treatments are no longer beneficial to them after the cost has increased due to that "customers get less insurance" change."
I think it possible to debate whether or not the patient's assessment is accurate, or if forcing them to make that decision is fair, but I (humbly) think that is the specific issue of disagreement.
But we know a priori the that interventions targeted at young people with terminal illnesses have a chance of added decades to life expectancy, but all other treatments add AT MOST a few years.
And in particular, we know that interventions targeting young people with diseases that are not fatal have a HUGE chance of doing more harm than good, since the harm could be up to decades.
And of course, interventions on old people can't possibly be doing more than a few months of harm or good to life expectancy since those people will die imminently no matter what.
Proponents of a treatment will argue, "This can save your life!" But how much life? Are you going to live for ever? No. But if you're a child, you might have seven, eight, or more decades left. But if you're 80, you have years or even just months left.
It works both ways--for a child, a damaging treatment harms or reduces more decades of their life. This was the insanity of mandating covid vaccines for teenagers who had a greater chance of being damaged by the vaccine, than being damaged by covid (see Dr. Vinay Prasad)
oh man I have so many jokes here, I've been waiting for this day... hold on.... something about how British people disproportionately die of cancers of the prostate and breast because they get too uncomfortable bringing it up to their GPs.
I think there's also a big factor that Scott practices medicine himself. So if someone says "half of medicine is a waste!" then he only thinks about the part he practices, which is probably not 50% waste. But I work in the part of "medicine" that isn't practicing medicine, and I will tell you: billions of dollars are spent that are labeled medicine but are not practicing medicine.
If you ARE a doctor, but you need to hire 10 new medical billing coders to comply with insurance billing or ICD10 codes or whatever, then you are gonna try to raise the price of the medicine you practice by 500k a year. Economists will come by and say "my, we now spend 500k more a year on medicine, and yet outcomes have not increased one bit, what a puzzler."
You can argue that the people collecting the data have figured out how to segment that out, but I will laugh at you for doing so.
We also mean two very different things when we say "medicine works."
There's also a ton of cost increases in the "quality" of medicine provided to me that do not make me healthier. Take the auto-injectors that some GLP1 drugs have. Maybe I as a customer love them, because I'm terrified of needles, and so they give me 1000$ worth of benefit. I might pay quite a bit more for that drug + auto-injector and consider myself better off, but it ain't going to show up in health outcomes. And the auto injectors are relatively expensive, and scarcity of building them has led to shortages of the drug package, even though the drug chemical is not in a shortage.
@Robin it seemed to me that the most important part of Scott's post for purposes of advancing the discussion, which you don't directly address, is here:
"So in the interests of getting a clearer understanding, I’ll pose Robin a trilemma:
1. Either we can’t distinguish between good and bad medical interventions, but the average intervention is net positive in expectation (in which case it seems like we should keep the amount of medicine we have now, since we assess each treatment equally and they’re all net positive)
2. Or we still can’t distinguish between good and bad medical interventions, but the average intervention is, after you count the monetary cost, net neutral or negative in expectation (in which case one should be equally skeptical of everything, including antibiotics and cancer treatment, and I don’t understand how saying this is a straw man)
3. Or we can distinguish between good and bad medical interventions, and we should throw out the bad ones and keep the good ones (in which case why does Robin keep saying the opposite, why does he call this a “monkey trap”, etc? And wouldn’t it be better for Robin to frame his position as “medicine generally works well, but there are some interventions that aren’t evidence-based enough”, which is the consensus medical position?)
If this is a false trichotomy, Robin should tell me how!"
Which of these options do you endorse, or do you think this is a false trichotomy? Based on this post I'd guess you endorse (2)? But I'm not sure.
#3, but. We have weak indicators of treatment quality which are sufficient for us to usefully cut medicine in half. The "consensus medical position" isn't willing to identify enough treatments as questionable to do this. And this doesn't imply the rest work "well."
