Is the "monkey trap" thing even real? Where the monkeys self entrap because they can't let go of a snack, to the point where they can be caught by a predator?
I've heard this a bunch of times and I've seen a video (but like, how long did the monkey actually remain stuck, I don't recall) and never was particularly skeptical. But somehow hearing it from an economist ("burly coolies", "cobra effect") makes me suspicious. This guy (https://samknight.com/?p=924) says he could verify none of the monkey trap stories.
Also, pre-1960 this kind of thing was referred to as a [derogatory term] trap. It's a fable included in the Talmud and Aseop's collection, among other sources.
Thanks. I recommend it, and a lot of the point of the analysis is not that it's a serious trap, it's a thing that young people fall for, and they learn to do better.
Bateson presented it as a challenge-- what *general* advice would be useful for the monkey?
Probably not the same derogatory term, but not at all out of the question that many of these sources use different derogatory terms for different groups.
That sounds plausible, but I can't find a source to back that up. The only thing I found is Urban Dictionary's "a bucket of fried chicken placed under a net." (Unless I'm substituting in the wrong "derogatory term." Is it the delightfully rhyming "[East Asian] trap" instead? That gets me Pacific War propaganda posters.)
"I even contacted a zoo and was assured that, while a monkey may fuss with the prize for a minute or two, it is most certainly smart enough to let go."
Maybe this distracts the monkey long enough to sneak up on it and kill it though.
Ironically, the boiling pot metaphor is exactly backwards in real life. If you slowly increase the temperature of a pot, the frog will jump out when the temperature gets too high. If you drop a frog in boiling water, it will die too quickly to jump out.
The point about the Idea of the frog being boiled slowly is that humans aren't likely to notice gradually worsening conditions. The contrast was between conditions that get worse very gradually compared to conditions that get bad quickly, but not so fast as to be immediately deadly. Of course the frog jumps out of somewhat hot water, but maybe a slow increase would fool the frog.
> The point about the Idea of the frog being boiled slowly is that humans aren't likely to notice gradually worsening conditions.
I don't think it works in even the most direct sense. Imagine you, a human, are in a room. It's getting hotter and hotter. Will you ever try to leave the room? Would it matter how slow the increase in temperature was?
Why imagine that frogs are fundamentally defective in this way, compared to humans?
Cold-blooded animals have to be able to discern slow increases of temperature in order to survive. Otherwise the rising heat of noon would kill every cold-blooded animal not already in shelter.
Biologists correct me if I'm wrong, but it's my understanding that every vertebrate animal experiences physical sensation, and can tell when a temperature isn't optimal for it.
I'd go further: even if you put a frog in a pot of water and don't turn on the heat, the frog will jump out. Frogs are pretty skittish, they don't wanna be in this weird metal cylinder.
I read about it in Gregory Bateson. It was described as a method humans used for catching monkeys.
Now I'm left wondering about the trap for humans described in one of the early NLP books. Scientists tested rats (mice?) on their ability to run a maze for a reward, and then on how long it took the rodents to stop running the maze if the reward was taken away. The rodents quit trying after some reasonable amount of time.
They tested the result using grad students and a giant maze baited with $5 bills in a basement. The grad students are still breaking into the basement in case there's a $5 bill.
A good allegory about lost purposes. This was from the 80s, so people weren't as cynical about grad school then. The story reliably gets a laugh, speaking of incentives. But is there any truth at all?
I was going to say I tried testing this myself in college, and neither worked very well, but I suppose it depends on what kind of fly you're targeting. Vinegar works quite well on fruit flies, but appeared not to with whatever kind of flies the ones I tested it on then were.
(shudders) The only thing that worked was hanging fly paper *everywhere*. And of course I had to repeatedly explain to the property manager that no, they weren't fruit flies, so vinegar traps weren't going to work.
I recently catched a lot of fruit flies using traps with wine and vinegar under perforated plastic foil. A mix of wine and vinegar seemed to work best. Did not try honey, though.
I also tried this at one point out of curiosity, setting out bowls of water, honey-water, and vinegar-water. The vinegar worked best, although I admit I have no idea what it was catching.
I suspect this is a case where an essential element of a tool that evolved in different times in the hands of different people has vanished from the telling. Imagine a world in which foot traps in their historical form have been so long vanished that children only know about live traps. The metaphor about a wolf chewing its own foot off to escape no longer makes sense, so they change it to a story about a wolf who ignores his hunger and does not go into a live trap. Still later someone in a discussion about incentives where this metaphor comes up quibbles about whether or not animals ever get hungry enough to go into live traps.
The monkey trap probably works great if its original form was anything like the DIY coon traps in rural America a century ago: it's a can with sharpened nails driven through at an angle so that the little paw fits through empty but not full. The goal is to cause the animal pain upon initial attempt at extraction, causing it to panic and disable its paw to where it can't let go. When you find it hours later its mangled forelimb is hooked on at least one nail and killing it with a club is an act of mercy.
Ah yes, I'd forgotten that. Is there a drawing of the trap in the book? I remember seeing a schematic for it in an old book. Perhaps one of the Foxfire series, but I doubt it (Foxfire would have snares). I think the same book had an illustrated how-to on butchering snapping turtles.
This one is, from experience, very true : crabs will kill each other if you put a bunch of them in a bucket. They will also kill each other if a small crab enters into the hole of a bigger one by the way.
I'm going to unironically say that this is the *worst* line in the post.
Robin Hanson is using a motte and bailey here where the motte, of course, is "medical care isn't good at the margins" and the bailey is "medical care is useless".
Scott's being a quokka here. He doesn't seem to ever be able to recognize that someone's acting in bad faith, so when he's presented with a motte and bailey, all he can do is mutter about how he "apologizes" for pointing to the bailey and "misrepresenting" it as the motte instead of calling out for what it is.
Robin's actual proposal is to cut medical spending in half. He was explicit about that from his opening essay at Cato Unbound, it wasn't a position he retreated to after his interlocutors pushed back. If other people agreed with him on that then he would no longer qualify as a contrarian on that issue!
So, does he have any strong opinions on *which* half ought to be cut? If not, I think this still resolves as "Hanson thinks all medicine is equally bad."
He has also explicitly said that some treatments work, but he doesn't propose to select which treatments work himself. He's an economist and his proposal is instead to cut spending, and other people would make downstream decisions on which things to cut. There aren't people arguing "We should spend money on useless treatments", so merely saying we should get rid of them doesn't move us from the status quo (similar to people saying they will lower the budget by getting rid of "waste, fraud & abuse").
But, at least according to some of the quotes in this article, he disagrees with people who think we should expend any effort figuring out which half to cut. And the first bit of this article is pointing out that (at least as Scott sees it), Hanson has made individual statements that seem to disagree with _any_ particular way of figuring out which half to cut.
If medicine is taxed rather than subsidized, people will simply make choices about which things is cut and I don't think Robin would object then. If people insist on first figuring it out before cutting, then I think Robin won't expect it to get cut.
>But, at least according to some of the quotes in this article, he disagrees with people who think we should expend any effort figuring out which half to cut.
Doesn't
>eg Britain spends only half as much per person as the US on healthcare, and Brits have _approximately_ as good health outcomes
imply that British NHS decision makers _already_ know where to cut?
And, presumably we (writing from the usa) could either
- copy their policies
- have them remotely make these decisions for us (presumably at a price)
- kidnap some of them and install them in our health care system
and similarly for individual decisions such as
>So my second question for Robin is: how do you recommend I proceed? Do I avoid going to the doctor for some specific subset of these categories, like 5-10?
Admittedly, this isn't _exactly_ right. The UK and USA populations are a bit different, and e.g. the fraction of medical care that goes to treating e.g. gunshot wounds probably doesn't match across the pond... Still, the British experience seems to imply that _most_ of the right choices are already known.
People keep saying that. He was very explicit, e.g. end subsidy, read Cochrane reviews. He didn't personally review every extant medical intervention, is that what you mean?
In countries with public healthcare, the government decides what to spend it in. The free market has nothing to do with it. In countries with private healthcare, courts decide what you can sue your insurance company for not paying for. The free market has nothing to do with it. So where exactly is he planning on implementing this?
I think you're right about the motte and bailey but wrong in your judgment of Scott's reaction. Assuming good faith is often an effective play in this sort of situation. If it turns out they are actually arguing in good faith and just communicating badly, it avoids you being a dick for no reason. And more importantly, even if they are acting in bad faith, it allows you to avoid *looking* like a dick when you politely press them over the "confusion," which is exactly what Scott does here.
Assuming good faith makes you look epistemically virtuous, but sometimes at the expense of causing a slight deterioration in collective epistemics. People end up thinking a view is more plausible, or a thinker more honest, than they should, and would if they had instead been exposed to a critic more willing to call a spade a spade.
If you assume good faith and they're doing it in bad faith, they're going to say "I'm sorry you're confused but..." and keep running with the scam.
Hanson has already responded to Scott's response, and his re-response is basically "here's some more motte that I've posted in the past, see, I have actual quotes"--he's doubling down on the motte and completely failing to address the *inconsistency* between his past motte-posts and his past bailey-posts.
I'm a fan of how Scott handles this sort of thing! If he wrote Robin off as acting in bad faith then I would have very little interest in reading this post, since Scott would have no reason to address any of Robin's arguments outside of pointing out contradictions. I think if as a writer you don't think the person you're addressing has a real position that can be understood, it would be a waste of words to write about them in any detail.
When you politely assume a person is smart, and you may just be misunderstanding them, you're leaving the door open for them to clarify themselves. In the best case you might understand their position if you restate it and they correct you, but in the worst case they make a more specific contradiction or weaken their argument and either way you prove your point.
You might think there's a downside to being a quokka (never heard that one before lol) but from where I stand it's the best way to address an opponent, and also selfishly I would not be very interested in reading Scott just dunk on someone's contradictory position without really getting into the weeds of what they believe and making an effort to steelman them.
If someone is not sincere, "the weeds of what they believe" is a mirage. It may not be possible to get what they believe from what they're saying because their words and their beliefs *aren't connected*.
Oh totally, but it's very hard to be 100% sure someone is insincere, and doesn't cost you much to assume (or pretend to assume) that they are acting in good faith.
The alternative to responding to someone as if they were sincere shouldn't be calling them out for being insincere, it should be ignoring them. If you think you possess overwhelming evidence that someone won't engage in good faith, a big breakdown of their position like this followed by a "you must be a contrarian" call out won't change their mind anyway. And if Scott is making this post to try to convince the audience that this is a nonsensical position, he might as well give the benefit of the doubt, to look to the audience like a fair and even minded evaluator instead of someone trying to find extra reasons to dunk on a bad argument by claiming that anyone making it must be arguing for the sake of arguing.
Agreed. I hadn’t heard of Hanson before now, and for me to be able to confidently dismiss Hanson as a kook or a troll, I need to be able to tell that Hansen genuinely has nothing useful to say. A discussion that left open the possibility that Scott and Hansen were simply talking past each other would be too ambiguous for me to conclude much of anything without reading more of Hansen’s writings. Attempting to seriously engage a troll may be unrewarding work, but it provides a record that allows other people to determine with relatively little work on their part that the troll is indeed a troll or a kook.
Based on the exchange so far, I’ve provisionally placed Hansen in the troll/kook category, but I need to see Hanson’s response to this latest post by Scott before before I can pass judgement with confidence.
Prof. Robin Hanson is no troll. Though known for surprising ideas. Same GMU-faculty as are Caplan, Cowen, Tabarrok. And very active and well known in EA/rationalist circles. A friend. Who is now surprised Scott did not just call him to clear stuff up beforehand.
Too many people think that once you learn about a fallacy, the only thing you should do is diagnose the fallacy.
It's usually much better to use your knowledge of the fallacy to press the real question, and leave the diagnosis of the fallacy implicit. I know it *feels* like you're letting the terrorists win when you don't call them out. But calling out isn't the goal - understanding is.
Indeed. I've rarely seen anything as low key vicious as, eg, "So this was exactly what I said in my post, except that Robin takes out my explanation and quotes only half of the section, so that I look like a moron who didn’t read the paper."
Specifically, the sentence seems subservient, but it's much more devastating than a direct "Robin is lying here". Politeness and respect can hide devastatingly sharp blades.
I endorse this in principle, but in practice it feels kind of weird because the rest of the essay goes on to say "...but actually, Robin Hanson *did* say what I'm accusing him of, here are the quotes."
This way of dealing with the views of others is sometimes referred to as the golden rule of hermeneutics.
That is, if you present the views of someone you disagree with so that their views seem obviously stupid and/or evil, you are probably misrepresenting what they believe. Since very few want to come across in the eyes of others as stupid and/or evil.
Instead, present your opponent's views so that they may possibly make sense to you.
That said, sometimes some people are really stupid and/or evil. (As Kurt Vonnegut remarks in The Sirens of Titan: The trouble with stupid people is that they are too stupid to understand that not everybody are stupid.) But if so, let that "show itself" through further communication. Do not have that as an ...eh.. prior. Since that will block communication.
A lot of medicine is useless, and fear of death makes people overspend on medicine, but a lot of other economic activity is also useless and motivated by arguably irrational reasons. Is it inherently more problematic that people overspend on medicine than on luxury cars for instance? If people did follow up studies on other kinds of consumption there would be a lot of results showing that it’s useless or worse at improving people’s lives.
Yeah, I want to be really careful not to make anything that looks like a medical recommendation here. There's a lot of overtreatment in back pain, but I bet there are specific kinds of back pain where surgery is effective and important. I would recommend starting off from a position of mild skepticism, but doing some research and asking your doctors frank questions and maybe getting a second opinion, and letting those things push you to yes if they all seem plausible.
Going to a back surgeon with a herniated disk and seeing on the intake form, " Is this your first, second, third, fourth, fifth.....fifteenth.... back surgery," was very eye opening.
I apparently got pretty lucky in my back pain adventure. I have (had? I was told it might eventually heal and I haven't had a check in a long time) a partially herniated disk that was, for a while, causing quite significant (although not debilitating) back pain. I saw a specialist who recommended physical therapy and, even though he was a doctor who performed both steroid injections and surgeries, told me that the level of pain I had was not enough to warrant something that invasive.
He was absolutely correct, because now, several years later, I am _mostly_ pain free (I'm unsure how much the physical therapy helped vs. it just got better with time), but apparently the fact that I wasn't pushed quite strongly towards a surgery or at least a steroid injection is somewhat surprising.
I do think that the MRI that my non-specialist doctor ordered was probably a waste. He should have just sent me straight to physical therapy and saved me the couple of thousand dollars on imaging that didn't (from my perspective) seem to impact my care very much.
This is not a "gotcha" - but isn't this statement functionally equivalent, from an economist perspective, to what Hanson is saying? You both agree there's a continuum of cost vs benefits on back pain treatments, and you both agree that actual dollar costs go up the more you lean on surgery/medicine vs non-medical interventions, you both agree that the efficacy of those costlier medical interventions varies a bunch, and that it is often hard to tell.
Hanson says "therefore, spend less on medical treatments of back pain" and you say "therefore, approach the use of medical treatments of back pain with mild skepticism, get second opinions and only accept treatments that seem plausible."
But these things seem very similar, in both the near and far way? "Medical interventions may be more costly/less effective than you think!" and "you should approach claims of medical intervention efficacy with mild skepticism, and demand a higher level of plausibility" result in the same actual actions by patients when confronted by back pain.
In both cases, in an economics context, you are both saying "when doing a pro/con on whether to get a medical treatment, you should put less weight on the pro side." Hanson isn't medical, so he can't provide *medical* recommendations on whether a given treatment passes the cost benefit test, (whereas you can) but he can look at aggregated data on outcomes, and make *economic* recommendations (i.e. based on the patient's desires).
At a high level, I think RH is thinking on the margin of the economic incentives: that if we convinced Americans to cut their aggregate medical spending by 100m, say by reducing medicare reimbursement rates across the board by a %, this would change every single one of their cost/benefit analyses for treatments. This would have a marginal impact on decisions, but would only actually change patient's decisions to seek medical treatment in a tiny % of cases: those where the patient's perceived net benefit from the treatment, minus costs, was greater than the small % change: and this (by definition) would be the (allowing for notoriously bad patient assessment) least beneficial cases: i.e. the guy saying "yeah my back hurts, but I've decided to lift weights and do PT from youtube instead of the now-more-costly-surgery". RH is saying that cost cut would, on the margin, likely make every patient better off (and also the payers, who have cut costs).
Obviously, yes, there would be some patients who go "100 bucks more to pay to treat my gunshot wound? No thanks!" and then die but RH is betting that those cases are rare enough that the benefit from the other vast majority of cases where patients do properly weight their cost-benefit will outweigh that.
Many more people have enough scientific literacy to investigate medical questions for themselves than actually do. If you are able to read literature and understand it well enough to have a meaningful back and forth with your doctor (to ask a question like "a systematic review found only low-quality evidence that people with my condition benefit from this surgery more than they would from physical therapy, why do you think it's recommended in my case") I feel like you should do that--especially for something as intensive and risky as surgery--rather than take a hearsay observation, even from someone you respect who is also a doctor, and run with it. This has got to be one of the most fruitful uses for a scientific education. I don't mean to impugn you specifically, especially not knowing your background, but I see people reasoning about their own medical care by lurching from one hasty generalization to the next all the time (I have surely done it too) and I'm really curious about what reasoning actually governs these decisions and why, besides laziness and time constraints, we rely on it anyway.
I think we ought not to treat back pain as a specific disease, because we'd get in the same trap as treating headaches as a disease. That's a symptom and not a cause. The cause is what's relevant. Try to under-treat a herniated disk(take a severe case), then we'll see. Don't make general statements about surgery. It might save your life some day.
Luxury cars are not generally paid for with other people's money. The enormous amount of medical spending that is paid for in really indirect ways makes it much more susceptible to perverse incentives.
I believe it's part of Robin's point, but we don't subsidize luxury cars and definitely don't guarantee that everyone can get a luxury car. Some people argue that medical care is a human right, and almost everyone even in the US agrees that certain care should happen regardless of ability to pay. I think if we allowed people to get whatever medical care they can afford (as they can with luxury cars), that he would be just as okay with that as someone cutting half of all medical expenditures.
That is a good point but in the Fear of Death and Muddled Thinking article he seems to concentrate most on how this is all a problem of human rationality, not one of subsidies.
The countries where health is subsidized more by the government are more keen on cutting down on costs. So there’s a certain paradox there.
The US might have one of the worst overall systems when it comes to cost incentives. I think the cost incentives in countries with universal systems is bad, but not nearly as bad as in the US. In universal systems there's at least pressure on the payer (government) to limit the care given.
Big picture, the US separates the people paying for procedures from those deciding when/how/who for those procedures. Employers pay for employees to get coverage, but can't decide how it's spent. The government mandates certain levels of coverage, but puts in few to no meaningful limits on that spending. Insurance companies pay for the care, but pass along costs to others. Hospitals convince patients they need care (right or wrong) while the patients only pay a small portion of the overall costs.
It's like if you were buying a car for your own use but someone else was paying for it. You might decide to be frugal and careful to get a car, but there's no compelling reason to do so. Some people would get exactly what they needed, but most people would instead fit into one of two buckets - 1) they would buy a car they otherwise would not have gotten, whether because they couldn't afford it or because they didn't need it, or 2) they would get more car than was needed for their purposes, at least getting a few upgrades to features but maybe even a higher level of car entirely.