"Again, if Robin’s claim is that medicine is only useless on the margins, we’re much closer to agreement. But I don’t know how that meshes with saying that maybe antibiotics don’t help, or that we can’t possibly distinguish marginal from core, or that health spending is mostly signaling (as opposed to a mix of people correctly spending money on health because they know it’s great and will help them, plus some extra from people not being scientists and not knowing which treatments are good or bad)."
Could you remind us why you think signaling theory needs to be invoked to explain medical overspending, as opposed to simple ignorance?
Simple ignorance is less plausible as a cause of trends that last longer and in more places. We need a reason for the persistence of that ignorance in the face of attempts to correct it.
> While I feel this quote is pretty clear, I also agree with Scott that he isn’t the only person to misunderstand me.
For what it's worth, I correctly predicted in [a comment on his original post](https://www.astralcodexten.com/p/contra-hanson-on-medical-effectiveness/comment/54674466) that you would write a response claiming that he misunderstood you for essentially the exact reasons you gave. So, it seems that at least one person (myself) interpreted your posts and papers correctly. This provides some evidence that it is the reader's fault for the misunderstanding, not the author's.
"Those facing smaller budgets consume fewer drugs and die more: mortality increases 0.0164 percentage points per month (13.9%) for each |${\$}$|100 per month budget decrease (24.4%). This estimate is robust to a range of falsification checks, and lies in the 97.8th percentile of 544 placebo estimates from similar populations that lack the same idiosyncratic budget policy."
I was one of your students in your health economics class at GMU decades ago. Count me in among the many people who have previously misunderstood your position. After your class, I became and remained *deeply* skeptical of almost every medical intervention. The only exceptions I made were instances where the mechanism of action were very well understood (similar to eyeglasses in the RAND study).
You suggest that you think a lot of medical spending is inefficient, and you suggest removing subsidies as a way of reducing spending. I get the public choice thinking that leads you to believe that spending will go down.
But I suggest that this is wrong. If subsidies were removed, spending will in fact go up, and will become less efficient.
The reason is that a reduction in government subsidies, all else being equal, will put that money in the hands of patients and/or insurance companies. There is little reason to imagine that either will demand less healthcare. Patients have little medical knowledge, and their medical spending will probably be less efficient than subsidised treatments. Insurance companies will not be seeing a major change, so I don't expect their behaviour to change.
In other countries, e.g. the UK, there is proportionally more government spending on healthcare, but total spending is lower. This evidence also suggests that government spending is not what is driving the high costs, and that a smaller role for government might have the effect of increasing total spending.
So... even if you're right about the problem, the solution just seems to be backwards.
Right. If subsidies were removed and people had to pay themselves, a greater percentage of medical costs would be rich people paying for treatment they don’t need and less would be poor people getting basic care. That would also shift the focus of the medical profession to providing feel-good therapy for wealthy people.
Of the money he's cutting from healthcare, 100% by definition is currently spent on healthcare. If people choose to spend 90% of that on healthcare once it's returned to them, that will be a cut of 10%. Hanson has listed nationalizing healthcare and having an agency control the spending as one possible approach though.
Remember those studies he likes where giving free health insurance doesn't lead to people using much more healthcare? They cut both ways. If giving people free stuff doesn't make them use much more, then there is no reason to believe that reducing subsidies would make anyone use less.
I understand the theoretical model that Hanson, and I guess you, are using, where reducing subsidies changes the public choice balance and automatically leads to less spending. But it is only a *theory*. It's a theory that I don't believe, and the empirical evidence adduced by Hanson contradicts.
My takeaway from the numbered list of mechanisms for selecting mechanism (which is probably my takeaway likely because it reinforces something I already believe) is that we shouldn't have a centralized actor deciding what medicine is "worth it" or not, especially if that centralized actor doesn't have a robust incentive structure that aligns with individual desires.
I think this is a classic case of Goodhart's Law, where as soon as you pick some metric for everyone in the world to follow for "what medicine is worth it", Pharma will, in short order, adjust their output such that they maximize that metric. Cochrane Review is a reasonable strategy right now because there is minimal incentive to corrupt it. If the FDA starts blindly following Cochrane Review I suspect that within a few years pharma execs would be rotating into/outof Cochrane management positions and there would be a whole lot of fine dining meet-ups with Cochrane members in attendance. Same if 90% of doctors focus on Cochrane Review recommendations.
For the others, they all work because there is no explicit/specific metric one can easily corrupt/optimize for. When comparing small vs big hospitals, it is non-trivial to corrupt all of the small hospitals in the world. Similar for asking your doctor for confidence intervals, asking your friends, etc. These are all fairly decentralized solutions, which makes them incredibly costly to corrupt.
Thought: If the helpful medicine is consumed by people who are paying attention, and the harmful medicine is consumed by people who are not paying attention, then arguing for reducing marginal medicine (which covers both helpful and harmful medicine) means that you are arguing for something that is harmful to people who are paying attention and helpful to people who aren't paying attention. This is going to make enemies with people who are paying attention, and it's not going to make friends with people who aren't paying attention (because they're not paying attention), so it's only a feasible policy to implement for people who can afford to oppose a large coalition of attentive people with no help from anyone else, which is ~nobody.
I thought you were headed down the path of marginal cost, which is problematic because every self-payer has a different healthcare price point based on their income, disposable income, etc. Most people can't afford some of the most basic care. Glad you went in other directions but are those directions better?
You seem most focused on the small hospital determinant so let's look at that. I guess it very much depends on what you define as a "small hospital" but how many procedures can we list that small hospitals don't do but have a clear benefit? I would bet the list is long but for the moment, let's just look at organ transplants. Maybe some small hospitals do kidney transplants, IDK, but surely most types of organ transplants only happen at "large" hospitals. Should we stop doing these? Are they unsuccessful? Should I ask if small hospitals do them before I replace my failing liver, heart, lungs, etc.?
It's also important to look at another aspect for which transplants are an illustrative if not also good example. Transplants had an awful record before cyclosporine but even after it improved long-term survival, the surgeries still had high failure rates, depending on the organ. That has improved. Would that have happened if we decided these surgeries were "marginal," which is what the data showed?
If you said cut medicine by 95% it would be easy to figure out what that best 5% is.
And in all probability this would be nearly optimal policy. After all, the best 5% of modern medicine is very different than going back to the 19th century or whatever.
It is not just *marginal medical treatment* that is somewhat obscure. So is *marginal non-medical-treatment*, viz., whatever would be purchased instead if a little less medical treatment were purchased, or whatever would not be purchased to allow a little more medical treatment to be purchased. The marginal effect of medical treatment overall can be thought of as the marginal direct effect of medical treatment *minus* the marginal direct effect of non-medical-treatment. The latter effect on health may be significant, even though the purchase or non-purchase is labeled “non-medical.”
True.. I meant to say that if one is worried by the difficulty in identifying *marginal medical treatment*, one should be equally worried about the similar difficulty in identifying *marginal non-medical-treatment*.
Given Hanson's celebrated disregard of social desirability, I kept reading this back and forth waiting to see some simple statement in the following vein:
"Lol don't give expensive treatments to old fat very sick people."
Don't the majority of medical costs come in the last six months of life? Americans in particular are often physically rotten by old age, full of stints and statins, plaques in the arteries and plaques in the brain. Tangles, too!
I've heard from old dying men that their doctors' standard of care was just a determination that the problems caused by some other specialist would kill him first - cancer doc hopes the heart will get him, or the kidneys, etc.
The 'death panels' of the original obamacare debate would solve this. What's that? You've had 3 heart surgeries and need a fourth? No, we're putting this needle in your arm, goodnight sweet prince.
To move the goalposts technically but not in spirit, I wonder what percentage is in the last 5 years of life? It doesn't matter to my point whether we're talking about the very old/sick who are going to die this year or the very old/sick who are going to die merely rather soon. In terms of life-years that's still where the fat is to be trimmed.