Well, yes. But there's a difference -- I think -- to taking what's essentially a placebo, or most likely to be, or at best "just in case, won't harm" or a vehicle for improving one's feeling of control and agency (I take vitamins + couple other supplements purely on this basis) and getting treatment that we believe will DIRECTLY improve physical state of the body in the desired direction. One reason for that being that many (maybe most) effective treatments have significant costs in effort, pain, time, potential side effects and if you live in the place like the US, also money.
So I can pop a 4x DRA of D3 daily because the cost is minimal and it gives me a good feeling and who knows, maybe leaving at a cloudy 57N I need it. But I'll not take a SSRI for moderate trauma induced anxiety because the cost (potential side effects and the dangers of long term use in this case) are not worth the possible benefits. IF I knew that that very same Sertraline was guaranteed to glue my broken brain back to where it was 6 years ago, I'd take it for sure.
I'll have massage because it feels good to be touched by another adult human, and it's relaxing but I'll cancel the appointment if I don't have money for it. Had I believed it was HEALING my hip arthritis I'd prioritise this spend. Etc etc.
Rarely, if ever, have I seen someone go to such lengths to prove something so obvious.
Huge respect for Robin Hanson, but this is an issue where he seems to have fallen into the contrarian trap of defending a stupid premise purely to show off how cleverly he can defend it. "Some ideas are so absurd that only an intellectual could believe them."
I think Hanson is a worthwhile thinker in the Rule Thinkers In, Not Out sense, but this kind of incautious contrarianism (and lack of clarity about his own central points) isn't at all atypical of him.
Yes. It's the unfortunately-familiar libertarian game of intellectual whack-a-mole. I'm frankly not seeing here any reasons to have huge respect for Hanson (which is ironic since I'm personally a cautious/skeptical health care consumer to the point of regularly exasperating my spouse on the topic).
This newest post clarifies that Scott's frustrated and almost frantic re-engagement with Hanson's childish intellectual bullying on this topic is being fueled by the situation with Scott's son. Every veteran parent grasps how that is feeling for Scott and his wife right now. Hanson obviously does not and never will.
The misshapen head link Scott supplied mentioned that it is more common with twins. Not speaking to his situation at all, but in general the significant rise in multiple births and hence the problems associated with them, could potentially be considered one of those ways in which medicine both gives and takes.
My understanding is that the primary cause of increased flat-headedness is a success of the movement to put kids to sleep on their back rather than their stomach in order to avoid SIDS.
I really don't think that labelling this kind of thing with _any_ particular political ideology is either A) true (in the sense that only one particular ideology uses this kind of strategy) or B) useful/helpful/elevating to the conversation.
I was going to say the same thing. To add, at an extreme of this phenomenon you get flat earthers: use a complex edifice of clever arguments to arrive at an obviously wrong conclusion, then make that conclusion a key part of your identity, and defend this identity at all costs.
I think that's a slightly different phenomenon; there are no flat-earth arguments that are "clever", at best they are confusing.
Modern medicine is sufficiently complicated that a logical case can, with effort, be constructed against it (albeit one that is pretty easy to demolish by someone of equal intelligence to the edifice's designer). The earth's roundness, on the other hand, has been settled science for thousands of years.
> Maybe we should skip it - I never see any adults with obviously mis-shapen heads out there
For anecdotal evidence, I had a friend in middle school with this issue. It's less noticable as an adult (possibly he learned to hide it better with haircuts or something? Or maybe it just got milder naturally), so you might not notice it passing him on the street but you probably would if you were sitting down for a conversation with him. I don't think the helmets were available back then.
I also know some people like this. It's not a big deal in the grand scheme of things, but I might have done double-takes and stared a little on first encounters.
Our son also had plagiocephaly, vociferously rejected attempts to sleep on his other side and ultimately we decided to go the helmet route
It worked fine, his head is not perfectly symmetrical but definitely not something you'd notice without calipers. It requires an extra dozen or so doctor appointments but there are so many appointments those first years I hardly noticed. Also need to give him regular baths and wash his helmet and hair so it doesn't smell terrible, but again, kids smell terrible a lot of the time anyway
Meanwhile one of my coworkers said she had a flathead and was never treated for it. I honestly hadn't noticed but she seemed kind of self conscious about it, so avoiding that also seem worth the $$$
I had a family member who was born with the same condition, got the helmet, and it made a significant visual difference at that age. There's no visible sign of the condition at this point, though he's a teenager now.
I don't know how well these findings have survived the reproducibility crisis, but there are definitely studies that show that facial asymmetry is perceived as less attractive even if it isn't at the point where it's something you'd consciously notice. For this reason, I too would buy the helmet.
I had a friend in varsity with this issue as well - his skull looked a bit like a muffin that had risen out of the pan. It had no functional effects that I'm aware of, and only become really noticeable if he cut his hair short.
My daughter (now 3 years old) had a misshapen head, probably from always sleeping with her head turned the same way. The doctor said "you can get a helmet, but I wouldn't". So we didn't, and now she looks fine.
He's a old school country doctor (obstetrician/pediatrician/geriatrician all in one, maybe 75yo) and as far as I can tell he generally would never treat himself for anything. When he does recommend treatment, I take it very seriously.
Fewer criticisms of economists or more weight to the criticism of economics? It would derail this thread to point out the flaws in most economic models, it might even derail open threads. So I’ll leave it at that.
I agree. Specifically, denouncing a whole profession as unreliable, though it might be interesting to look at reliability of economists in general. It would be a hard thing to test.
>though it might be interesting to look at reliability of economists in general. It would be a hard thing to test.
Agreed on both counts.
<mild snark>
Though if one wanted to look at a profession where reliability would be _really_ hard to test, consider what a double-blinded RCT to test the reliability of clergy would require... :-)
You would have to know something about economics to be able to point out flaws in the models. I have suggested this to you before, and will do so again, learn some economics. It will enrich your intellectual life. Try marginal revolution university's microeconomics 101 course
I would prefer to see fewer criticisms of anything or anyone that take the form of a single sentence saying in essence, "you can't trust *those* types", without elaboration or justification.
I think the problem here is neither of you has framed this in terms of computational complexity. You seem to be objecting to Hanson’s lack of specific actionable advice at the macro or individual level. But this doesn’t make his fundamental claim wrong!
Health care is computationally opaque. The effects of anyone one purchase aren’t obvious or legible; they are subtle and interact combinatorially. The placebo effect plus the finance model of insurance mean there’s some psychological boost for _any_ intervention, and a while supply of people getting paid for you to keep using it.
Suppose you model health interventions as being probabilistic in nature - what works for some, doesn’t work for others. You can’t know what will work in advance. If you find yourself with problems, you go to a doctor, and the doctor says “well, intervention X works for 80% of people who try it with this condition.” So you try it, and it seems like things are a little better… maybe? Plus insurance covers it so you keep doing it. If most people just blindly follow doctors advice, you should expect that roughly 20% of the people taking that intervention aren’t benefiting from it.
My own experience has been, after a decade of intense focus on my health, picking up and dropping numerous approaches and interventions, I’ve finally found the right diet that works for me, and habits that on net make me better off. My wife did none of this and has overall about the same health as me.
As I type this I’m using lamisil to treat a rash I got on my foot from spraining an ankle and then wearing a boot, the sprain came from exercises I was doing to try and strengthen muscles I thought were too weak. I wish I could say this was a one off case but this general pattern has been the story of the last ten years of my life, when I focused immensely on trying to improve my health.
What I’ve discovered only in the last year or so, is that what I pay attention to and how good I feel about the present moment seem to be the single biggest lever I have for improving my health outcomes. Regular prayer and gratitude and trust in my body to manage itself seem to be working better for me than attempting to tinker in a complex system I don’t understand, and then incurring the anxiety of knowing I can’t control this very important thing. No amount of dietary fidgeting is as effective as increasing my sense that my body is a gift from God so maybe don’t stuff it with junk food.
So, yeah, if there is an answer here, I think it’s leaning into the placebo effect and intentionally trusting reality itself to be sufficiently, without knowing the specific details of how that will be. Google SRE has this motto, “hope is not a strategy,” to which I respond, bullshit, hope is the optimal strategy for navigating a complex system when your navigation mechanism (ie brain) tries to reduce the error in your valence assessment of that system. Mental health is, imo, a necessary precursor to good physical health.
I don't think Hanson is unaware of the computational complexity angle. The solution he most favors, according to his latest response, is to parallelize the computation by incentivizing individual insurance companies to "compute" what spending should be cut (and he ostensibly achieves *that* by mandating that medical insurance be coupled with life insurance).
Let me try to steelman the anti-healthcare position here.
I have chronic heart disease. Based on some early symptoms and family history, I was put on several medications (beta blockers, ACE inhibitors). Sounds great, right? However, I have developed chronic fatigue in the years since, which is 10x worse than my (mild) heart problems.
Years into this, I find out that the heart medications I've been taking have extremely strong system-wide effects: reduced activation of the sympathetic nervous system (beta blockers), increased inflammatory response and lowered pain threshold (ACE inhibitors), and even reduction in the efficiency of respiration (apparently, lowered heart function, which my meds induce, can lead to slight respiratory alkalosis even as oxygen saturation read as "healthy"). My cardiologists never mentioned any of these effects to me. These don't correspond to the typical image we have of side effects (rare, acute complications, e.g. "rashes in less than 5% of patients"). Rather, this is cumulative damage to half a dozen vital subsystems, throwing homeostasis way out of whack.
My point is that some outcomes are relatively easy to measure and correlate (early start with heart medications reduces incidence of heart failure), while diffuse downstream effects that sap your vitality and make your life worse are extremely easy to miss, in all but the most egregious cases. If we assume that such things are systematically happening in the treatment of many diseases that aren't immediately life-threatening, we can end up with a picture where lots of people walk around saying that modern medicine has saved their lives many times over (I used to say such things in my younger days), while simultaneously the health of the population gets mysteriously worse, in ways that are easy to dismiss with a "better screening" pseudo-explanation.
Yeah, I see this kind of thing in my practice too.
I think we really need to spread the announcement "If you take a medication and feel much worse, then unless this is part of the plan (eg everyone knows they'll feel worse on chemo, but it's worth it), tell your doctor and unless they have some great counterargument consider stopping the medication."
Did you eventually stop the medication, and did your fatigue improve?
A cardiologist eventually took me off the beta blocker. I'm considering stopping with the sacubitril/valsartan (similar to an ACE inhibitor). Based on everything you've written on predictive coding etc, I freak out 10x more about things that mess with homeostasis than I used to.
Note that the heart medications arguably *are* in the category of "feeling worse is part of the plan". Reduce cardiac output, get low blood pressure and a more sluggish body in exchange for protecting the heart. The really bad fatigue problems showed up a decade down the line.
Also, the cause-effect here is not clear cut at all. I elided a bunch of stuff (reduced excercise due to covid lockdown + some injuries, stressful job, fibromyalgia).
ETA: my fatigue is improving now! I tried many things, what seemed to work is a combination of breathing and relaxation exercises, improving posture, deep meditative states to better feel where the body is tense, learning a lot about anatomy and becoming better at loosening specific muscles.
Lo ok at it this way: If a heart condition makes you 2x as likely to die of a heart attack in any given year. What does that mean?
When you’re young instead of a 1 in a 100000 chance of dying of a heart attack per year, it’d be 1/50000. And when you get old it’d be more like 2/3 instead of 1/3.
The heart disease takes about one year off life expectancy, but has a HUGE impact on the proximate cause of death.
So if the pills cured the heart problems perfectly they’d only a year to LE, and if they aged you even a little it’d easily wipe out the gains.
Do you see? Everyone dies of aging, aging is five orders of magnitude more consequential than your little heart disease, and all the diseases (and medicine) in the world just shuffle cause of death around.
Except for the few really bad diseases that can easily kill a young person.
But the rest of it is obviously just shuffling around cause of death.
Just think for one second! You’re heart disease won’t kill you when you’re young, it’ll kill you sometime after you start to age. And how likely it is to kill you is a function of your aging.
Which is to say the heart disease sin’t changing when you die, but what the proximate cause is!
Note that the article, in addition to citing this, actually disagrees with the position, kind of like HIV doesn't actually kill people, just allows something else to kill people which otherwise wouldn't. Which is kind of like saying guns don't kill people, but bullets do.
Aging is almost exactly like HIV (before they had the medications they have now). It's a total death sentence.
We take it for granted, but it is responsible for >99.9% of the lost life expectancy. You might expect to live to 100 000 without it.
If an 82 year old dies of cancer, the aging was the cause of the cancer in the first place, and the aging was likely to kill them shortly anyway.
If you allowed aging as a cause of death it would be the primary CoD for nearly every single person 65+, with their proximate CoD being at most a small contributing factor.
Something else I learned from the article is about cellular apoptosis, a normal way for cells to die. We have much more to learn about aging, but I agree with the decision not to include "old age" as a cause of death. It may make things more complicated in some cases, but, as far as I see it, old age merely makes more and more maladies more and more likely as time goes on. Maybe a combination of them can kill someone whereas taken individually they are survivable.
But putting "old age" on a death certificate would be as informative as listing "suicide", as that isn't the actual manner of death. HIV, however, does seem like a reasonable cause of death to use, even though another disease actually caused the death.
It may be that old age itself can be conquered. Redwood trees can live for thousands of years (https://en.wikipedia.org/wiki/Sequoioideae). What is the difference in how they live?
But the point is that the benefits of the medications are bounded and the potential harms are (nearly) unbounded.
The best case for the benefits are: return this patient to being on track to dying of old age, and the harms are are: this might accidentally be aging the person.
So as long as the dangers of the heart disease are much smaller than the harms of aging (how on Earth could they not be, aging is 10x worse than AIDS) then the patient has FAR more to lose than to gain.
That’s the underlying reason people have been having this debate since the dawn of time.
I wonder how much it's consistently possible to tell this. I'm not questioning your experience, just wondering how generally applicable it is. I've been on beta blockers for many years, but I have no idea if it's sapped my general vitality. I am slower and more tired than I was years ago, but not cripplingly so, and I'm also a lot older.
For what it's worth, I finally stopped reading Overcoming Bias after some anti-medicine post where I tried to find a charitable reading that didn't imply medicine didn't work at all and didn't find a reasonable one.
We used the helmet for child #1's mildly flat head. Not sure she wouldn't have normalized naturally, but feeling like we had some agency and doing 'what we could' did certainly bring us peace of mind. We have no regrets.
The current version of the helmet dilemma for us is atropine eye drops for slowing myopia progression. Useless treatment at the margin? Or maybe will it make a meaningful impact on our daughters QoL...
"We used the helmet for child #1's mildly flat head. Not sure she wouldn't have normalized naturally, but feeling like we had some agency and doing 'what we could' did certainly bring us peace of mind. We have no regrets. "
NOOOoooooo, don't say this, now you're proving Robin's signaling theory right!
I don’t want to over-react to a sly wry joke, but if Robin were saying medicine is 100% signaling, and you were saying it is 0% signaling, my guess is that you’d both be wrong.
I don't think anyone's argued that 0% of medical expenses are signalling; certainly Scott has previously discussed cases in which he fully agrees that the requested care was pure signalling ( https://slatestarcodex.com/2013/07/17/who-by-very-slow-decay/ )
You're probably already aware, but dropping a comment in case you hadn't heard about the ongoing, promising research linking lack of bright light exposure in infants and children to the development of myopia. A search for "early light exposure myopia" or similar will bring up a host of studies, or a more casual overview from Wired: https://www.wired.com/story/taiwan-epicenter-of-world-myopia-epidemic/
No, we followed the "back is best" advice. Another intervention with questionable evidence base, but fear of death is an even stronger motivator when it comes to your children.
Re: sleeping on back, best evidence shows that it reduces the odds of SIDS by 1/1000 over the childs first year, at best. In other words, if nobody followed the back-to-sleep advice, we'd have (at most) 0.1% of children die from SIDS compared to the counterfactual.
With the caveat that thinking about very small probabilities is very hard...:
Isn't lowering risk of death for a newborn in their first year of life by 1 per 1000 kinda huge? Thankfully, few babies die of preventable causes anymore in Western countries - the SSA says the prob of death under 1 is about 0.5-0.6 per 1k. In reality, a large fraction of these deaths are probably due to differences in baby "type" than anything parents can change by behavior/action - babies born with conditions such that they survive only a short time after birth, or have conditions such that the first time they catch anything, they die. If this is half of deaths under 1, then a sleep on back vs front SIDS margin of 1 per 1k would amount to a 1/3 to 1/2 increase in *aggregate* death risk under 1 for 99% of people. That's pretty large.
Or if you want it in expected life year terms, a 1 per 1k risk of losing 80 years of life is 0.08 life years per person. By comparison, we had a society wide freakout over a pandemic a few years ago that killed something like 4 people per 100k , people whom on average would otherwise have had something like <10 years left to live. (And I don't think ELY is quite the right metric: 1 baby dying at a few months old from SIDS or some other preventable cause is devastating in a way that 10 elderly people dying at 78 rather than 82 is not).
1/1000 is a *very* high end estimate. In reality once you control for smoking, mattress firmness and the presence of choking hazards in the infants bed it’s likely to be closer to 1/3000 for prone sleep alone. Note that Back to Sleep focused on multiple things, not just prone sleep alone.
Is a 1/3000 chance of dying due to prone sleep worth the 16-45% chance of developing a head deformity? Most head deformities go away on their own but they do cause stress for parents and potential stress for the child if it’s not cured.
I was just using the numbers you stated! (You did say "at best" and "at most", but still, its the number you gave!)
Anyway, I think 1/3k is still pretty non-trivial.* This is still a very sizeable chunk of preventable child deaths. "Is a 1/3000 chance of dying due to prone sleep worth the 16-45% chance of developing a head deformity?" 16-45% causal effect??? Since WAYYYY fewer than 16-45% of young kids are walking around with one of those head helmets on, either the vast vast majority go away on their own / they are pretty minor to begin with, or the vast majority of kids are walking around with them untreated and few people seem to mind/care. Since I don't notice many kids with hugely misshapen heads I'm guessing "correct not to notice/mind/care" seems quite justified.
As for back of envelope math: if VSL = $10m, 1 per 3k chance of death = $3300. So as long as prob increase of misshapen head x min(cost to fix, loss if not fixed) is below this, seems probably worth still doing.
Or to put it a different way, I am (a) completely open to the idea that most medical interventions have much smaller preventative effects than advocates often like to claim, and (b) that much medical "expert advise" is ad hoc and not really based on sound evidence (which is how it can completely flip on a given topic so often), but this conversation isn't doing anything to dissuade me from discarding this particular piece of 2020-era "expert medical advice".
(Off topic: I know about, but don't really understand the claimed effect of smoking on SIDS. Parent smokes, so... kid dies in their sleep. Is the claim a direct causal one - kids lungs are messed up in some way that makes situations that otherwise would be survivable, deadly? Or is it "smoking parent of young kids is probably correlated with social class which in turn correlates with a whole basket of stuff wrt parent actions that we don't observe, and these are what drives SIDS. I've always assumed its the latter, but I don't have any evidence for it beyond "social class almost always predicts this stuff which suggests the direct cause isn't properly observed/controlled for" and "the claimed effect seems underspecified").
>My wife and I recently took my four-month-old son to the pediatrician. The pediatrician said he had a mis-shapen head, and referred him to a head specialist for a second opinion. The specialist said yup, looks pretty mis-shapen, and referred us to a helmet-maker. The helmet-maker said yeah, definitely mis-shapen, and wants us to pay $300 for a helmet to correct it so my son doesn’t get stuck permanently looking like Frankenstein when he grows up.
> (there are some studies of this intervention, which are neither obviously wrong nor obviously unimpeachable. They say the helmet works, but not necessarily better than “repositioning therapy”. Our son refuses repositioning therapy, so for us it’s the helmet or nothing.)
>That helmet is probably our “marginal” health care expense, in the sense that it’s less obviously important than the other two things we’ve used healthcare for this year (childbirth, a scare with our son’s breathing). So, if we’re trying to cut the marginal health care expense, should we skip the helmet? Maybe we should skip it - I never see any adults with obviously mis-shapen heads out there, and surely they didn’t all get $300 helmets as kids. Maybe it’s all a racket.
What does "net positive" mean here? I expect the marginal medical spending to always be net positive, but approaching zero as spending increase. Isn't the case being made by Hanson that medical spending on the margin has quite high opportunity cost? To me, this seems like a very defensible claim.
I also expect cutting expenses would preferentially defund ineffective medical interventions over effective ones just by the system naturally "wanting" to help. I expect this to be true for both individual spending (in expectation) and for public spending.
I think it is likely that pretty much everyone has someone within a couple degrees of separation from oneself that is benefitting from the least effective half of current medical spending. In Scott's case, that is his son. Likewise, I think it is likely that pretty much everyone has someone within a couple (or three) degrees of separation from oneself that would benefit from the medical interventions that would have been made possible by doubling the current medical spending. This is a question of opportunity cost.
Rather than fiddling with the dial of how much money is allocated to healthcare, I think the more effective use of relevant political capital (in the US at least) would be to focus on sorting out price ambiguities, perverse incentives, and overwork.
Requiring that hospitals charge everyone the same price for any standard procedure, and spell out that price as part of getting consent, with no slack for unexpected complications, might motivate administrators to look for ways to actually improve efficiency rather than just gouge and obfuscate.
On the sleep-deprivation side of things, how about an overtime formula which adds 50% to hourly pay rates for all medical personnel after every 40 hours worked in a week, and the "week" doesn't officially end (resetting the combo multiplier) until off-duty for 36 consecutive hours? Budget decisions about bringing in enough staff to avoid burnout would get easier.
Is it magic that, "would preferentially defund ineffective medical interventions," or do you know of a mechanism? It seems that you are claiming that there is a mechanism that favors effective over ineffective interventions but only at an absolute spending level that is less than what we currently spend. Seems unlikely that such a mechanism would be tied to absolute funding levels.
I do not think I am claiming what you are claiming that I claim. I think there is a mechanism that (in expectation) preferentially funds effective over ineffective interventions (also) at our current level of spending:
- For individual spending "on a budget": The individual has some idea of which medical interventions have low opportunity cost.
- For public spending under spending cuts: Governments/hospitals/doctors do triaging, governments/hospitals/doctors have some idea of which medical interventions have low opportunity cost.
The case being made (at least in my comment) is that at our current level of spending (it might be the case that) even among the effective medical interventions most have high opportunity cost.
Let me know if the above clarification clears things up or not. If not, I might be misunderstanding something.
EDIT: Maybe I see where I went wrong. See my subcomment.
Hmm, I was thinking about this a bit more and it is indeed more complicated than I made it out to be. Any effective medical intervention would likely be low opportunity cost (the medical condition they fix would have an associated cost and the resources used to cure the condition would unlikely to be better used somewhere else). However, if for every effective medical intervention we defund we also defund 9 ineffective ones (since we don't know exactly but at least have some idea of which ones are effective and which ones are ineffective) then the alternative use of the money saved could (maybe) have a better use.
But isn't this just making the rather banal statement that we spend money very inefficiently (and not just on healthcare). In that case, the political issue that grabs all the headlines (spend less) becomes completely divorced. It is pretty easy to show that there are tons of people out failing to be treated for conditions we can very effectively treat. So, why don't we spend more efficiently and save more lives? Spending wisely is hard!
>I can’t figure out how these claims fit together coherently with what he’s said in the past.
IMO Hanson's sincere opinions are actually only a little outside the mainstream but he likes trolling by saying things exaggerated to the point of ridiculousness. Therefore he is not coherent overall but he apparently respects you and wants you to engage only with his sincere opinions, although he can't say so explicitly.
I recognize that saying someone is often not debating in good faith is a serious accusation, but I used to read a lot of what Hanson wrote (yes, I own his book) and it's the conclusion I came to.
A more polite way of saying that he likes trolling by saying things exaggerated to the point of ridiculousness is ton say that he uses hyperbole for rhetorical effect.
If you're concerned with the "so what?" question, his clear answer is: bundle healthcare with life insurance, to give providers the incentive to keep you alive. There are a couple of kinks with this idea that he's worked out, of course there are practical challenges that we don't know yet how they will play out. But the upshot of it is: this healthcare-insurance bundle provider would have the right incentives to figure out which half of medicine is bad, and which is good (i.e. efficacious).
It seems there is a bigger question, one that Robin does not address at length, about the efficacy of clinical research. I'm not deep into the details, but the replication crisis may affect it. The incentive for pharma after a successful clinical study that leads to FDA approval is to not ever question the results and sell as much of the drug as they can. As far as I understand, they have no liability after approval. Doctors also don't seem to have the right incentives (which I can't describe from experience or with sufficient detail, but from what I understand it also has to do with liability / malpractice lawsuits).
"But the upshot of it is: this healthcare-insurance bundle provider would have the right incentives to figure out which half of medicine is bad, and which is good (i.e. efficacious)."
Ah yes: the infallible market will solve it all by magic, and there is no way it could go wrong. Mastectomy is very efficacious in reducing risk of breast cancer, so everyone (male and female) should have mastectomies!
I think the problem of perverse incentives is the driving force behind pseudo-therapeutic bloat. The question for the medical system isn't whether it's good at producing benefit to patients, but whether it's even designed to do so. A good example is the cancer question. Clearly some cancer treatments work, but just as clearly many are very expensive and do not work. Yet every year NEW expensive therapies that don't work are approved and pushed out to patients, proving this is an ongoing problem, not just an observed phenomenon from years past.
I recommend Vinay Prasad's book "Malignant" for a clear-eyed view of the story in cancer, which in my experience is representative of the systemic problem. It's hard not to conclude cancer research - and much of medical research in general - is more a profit-generating machine for pharmaceutical companies than it is about developing effective therapies.
Now don't get me wrong, the world's medical researchers aren't mad scientists, rubbing our hands together as we watch the money flow. We're really trying to get this thing right. And I think this is part of why the question of efficacy is so difficult to answer for an outsider who isn't involved in helping craft new interventions. Everyone is working to build a better world here, and sometimes we even succeed! But the incentives are driving the car. And the incentives all point to "generate profit from drug sales", without the stabilizing element of "... by selling therapies people want/need."
The problem, IMO, is that the consumer is decoupled from the payer, both in the US and most of Europe. There's an incentive to get drugs approved, not to produce drugs people want for prices they're willing to pay. Here's a good summary from Prasad's YouTube channel:
Relevant quote: "It's a financial transfer of money from all the people being taxed to all the people who own stock in the companies. And the patient is only deriving some small sliver of that money (the cost of the product plus R&D) and may not be deriving any of the benefit. In some senses it's a financial product, with the rare adverse event of improved health outcomes."
I think this is a good description of what's really going on here. The reason it's hard to disaggregate which treatments are effective and which are bunk is because there's a financial incentive to ensure we can't tell the difference.
I mean, the best one would just be to take a population and ban half of them from acessing medical care for about a century. Obviously not feasible to do, of course.
"If I thought medical care was mostly effective and just needed to be trimmed around the margin, and my readers were posting that I thought medical care was “useless”... I would be horrified and try to clear it up as quickly as possible."
Cutting medicine by 50 percent seems quite a bit more than it being "trimmed around the margin", that when Hanson talks about marginal effects, those are pretty big margins. Moreover one of his big claims isn't merely that a large amount of medical utilization is wasteful, but rather that stress, air pollution, diet, exercise, sleep quality, smoking, excess alcohol consumption etc. seem much more important. In which case seemingly flippant comments about healthcare seem fairly reasonable, that on a societal and individual level it seems much more important to emphasize the more robust correlates of health, as opposed to medicine. Also lots of the examples you link seem misleading eg (https://www.lesswrong.com/posts/yEGZcJtPS4L4qsB4f/testing-hanson-s-hypothesis-about-the-uselessness-of-health) and the last 3 tweets, that as best as I can tell they are consistent with someone actually understanding Hanson's views, "2. We over consume healthcare in the USA" does not mean that all healthcare is waste, obviously. etc. etc.
So, relative to my health, I do consume a lot of health care, and much of it has proven to be useless.
A few years ago, my doctor noticed that I had some elevated liver enzymes. After getting this result for several consecutive lab tests with no other explanation (negative for hep, etc) he had me do a liver ultrasound. In that liver ultrasound a polyp on my gallbladder was noticed. Since then he's had me repeat that ultrasound every 6 months for the last 4 years or so, during which time there has been no change in this polyp. I feel fine. Seems like a lot of wasted healthcare to me.
Then my labs started to show an abnormally low thyroid hormone. Again, after a few consecutive labs, he ordered an ANA test lab. That came back positive, and so he did a whole (expensive!) ANA test panel. They all came back negative. So I have unexplained high liver enzymes, unexplained sub-clinical hypothyroidism, unexplained positive ANA, and I feel fine. Thousands of dollars have been spent to tell me that, "Yes, you are indeed fine."
The counter-point to this, I suppose, is that I take a statin for high cholesterol, and if I hadn't learned of that need from similar lab tests, my risk of heart attack would be much greater, so just the fact that you feel fine doesn't mean that there's no medical need for treatment. But I wonder if there's some "marginal healthcare" identification to be achieved by identifying conditions that are extremely unlikely to occur while "feeling fine" and completely ignoring that they even exist unless a person exhibits bodily distress that would indicate those conditions.
+1. It's not fun when you find out all at once that you have stage IV cancer. This kind of ongoing screening ends up finding something in like 5% of people, which makes it well worth it.
I heard that finding it this way still does not allow to treat if, because aggressive cancers are hard to treat, and slow cancers will be found without excessive tests.
Some slow cancers, most famously pancreatic cancer, are known for producing essentially no symptoms until metastasis is well underway (stage III - IV).
Is there a screening test for pancreatic cancer? There is no single diagnostic test that can tell you if you have pancreatic cancer. Definitive diagnosis requires a series of imaging scans, blood tests and biopsy—and those tests are typically only done only if you have symptoms
My understanding: that means basic screening is unlikely to help. Biopsy is invasive, and can cause side-effects for healthy people
"Feeling fine" isn't an indication that you are indeed fine. I'm very glad your health problems never developed into anything serious, but it's understandable that your doctor would be concerned about unusual results with no obvious cause.
Better to have four years of tests that turn out "you're okay" than presume "well it's probably nothing", do nothing, and then whoops turns out it's cancer.
"Gallbladder polyps (GBPs) are generally harmless, but the planning of diagnosis and treatment of the GBP is of clinical importance due to the high mortality risk of delays in the diagnosis of gallbladder carcinomas that show polypoid development.
...Polyp size is the most important predictor for neoplastic polyps. A diameter greater than 10 mm is a risk factor alone. Because of the possibility of malignancy of polypoid formations containing risk factors, surgery should be recommended first and follow-up should be recommended in clinically nonsymptomatic polypoid formations after the differential diagnosis. Especially in non-neoplastic and polypoid lesions below 10 mm, a “watch-and-wait strategy” should be recommended every 6-12 months with USG. "
Yes, as I wrote above it’s a matter of onus. Would you rather save the system or yourself some money and risk missing a serious but treatable condition or carefully follow up on potentially worrisome signs, symptoms and test results based on good medical judgement?
The best Takes are immediately antagonising and difficult to precisely define.
When the Take gets challenged, the Artist starts shapeshifting around each specific argument, and "I never said x" is one of the Artist's favorite defenses.
In the end, "medicine doesn't work" is a million times more successful than "medicine doesn't work as well as people think", so that's the Take that sticks.
I think your assumption that “either we can’t distinguish between good and bad medical interventions or we can’t” is too simplified. You can get around the trilemma if there exists a class of clearly identifiable good treatments and another class of treatments where some are good and some bad, in total they have negative value and they are indistinguishable.
I think the simpler explanation is that we won't in practice distinguish between good & bad treatments that way. If Hanson is right that medicine is considered quasi-sacred and sought ought for reasons of signalling, and that people will flat out ignore information about which hospitals have better outcomes, then expecting people to cut useless medicine while keeping the good stuff is to just misunderstand people's motivation. It's all "good", so people just won't cut it.
We already do distinguish between good and bad treatments that way. Some treatments are compulsory for a hospital to provide if you show up at the hospital needing them. That is, mostly, treatments that are clearly good (and have some other characteristics like urgency). Most treatments aren't like that, in both senses.
The big one I hear about is over-medication, where medication is given, and more medication is given for side effects (sometimes recognized as side effects), and the whole medication load doesn't get evaluated for quite a while.
Does getting the medication regime evaluated count as more medical care or less? Does getting off medications count as medical care, since some of them really need to be tapered carefully?
There also the matter of aggressive end of life care. I don't know whether it's half of medical care, but it's a lot.
I'm with you about the whole medication load not getting evaluated often enough.
About drugs to treat side effects though, I want to speak up for that as not (maybe even mostly) being over-medication. Drugs prescribed to treat primary conditions often have quality-of-life reducing side effects but are necessary for one's health and so the side effect managing drugs are an essential companion for treating the primary condition. Cancer would be the prime example, but we see it with depression and a bunch of other conditions.
It does at one level seem absurd that we have to take drugs to manage the drugs we take, but at another level, it makes total sense and I think should be seen as part of the primary treatment regime. If the hormone suppressing drugs you take to prevent cancer recurrence cause insomnia, then the drugs you take for insomnia are an essential part of the treatment plan.
I don't think you probably disagree with any of this, so thank you for letting me put this in here anyway.
Ah, once again when I ctrl-f for keywords in a point I'd like to make I find the name of Nancy Lebovitz, followed by my point. This is more flattering to me than it is to you. Aren't you a Less Wrong OG?
Anyhow, I just left a similar (but uglier-worded) comment on Robin Hanson's reply to this reply. I really wish everyone could just get together and shout at both of them about the problem not being $300 helmets for newborns but the 4th open-heart surgery for an old fat man on dialysis who is already scheduled for chemotherapy next month.
Putting aside your terrible gun trial analogy, it might be worthwhile to go over some of the correlational evidence that Hanson has alluded to over the years, if you are so hung up about insurance experiments. Here's something from RCA "Objectively read, i.e., studying the actual data instead of weak characterizations of it, the weight of the evidence points strongly against differences in health care access, especially via health insurance, being a significant contributor to broad differences in health outcomes. We don’t find this at an international level amongst broadly comparable countries; we don’t find it domestically between counties, states, MSAs, etc; and we don’t find this with more carefully controlled experimental evidence.
I do not mean this to suggest that health care is totally irrelevant, but rather that the entire developed world gets almost all of the bang for its buck vis-a-vis broad mortality measures in the first few dollars spent. This sort of high impact medicine is broadly similar everywhere in the developed world (antibiotics, vaccines, trauma medicine, etc) and tends to be targeted at cases where its apt to make a real difference (those with obvious need), so the differences in broad averages between places in the developed world largely come down to genetics and lifestyle. The US could probably cut spending in half with little obvious effect on overall life expectancy (not the same as specific outcome measures), but the same is almost certainly true for most of the developed world too!" (https://randomcriticalanalysis.com/2016/11/06/us-life-expectancy-is-below-naive-expectations-mostly-because-it-economically-outperforms/)
What did you think was terrible about the gun trial analogy? I thought it was pretty apt--let's take a case where we already know the effect of an 'intervention' and imagine we were measuring that effect using either the controlled intervention-specific trials used in medicine or the population-level varying incentive trials of economics, where many additional variables intervene so that statistical power is reduced and inferential distance is increased. The one point of breakdown I see is that 'gun insurance' would be analogous to, like, cardiac stent insurance or something, because medicine comprises many more and varied things than does gun ownership. But that only strengthens the case that a null result in a hypothetical gun insurance voucher trial wouldn't surprise you even if specific interventions (like getting shot with a gun) have clear and knowable effects.
Did you even read Scott's article? Scott didn't disagree with this, "The US could probably cut spending in half with little obvious effect on overall life expectancy," and he very clearly made the case that this is 'obviously not true' for all methods of deciding which "half."
If we assume that the value of healthcare is, on average across the whole spectrum of healthcare solutions, neutral (some harm, some help), then we should be able to cut it in half by the first letter of the therapy/medicine and still end up on average net neutral. We can then cut that in half again, and again, to the limit of cutting it all and still have net neutral medicine.
I think what Hanson is trying to claim is that we have data that suggests that people who have more money *do* have better outcomes (in aggregate) even after you tease out medical spending, so we should be willing to make the trade of "just remove all medicine and have people spend more of their money on other things, and we'll end up net positive in terms of health outcomes."
While one *could* do what Scott suggests and try to separate the good treatments from the bad, we have already tried this (e.g., FDA) and it doesn't seem to work (at least not today). So rather than trying to solve the same problem in the same way (decide what medicine is good and what is bad in a centralized manner), we should just scrap all of medicine, let people be healthier because they have more money, and then separately maybe experiment with some new novel systems (in controlled/small settings) that are incentive aligned for health outcomes (like his health insurance ideas).
The trouble is that biology is complicated. A common test for one patient will do no good because they really don't have anything wrong with them, but for another patient it will uncover the problem. Since you don't know which patient is which, you order tests for both patients.
Sure, people have a lot of unnecessary tests and procedures which probably don't do anything for their general state of health, but US medicine has the latest drugs and tech, the population expects a lot of intervention to sort out what ails them, and health insurance is there to pay for it (in theory).
Take gastroenteritis - most people will get better at home, but if you need the sick cert for work, you have to go to the doctor. And just maybe it's something more serious, so you should go. It would be easy to say "Okay, we should cut out all trips to the doctor just because your stomach is sick and you've been vomiting" - until it turns out that it's a symptom of something more serious.
I fully agree that some therapies are useful. I also fully agree that some therapies are useful for some people and not useful for others.
The argument here though is that the system we have for figuring out which is which (i.e., the FDA) has resulted in an aggregate outcome where you are not obviously better off blindly accepting all healthcare the system recommends than if you accept no healthcare the system recommends. Since healthcare costs are very far from free, it is possible that the opportunity cost of having a healthcare system *at all* ends up net negative for society.
An individual can certainly do better than the aggregate by doing their own research and picking out the good healthcare advice and ignoring the bad, but the vast majority of people either get "no healthcare" or "blindly get all the healthcare".
When we are talking about healthcare for all of society, we need to look at the system as a whole and its aggregate outcomes, not whether the system has produced any individual good/bad outcomes. The system we have in place isn't working, so the argument is that we should get rid of it in favor of the far cheaper system that just does nothing (and achieves a similar aggregate outcome).
One can (and many do, including Scott to some extent in these posts) argue that the current system is not net-negative/net-neutral in aggregate, and I think that is a worthwhile debate. However, even taking Scott's aggregate claims at face value I'm not sure it is worth the cost to society. The quibbling seems to be over very marginal benefits in aggregate (e.g., single digit percentages) while US healthcare spending is measured in double digits of GDP.
My reading of his argument as to why he suggests half is for two reasons:
1. Scrapping the whole thing is less palatable to humans, because humans have a bias towards "doing something" over "doing nothing", and half of something is still something.