Certainly if the debate is "does medicine provide a good value in predicted life-years for the cost?" then your best bang for the buck will be cutting spending on extremely unhealthy old people. Scott says, truthfully, that we aren't great at assessing which treatments "work" - but actually we do a very good job of assessing which treatments are likely to be very costly and not provide a great deal of extra life to which patients, based on their current age and health status. If it's "does medicine work" then we need to talk about the definition of "work."
Re: old age, there's also so much more you need to think about on both sides rather than just "how much $" and "how many more years". The data that Scott and Robin are debating have no entries for "we spent 10,000$ on a treatment that actually caused the patient to die 10 days earlier, but wasn't in desperate pain the whole time and could say goodbye to their family" Or "we spent 30,000$ on a treatment that meant the patient lived 10 days longer but they were in excruciating pain the entire time."
At the highest level of abstraction, that medical payments do not seem to improve medical outcomes is merely step 1 that only inevitably leads to step 2 of "therefore we should spend less on things that are labeled medicine." Robin says that we can therefore cut medical spending at the top, and let people decide to do with that new cost-benefit analysis, which may be true, but the fact is that despite what Scott says, we *can* assess the effectiveness of billions of $ of medical (and "medical") costs vs the benefits they provide, not only in life years, but quality of life, pain and many others. Plucking the low hanging fruit there can provide the real world evidence to justify reallocating resources to better things.
Even a fifth is a lot. Does the article distinguish "last year treatment" between "treatments we give lots of people with the usual result and expectation of adding lots of healthy years, but with some people dying soon anyway" and "treatments we give with high expectation the patient will still die soon"?
In the first case, its arithmetically possible we could spend half of all health spending on that, with 40% of those patients dying soon but 60% living many healthy additional years, so perhaps worth it. The second case is much harder to justify, and there's a massive margin that if cut wouldn't affect average life span.
One could consider the pressure in the other direction. The decision of which treatments to pay for can become very political and health bureaucrats responsible for making these calls sometimes complained about public, political, or media pressure to make exceptions to the usual cost- benefit criteria and approve treatments that didn't make the cut because regardless of the little bang for the buck the patients or disease type aroused greater than typical levels of sympathy and support. That motivated proposals to provide some political insulation to entities like NICE / IPAB. Each such treatment operated in the manner of Olson's "concentrated vs diffuse interests" and "special interest" treatment exceptions are individually small but aggregate into massive costs. This is also what Effective Altruists complain about when they criticize charitable giving to popular or sympathetic causes where the money could save many more lives if more efficiently allocated.
Now consider that whatever system eventually succumbed to these kinds of pressures and granted most of the too-costly-but-popular treatments, spending a lot of extra money but with little to show for it. Now imagine realizing that a ton could be cut without affecting average outcomes much, and then trying to run this process in reverse. Every single thing you can mention with specific identification is going to make you the target of a pile-on calling you a heartless monster who doesn't care about killing the members of some very sympathetic population.
Personally I have a lot of sympathy for someone put on the spot to make such specific identifications and can understand the need to keep discourse at a more general level.
Scott Alexander is a very honest dealer. In perhaps literally anyone else I would have regarded introducing, as he did, a $300 helmet for an infant with possibly 75 years ahead of him into a conversation about medical costs utterly disingenuous. No, it's the old fat sick people, Scott, ffs.
I think the key point of contention is contained here:
"The marginal unit of health care doesn’t come clearly marked. If we want to cut the marginal unit of health care (for example, following Robin’s recommendation to cut health care in half) we need to cut specific things. If you would otherwise get ten treatments in a year, you need to cut out five if you want to halve health care like Robin suggests. Which five? You could make the decision centrally (the medical establishment decides some interventions are less valuable than others, and insurance stops covering those) or in a decentralized free-market way (customers get less insurance, increasing the cost of medical care and causing them to make harder trade-offs about when to get it), but somebody has to make this decision at some point. On what basis do they make it?"
Scott is essentially saying "we can't identify which 5 treatments fail the cost/benefit analysis" which is true at the aggregate, nationwide level, but I think you are saying "but patients CAN sufficiently identify, for themselves, at least on average, which 5 treatments are no longer beneficial to them after the cost has increased due to that "customers get less insurance" change."