2. If you cut it in half, there is a reasonable chance that any method of cutting *other* than random selection will result in aggregate quality of the system improving (even if only marginally). This assumes that we have some mechanism of cutting in half that isn't incentive backward-aligned, as one could certainly imagine some incentive structures that result in medical care actually getting worse with a not-random cutting. e.g., cut antibiotics and keep <insert obviously bad therapy here>.
I thought he clearly said that the RCT evidence he bases his conclusion show that variance in the amount of treatment got didn't seem to be associated with better health, so this justifies cutting within that margin. If there were studies of people prohibited from getting any healthcare, that could potentially justify cutting it all. I believe Tyler Cowen referred to that sort of thing when he brought up "studies of Christian Scientists and Amish". https://marginalrevolution.com/marginalrevolution/2024/04/sunday-assorted-links-466.html Of course, Scott himself has written that the Amish don't get zero medical care either. https://slatestarcodex.com/2020/04/20/the-amish-health-care-system/ Instead they spend 1/5 as much as the average, so Hanson's proposed 1/2 is still less of a cut.
Something interesting about the Amish is that they are spending on healthcare that is output from a different system. I don't think it is impossible to have a healthcare system that is net-positive in aggregate, I just think it is a reasonable argument (debatable whether true or not!) that the modern western system is not such a system.
It is certainly possible I'm misunderstanding Hanson's arguments, but I don't think what you have said here necessarily supports that. One can believe that "cut it all" is reasonable while simultaneously recommending "cut it in half" because that is more palatable to your target audience. Also, one can imagine ways to cut it in half that result in a significant increase in healthcare quality in aggregate, thus one can be hopeful that such a cut is achieved.
You can imagine what's going on in Hanson's head that he's not saying, but I'm saying that he explicitly lays out his 50% as an extrapolation from the variance in those RCTs. So if anyone wants to argue against cutting 100%, that's not countering an argument he actually made.
>I might add that studies of Christian Scientists and Amish also dent one’s faith in the very high value of medical care. That said, in my own life medical care has only done me good.
am I misreading and overlooking a link to a study of Christian Scientists?
Christina the StoryGirl had a horrifying description of what she saw first hand in
I think he's just explaining that he's not going to bother quoting Scott quoting himself when he can instead summarize that quote by "none of which have me saying all medicine is useless on all margins".
So, my uncle was apparently born with a misshapen head (firstborn, difficult birth, pointed head). The doctor just said, here's some oil, massage his head every day. Apparently this worked, and he had a university career and is a great-grandfather, so there don't seem to have been any significant side effects. Sample size 1, of course. I wonder if this is still considered a good treatment, or if it's just for a different kind of case. This would have been in 1934/5, so it's quite plausible that doctors today would be horrified by the idea.
Unfortunately we can no longer ask his mother how she did it.
Robin Hanson is very smart. Smarter than me. And also a very original thinker. So I'm always a bit surprised when he is somewhat regularly wildly, obviously wrong and overconfident in the way that economists seem to be.
The "medicine doesn't do anything because look at this data I cherry picked" is one such weirdly wrong thing. It reminds me of a theoretical physicist who says something like,
according to this model, whenever a particle is observed two new universes are created." I'm like, "I believe that's what the math says and I believe your match is correct. I don't have the knowledge to tell you why it's wrong, but it's simply obviously wrong." Robin is sort of like that guy, but instead it's a rape auction or something.
I think "original" does cover it. Like the old joke about the book review: "This is both original and good, but the parts that are good are not original, and the parts that are original are not good".
We can and do have evidence of *coherent* superpositions, eg. from quantum computing. But that doesn't add up to Many Worlds in the popular sense.
There is an approach to MWI based on coherent superpositions, and a version based on decoherence. These are (for all practical purposes) incompatible opposites, but are treated as interchangeable in Yudkowsky's writings. "The decoherence (a.k.a. many-worlds) version of quantum mechanics ....". Decoherent branches are large, stable, non interacting and irreversible...everything that would be intuitively expected of a "world". But there is no empirical evidence for them , nor are they obviously supported by the core mathematics of quantum mechanics, the Schrödinger equation.
Whatever is going on inside a Quantum computer isn't classical computation (using the same number of particles) as Deustch correctly states. However, it also isn't a set of *decoherent* worlds, since quantum computing works via coherent superpositon. (It also isn't a set of macroscopic, irreversible, objective , etc, worlds). So a QC is neither a single world *nor* multiple worlds, for reasonable definitions of "world". Quantum computing is evidence for quantum m superposition, but simultaneously evidence against the popular idea that irrevocable splitting into non-interacting worlds occurs at every elementary interaction or observation: that would predict that coherent superpositions don't exist, for all practical purposes,.since they instantaneously decohere...and QC therefore doesn't work.
"David Deutsch, one of the founders of quantum computing in
the 1980s, certainly thinks that it would. Though to be fair, Deutsch
thinks the impact would “merely” be psychological – since for him,
quantum mechanics has already proved the existence of parallel uni-
verses! Deutsch is fond of asking questions like the following: if Shor’s
algorithm succeeds in factoring a 3000-digit integer, then where was
the number factored? Where did the computational resources needed
to factor the number come from, if not from some sort of “multiverse”
exponentially bigger than the universe we see? To my mind, Deutsch
seems to be tacitly assuming here that factoring is not in BPP – but
no matter; for purposes of argument, we can certainly grant him that
assumption.
It should surprise no one that Deutsch’s views about this are
far from universally accepted. Many who agree about the possibil-
ity of building quantum computers, and the formalism needed to
describe them, nevertheless disagree that the formalism is best inter-
preted in terms of “parallel universes.” To Deutsch, these people are
simply intellectual wusses – like the churchmen who agreed that the
Copernican system was practically useful, so long as one remembers
that obviously the Earth doesn’t really go around the sun.
So, how do the intellectual wusses respond to the charges?
For one thing, they point out that viewing a quantum computer in
terms of “parallel universes” raises serious difficulties of its own.
In particular, there’s what those condemned to worry about such
things call the “preferred basis problem.” The problem is basically
this: how do we define a “split” between one parallel universe and
another? There are infinitely many ways you could imagine slic-
ing up a quantum state, and it’s not clear why one is better than
another!
One can push the argument further. The key thing that quan-
tum computers rely on for speedups – indeed, the thing that makes
quantum mechanics different from classical probability theory in the
first place – is interference between positive and negative amplitudes.
But to whatever extent different “branches” of the multiverse can
usefully interfere for quantum computing, to that extent they don’t
seem like separate branches at all! I mean, the whole point of inter-
ference is to mix branches together so that they lose their individual
identities. If they retain their identities, then for exactly that reason
we don’t see interference.
Of course, a many-worlder could respond that, in order to lose
their separate identities by interfering with each other, the branches
had to be there in the first place! And the argument could go on
(indeed, has gone on) for quite a while.
Rather than take sides in this fraught, fascinating, but perhaps
Because the math doesn't say anything about other universes, its only job is to help you predict experimental results. All that "interpretation" stuff is philosophy, which may or may not be "correct" in some sense, but it most certainly isn't straightforwardly implied by our current math.
People who try to convince me of--let's call them Quantum World curiosities--I think are doing card tricks so well that they fool themselves. That is card I was thinking of, that is my signature in an unopened pack of playing cards. It somehow got into my jacket pocket. I am astonished but I don't believe you are doing magic.
Almost every reasonable sounding and surprising "infinite worlds" theory still ends up somewhere like this:
Me: So there's infinite worlds?
Quantum Theorist: Yes.
Me: That means a universe exactly like this one exists and nothing else is different except on election day a majority of Americans will write in my name on their Presidential ballots and I will become President despite being almost totally anonymous?
Theorist: Infinity is infinity. So every possible thing that *could* happen *will* happen. There's also a universe where Margot Robbie invents teleportation and the first thing she does is teleport into your birthday cake naked. Yes.
Me: I don't know exactly why that's wrong but I know that it's wrong.
Maybe there's some sort of quibble where there's not INFINITE universes or it's one infinitely big universe or whatever. But as far as I can tell every one of those situations leads me to a place where informed horse sense allows me to be confident that the math is incomplete. The same way I know that medicine has a positive impact without studying it. If the publication is rigorous and it says vaccines don't move the needle, I don't know how they got it wrong, but I know it's wrong.
1. The math doesn't say that. 2. There seems to be significant confusion as to how math is used in physics. 3. Considering how absurd the many worlds theory is, it is almost nitpicking to point out that it doesn't even give the experimentally observed probabilities. 4. There are theories that agree with experiment and are not absurd like many worlds or unprofessionally vague like Copenhagen.
Robin doesn't seem to be selectively cherry-picking studies that fit his conclusion, rather he's got criteria for taking studies seriously that eliminate most without further discretion on his part.
Observation is only important under the Copenhagen interpretation, whereas I believe Robin rejects that in favor of the Everett interpretation in which universes are created regardless of observation.
I think the same as you about Hanson, except that... of course he might be right about something.
I have a couple of ways of reminding myself how wrong I must be. One is religion: I'm not religious, but that means I am disagreeing with most of the brilliant thinkers in Europe that I revere over the last 2,000 years. If I am certain that Descartes was wrong about god, then it's certain that I'm wrong about much, much more.
Another is an idea from a philosopher I like called Eric Schwitzgebel, that he sometimes calls crazyism. He looks at lots of philosophical questions which remain open, things like dualism, infinite universes, the objective existence of morality, and so on. He notes that if you pick any possible answer to these questions and work out all of the implications of that answer, you always end up with at least some conclusions that are "crazy". And yet, one of these answers must be true! Whatever the truth of these unknowns, it's going to mean something crazy. Therefore, the fact that an idea is crazy, or has crazy implications, should not be seen as an argument against it.
Worst case scenario: Hanson is an interesting thinker who has spewed out a lot of interesting ideas into the world, and they're all wrong. That still makes him massively valuable, simply for generating and elaborating these ideas. We shouldn't look to public intellectuals to be *right*. Often, that's not their job. Their job is to make and explore new ideas.
If Hanson said "modern medicine is useless" then I agree, that would be a dumb thing to say.
The problem is, Hanson has said so many dumb things in the past, it's easy to believe he *did* say this as well.
"Also, we can look today at how the death rates of individual households correlates with the water sources and sewer mechanisms used by those households. "
I would love to dump this fool down into the house I was reared in as a child, the same kind of house that other neighbours and family members were raised in, and let him see if it makes a different to have running water and sewerage. When was the last time he took clothing out of a wardrobe to find it covered in mildew, for example? Frost on the *inside* of the window pane? You get your drinking water from a pump in a field? You can't ask my parents now since they're dead, but their siblings who died young from "we just don't know what medicine is good or bad, or why some people live and some people die", is that it, Robin?
How does anyone take this man seriously? *Why* does anyone take this man seriously?
It's useful to look at the rest of the paper and not just the line you quoted:
"Most of the obvious theories have serious problems, you see. For example, exercise,
smoking, social status, and urban living appear to have large effects on individuals. But the time trends for exercise, urban living, and smoking have been in the wrong direction - those trends would predict decreasing lifespan. And since social status is usually thought to be relative to contemporaries, it is hard to see how average social status can increase with time.
Finally, while the biggest advances in nutrition, medicine, and public health seem to
have occured during the first two thirds of the twentieth century, death rates have fallen
just as fast during the last third of the twentieth century. Perhaps some new influence rose in importance just as those other influences became less important, but if so it seems a remarkable coincidence that the total rate of improvement has remained pretty steady."
Hanson's point is that if improved water/sewer infrastructure was the key to longer living, lifespan should increase after these improvements and then level out. Yet lifespan keeps increasing at a steady rate generations later.
"Hanson's point is that if improved water/sewer infrastructure was the key to longer living, lifespan should increase after these improvements and then level out. Yet lifespan keeps increasing at a steady rate generations later."
When was the last time you heard of someone dying of cholera? It used to be very prevalent:
"Typhus was not the only disease to kill tens of thousands of people in Ireland during those decades. More than 50,000 people died from cholera during the 1830s, according to the 1841 census report, with more than 46,000 of those deaths occurring in three years of epidemic, between 1832 and 1834.
Cholera is a bacterial infection caused by water and food contamination. Acute, sudden diarrhoea can cause severe hydration and death. The World Health Organizastion primarily links cholera with poor access to clean water and proper sanitation.
The census report from 1941 [sic] notes: “Cholera prevailed most in the towns of the civic districts; and ... in 1833 it had spread throughout the country at large, and prevailed most in the rural districts.”
Dublin had many cases. A news story in the Belfast Newsletter, April 3rd, 1832, detailed a Dublin case before the central board of health, before which “reports were laid before them of the deaths of a man and two women, in the neighbourhood of Summer-hill, and of the sickness of a young woman who had been at the wake and funeral of the man”.
The census report identifies Sligo, Drogheda and Belfast as hotbeds for the disease. The 1832 outbreak in Sligo has since been associated with the works of Bram Stoker, whose mother apparently related vivid descriptions of the devastation to her son. A report in the Tralee Mercury on September 1st, 1832, also notes some "melancholy details" of the illness, which "broke out with fearful violence" in the town: "In almost every case where the sick person died in his own house, three or four of the same family have been attacked, thus paralysing every exertion and rendering them incapable of any effort to subdue the virulence of the plague."
Cholera, that decade, accounted for one in every 7.36 deaths from all causes; by comparison, influenza killed 10,575 people, or one in 112.28 deaths."
Typhus and typhoid fever are also spread due to poor sanitation. People died from those in Ireland, and from diphtheria:
"Diptheria is caused by a bacterium called Corynebacterium diphtheria and is transmitted in respiratory droplets. Diptheria was once a highly significant illness and a major cause of death in children. Diptheria death rates range between about 20% for children under 5 and people over 40, with a slightly lower rate of between 5-10% if the infected person is aged between 5-40. Symptoms include loss of appetite, sore throat and fever. In 1948, 521 cases of diphtheria were reported in Ireland and due to the introduction of the combination diphtheria/tetanus/pertussis (DTP) vaccine in 1952/1953, the case numbers declined progressively until none were reported in 1968. After a gap of nearly 50 years, one case was reported in 2015 and another in 2016."
Vaccines are a great boon, but it's easier to be hygienic when you have a reliable source of plentiful, clean water and sewerage system to take foul water and waste away. Contamination of water sources still kills, even in developed countries:
"“Water sources in Ethiopia are regulated, but private wells in Ireland are not,” said Dr Hynds.
In his first PhD, he looked at hundreds of wells across Ireland, and during nearly two years of taking a sample every month, he got to know many of the families.
“One of these wells belonged to a dairy farming family. Their grandchild got sick from water from the well. He got hemolytic uremic syndrome from a STEC infection,” he said at a recent UCD Earth Institute event on the Hidden Hazards of Unregulated Private Wells.
“He then went into renal failure and got a 400pc blood transfusion. Every drop of blood in his body was replaced four times and he died.
“So in a country like Ireland, a rich country, we had a child die from a water-borne infection, and that was not on the Six One News.
“I’m an epidemiologist, and in this day and age, this is simply not good enough.”."
Hanson is like someone claiming that "You say these Irish peasants are dying of 'famine' by 'starvation', and yet once they get fed, we see that people are still dying afterwards! So your obvious theory has a serious problem!"
"Finally, while the biggest advances in nutrition, medicine, and public health seem to have occurred during the first two thirds of the twentieth century, death rates have fallen just as fast during the last third of the twentieth century. Perhaps some new influence rose in importance just as those other influences became less important, but if so it seems a remarkable coincidence that the total rate of improvement has remained pretty steady."
That's the low-hanging fruit being picked; first we get huge boosts from solving obvious problems like poor sanitation, lack of effective medicine, and insufficient knowledge. Then, as we clear those problems out of the way, we learn more, improve more, and build on the successes of the past. Now you're much less likely to die before you reach your fifth birthday, and since there is now better medicine, more access to public health services, education, and all kinds of attempts to motivate the general public to eat healthy, exercise more, stop smoking and drinking and engaging in risky behaviour, these too pay off. Healthier people have healthier children who live in healthier conditions and are exposed to less dangerous factors at work, so there's a slowly improving increase in lifespan. We're not roofing our houses with asbestos and painting the walls with lead paint anymore, Robin.
Also, hmmm, lemme think - was there anything in particular that happened in the first two-thirds of the twentieth century that might have bumped up the death rate? Maybe even happened twice? Like some global martial phenomenon? Nope, can't recall anything of that sort!
I had the same reaction as you when I first read this. What kind of moron claims water quality doesn't affect health outcomes? Look at the work John Snow did on Cholera in England during the mid-19th century. At the broad street pump site, there were two buildings that had private wells, and the mortality rate was a huge margin lower than the surrounding buildings. The chart with the residents of two different water companies was also critically illuminating; the company which piped water upstream (clean) had Cholera rates less than the London average, yet the company that piped downstream (sewage contaminated) had rates many times higher.
After reading Hanson's reply to this post, his position seems to be that medicine works but half of it is junk and the medical establishment is incapable of separating the two. Also, I think the point of the paper quoted above is the question about why lifespan continues to increase at a continuous rate despite the factors supposedly influencing it happening at discrete intervals. My position on Hanson has gone from 'moron' to 'bad at communicating what he really means.' Although in fairness, I never heard about him before these posts by Scott, so maybe that lens makes things seem worse than they are.
As a final note, I hope the excess mortality from war is obvious enough that people exclude it from their data on life expectancy.
For some conditions, we have treatments that are really, really effective. Antibiotics for example.
I guess I could entertain a claim that doctors try to treat everything, and the distribution has quite a tail of things where they actually don't have an effective treatment.
I was once in hospital (outpatient only) where the consultant was reduced to making up a diagnosis with "idiopathic" in its name. This is like, sure it looks like there's something wrong with you, but we have just run off the edge of our diagnostic flowchart here, and are in the land of rare diseases. NB: i am not criticizing the doctor here .. i am quite in favour of being given an honest "no idea what this is" is that's what the doc is actually thinking.
And as for random inexplicable stuff .. one time, I get severe pain in the extremities, like in my fingertips. it Felt like the sort of thing that deserves immediate medical attention. (I guess something with similar symptoms is the onset of diabetes and you're getting crystals in the extremities. Tests reveal it was not diabetes.) Doctor is like, "have you had a recent tetanus shot? here, have a tetanus shot right now.") Symptoms went away after a few days. No idea what it was. Shrug.
Total cost of that was presumably a tetanus vaccination, a blood glucose test, and ten minutes of a doctor's time, to (probably) no net benefit to the patient.
>That helmet is probably our “marginal” health care expense, in the sense that it’s less obviously important than the other two things we’ve used healthcare for this year (childbirth, a scare with our son’s breathing).
I think this one parenthetical here contains one of the clearest examples of a major medical expense we probably *could* cut with minimal loss. Child delivery is *much* more expensive in the United States than most countries (https://www.statista.com/statistics/312026/total-hospital-and-physician-costs-of-a-normal-delivery-by-country/). How much safety are we buying with that extra expense? Well, the CIA factbook figures on maternal mortality rank the US second-to-last out of the selected countries (https://www.cia.gov/the-world-factbook/field/maternal-mortality-ratio/country-comparison/) better than South Africa, but behind Kazakhstan, Chile or the United Arab Emirates. And it's not like poor countries simply get more bang for their buck due to lower costs of labor, costs in Germany are lower than costs in Chile.
Unfortunately, as a consumer, I don't think you have the leeway to buy the services/price package of German child delivery. But if we looked at questions like "why does American child delivery cost so much more than German child delivery?" I think we might be digging into a pretty rich seam of low value marginal dollars spent on medical care.
Maybe one of my disagreements with Robin is that I doubt the medical waste forms a clear category that patients can just opt out of. For example, during postpartum, nurses came to check up on us a bunch of times. My wife was very tired and asked them not to do that. They said they had to (I assume this was a legal liability thing).
I don't think Hanson said "We should cut spending by getting individual patients to opt out of treatment". But if, say, we taxed rather than subsidized medical care, many such practices would go away.
"Unfortunately, even if you believe everything that I have said, your behavior will probably not change much as a result. You will still spend nearly as much on medicine for yourself and your family, and spend much less effort on the more effective ways to increase lifespan. After all, your sick family would consider it the worst kind of betrayal if you did not “do something,” and give them all the medicine that your doctor recommends (Hanson, 2002). Alas, the problem of the fear of death muddling our thinking is so much worse than we imagined."
This is one of the places I think Hanson is really off-base in his views on medicine; I think if you cut out all signalling-based medical expenditures, it'd probably only cover a little bit of the distance between US healthcare spending, and spending in countries like, say, Kazakhstan. I think a larger portion of the price tag probably comes from CYOA procedures based to limit legal liability. You and your wife didn't deliberately seek out a hospital which would send nurses to check up on your wife repeatedly despite your expressed wishes because you wanted to signal the intensity of your commitment to your children. The hospital insisted because they wanted to minimize their risk of being sued.
I think it is fun to think through this. The patient finds this annoying, but perhaps it is net helpful. You never know when you are going to catch a furtive co-sleeper (or something).
Or perhaps this is useless with healthy patients, but it's difficult to train nurses to only check on red-flagged rooms with consistency and net positive to just say "check them all". So the healthy patient pays the price to on net help the unhealthy one.
Or perhaps this causes healthy patients to prefer to leave AMA (like we did) in order to avoid the constant intrusion, but there are a group of healthy patients where maybe something happens and it would have been better for them to stay at the hospital. So on net it goes back negative compared to "only check on red-flaggers".
Or maybe... four more levels down. Is someone going to do a RCT on this? Will that RCT be designed in the same way this hospital's training and staffing levels are designed? Will following that RCT from some other hospital during normal staffing work the same in this hospital dealing with understaffing? Or from a nurse's supervisor who just doesn't care? Or from a local wave of furtive co-sleepers?
I don't know, but fun to think about. How do you collect evidence in a world with ethics advisory boards and administrators that are just trying to survive one day to the next and maybe were never really a facts person to begin with?
"However, these figures are completely wrong, and they have been known to be wrong for many years now. The U.S. National Center for Health Statistics, the branch of the Centers for Disease Control (CDC) charged with collating health and vital statistics, has published three separate reports elaborating in excruciating detail on one crucial fact about U.S. maternal mortality: It is measured in a vastly more expansive way than anywhere else in the world.
As a result, U.S. maternal mortality is overestimated by two to three times. Properly measured, the real U.S. maternal mortality rate in 2019 was 9.9 maternal deaths per 100,000 births, which would put it at 36th place—still not impressive by comparison, but somewhat better than Canada and a bit worse than Finland or the United Kingdom.
...
Historically, the United States and most countries have tracked maternal mortality using data based on the cause of death listed on death certificates. When a person dies and the cause is assessed by an examiner of some kind, certain “maternal” causes, such as “eclampsia” or “obstetric trauma,” are commonly tracked. If a woman has died due to one of these pregnancy-related causes, she is listed as a maternal death. This process is fairly straightforward and has been widely adopted across many countries.
But in 2003, the U.S. CDC decided to launch an improved death certificate form. Among the various changes proposed was the addition of a checkbox, wherein whoever filled out the paperwork would identify if the deceased had been pregnant in the last 42 days or the last year. The reason for this checkbox was that the CDC believed (correctly, as it turns out) that in only measuring “maternal causes of death,” it might be underestimating the true health hazards of pregnancy. Pregnancy might alter the course of other diseases and conditions or interact with them in important ways.
The CDC anticipated that the checkbox would increase measured maternal deaths; it did not anticipate just how much it would increase them. As it happens, the CDC’s own reporting, which I have confirmed elsewhere, shows that the addition of the checkbox approximately doubled maternal mortality rates.
You might think a sudden doubling in maternal death rates would be obviously flagged as a data issue to correct, but this turns out not to be so. Because the United States has a federal system, individual states added the checkbox in different years. While individual state maternal deaths showed sharp level shifts, the national maternal death count drifted upward gradually as states added checkboxes to their death certificates: California in 2003, Florida in 2005, Texas in 2006, Ohio in 2007, Tennessee in 2012, etc.
In 2018, further modifications were made to the data-processing protocols used by the National Center for Health Statistics for pregnancy-related checkbox deaths, leading to more thorough inclusion of them. The result was a massive but gradual artificial inflation of maternal mortality.
This doesn’t mean that the American way of measuring death is wrong. It’s just quite different from the countries that it’s being compared to...
But the U.S. case is particularly beguiling, since the United States now tracks all deaths of women who were pregnant, not only women who gave birth. Women who miscarried early or had abortions—whether officially reported or not—are also counted in the checkbox method. As a result, the United States may be the only country in the world where central vital records systems track all pregnancy-related mortality, not just maternal mortality."
This does a lot to explain why listed US outcomes are worse than countries like Kazakhstan or Chile, but it still leaves the US with similar outcomes to countries with similar standards of living (better than some, worse than many) despite spending dramatically more money on childbirth services. Even without the US doing so anomalously poorly in its actual outcomes, the figures suggest a lot of dead weight spending.
Cut all antibiotic use for ear infection in children or cut it in half. To paraphrase Robin Hanson "Ear infection treatment is not about curing ear infections".
One type of ear infection in children accounts for 10% of all antibiotic use in Iceland (From the paper 20% of antibiotic use in Iceland is consumed by children under 7 of which 50% is used to treat a type of ear infection.) Relative to other Nords Iceland has a larger problem with antibiotic resistance. - https://www.ncbi.nlm.nih.gov/pmc/articles/PMC3442319/.
Instead ear infection treatment is about -
...“the lack of stable doctor–patient relationships due to lack of continuity in medical care. Pressure from patients in a stressful society, the physician's work pressure, the physician's own personality, particularly the earnings incentive and service mentality and, last but not least, the physician's lack of confidence and uncertainty, resulting in use of antibiotic prescriptions as a coping strategy in an uncomfortable situation” Petursson P. GPs’ reasons for “non-pharmacological” prescribing of antibiotics: A phenomenological study. Scand J Prim Health Care. 2005;23:120–5.
Have you ever had an ear infection? I'm not trolling, it's a serious question.
The pain is.... I don't know what to compare it to, I've never experienced anything even close. And it was a bacterial infection, and maybe you're telling me it would have just cleared by itself, or maybe I'd go with what the doctor told me, that it was quite dangerous because of its proximity to brain, here's a prescription, you should notice an improvement withing 24-48 hrs., and I did. Take the pills and noticed an improvement.
I do not recall if I have had an ear infection. Probably. Given five out of six children have at least one ear infection before their third birthday. I may have had one but have no a memory out it.
There is a difference between your personal experience and what makes for good medical policy.
Does an ear infection cased by virus feel different than an ear infection caused by bacteria?
It my understanding that in 80% of the cases a virus is the cause of ear infection in children. To confirm bacterial infection requires a test. Did you receive a test?
Most middle ear infections resolve in 2 to 3 days.
Depending on when you went to a doctor it could be that it cleared on its own, if no bacterial test was taken it could have been a virus. If it was a virus the antibiotics were likely useless.
Did you finish the course antibiotics or did you stop taking them after you felt better?
Many people stop taking antibiotics before finishing their course of antibiotics. Failing to finish antibiotics increases the risk of antibiotic resistant strains of bacteria developing.
In Sweden the average child has about 5 courses of antibiotics in childhood while in American it is closer to 12 courses of antibiotics. Swedish children have lower levels of childhood obesity some have suggested that part of this may be due to over prescription of antibiotics in children. The conjecture relates to the effect antibiotics use has in the fattening of livestock.
Have you or a loved one died from a antibiotic resistant infection? At the margins a child given unnecessary antibiotics does not increase the likelihood. A policy that over medicates with antibiotics in aggregate does lead to increased death due to antibiotic resistant strains of bacteria.
I…. am not arguing against antibiotics overuse being the case? But you proposed “cut all antibiotics” for ear infections, and I don’t think it’s a medically defensible position. Some ear infections are viral, some are bacterial.
FWIW yes mine was bacterial, yes I finished my bottle, yes I had a friend of a friend “almost” die of a resistant infection (2-months stay in a hospital, intubated at some point).
I agree. I should have said cut all antibiotic use for ear infections in children without a test to confirm bacterial infection. Given that bacterial test takes one to five days. That most middle ear infections resolve in 2 to 3 days. That viral infection is the leading cause of ear infection in children and that getting a followup appointment can be difficult to go over test results. In practice it may amount to the same thing. Also I softened my initial assertion "or cut it in half" based on the lower occurrence of antibiotic use in Swedes during childhood compared to Americans and the two populations are about as healthy.
Even if the following were true "More people are harmed by drug resistance bacteria caused by the current level of prescription of antibiotics for child ear infection than are helped by antibiotics for ear infection" that still isn't medically defensible argument for cutting all antibiotic use for ear infection in children. Medicine concerns itself with the patient not the society. This was attempt to argue the marginal utility of medicine in this one narrow avenue to society as it practiced may be zero or near zero.
“ Medicine concerns itself with the patient not the society.”
Couldn’t have said it any better! I think you hit on the key problem with Hanson’s arguments, and why in general it’s so hard to cut medicine. We go to great lengths to reduce suffering of our loved ones, rationality be damned.
Mate, I had an ear infection that I didn't get antibiotics for, as a young adult, and I was *crying* with the pain of it and dosing myself up on over-the-counter painkillers.
Imagine a young child screaming with pain. Are you going to be the parent that goes "Just tough it out, Junior, we're not going to the doctor for something this trivial?"
Even if they're only doing it for "I can't take another day of the screaming for 24 hours solid, please make it stop doctor" reasons, I think those are valid.
Not all ear infections are that bad, and the old-fashioned treatment used to be warm up some olive oil (which was obtained from the chemist), pour a little in the ear, plug it with cotton wool and let it act that way. Presumably the heat killed the infection. You can also now buy over the counter ear drops. But that really bad ear infection? I was lancing pus out of the ear every night for a week and believe me, warm olive oil would not have cut it.
What a parent should do for a sick child is different than what good medical policy should recommend with respect to antibiotic use for ear infection.
I would not presume to know the best course of action for parents in a given circumstance.
What I do know is that about 80% of all ear infections have viral causes. In most cases doing nothing would lead to resolution, folk remedies would be just as effective as antibiotics as would placebo. It is my understanding that antibiotics do not have antiviral action.
To confirm bacterial infection requires a test. A bacterial culture test can take between one to five days. Most middle ear infections resolve in 2 to 3 days.
If one has a policy to give all children antibiotics without regard to the nature of the infection you risk doing more harm than good. The paper I linked to talks about that effect in one population. It also has a link to recommendations for when to give antibiotics by referencing a guideline. In the practice of medicine these guidelines are often ignored. https://www.nice.org.uk/guidance/ng91
I am aware that antibiotics do not have any effect on viruses. I am aware of over-prescription. But a blanket ban on "we won't prescribe antibiotics" isn't going to improve health, either, because sometimes it *is* a bacterial infection, it's bad, and it does need treatment:
"The body’s immune system can often fight off middle ear infection on its own. Antibiotics are sometimes not needed for middle ear infections. However, severe middle ear infections or infections that last longer than 2–3 days need antibiotics right away."
'Parents pressure doctors to over-prescribe' is true. But so is 'the child has been sick for over 3 days and pus is oozing out of the ear'.
We're in agreement that there should be guidelines, and that there is often social pressure to prescribe antibiotics when those do no good. But there is the lovely shiny theory of "let us reduce over-prescription by issuing guidelines all doctors should stick to" and the reality of "frazzled parent with screaming child turns up in your surgery looking for something". I am not surprised a doctor might think "Well, maybe it is bacterial, writing a prescription won't hurt".
I would agree that a blanket ban on antibiotic prescriptions won't improve health. Which is why I did not make that assertion. I am weakly confident that some kind of prohibition, ban, tax, disincentive or policy enforcement that results in the cutting in half the average childhood prescriptions in the US from 12 to the Swedish average of 5 would not make American children less healthy. I am weakly confident it may do good on net for overall societal health and from the Icelandic paper even some children themselves .
It has been thirty years since the effects of over prescription of antibiotics have been warned about in the literature. At least two generations of doctors have now been educated knowing this to be the case. Yet the problem persists.
We are only now beginning to engage in research projects on how to stop doctors from continuing to engage in a practice, that they know most of the time does not work. Watchful waiting is a meaningless platitude once the patient parents demand a doctor do something.
This is a small issue where the problem of how medicine actually practiced is broken. Where we know it is broken and we know how to fix it have known for quite sometime but have had limited success. I do not believe this is an isolated issue. Incentives innate in medicine lead to wasteful use of medicine. Even when we know which parts are bad we can't cut them.
I don’t how much but a lot of medical spending is done to rule out worrisome conditions. Tests with negative results will dilute the medical effectiveness of the money spent testing. Eliminating these tests would be a net positive in theory except that you can’t know which cases to desist from testing in advance. The onus (justifiably) is to maximize health rather than save money.
I feel like the way to square the circle here (though I don't think this is Robin's stated position from the quotes) is that a lot of the end of life interventions we make are probably of very low value (both realized and in expectation) and are sometimes instituted by families trying to show they care. Related: Atul Gawande's "Being Mortal" does a good job covering why the way we approach death medically is likely neutral / negative for patients and very costly.
His wife is a hospice nurse, so I believe he has discussed that specifically. Aside from signalling as an explanation, I recall that he's also written that we just have an aversion to thinking about death.
There's an additional dimension not in conversation here which is where treatments for some conditions have unequivocally improved outcomes over time. Nature: "The number of people who die after a breast cancer diagnosis has decreased by two-thirds since the 1990s, a study of more than half a million women in England has shown."
I think medicine should be spending more on the kinds of lifestyle changes that make a huge difference. On cancer recurrence for instance, there's good research showing exercise and diet make a huge difference to reducing recurrence rates. We spend a quarter of a million dollars treating someone for cancer and then spend nothing on making sure they do the things that prevent more cost later and that make a huge difference to mortality rates.
But exhorting people to make these lifestyle changes doesn't work. What's it look like to take these numbers seriously and provide the kind of healthcare system support for diet and exercise that we do for getting someone through chemo? The thing is, it won't be money to pharmaceutical companies, it would be for salaries for huge teams of coaches, fitness instructors, dieticians, and recreational facilities. What's it look like to not treat these as optional add-on things but as a central part of the medical treatment?
Instead of bemoaning people's unwillingness to make the kind of health-oriented changes that make a difference, I think we should be incorporating them more into medicine as part of treatment. And then medicine would be even more effective for the people it's already spending money on.
"Public health measures? Sanitation? Clean water? Robin again:
If medicine for treating individuals is not quite the miracle we have heard, does public health make up the difference? Have not we all heard how the introduction of modern water and sewer systems greatly improved our ancestors’ health? Well, a century ago the U.S. cities with the most advanced water and sewer systems had higher death rates than the other cities. Also, we can look today at how the death rates of individual households correlates with the water sources and sewer mechanisms used by those households. Even in poor countries with high death rates, once we control for a few other variables like social status we usually find that water and sewer parameters are unrelated to death rates. Well we must live longer now for some reason, right? Yes . . . but the fact is that we just do not know why we now live so much longer.
(Robin seems more bullish on sanitation elsewhere, so maybe he’s changed his mind? Like I said, I have trouble fitting all his statements into a coherent model.)"
End quote.
One of the issues in this debate (which Hanson has observed in the past) is that there doesn't seem to be a widely agreed-upon definition of "medicine". Personally, I don't really think that civil engineers and site work contractors designing and building modern sewer infrastructure counts as medicine.
This is especially true of "public health", which, we learned in the pandemic, can be defined to include basically everything that human beings do. In my region, local health inspectors (previously considered minor bureaucrats who inspected restaurant grease traps) turned out to have broad dictatorial powers to shut down businesses, churches, and schools, and to make extremely granular rules about how people could behave.
If "public health" is part of "medicine", and if it covers basically all human activity, then medicine is actually responsible for all of the gains in life expectancy that have occurred over the past few centuries.
> In my region, local health inspectors (previously considered minor bureaucrats who inspected restaurant grease traps) turned out to have broad dictatorial powers to shut down businesses, churches, and schools, and to make extremely granular rules about how people could behave.
Bret Devereaux had a series on the structure of the government of Rome, which included the fun observation that the tribune of the plebs was an important position since quite early on with well-understood powers that were used in traditional ways. This lasted for centuries.
Then one guy was elected to the position and used the powers - which everyone agreed on - in nontraditional ways. At that point, the system broke down.
"Well, a century ago the U.S. cities with the most advanced water and sewer systems had higher death rates than the other cities."
Gosh durn it, this is the kind of careless statement that makes me go "Sources!" Where is he getting these figures, does he explain?
Because if he was writing in 2002, "a century ago" would be 1902 and I'm betting the "cities with the most advanced systems" were also the largest, therefore the higher death rates. I'm just pulling that out of my backside, but since we seem to be in the land of "simply make an airy claim", then I'm as justified.
Some stats nerd in 1906, God bless 'em, did a whole "The General Death-Rate of Large American Cities, 1871-1904: [Tables for Individual Cities]". Take it away, Mr. Frederick L. Hoffman:
Taking the 1902 figures, we find a few samples (I could give a lot more):
Mortality Rate per 1,000
New York City (Manhattan, Bronx) 19.2
Chicago 14.3
Philadelphia 18.1
New York City (Brooklyn Borough) 17.8
St Louis 18.3
Boston 18.7
Cleveland 15.9
Buffalo 14.6
San Francisco 19.9
Cincinnati 17.3
Pittsburg 21.3
Detroit 15.6
He even gives us a handy comparison between races:
Baltimore – white population 17.5
Baltimore – black (“coloured”) population 29.5
New Orleans - white population 18.8
New Orleans – coloured population 30.3
Washington D.C.
District of Columbia – white population 16.3
District of Columbia – coloured population 30.1
Wow, just look at the difference in rates there, all due to the "coloured" people having such over-abundance of advanced water and sewer systems! Right, Robin?
So how do these rates stack up against US mortality at the time? Well, again bless the stats nerds of the US Census because they gave us Life Tables for 1890 to 1910:
Total (reported) deaths for 1902 were 318,636 on a population estimated to be 20,408,867 in 1901 which - if my shaky maths can be believed - is a rate of 15.6 per 1,000. The cities aren't doing too bad by comparison with that, I think; some are higher, some are lower, but the figure isn't all deaths, just reported deaths, and covers the entire country where some centres of population are going to be much smaller than New York and Chicago.
Yeah, I would still like to see where Hanson is getting his comparison figures for "cities with better sewerage versus cities with worse and comparable death rates".
Seriously, this is a fascinating account, and *may* give some colour to Hanson's argument, on the grounds that improved facilities led to demands for luxuries such as indoor lavatories, but the existing sewerage systems were inadequate to deal with the new demand (and output of waste). So yes, "better plumbing" did mean "more disease" BUT only until the city officials caught up with the necessity for "better sewerage" to go along with the better plumbing. So I can't agree with the totality of Hanson's comment there.