I think it possible to debate whether or not the patient's assessment is accurate, or if forcing them to make that decision is fair, but I (humbly) think that is the specific issue of disagreement.
But we know a priori the that interventions targeted at young people with terminal illnesses have a chance of added decades to life expectancy, but all other treatments add AT MOST a few years.
And in particular, we know that interventions targeting young people with diseases that are not fatal have a HUGE chance of doing more harm than good, since the harm could be up to decades.
And of course, interventions on old people can't possibly be doing more than a few months of harm or good to life expectancy since those people will die imminently no matter what.
Proponents of a treatment will argue, "This can save your life!" But how much life? Are you going to live for ever? No. But if you're a child, you might have seven, eight, or more decades left. But if you're 80, you have years or even just months left.
It works both ways--for a child, a damaging treatment harms or reduces more decades of their life. This was the insanity of mandating covid vaccines for teenagers who had a greater chance of being damaged by the vaccine, than being damaged by covid (see Dr. Vinay Prasad)
Dr. Ezekiel J. Emanuel
What about Brits?
oh man I have so many jokes here, I've been waiting for this day... hold on.... something about how British people disproportionately die of cancers of the prostate and breast because they get too uncomfortable bringing it up to their GPs.
I think there's also a big factor that Scott practices medicine himself. So if someone says "half of medicine is a waste!" then he only thinks about the part he practices, which is probably not 50% waste. But I work in the part of "medicine" that isn't practicing medicine, and I will tell you: billions of dollars are spent that are labeled medicine but are not practicing medicine.
If you ARE a doctor, but you need to hire 10 new medical billing coders to comply with insurance billing or ICD10 codes or whatever, then you are gonna try to raise the price of the medicine you practice by 500k a year. Economists will come by and say "my, we now spend 500k more a year on medicine, and yet outcomes have not increased one bit, what a puzzler."
You can argue that the people collecting the data have figured out how to segment that out, but I will laugh at you for doing so.
We also mean two very different things when we say "medicine works."
There's also a ton of cost increases in the "quality" of medicine provided to me that do not make me healthier. Take the auto-injectors that some GLP1 drugs have. Maybe I as a customer love them, because I'm terrified of needles, and so they give me 1000$ worth of benefit. I might pay quite a bit more for that drug + auto-injector and consider myself better off, but it ain't going to show up in health outcomes. And the auto injectors are relatively expensive, and scarcity of building them has led to shortages of the drug package, even though the drug chemical is not in a shortage.
Who is he? Scott Alexander or Robin Hanson?
@Robin it seemed to me that the most important part of Scott's post for purposes of advancing the discussion, which you don't directly address, is here:
"So in the interests of getting a clearer understanding, I’ll pose Robin a trilemma:
1. Either we can’t distinguish between good and bad medical interventions, but the average intervention is net positive in expectation (in which case it seems like we should keep the amount of medicine we have now, since we assess each treatment equally and they’re all net positive)
2. Or we still can’t distinguish between good and bad medical interventions, but the average intervention is, after you count the monetary cost, net neutral or negative in expectation (in which case one should be equally skeptical of everything, including antibiotics and cancer treatment, and I don’t understand how saying this is a straw man)
3. Or we can distinguish between good and bad medical interventions, and we should throw out the bad ones and keep the good ones (in which case why does Robin keep saying the opposite, why does he call this a “monkey trap”, etc? And wouldn’t it be better for Robin to frame his position as “medicine generally works well, but there are some interventions that aren’t evidence-based enough”, which is the consensus medical position?)
If this is a false trichotomy, Robin should tell me how!"
Which of these options do you endorse, or do you think this is a false trichotomy? Based on this post I'd guess you endorse (2)? But I'm not sure.
#3, but. We have weak indicators of treatment quality which are sufficient for us to usefully cut medicine in half. The "consensus medical position" isn't willing to identify enough treatments as questionable to do this. And this doesn't imply the rest work "well."
It could be useful to edit this into the article.
Done
Ah, thank you! That does clarify a lot
"Again, if Robin’s claim is that medicine is only useless on the margins, we’re much closer to agreement. But I don’t know how that meshes with saying that maybe antibiotics don’t help, or that we can’t possibly distinguish marginal from core, or that health spending is mostly signaling (as opposed to a mix of people correctly spending money on health because they know it’s great and will help them, plus some extra from people not being scientists and not knowing which treatments are good or bad)."
Could you remind us why you think signaling theory needs to be invoked to explain medical overspending, as opposed to simple ignorance?
Simple ignorance is less plausible as a cause of trends that last longer and in more places. We need a reason for the persistence of that ignorance in the face of attempts to correct it.
> While I feel this quote is pretty clear, I also agree with Scott that he isn’t the only person to misunderstand me.
For what it's worth, I correctly predicted in [a comment on his original post](https://www.astralcodexten.com/p/contra-hanson-on-medical-effectiveness/comment/54674466) that you would write a response claiming that he misunderstood you for essentially the exact reasons you gave. So, it seems that at least one person (myself) interpreted your posts and papers correctly. This provides some evidence that it is the reader's fault for the misunderstanding, not the author's.
New paper just dropped:
https://academic.oup.com/qje/advance-article-abstract/doi/10.1093/qje/qjae015/7664375
"Those facing smaller budgets consume fewer drugs and die more: mortality increases 0.0164 percentage points per month (13.9%) for each |${\$}$|100 per month budget decrease (24.4%). This estimate is robust to a range of falsification checks, and lies in the 97.8th percentile of 544 placebo estimates from similar populations that lack the same idiosyncratic budget policy."
I was one of your students in your health economics class at GMU decades ago. Count me in among the many people who have previously misunderstood your position. After your class, I became and remained *deeply* skeptical of almost every medical intervention. The only exceptions I made were instances where the mechanism of action were very well understood (similar to eyeglasses in the RAND study).
I'll bite on this.
You suggest that you think a lot of medical spending is inefficient, and you suggest removing subsidies as a way of reducing spending. I get the public choice thinking that leads you to believe that spending will go down.
But I suggest that this is wrong. If subsidies were removed, spending will in fact go up, and will become less efficient.
The reason is that a reduction in government subsidies, all else being equal, will put that money in the hands of patients and/or insurance companies. There is little reason to imagine that either will demand less healthcare. Patients have little medical knowledge, and their medical spending will probably be less efficient than subsidised treatments. Insurance companies will not be seeing a major change, so I don't expect their behaviour to change.
In other countries, e.g. the UK, there is proportionally more government spending on healthcare, but total spending is lower. This evidence also suggests that government spending is not what is driving the high costs, and that a smaller role for government might have the effect of increasing total spending.
So... even if you're right about the problem, the solution just seems to be backwards.
Right. If subsidies were removed and people had to pay themselves, a greater percentage of medical costs would be rich people paying for treatment they don’t need and less would be poor people getting basic care. That would also shift the focus of the medical profession to providing feel-good therapy for wealthy people.
Of the money he's cutting from healthcare, 100% by definition is currently spent on healthcare. If people choose to spend 90% of that on healthcare once it's returned to them, that will be a cut of 10%. Hanson has listed nationalizing healthcare and having an agency control the spending as one possible approach though.
Remember those studies he likes where giving free health insurance doesn't lead to people using much more healthcare? They cut both ways. If giving people free stuff doesn't make them use much more, then there is no reason to believe that reducing subsidies would make anyone use less.
I understand the theoretical model that Hanson, and I guess you, are using, where reducing subsidies changes the public choice balance and automatically leads to less spending. But it is only a *theory*. It's a theory that I don't believe, and the empirical evidence adduced by Hanson contradicts.
No, he has emphasized that those studies find that utilization increases.