"Population Growth
During the nineteenth century, there was considerable urban population growth in the United States. In 1820, less than 5 percent of all Americans lived in urban areas (cities with a population larger than 8,000), but by 1860 the percentage increased to 16 percent and by 1880 had risen to 22.5 percent. From 1820 to 1880, most major cities in the United States experienced considerable growth. For example, during this time Boston’s population increased eightfold, New York City’s tenfold, Philadelphia’s thirteen fold, and Washington, D.C.’s fivefold. As a result of this increased population density in urban areas, the decentralized privy vault-cesspool wastewater management systems became overtaxed. Mitigation measures included increasing the cleaning frequency and constructing additional privy vaults and cesspools. The improvements, however, only slightly reduced the periodic overflows and development of nuisance conditions. The privy vault-cesspool system, as it existed then, was inadequate to handle the increased amount of wastewater. The centralized water-carriage sewer system, on the other hand, was being promoted as the management alternative for urban areas with increasing populations.
Public Water Supplies and Water Closets
Another major cause of the abandonment of the decentralized privy vault-cesspool system was the increased construction of piped-in water-supply systems. More and more during the middle of the nineteenth century, potable water supplies were being piped in because local water sources were contaminated, frequent disease outbreaks were occurring, and water quantities above what was available locally were needed for fire fighting and street flushing.
Water-supply systems were constructed in most of the major U.S. cities in the early to mid-nineteenth century, and by 1860, the 16 largest cities in the nation had waterworks.
Piped-in water supplies influenced wastewater management in two ways. First, water-carriage waste removal required a copious supply of water, and the introduction of a piped-in water supply made water-carriage sewer systems viable. And second, the improved standard of living for urban dwellers in the nineteenth century coupled with the availability of water led to the implementation of modern plumbing fixtures and a concomitant increase in wastewater production. The water closet probably had the most significant effect on wastewater management compared to the other plumbing fixtures because it increased not only wastewater quantity, but also the quantity of fecal matter in discharges. The high level of fecal matter being discharged with the wastewater heightened the risk of disease transfer and outbreak, but this was not understood at the time.
The increased wastewater levels overwhelmed the privy vault-cesspool system, but few municipalities planned for, or constructed, additional wastewater management infrastructure. Residents had two ways of addressing the increased wastewater being produced: (1) continue to discharge to an existing privy vault or cesspool, or (2) create an illegal connection to a storm sewer or street gutter. Both choices were ineffective solutions because neither the privy vault-cesspool system nor the storm-sewer system were designed to accommodate the increased wastewater. Instead of addressing infrastructure needs, municipalities implemented ordinances to mitigate the problems created by the increased wastewater quantities. One such ordinance was instituted in Boston during 1844 that prohibited the taking of baths without a doctor’s order. Municipalities also tried to prohibit the discharge of fecal matter to the sewer system. Bans such as these were in effect in Boston until 1833, in Philadelphia until 1850, and in New York until 1854, at which time sanitary connections to sewers became required. The enforcement of imposed wastewater discharge limits and the prevention of illegal sanitary connections to the storm-sewer system was difficult for a municipality. Privy vaults and cesspools continued to overflow, while the connections to the storm-sewer system also resulted in sanitation problems. In most cases, neither the privy vault-cesspool system nor the uncoordinated sewer system were able to handle the increased quantity of wastewater. In many American cities (e.g., New York City), physicians, public health officials, and the general public demanded action to address the wastewater management problems created by the influx of piped-in water, and most supported the implementation of centralized water-carriage sewer systems.
Public Health
…In the United States, repeated cholera epidemics and other disease outbreaks gradually influenced municipalities to improve sanitation practices. Between 1832 and 1873, numerous American cities were afflicted with major outbreaks of disease, including cholera in 1832, 1849, and 1866 and typhoid in 1848. The causes of the outbreaks were attributed to a variety of reasons including unsanitary conditions and punishment from God. The experience gained from the epidemics improved the understanding of cholera and other diseases and their corresponding etiology. A cholera outbreak, following the Civil War, provided a chance to practice some of the prevention techniques based on improving sanitary conditions and disinfecting the waste products of infected individuals. The relative success of those measures indicated that the effective management of human wastes was an important component in protecting public health. The search for an effective method of protecting public health by managing human wastes invariably encouraged the construction of water-carriage sewer systems."
It feels like there are a number of different arguments one could be making here and I'm not sure which one you are making.
1) Are you just saying that Robin believes we can't distinguish useful treatments from useless ones therefore it's impossible to cut spending without cutting the good things?
Seems plausible but it doesn't entail that modern medicine is on net neutral only that benefit is maybe half of what it could be.
2) Are you saying Robin believes 1 plus a claim that the costs of harms of the non-working medicine are so large as to render the net impact of medicine zero?
3) Or are you suggesting that Robin believes that the marginal impact of extra medicine is zero or negative? This seems more plausible but like a different claim -- and again it doesn't even mean that all of it is bad only that we can't effectively sort the good stuff.
Ok, fair. My sense is that Rpbin definitely means 1 and likely means 3 but probably not 2. And I think he's unsure about how much net life extension is provided by medication even if it's net positive.
However, it can sometimes be a bit frustrating to figure out Robin's position since he likes to just toss out ideas (maybe that's something more people should do but it can be frustrating).
Part of the disagreement here is that Scott (as a psychiatrist) is a lot more focused than Robin on identifying specific actions that people should do given their views on healthcare. Robin doesn’t really give actionable advice to individuals as far as I’ve noticed. He’s more into abstract comments on society
Perhaps the amount of money “society” spends on healthcare for an individual should have an annual cap that decreases with age. Instead of what we do now, which is more or less the opposite.
I am, in fact, putting Hanson's recommendation into practice!
• My feet and legs are extremely swollen, and have been for some months
• It has recently become so bad that when I went to my monthly appt. with the head-shrinker (mostly, this is for buprenorphine; government-mandated bookkeeping more than anything else, I think), I — having started to forget it isn't normal to have feet that look like balloon-animal versions of themselves, and hence accidentally worn sandals — took quite literally a single step into the office, and one of the nurses said: "Hello, Mr. K—... look at your feet. Oh, my God. Come with me!"... and then I had to lie a bunch to all of 'em to minimize it as much as possible.
• They have demanded a bunch of stuff: tests, "I forbid you to go back to work and you need to keep your legs elevated in fact elevate them right now I'm not sure I've ever seen it this bad before if this isn't gone 100% by tomorrow you need to blah blah".
• I have done none of this. They sent me for blood and urine test right then and there, for kidney function; all normal; I think the next step was supposed to be cardiovascular checkup or something, but I just told them it went away and was probably just a bad reaction to some sketchy "pre-workout powder."
Ha ha! Fools! It is worse than ever, if only you knew!
Now, obviously, this sounds like the actions of a crazy person. No! It's not actually that bad: I think I know the reason this is occurring, and it would probably go away if I did One Simple Trick... but also, I don't care too much if I bite it, heh.
----------------------------------------
(In fact, I'd honestly been close to making that "I care a lot if I kick the bucket—as in, I will be upset if I *don't*"... but my ex-wife was very upset at the idea, and I've made her cry enough tears in this dumb life of mine... so I won't hurry it up — but I won't try to prolong it, either.¹
Unless she comes and checks on me again like she's been threatening. Then she'll see my feet and force me to stop my self-destructive lifestyle and actually go to the doctor again and a bunch of other stuff that Hanson and I are totally against.)
---------------------------------------
(...right? Robin? This IS what you'd recommend, right? I've made sure to leave a bunch of notes around saying stuff like "To My Notoriously Litigious Mother: If anything happens, please know, I am taking medical advice from Robin Hanson, so probably there was nothing to be done" and "Note to self: Hanson will be angry if you mention chest pains and arm numbness to Dr. again" and "LAWYER APPT: NEED SOON, HANSON DEMAND ONLY BENEFICIARY" etc etc.)
...I crack myself up sometimes, heh heh heh. But seriously, I do wonder what the Hansonian view here is. Is THIS still probably net-neutral in EV, to go get checked?
I mean, hey, ignoring it HAS worked so far; for a while, about half a year ago, the swelling wasn't as bad but I was having *way* worse symptoms; but they all resolved!
(...I suppose that if I suddenly stop commenting, well, y'all know how this experiment went then—)
~ footnotes which will inevitably become longer than the main post and which contain nothing that really even needed to be footnoted ~
---------------------------------------------
¹: Destroying your marriage to screw a pretty face turned out to be a bad idea, shockingly! You just end up stuck with someone who grates on you because she isn't your ex-wife, who was the only woman who has ever truly known you and still loved you for it; and then, considering how things started, you can hardly complain when your lover does to you what you did to your sweet innocent wife.
god damn it.
...and now I'm feeling like I can't face going to work at another stagnant and dull position, so just living off savings—but... who's going to give me a chance to change careers at this point, y'know?
"Hey guys trust me my degree is in a completely unrelated field but I would be a real good developer/data scientist/whatever, I swear! Take a chance and see!"
When I'd first graduated, I remember thinking "well, this should be easy as long as I can get someone to gimme that first look" — ...lol, dream on, moron! No one gave my résumé another glance, I expect; must have sent that damn thing out a thousand times.
Only got to such a good position as I recently had by a sheer chance meeting, in fact, not due to any application: sat next to the 'Fleet Director' of an oilfield company, on a flight back from Houston. He'd usually have been in the company jet (oilfield company: of course there's a jet!), but had let some other division charter it and just requested a ticket on a regular flight instead... luckily for me.
During the short hop over to WTX, I had joked with him about my situation, which at the time was getting pretty dire — though of course I didn't let on just how bad; said something like "someday I'll be able to get the GOOD bread, haha, the $6 kind you know" — but he's got a surprisingly soft heart, for an Important Executive™, and read between the lines.
(It seems to have preyed on his mind for the entire next week, as I later learned — he kept mentioning it to people, apparently: "Wanted to try the damn $6 bread... had to put the 'good' bread back on the shelf, and just-married too... poor fucker...")
So, though I'd left my number with him, I never expected anything to come of it — but after that week had passed, got a 9 am call from his office: "Mr. M____ said you're going to be our new Fleet Analytics guy, so when can you come up here for orientation—?"
I remember how my (then-not-ex-yet-) wife and I hugged each other and jumped around that shitty little apartment for a while, once I hung up and told her. She was so proud for/of me. "I told you! I told you you were a charmer!" "Only you think so... well, you and Mr. M___ I guess! Get your shoes on, baby — we're not having ketchup sandwiches again for lunch, not today!"
I miss her a lot. Can't expect that kind of luck again, though, heh... neither with love, nor labor.
The majority of those tweets/screengrabs are at least broadly consistent with what Robin says he says, so it's not obvious why you'd demand he leap in to correct them. Note the first few words of the LessWrong article are "Robin Hanson has hypothesized that much health care is useless...", i.e. "much" not "all".
Never having read Hanson, Is it possible that the core disagreement is around the word marginal? You seem like you’re using it for the IMO standard meaning of “least valuable” but I medicine I could kind of imagine it meaning “final”. So your sons helmet isn’t marginal because it will benefit him for years, but a lifesaving surgery on a 90 year old might be “marginal” if they die or something unrelated two weeks later.
In that regard, I could almost, if I squint at it, understand the argument “cut healthcare spending by 50% by investing dramatically less in end of life care”.
To be fair that seems like an easy thing to just say if that’s his actual belief, but it seemed like the kind of off chance thing that could be possible.
Edit: Here's an exchange that happened in the comments on the response to this post:
Commenter: 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.
Hanson: No, most med costs don't come from last six months of life.
1. Poor people without insurance don't get regular checkups and ignore medical conditions more often, but eventually a disease will get serious enough that they're forced to go to the doctor.
2. Therefore, chronic diseases that would eventually kill someone usually get detected and treated regardless of insurance/wealth status.
3. Therefore, although medicine "works" in the sense that early checkups catch the disease earlier, allow better and less invasive treatment, improve quality of life, etc, this doesn't show up in lifespan data because our medical system was already pretty good at making sure people don't literally *die*. The marginal medical dollar might improve healthspan but this is harder to measure.
What about the claim that the Obamacare study reported the one cohort with significance?
Generally it's pretty clear that Hanson was laying a monkey trap where unsuspecting bloggers would critique his statements that healthcare should lowkey be abolished and then he would respond by saying "haha that's super preposterous and now you must agree with my real argument which I will now reveal and pretend it was the main one all along"
However it did appear to be a real argument that the Obamacare study picked a cohort
Robin's monkey trap analogy was that waiting to decide which parts of healthcare to keep was an obvious mistake because it's so clear that marginal health spending is not valuable at all and perhaps impossible to demonstrate which 50% is good
Rhetorically speaking this is saying "lowkey" abolish healthcare, or "medicine doesn't work", especially since the studies supporting it don't have statistical power!
No, it's not saying abolish healthcare. If the monkey indiscriminately let some of the food out of its paw, it's paw would take up less space and could perhaps be withdrawn. If the monkey doesn't let go of anything until it's sure the food it releases is worse than average, then it's paw will remain stuck.
The key difference is paw open/closed. I mean, try it with your own hand: open hand has a lot of mobility, and can nicely curve to reduce its min. width. Once you close your fist, loosely or tightly, the knuckles spread out, the palm widens, and the min. width is significantly larger. So as long as the palm is curled, the exact amount of food kept is immaterial.
I feel silly just typing this. Hanson's arguing for silly things using clever silly arguments, and refuting them is like wading through mud.
Is it signaling if you do whatever (coincidentally expensive) thing the ER proposes for your kid "out of an abundance of caution" because you don't want them to call child services on you for endangering your child?
"It seems to me that if we were to cut medicine in half, figuring out which half to cut would be among the most consequential decisions in history."
This seems a bit over-dramatic. Doctors pick what's most important routinely when triaging patients.
If you have lots of time and resources then that cut is to be carefully sewn or maybe you get the plastic surgeons involved if it's on their face. If you're short of time and resources? Stop the bleeding, keep it clean, treat any infections.
That list of 10 things seems pretty easy to mostly cut in half in a not-too-controversial way while retaining the biggest benefits.
If doctors lost a little bit of money (rather than gained) every time they sent a patient for an unnecessary test or prescribed a useless treatment then they'd quickly prioritise the most vital tests. They're not helpless.
> If doctors lost a little bit of money (rather than gained) every time they sent a patient for an unnecessary test or prescribed a useless treatment then they'd quickly prioritise the most vital tests. They're not helpless.
The payment-for-procedure system causes a lot of problems.
"If doctors lost a little bit of money (rather than gained) every time they sent a patient for an unnecessary test or prescribed a useless treatment then they'd quickly prioritise the most vital tests. They're not helpless."
> This is why I find this conversation so frustrating. Mention a medical treatment to Robin, even one of the “good ones” like cancer or antibiotics, and he’ll try to argue that maybe the evidence it works is being misinterpreted, and in fact it’s unclear how well it works. Then I say he thinks this stuff might not work, and he accuses me of straw-manning him.
"Cancer" isn't a medical treatment. In fact, we know that a huge amount of medical spending on cancer is deleterious. Doctors with cancer accept much less treatment than laymen do, and when you ask them why, they'll say that it's because of their experience watching what happens to cancer patients.
I've seen an article (not by a doctor) that argued in apparent seriousness that under the modern American medical system, anyone coming down with cancer dies in poverty after the system confiscates all their money, and that this is a desirable state of affairs. I take some issue with that.
When my grandmother had cancer, my mother (a doctor) specifically advised that she not accept most of the treatments she was offered. The awfulness and unhelpfulness of most cancer treatment isn't some big secret.
At the same time, survival rates keep going up. When my 𝘴𝘪𝘴𝘵𝘦𝘳 had cancer, my mother (still a doctor) was very pleased to note that a lot of progress had occurred in cancer treatment 𝗳𝗼𝗿 𝘆𝗼𝘂𝗻𝗴 𝗽𝗮𝘁𝗶𝗲𝗻𝘁𝘀.
Young patients are ideal for several reasons. Most prominently, (1) there is less wrong with them, (2) they recover faster (from everything), and (3) even stipulating that life while being treated for cancer isn't worth living, their treatment will presumably give them a lot more of the good kind of life with the cancer in remission and treatment sporadic or on hiatus.
But most cancer treatment isn't delivered to young patients for the obvious reason that there are very few young people with cancer.
Anyway, looking at what medical treatments doctors obtain for themselves and their family members will surely 𝗵𝗲𝗹𝗽 with judging which ones are likely to be harmful, in proportion to how common the problem being treated is.
There are other tells. Another essay I've read, by a doctor, pointed out that the elderly patients who get expensive long-shot (or: deleterious) medical care don't live near their children. It's the absentee children who insist on it for what seem to be reasons of guilt.
It would definitely be useful for many patients to be aware that a lot of medical treatments, particularly during end-of-life care, may have limited benefit and dramatic downsides.
It is also useful to be aware of the many psychological factors at play among physicians, patients, and patients' loved ones (including "signalling") that lead to treatments being undertaken that, if certain conversations had been had, might have been turned down.
Among other things, this may be as simple as the patient feeling comfortable to ask questions like, "Would you do this [additional course of radiotherapy] if you were in my position?".
Scott, in terms of it being a challenge to "fit claims together coherently" with what someone has said in the past: you yourself have written vividly on the horrors of excessive end-of-life care. Shouldn't the practical relevance of Robin's claims for the individual be fairly intuitive?
I recognize there are issues wrt futile care with cancer diagnoses, but in the last few years, I have seen three older relatives get diagnosed with serious diseases--two with cancer, one with another blood disorder. In two cases, they tried a treatment course, found it wasn't working, and advised ceasing all treatment except palliative care; in the third, by the time the cancer was diagnosed, the doctor advised going straight into hospice.
Yeah, you're strawmanning him; he listed a bunch of ways to decide what medicine to cut, e.g. ending subsidy as a government strategy and Cochrane reviews as a patient strategy; pretending he didn't is weird.
There's a huge "public choice" problem here. It may be true that a smart, prudent individual could distinguish which parts of medicine are "good" and which are "bad." That's does not *at all* imply that consensus could be generated sufficient to push those determinations through the political and bureaucratic process need to get them implemented at the level of the federal government. (It is a major, common rationalist fallacy to elide this distinction. "We should just do X!")
I may do good research and statistical analysis and decide that spinal fusion surgery for degenerative disks clearly falls into the "bad" category. This does not change the fact that there are many back surgeons who not only perform but will vociferously counterargue for this procedure. Rinse and repeat for many, many "bad" treatments.
I take Robin's "just cut half" argument to be partly a response to this fact. It's not that literally no one can distinguish good from bad. It's that it's hopeless to rely on those distinctions being made and accepted in advance of taking ameliorative action.
This is a good argument in favor of the decentralized/free-market approach to trimming health care, as opposed to the centralized/"medical establishment decides" approach.
I agree that simply throwing up our hands and saying that distinguishing good from bad is hopeless is not a good recommendation.
I don't know if this directly affects the way you refer to it in the argument, but it seems salient to the overall discussion that plagiocephaly is itself the product of a societal medical intervention: one of the side effects of the Back to Sleep campaign from the 90s targeting SIDS.
To me Back to Sleep is a microcosm of the issues with modern day medical science which make me personally skeptical of the value of the marginal dollar of medical spending. You have a very legible outcome targeted aggressively through public health interventions without any mechanism to weigh the benefits with any less legible costs. (Babies also seem to sleep more deeply on their stomachs - this is one of the proposed mechanisms for the success of Back to Sleep, but clearly there could also be drawbacks to babies sleeping less deeply.) And once the clarion call of the public health campaign has sounded, scientists, journals, and especially media outlets trip over themselves to frame every nuance so that it comes in line with the official orthodoxy.