My takeaway from the numbered list of mechanisms for selecting mechanism (which is probably my takeaway likely because it reinforces something I already believe) is that we shouldn't have a centralized actor deciding what medicine is "worth it" or not, especially if that centralized actor doesn't have a robust incentive structure that aligns with individual desires.
I think this is a classic case of Goodhart's Law, where as soon as you pick some metric for everyone in the world to follow for "what medicine is worth it", Pharma will, in short order, adjust their output such that they maximize that metric. Cochrane Review is a reasonable strategy right now because there is minimal incentive to corrupt it. If the FDA starts blindly following Cochrane Review I suspect that within a few years pharma execs would be rotating into/outof Cochrane management positions and there would be a whole lot of fine dining meet-ups with Cochrane members in attendance. Same if 90% of doctors focus on Cochrane Review recommendations.
For the others, they all work because there is no explicit/specific metric one can easily corrupt/optimize for. When comparing small vs big hospitals, it is non-trivial to corrupt all of the small hospitals in the world. Similar for asking your doctor for confidence intervals, asking your friends, etc. These are all fairly decentralized solutions, which makes them incredibly costly to corrupt.
Thought: If the helpful medicine is consumed by people who are paying attention, and the harmful medicine is consumed by people who are not paying attention, then arguing for reducing marginal medicine (which covers both helpful and harmful medicine) means that you are arguing for something that is harmful to people who are paying attention and helpful to people who aren't paying attention. This is going to make enemies with people who are paying attention, and it's not going to make friends with people who aren't paying attention (because they're not paying attention), so it's only a feasible policy to implement for people who can afford to oppose a large coalition of attentive people with no help from anyone else, which is ~nobody.
I thought you were headed down the path of marginal cost, which is problematic because every self-payer has a different healthcare price point based on their income, disposable income, etc. Most people can't afford some of the most basic care. Glad you went in other directions but are those directions better?
You seem most focused on the small hospital determinant so let's look at that. I guess it very much depends on what you define as a "small hospital" but how many procedures can we list that small hospitals don't do but have a clear benefit? I would bet the list is long but for the moment, let's just look at organ transplants. Maybe some small hospitals do kidney transplants, IDK, but surely most types of organ transplants only happen at "large" hospitals. Should we stop doing these? Are they unsuccessful? Should I ask if small hospitals do them before I replace my failing liver, heart, lungs, etc.?
It's also important to look at another aspect for which transplants are an illustrative if not also good example. Transplants had an awful record before cyclosporine but even after it improved long-term survival, the surgeries still had high failure rates, depending on the organ. That has improved. Would that have happened if we decided these surgeries were "marginal," which is what the data showed?
If you said cut medicine by 95% it would be easy to figure out what that best 5% is.
And in all probability this would be nearly optimal policy. After all, the best 5% of modern medicine is very different than going back to the 19th century or whatever.
It is not just *marginal medical treatment* that is somewhat obscure. So is *marginal non-medical-treatment*, viz., whatever would be purchased instead if a little less medical treatment were purchased, or whatever would not be purchased to allow a little more medical treatment to be purchased. The marginal effect of medical treatment overall can be thought of as the marginal direct effect of medical treatment *minus* the marginal direct effect of non-medical-treatment. The latter effect on health may be significant, even though the purchase or non-purchase is labeled “non-medical.”
The studies of marginal medicine effects must already be including that effect.
True.. I meant to say that if one is worried by the difficulty in identifying *marginal medical treatment*, one should be equally worried about the similar difficulty in identifying *marginal non-medical-treatment*.
Given Hanson's celebrated disregard of social desirability, I kept reading this back and forth waiting to see some simple statement in the following vein:
"Lol don't give expensive treatments to old fat very sick people."
Don't the majority of medical costs come in the last six months of life? Americans in particular are often physically rotten by old age, full of stints and statins, plaques in the arteries and plaques in the brain. Tangles, too!
I've heard from old dying men that their doctors' standard of care was just a determination that the problems caused by some other specialist would kill him first - cancer doc hopes the heart will get him, or the kidneys, etc.
The 'death panels' of the original obamacare debate would solve this. What's that? You've had 3 heart surgeries and need a fourth? No, we're putting this needle in your arm, goodnight sweet prince.
No, most med costs don't come from last six months of life.
OK, you're right. Here's a literature review showing 21% in last 12 months of life: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC6610551/
To move the goalposts technically but not in spirit, I wonder what percentage is in the last 5 years of life? It doesn't matter to my point whether we're talking about the very old/sick who are going to die this year or the very old/sick who are going to die merely rather soon. In terms of life-years that's still where the fat is to be trimmed.
Certainly if the debate is "does medicine provide a good value in predicted life-years for the cost?" then your best bang for the buck will be cutting spending on extremely unhealthy old people. Scott says, truthfully, that we aren't great at assessing which treatments "work" - but actually we do a very good job of assessing which treatments are likely to be very costly and not provide a great deal of extra life to which patients, based on their current age and health status. If it's "does medicine work" then we need to talk about the definition of "work."
Re: old age, there's also so much more you need to think about on both sides rather than just "how much $" and "how many more years". The data that Scott and Robin are debating have no entries for "we spent 10,000$ on a treatment that actually caused the patient to die 10 days earlier, but wasn't in desperate pain the whole time and could say goodbye to their family" Or "we spent 30,000$ on a treatment that meant the patient lived 10 days longer but they were in excruciating pain the entire time."
At the highest level of abstraction, that medical payments do not seem to improve medical outcomes is merely step 1 that only inevitably leads to step 2 of "therefore we should spend less on things that are labeled medicine." Robin says that we can therefore cut medical spending at the top, and let people decide to do with that new cost-benefit analysis, which may be true, but the fact is that despite what Scott says, we *can* assess the effectiveness of billions of $ of medical (and "medical") costs vs the benefits they provide, not only in life years, but quality of life, pain and many others. Plucking the low hanging fruit there can provide the real world evidence to justify reallocating resources to better things.
Pain outcomes were included in the RAND studay.
Ah ok, that's good to know! I guess what I was thinking of were the charts showing increased medical spend against just life exp.
Even a fifth is a lot. Does the article distinguish "last year treatment" between "treatments we give lots of people with the usual result and expectation of adding lots of healthy years, but with some people dying soon anyway" and "treatments we give with high expectation the patient will still die soon"?
In the first case, its arithmetically possible we could spend half of all health spending on that, with 40% of those patients dying soon but 60% living many healthy additional years, so perhaps worth it. The second case is much harder to justify, and there's a massive margin that if cut wouldn't affect average life span.
One could consider the pressure in the other direction. The decision of which treatments to pay for can become very political and health bureaucrats responsible for making these calls sometimes complained about public, political, or media pressure to make exceptions to the usual cost- benefit criteria and approve treatments that didn't make the cut because regardless of the little bang for the buck the patients or disease type aroused greater than typical levels of sympathy and support. That motivated proposals to provide some political insulation to entities like NICE / IPAB. Each such treatment operated in the manner of Olson's "concentrated vs diffuse interests" and "special interest" treatment exceptions are individually small but aggregate into massive costs. This is also what Effective Altruists complain about when they criticize charitable giving to popular or sympathetic causes where the money could save many more lives if more efficiently allocated.
Now consider that whatever system eventually succumbed to these kinds of pressures and granted most of the too-costly-but-popular treatments, spending a lot of extra money but with little to show for it. Now imagine realizing that a ton could be cut without affecting average outcomes much, and then trying to run this process in reverse. Every single thing you can mention with specific identification is going to make you the target of a pile-on calling you a heartless monster who doesn't care about killing the members of some very sympathetic population.
Personally I have a lot of sympathy for someone put on the spot to make such specific identifications and can understand the need to keep discourse at a more general level.
Scott Alexander is a very honest dealer. In perhaps literally anyone else I would have regarded introducing, as he did, a $300 helmet for an infant with possibly 75 years ahead of him into a conversation about medical costs utterly disingenuous. No, it's the old fat sick people, Scott, ffs.