Note that I claim no certainty whatsoever that the costs of back sleeping outweigh the benefits for the population at large or any slice of it. But I am confident that our system of research, public health, and primary care is not equipped to meaningfully investigate the possibility. The further question of how this relates to marginal medical spending at a patient level is complicated, but I think it's always worth considering the possibility that any particular recommended medical intervention may be myopically aimed at one legible metric.
I read about a third of this and shut it down due to boredom.
Modern medicine is a business and a big one at that. It’s not in existence to help or cure anything or anyone which is not sayin there are no caring physicians and nurses in existence. There are plenty of them. There are also the type that are in the medicine business for the money. Those types believe in prescribing pharmaceuticals for practically everything. Good physicians are limited by government and big business. Government tries to justify this by saying they are protecting the people from crazy quackery. That’s fine, except they take it too far. A lot of doctors’ hands are tied when it comes to how they treat their patients. This is a travesty. The government also gives a lot of criticism and objections to homeopaths…real medicine. Another travesty.
Bottom line is that Big Medicine is running the show and robbing doctors from doing not only what they’ve been trained to do, but also not allowing them what their intuition combined with their training and common sense guides them to do.
So this therefore causes much confusion and misery for doctors and especially their patients.
What’s more, this topic spills over into the ridiculous high cost that government imposes on everyone, especially the patients. I won’t call it “healthcare” because the truth is that the Big Business is not about care, only money.
Ha ha ha! You damned fool, you've fallen right into our trap! Watch in awe as we eloquently explain the mechanism behind homeopathy, adduce dozens of large-scale pre-registered double-blinded studies, and conclude by rendering trivial the objection "but why isn't seawater and stuff like super potent homeopathic poison-cures?" with plausible and simple corollaries of the initially-explained physical mechanism.
Take it away, Majo! Now is when all the rigorous coursework in "Evidence, Application, & Terminology: the Study of Homeopathic Ingredients & Theory" pay off!
...Majo?
...
...look just hold tight man he's got the goods alright this is gonna blow your little establishment-drone MIND with some non-Illuminatized hyperrreality
I studied homeopathy and have been using it since 1992 on myself and my family. I've attended the National Center's weekend seminar in '95 and learned much more about it, very impressed. My personal physician is a homeopath. I have in my repertoire aka., medicine cabinet, many remedies and a few helpful guidebooks written by reputable homeopaths, plus a repertory of remedies and Materia Medicas. (The remedies never expire if safely kept). I don't care what you choose to believe or think you know, nor do I wish to debate the subject. I do find your closed mindedness very sad indeed. You’re making blanket statements out of ignorance. I assume you are one of those who runs to get an RX filled at your friendly local pharmacy store as your first line of healing anything.
I sincerely recommend you reading up on how and when homeopathy was discovered. It’s a fascinating subject. The best source is NCH or National Center for Homeopathy.
I know how and when it was discovered. The theory behind it is bunk. Your assumptions about me and what I know and what I do are classic case of projection.
One minor thing - of course the "remedies" never expire, they are pure water.
What about the value of this as a precautionary principle, or a bravery debate? Can't counseling something like skepticism towards the medical establishment be valuable? I'm being charitable by imagining a steelmanned version of this argument that encourages people to take a more critical look at, say, the correlation between medical spending and medical efficacy specifically. Maybe more costly treatments are not sufficiently more beneficial? Something like this seems to be at the heart of positions like Hanson's.
> What about the value of this as a precautionary principle
Easy question. Precautionary principles don't have value. Anyone invoking a "precautionary principle" is a moron who isn't capable of following simple logical chains of thought.
The precautionary principle answers "no" to every question of the form "should I do X?". In addition to carrying literally zero information in the entropic sense, this means that it generates irreconcilable logical conflicts as soon as anyone rephrases a question.
Okay, so you are fundamentally misunderstanding what people mean when they use the phrase 'Precautionary Principle'. But the fact that you've introduced information theory to a high-level discussion about basic heuristics leaves me feeling like it would be worthless to try addressing it. Consider this a human check and get back to me if I'm wrong or you were just having a bad day or something.
I haven't misunderstood anything. I gave you an accurate definition of the Precautionary Principle: everything is forbidden. That is both the definition people give and the meaning they intend when they invoke it.
This isn't new news or anything; you can see the problems pointed out on wikipedia. Check it out:
> As Michael Crichton wrote in his novel State of Fear: "The 'precautionary principle', properly applied, forbids the precautionary principle."
> If all that the (weak) principle states is that it is permissible to act in a precautionary manner where there is a possible risk of harm, then it constitutes a trivial truism and thus fails to be useful.
> If we formulate the principle in the stronger sense however, it looks like it rules out all courses of action, including the precautionary measures it is intended to advocate. This is because, if we stipulate that precaution is required in the face of uncertain harms, and precautionary measures also carry a risk of harm, the precautionary principle can both demand and prohibit action at the same time. The risk of a policy resulting in catastrophic harm is always possible.
> The strong version of the precautionary principle, in that "[i]t bans the very steps that it requires", thus fails to be coherent.
If anyone uses the words "precautionary principle" other than to point out how stupid it is, there are only two possibilities:
(A) That person is, themselves, extremely stupid.
(B) That person is completely unconcerned with whether the things they say are true or false.
I’m very sympathetic to the concept of avoiding “low value” care. I’m a medical conservative practicing with a constrained vision (or at least striving at that).
But 2/5 Hanson prescriptions (low spending region, and small hospital, practices) simply removes “expensive” care. I would argue “expensive” does not necessarily equate to “low value”.
He also doesn’t precisely define how he would weigh those 5 steps, in order of importance (unless his number 1 to 5 is in that order or reverse order).
Finally, his suggestions can be operationalized on an individual level, but how do you apply that on a system level? The FDA already screens for “effectiveness” (admittedly doing a crap job at it) and Medicare and CMS does some attempt at covering things that are “cost effective” (also not the best at what they do). Where should it go from there?
He thinks such expensive care generally isn't valuable because there don't appear to be better health outcomes associated with higher spending regions/hospitals.
In the linked post he mentioned Cochrane reviews as one more way to determine which things to cut.
I think there's a fourth possibility that does not fit in your trilemma: antibiotics today might be more expensive than antibiotics 50 years ago without being any more effective today than those antibiotics were 50 years ago. (This is the problem of antibiotic resistance.) Perhaps some parts of medicine are an adversarial game, where you spend more and more over time to stay in place against an adversary that keeps adapting to your spending. The problem is not effectiveness but the cost of being effective.
This picture is compatible with both "increased spending is not at all correlated with better outcomes" and with "cutting anything will kill lots of people". The options here are "keep spending more and more money to stay in the equilibrium we're in" or "cap medical spending, regress to a worse equilibrium".
(Incidentally, I believe Hanson might suggest cutting out medications in order of development cost, not alphabetically.)
> Perhaps some parts of medicine are an adversarial game, where you spend more and more over time to stay in place against an adversary that keeps adapting to your spending.
"Adversarial game" doesn't mean this. You want "arms race" or, if you're feeling fancy, "Red Queen's race".
You need > 1% credence in the _combined_ proposition that both [son will have misshapen head without $300 intervention] AND [intervention will reliably fix this].
The problem is you and Hanson are (as of right now) arguing past each other. You're pointing at a study that shows if you point a gun at someone and pull the trigger they're really likely to die. Hanson is pointing at a study that says the number of guns in a society doesn't affect how many people die each year. So when you ask the question "Do guns kill people?" you answer yes and he answers no and you're both right.
You are in fact missing a plank of evidence. If a treatment is both available and effective it seems to casually follow that it has a positive effect on the population but Hanson says that looking at the end result on the population the casual chain you assume doesn't exist.
To provide yet another analogy, seatbelts are effective and available, it's easy to show that if you're crashing a car at any speed wearing a seatbelt is much better then not wearing one. But if you make them mandatory to try and apply the effects to the population and then everybody starts driving 30 mph faster you could watch more pedestrians die and even have more auto fatalities if a 40 mph crash without a seatbelt is less deadly then a 70 mph crash with a seatbelt.
Now mind you, I think you're instincts are actually the correct ones and Hanson is putting too much faith in those three studies which only kind of support his strongest statements but proving it might be difficult.
I mean, if people are overconsuming healthcare and this is economically inefficient presumably the thing to do is tax healthcare and reduce taxes elsewhere to offset the revenue, and the consumers will decide which healthcare options are marginal.
Is this discussion suffering from equivocation on the definition of "interventions", i.e. interventions as classes of treatment versus interventions as specific, individual medical decisions?
It is clear in some sense that antibiotics are a good class of interventions and can be extremely effective in certain circumstances.
We can hold that belief while also thinking it is possible that antibiotics are very often over-prescribed and generally used in circumstances (e.g. viral infections) where their absolute impact is zero or even negative (due to disrupting the gut biome).
It is then also credible that there are policy models available, such as increasing the cost of antibiotics for the patient, that could result in antibiotics being deployed more efficiently.
It's also credible that individual education, regarding when and where antibiotics are useful, is beneficial. It is observable that attitudes towards antibiotic use, and how likely patients are to ask for a prescription, varies substantially across countries.
Analogous comments hold for other types of medical interventions, including for cancer.
The comparison to the UK should probably mention that most healthcare in the UK is provided by the National Health Service (NHS). The NHS will (generally) only provide treatments recommended by NICE, an institute that does very explicit and rigorous cost-effectiveness estimates which form a large portion of the basis of their advice. Treatments which aren't cost-effective are therefore rare in the UK.
We can observe that the US spends more (% GDP, expenditure per capita, whatever metric) than any other first world nation, but is distinctly middling in terms of “health outcomes”. So it’s not difficult to infer that maybe the US is not getting nearly as good “bang for buck” as most comparator nations. Unless US docs are particularly inept (unlikely), or the US uses lesser versions of treatments (for identical conditions) than elsewhere (impossible), it seems to stand to reason that much/most of the money is spent without commensurate outcome benefits. Now, there is the minor issue of substrate….the average American is huge, sedentary, and eats crap…so that US dollar is spent caring for a more unhealthy person, on average, than most other comparable places. Still, even if it seems plausible that it should cost more to keep the average American from keeling over, it seemingly shouldn’t cost that much more. It’s not hard to accept the general Hanson point that there is a fair bit of wastage going on.
But Scott is asking about where the rubber meets road (and I would expect nothing less). “Cutting at the margins” is pointless hand waving unless you specify where those margins are, how you arrived at them, and what you would be cutting as a result. Then hopefully return to measure the net benefit (or harm) of what you have wrought.
And as Scott points out here, there are many devils in those details. And it’s not cut and dried. Antibiotics for bacterial infections are essential and potentially life-saving, ….and are also ridiculously overused. We need for-profit industry to innovate (and recoup investment and make profit) while ideally making sure that what they are making actually works and is only used where indicated (as opposed to usage creep and therapeutic fashion). And the best RCT only gives an average treatment effect on a population level…and an ARR of 10%/NNT of 10 would be a blockbuster in my field but still means you are wasting your time and $ on 9 out of those 10 people without any a priori ability to determine which category Mrs. Smith sitting in the office slots into…and that on an individual level, any treatment effect (on hard outcome) is categorical (ie either it works, or it doesn’t; either you’re alive, or you’re not) and not a percentage.
So what Hanson sells is not objectionable per se….but is useless drivel unless he brings some meat and details to his prescription. Otherwise it’s just pointless whinging.
> Every study on the marginal effect of medicine has some way it operationalizes “marginal medicine” for the purpose of that study. In geographic variation studies, it is the medicine done in places that spend more on medicine, but not in places that spend less. For studies that compare large to small hospitals, it is the treatments done in large but not small hospitals. For experiments that vary the price of medicine or insurance, it is the medicine chosen by subjects who faced lower prices, but not chosen by those who faced higher prices. I remember at some point also suggesting using treatments with a lower Cochrane Review rating.
"Unless US docs are particularly inept (unlikely), or the US uses lesser versions of treatments (for identical conditions) than elsewhere (impossible), it seems to stand to reason that much/most of the money is spent without commensurate outcome benefits."
Or if US doctors are about as good as doctors elsewhere, but we pay them much higher salaries. Which, as I understand it, is the case.
I think Robin would respond that the real issue is that you've listened to doctors telling you to not allow your child to sleep on their stomach, which in turn created the need to wear a helmet to correct for the head deformation caused by forcing the baby to unnaturally sleep on its back instead of its stomach. Obviously some children _do_ naturally sleep on their back so I don't know if that's the case for your family, but in general the helmet thing is usually due to the overconsumption of medical advice that Robin Hanson warns against.
The SIDS hysteria would be a great example of medicine overconsumption in general. It had good advice about not smoking next to your child and not keeping tons of blankets (and other sources of suffocation) near the bed but then overreached and demanded that parents never share the bed with their baby, as well as never let them sleep on their backs. Meanwhile in Japan almost all mothers share the bed with their baby and their SIDS rates are very low. But the medical establishment will never admit to these Japanese statistics and keeps on repeating S-I-D-S over and over again to scare parents into submission.
The problem is that its quite hard to disentangle the effects of every individual recommendation. It makes sense to avoid putting soft objects or using a soft mattress in the baby's crib (avoids suffocation) and not sleeping next to the baby while intoxicated (risk of choking them with your body weight).
But in Japan and South Korea babies (almost) always sleep with their mother in the same bed and are not forced to sleep on their backs. Yet their SIDS rate are lower than in the West. SIDS researchers completely ignore this fact or dismiss it as some sort of a statistical fluke. But they're likewise unable to prove that _specifically_ sleeping on the back is of much use.
The article itself seems to confirm the evidence is very weak? To quote:
"Conclusions
Low- to very low-certainty evidence suggests that supine sleep position may reduce the risk of SUDI (0-1 year) and SIDS (0-1 year). Limited evidence suggests that supine sleeping probably delays short-term ‘gross motor’ development at 6 months, but the effect on long-term neurodevelopment at 18 months may be negligible."
And then:
"Of 48 observational studies, 46 were considered to be at serious risk of bias, mostly due to confounding and misclassification bias."
And then:
"An increase in positional plagiocephaly at 2-7 months of age with supine sleep position is possible (OR = 2.77, 95% CI = 2.06-3.72; 6 studies, 1774 participants)."
There's pretty good evidence of supine sleep causing positional plagiocephaly and we have a clear mechanistic explanation for it. I don't think we have good evidence for prone sleep *on a firm mattress* causing SIDS, nor is there a mechanistic explanation for it. Yes, prone sleep on a very soft mattress or on a soft blanket could cause the baby to choke during sleep, but how would this happen on a firm mattress?
"Co-sleeping was recorded for 61%, and the prone position was found for 40% of cases at discovery. Mother’s smoking habit exhibited an odds ratio of 4.5 (2.9–5.8)."
It shows that smoking near babies is indeed a bad idea but there's no correlation between prone sleeping and SIDS.
I agree the evidence is weak, but it seems to be leaning that direction pretty clearly, especially when combined with the epidemiological evidence from the campaigns.
What makes you think the evidence for the plagiocephaly is better than the evidence for the SIDS? There are fewer plagiocephaly studies and they don't grade them for bias.
Do we know what percent of general infants in Japan sleep prone? I don't think you can make a conclusion until you know that.
It seems like in every country that has had a Back To Sleep campaign (many), SIDS rates have fallen dramatically; did they all campaign about other risk factors at the same time?
(and from what I could find, some other risk factors like cosleeping didn't seem to decrease during the campaigns anyway)
“There is little question that the advice prevented deaths from SIDS, but it was noted as early as 1996 that there was a head-shape side effect. In the pre-Back-to-Sleep period, positional plagiocephaly was thought to occur well under 1% of the time. More recent updates suggest the current share is somewhere between 16% and 45%”
(Incidentally she recommends helmets as a secondary approach)
> Do we know what percent of general infants in Japan sleep prone?
“They also reported that more than one-half of parents did not place their infants on their backs when they noticed that the infant was sleeping prone”
A study from 1999 reported a 38% prevalence in the control group (but an 81% prevalence in the SIDS group, though note that the study size is smaller):
“The proportion of prone sleeping was 81.0% and 38.3% in the SIDS and control groups, respectively, and an odds ratio was calculated as 10.4 (99% confidence interval, 3.9 to 37.6).”
> It seems like in every country that has had a Back To Sleep campaign (many), SIDS rates have fallen dramatically; did they all campaign about other risk factors at the same time?
“The pamphlet encouraged parents to do the following to lower the risk for their baby:
Sleep the baby on the back
Do not smoke near the baby or during pregnancy
Breastfeed as much as possible
Do not leave the baby alone when asleep
The pamphlet urges parents to sleep the baby in the same room with parents, to avoid leaving the baby alone, to avoid sleeping the baby in unusual places like a sofa, and to avoid soft toys and pillows in the sleep area.“
So, yes, the campaign focused on several things at once.
Counter-counterpoint: the Dutch data is hard to benchmark against Japanese data because Japanese beds are a lot more firm on average. Prone sleeping would also be more dangerous in the presence of pillows, toys and blankets which weren’t a focus of the initial 1987 campaign. This would also explain why Japan had a lower SIDS rate than the Netherlands before the Back to Sleep campaigns have started: there’s just less soft objects for infants to choke on in Japan.
> That helmet is probably our “marginal” health care expense, in the sense that it’s less obviously important than the other two things we’ve used healthcare for this year (childbirth, a scare with our son’s breathing).
Maybe I'm way off, but isn't childbirth in hospital both expensive and not that outcome-improving in normal circumstances, making it a likely candidate for marginal health care expense? Maybe not in your particular case.
I don't think it's randomly distributed, though (that is, for some people the expected split is 9995/5 and others it's 80/20). I double checked and of course the stats are super noisy: the midwife associations will use stats that don't control for base risk factors, so they can claim that home delivery is safer than hospital births, but attempts at controlled trials for people with low expected risk to start with (see https://www.jogc.com/article/S1701-2163(16)00089-X/fulltext ) find a 0.2/1000 increase in Ontario and 2.1/1000 in Oregon. At a statistical value of life of $9M, that means being willing to spend between $1800 to $18900 more (Though $9M is an old estimate for adults, and children tend to command a 1.5 to 3x premium, so you may want to adjust further).
BTW, in a perfectly designed system, shouldn't we *expect* that it's net neutral on the margins? A perfect system would include all the net positive stuff, exclude all the net negative stuff, and be neutral towards the net neutral stuff, that would therefore be specifically what exists on the margins?
You expect that it is effective commesurate with its cost, net neutral with respect to the value of a dollar spent, not that it is has zero measurable effect but costs a lot of money. The claim by Hanson is the latter.
Gun analogy, yes the conclusion is that adding more marginal guns is not signficantly dangerous, and that reducing or increasing the number of guns will not change the murder numbers significantly. More murders happen without guns that with even for the population with guns; the main determinant in the murder rate is how many murderers are in the pop rather than whether they are armed. Hell maybe arming more people reduces murder rate by deterrence more than it enables murderers?
It's not that "guns can't kill you" just like Robin never said "no advanced healthcare treatment purchased at the far margin of expense has ever helped anyone ever". Complete strawmanning, are you even trying to be fair here?
To the extent that a large fraction of healthcare spending is wasted -- fails to buy extra QALYs, or does buy some QALYs but at a price wildly higher than broad social consensus could agree is reasonable -- it seems like a very large share of that wasted spending is probably measures taken in the last 2-3 years of people's lives, where everyone kind of knows that they're buying days at a time, or at best a few months, at very large expense. This is why we ought to be doing a lot more end-of-life counseling, and getting people to actually write down some kind of objective standards for deciding when they're done -- when their family (and/or subsidies from the state) should STOP spending money on trying to buy them a tiny bit more time. But of course, when that was proposed as a cost-saving measure to make the insurance subsidies in Obamacare work out better, it was caricatured as "death panels". In this sense, I do think Hanson is correct, that people feel like if they fail to spend every last dollar possible to "save" their loved one's life (for a few days), then that means they didn't really love them. But this is a far cry from his casino metaphor.
(And for the record, yes, I have sat down with my spouse and a lawyer and put medical directives in writing.)
(I originally wrote that the vast majority would be end of life care, but upon some thought decided I couldn't justify a claim quite that strong, considering stuff like the heart attack care comparison that Atul Gawande presented, some years back, where in neighboring counties you saw one of them following a standard care regimen that cost drastically more -- and profited the hospitals much more -- but did not improve outcomes. I think there is quite a bit of wasted spending that looks like that as well. Actually, my aunt wrote an entire book about that kind of wasted care in the ob/gyn field: https://www.amazon.com/Optimal-Care-Childbirth-Physiologic-Approach/dp/1598491326/ )
I think it would be interesting to know to what extent it's *actually* true that vast amounts are spent trying to prop up people who will almost certainly be dead within a year anyway.
Everyone seems to have this intuition, but is it actually true, or are we actually just talking about ~10% of total medical spending?
How does medical spending break down between these categories of people?
a) Generally-healthy non-ancient people who can be made healthy again
b) Decrepit ancient people whose lives are being artificially prolonged at massive expense?
c) Chronically ill people who may continue to live for a long time but never truly be healthy?
I feel like this third category might be the really big money sink.
Your category (c) can spend a lot of money on a per-individual basis, but those people are rare. People in category (b) aren't rare and also spend a lot of money per individual.
I think Hanson is wrong about the specifics but his main point is right. Singapore has world class healthcare and only spends 4% of it's GDP on it. So it's possible in principle to deliver the same quality with half the cost or less.
It's probably not possible in practice though. In Singapore, healthcare costs are controlled via the same mechanism of almost everything else - markets. Patients pay for healthcare out of their own healthcare savings accounts and it's up to them to decide which interventions are worth it, and it's then up to providers to convince them what interventions are necessary. Of course, this introduces the information asymmetry problem, but going off outcomes, it doesn't seem to be worse than the US system.
Westerners are used to having almost everything paid for at point of service, either from government health plans or private insurance so there is no incentive to ration. Taking that away is politically impossible.
It's not markets in the sense that all payments are voluntary , or that nothing is taxpayer funded.
"Financing of healthcare costs is done through a mixture of direct government subsidies, compulsory comprehensive savings, national healthcare insurance, and cost-sharing."
After looking up the numbers, Singapore's healthcare spending has changed more than I expected over the last decade with government spending going up and private spending going down. However, compared to most Western countries Singapore still has a much higher proportion of spending either from out of pocket or from health savings accounts (around half). That's still enough to have a large effect of how people consume healthcare. Instead of only considering the benefits when considering an intervention, patients have to consider costs and benefits.
Re the Britain point, one needs to control for the fact that people there are healthier than Americans (less diabetes etc). British people receiving US-level medical care might well become healthier.
"In the aggregate, variations in medical spending usually show no statistically significant medical effect on health"
One needs to track in each study how an "effect on health" is defined. For example, it shouldn't be necessary for healthcare to expand lifespan in order to be qualified as useful. I recently started an allergy shot treatment a few months ago. This will almost certainly have a negligible effect on my lifespan, but the increase in my quality of life has been dramatic. And insurance complicates the analysis for cases like this.
The only reason that I got this in the first place is that I am graduate student, so the insurance I get is heavily subsidized, and it covered the allergy shot treatment. Given the option between taking the cash and doing the allergy shot treatment, my past self would have chosen to take the cash. (Knowing what I do now, I would tell a parallel-universe version of myself that the price tag is worth it).
A significant fraction of medical expenses are arguably like this. Painkiller medication, palliative care, hospice care, etc. might have no effect on "physical health" but are still worth the price tag, and may be covered by insurance. The helmet for your daughter might just be cosmetic but still worth it. Most people in the United States get braces, women (and others) spend money on cosmetic products, folks get plastic surgery, people get hair transplants, and the list goes on and on.
So imagining you have some study where you give people better insurance, and you find out that people used the insurance, but it didn't increase their "health." When you look closer you find out that the types of extra treatments people received were not the type where we would expect an increase in "health" anyway. However, people in general reported that knowing what they do now, it would have been worth it to pay out of pocket for the treatment they received because of the increase in quality of life!
In this post, Scott goes "near mode" (his term: I'd say something like "lets the rubber hit the highway") by asking what Robin would have him do in actual medical situations.
One time when I posted a query about what someone giving out contradictory indications would have me do in the situation in question, I got a lot of severe replies saying it was my decision and I didn't have to do what the other person said. I KNOW that: my question was, what would he -advise- that I do? I don't know why that wasn't clear, and I wish Scott better luck with the same mode of reasoning.
He actually does explain it in the new response he wrote:
> My key point was and is that each of these operationalized definitions of “marginal medicine” offers a concrete way to avoid marginal medicine. As an individual considering various possible treatments, here are five ways:
1. Ask about a treatment’s Cochrane Review rating,
2. Ask if a treatment is done in low spending geographic regions,
3. Ask if treatments are done in small hospitals,
4. Ask your doctor how strongly they recommend a particular treatment; decline if recommendation is weak. (I’ve done this.)
5. Ask yourself and associates if you would be willing to pay for them out of your own pocket, if insurance did not cover them.
Maybe even better to ask several of these questions, and average their answers.
As a wonk considering various possible policies, you can also consider regulating or subsidizing/taxing based on these indicators. Or consider policies that make more patients face higher personal prices for treatment. When I said “most any way to implement such a cut” I had in mind these sort of options; most any should help. Though my favorite option is still creating agents who face strong direct incentives.
I completely agree with you about investigating any proposed treatment via Cochrane reviews, etc. , but I really think it's a terrible idea to ask a doctor about Cochrane reviews of something they're recommending. Doctors often *hate* it when the patient believes he's knowledgable, and hate it even more when the patient puts them on the spot. And most doctors are not going to KNOW whether there was a Cochrane review and what the upshot of it was. I have no problem being assertive with doctors, but I have stopped asking them questions like the one suggested because it almost always backfires. You don't get your question answered, and the doctor's attitude towards you is going to be less friendly and helpful forever after.
Not that I don't recognize where you're coming from in these experiences, but I think I've had better luck than this at getting the information I want out of doctors, to the extent that they have it. I also wouldn't ask these specific questions, except #4, which seems fine and I think people use variations of this pretty often ("would you do this if it were yourself/your spouse/your kid/your dog") without wasting one of the fifteen minutes that have been allocated to the visit or offending the doctor's ego. Could be luck or practice or some kind of weird doctor soft skill voodoo but now I'm wondering whether the approach can be systematized.
I dunno, I am generally pretty good at tactfully raising issues without provoking conflict. I once said to a doctor who recommended treatment X, "I've read about treatment Y is being used for this? What do you think of trying that?" And she glared at me and said "So you know everything and I know nothing!" Another muttered something about "Dr. Google." One glorious exception: a smart, cheerful dermatologist my daughter and I both see. I asked about a treatment for hair loss, which my daughter has, and the derm said "Yes! They're doing that in Philadelphia now. It looks promising. Let me see if I can remember the name of the guy heading the research." And she got on the office computer and found it for us. Agreed we should try the new thing if the present one wasn't effective enough.
The gun analogy does not seem very informative. Gun advocates in the real world do claim that guns can be used in self-defence or as a deterrent (preventing murders), and that guns only substitute for other weapons (ie murders that happen with a gun would have happened anyway with a knife).
This is why it is very reasonable to prefer to look at the society-wide, aggregate effects of gun prevalence.
The same is true for medicine: something helpful in some circumstances may be outright harmful, useless, or negative net value in others.
And this is of course only if one accepts that the evidence in favor of medical treatments in general is as compelling as a hypothetical RCT on “can a gun kill you?” would be.
On which, regarding p-values I hope it is true that standards of evidence are improving. But surely most medical treatments, and especially those that drive the bulk of medical spending, have been approved historically. Recent uplifts in evidence standards presumably then aren’t of much relevance? Or am I missing something?
This topic has made me think about my own medical care, which I generally walk around thinking has been decent, and definitely worth it, though of course not perfect. But I just sat down and did an inventory of all the medical problems and care I've had, and rated each class of care. I gave pluses to things that had undoubtedly helped me, and minuses for bad features of the procedures, including stress, expense, pain, and iatrogenic features. And when I added up all my ratings, medical care was a net negative.
Lots of caveats: I did not include immunizations in my accounting, and I regard them as an unmitigated positive. And of course I’m an n of 1, so I’m a datum not data. And I have had good health all my life, so none of my medical treatment has saved me from something that could have ruined my life or killed me. If it had, my ratings might be very different. And of course I understand there’s lots of room for disagreement about how heavily to weigh various good and bad features and consequences of medical procedures, and that my accounting doesn’t prove anything. Still, the result surprised me, and I’m curious whether anyone else using a system like mine ends up with a negative evaluation.
*Orthodontia as a kid. My teeth really did look pretty awful before I had braces, and afterwards were aligned quite neatly. On the other hand, my dentist used an old style of braces that involved placing metal bands around the teeth close to the gum line. I now have an exceptionally bad case of gum recession, which I’m told those bands contributed to. However, I do various dentist-recommended things to keep the exposed surfaces from developing cavities, and so far none has.
So overall: + -
Routine care:
*Cleanings: I believe the Dental Authorities that scraping off plaque is important, so cleanings are entirely a +
*I have had about 4 fillings and one root canal. I’m sure those 5 teeth would be gone or else hurting like crazy without the dental care, so that’s entirely a +
Overall ++
Non-dental medical stuff
* Had 6 or so warts on my hands as a kid. Doctor tried dripping some acid on the biggest one, and that hurt like crazy for a whole afternoon. A few months later all the warts disappeared on their own.
Overall. -
*Had my first UTI as a teen, did not know what it was so did not go to the doctor. Eventually I developed back pain and a fever when the infection reached a kidney and I went to the ER, where they gave me antibiotics that wiped out the infection. Seems to me that without medical care that infected kidney might have made me quite ill.
Overall +++
Birth control
Pills worked but made me gain 25 lbs in a year so I stopped them + -
Had an IUD inserted and a few days later developed a uterine infection with discharge and fever. Device removed. -
Diaphragm: It worked, but after getting it I had maybe 20 uti’s over the next couple years, and had to take many courses of antibiotics. Also visited a urologist whose exam was by far the most painful thing a doctor has ever done to me, and exam result did not find any abnormalities. Eventually I figured out that my diaphragm was a size too large, and was stretching me in a way that made it easier for debris to enter my urethra. - - -
Overall: + + for no unwanted pregnancies, - - - - for quite significant iatrogenic problems
Fertility treatment:
*Had 8 cycles of IVF, which were extremely stressful and did not work. During the first cycle the anesthesiologist made a mistake when knocking me out in order to retrieve my eggs. There are 2 drugs that I know of used in that kind of anesthesia: One that paralyzes you and one that knocks you out. He gave me some of the paralytic when I was fully conscious. My eyes went out of focus and I felt myself becoming unable to move. AND he put a mask over my nose and mouth and there was no flow of air coming through it. I have a vivid memory of trying hard to inhale, getting no air, and feeling the edges of the mask pushing deeper into my face because of the suction I was causing. I was terrified, and clumsily waving my arms around and trying to say “knock me out,” but could only produce incoherent sounds. (Afterwards I described what had happened to some anesthesiologist higher on the hierarchy than the one who’d tortured me — who was maybe a resident? — and he said he believed me, and would make sure nothing like that happened during any further surgeries I had with their group.).
Fertility treatment overall: - - - -
Some kind of post-viral syndrome after the flu
I had it for several *years*: Hypersomnia, exhaustion, exercise intolerance, pretty bad joint aches and several joints that were extremely tender, fever of 99.5 that came and went. My internist tested me for everything we could think of, and all tests were negative, then she kind of washed her hands of the matter, and began to imply there was no point in using time in my appts with her to discuss the problem. She never even told me that post-viral syndromes are a thing.
- -
Prophilactic scans and exams:
*Blood work: Does not bother me, but they never find anything except stuff whose validity seems kind of questionable to me, such as low vitamin D levels. On the other hand, I do like knowing that everything’s in a good range, even my cholesterol — and that one’s important to intervene with (I think) if abnormal.
*breast scans: Have had a bunch, about 3 of which involved some drama: 2 callbacks for further scans, one for a needle biopsy (which was aborted when the doctor found the spot in question was a cyst that popped and drained the minute his needle entered it)
*colonoscopies: Of course I loathe them, like everybody else, and they never find anything. I read recently that the pre-scope purge disrupts the gut microbiome, & that concerns me some. Also, staying up all night purging then getting knocked out and probed with a plumber’s snake screws me up so much that I don’t feel normal for a couple days afterwards. So counting the pre-scoping diet, where you can’t eat anything except shit like jello that I never eat and the post-scoping malaise, every colonoscopy takes a week of wellbeing out of my life. Oh yeah, one time I got a bill from the place I had the colonoscopy — turned out my insurance covered the procedure, but not the place where the colonoscopy was done.
*bone density: Docs push these, and I’ve had some, but now doubt they’re worth having.
*CAT scan fishing expedition: Once had a weird pain in my lower abdomen, and doc ended up sending me for a CAT scan, which found nothing in the area that hurt, but an “incidentaloma” in my upper abdomen on the opposite side: A duct leading out of my gallbladder
was dilated to an abnormal size. Doc who’d ordered the scan wanted to do the mother of all endoscopies. It had a 20 syllable name that I’ve forgotten — involved sticking a tube with a camera down your throat, but then once in the stomach they fucking *keep on going* into the gut somehow until they can get a look at the gall bladder. I looked up the procedure and found it caused pancreatitis 5% of the time. So I suggested waiting a couple months then rechecking that duct via ultrasound. Doc said OK and the duct was normal on the recheck.
So these exams earn a LOT of minus points. I’m going to give them 4 minuses, because of how goddam many of them I have had, just because I usually get whatever my internist recommends, and because of the stress, expense, discomfort and, in the case of the threatened gut probe, danger they have caused me. I’m not sure how many pluses to give the exams. I’ve read recently that both mammograms and colonoscopies save relatively few lives. Also read something pretty persuasive about how noisy the data from bone scans is, and how it’s not possible to compare readings from 2 different machines. And besides, I don’t trust the drugs given for osteoporosis. Blood work is easy, and I have the feeling that it would pick up a lot of subtle things that are taking a bad turn, so is worth done. (But I’ve never seen data about how many lives or days of serious illness are avoided by getting bloodwork, so I may be wrong.) So I’m going to give this mass of time-consuming and unpleasant prophylactic exams 2 pluses.
Wow, I'm sorry you've suffered so much with the contraception and IVF! That "conscious but paralyzed" part is the stuff of nightmares. I'm going to have my first colonoscopy in a couple of years, so that's something to look forward to :)
Here's my n = 1: I've had lots of treatments in my life, mostly for small things, but I had an appendectomy at age 10, without which I almost certainly would have died a painful death of sepsis, so that right there makes my lifetime health care use a ++++++++++++++++++++++++
Also, I had scarlet fever as a baby, I have no memory of this but my mom told me I was horribly sick and had to be taken to the hospital, so there's a good chance I would have died at that point without modern medicine.
For me, the most painful medical procedure ever is a close tie between having a wart frozen off my thumb and having my stitches removed after the appendectomy (it was many years ago in Poland, and they didn't use self-dissolving stitches then). The nurse was simultaneously digging the stitches out of the scar on my abdomen with a metal instrument and pouring a disinfectant on them, and telling me to stop crying.
Also, the use of epidurals in childbirth is a +++++++ for modern medicine. Unmedicated childbirth feels like someone slowly ripping your lower abdomen apart with a dull knife, repeatedly, over the course of hours. No. Thank. You.
Yes, if your life has been saved by medicine the math is different. Still, maybe it would make sense to separate out the life-saving responses from the rest. I actually do not doubt that in most US ER's someone with a common, life-threatening problem will get life-saving treatment. Seems like maybe where health care falls down is with more routine things, where there's a lot of ineffective investigation, and some treatments that cause harm. That was my experience, and I think it's likely to be a pretty common harvest from the routine stuff. What do you think of your medical care aside from the 2 emergency interventions when you were a kid, and the pain epidurals saved you? How'd the doctors do with more everyday things?
As for my bad ob-gyn experiences: The thing is, I actually had a pretty well-running reproductive system. I never had bad cramps (except for the year when I was on BC pills), or any of the other things things that are pretty common -- polycystic ovary syndrome, endometriosis, infertility at a young age (mine was just due to being in my 40's), severe PMS, etc. ALL of my bad experiences were due to medical interventions. I'm especially indignant that over 20 visits for UTI's it never occurred to any doctor to check the fit of my diaphragm when I told them that the infections invariably started the day after I'd had intercourse. I doubt that an over-large diaphragm is a rare problem, and with no specialized knowledge about fitting the damn things or about UTI's *I* eventually figured it out: Wait, I can actually feel the bottom of the loop right inside, and even protruding a tiny bit. Oh, in fact there's a stretching feeling where the diaphragm is pushing its way out. Wait a minute, could it be that . . .
I'd be long dead without modern medicine, so I have a special dislike for Hanson's elaborate arguments that - to me - amount to telling me I don't deserve to live.
Drosophilist would also be dead without modern medicine. I wonder if it's more in the "routine care" that medicine falls down. Note that the contingencies and rewards in routine care are very different from those in treatment of grave illness or injury. In the latter, there's quick feedback, visible to all, regarding whether the medical interventions succeeded. In the former there's not, and of course the medical system needs to earn its keep by doing *something*. I keep running across articles making pretty persuasive arguments that lots of routine stuff has little or no benefit: yearly physicals, various screenings and tests.
So my case is an interesting illustration why there may be a statistically sound case for routine care being less than useful, but in this case it was literally life-saving. A malignant melanoma was found during a routine-care visit, and it was just in time to still be at a stage where it didn't have a chance to metastasize - barely. The GP literally cut it out right there and sent to the lab, and I was under the knife for a wide excision a couple of days later.
Regarding the gun analogy, it seems to me that a lot of medical studies are "apply medicine, see if they get better afterwards, control for placebos, etc." like seeing how many people die after getting shot by a gun. Are there any medicines we can test in the manner of "actually looking at what a bullet does to the human body?" Like analyzing how living tissue reacts to the drug that's supposed to treat it (or the bad thing that's supposedly harming it)?
Yes, many. But it's too easy to do this (at least on a superficial level) even for things that end up not really working, so people don't take it very seriously.
One model that Hanson could have could be something like the following.
Conceptually, we can think of the process by which we decide which medical treatments to do as something like this:
1. Rate all the treatments available on a scale of how worthwhile they are, from 0 to 100.
2. Do all the treatments above some sort of threshold, say 50.
Now of course our system doesn't do that explicitly (although we might make better decisions if the system did!) but the way the system does work could be factored out that way. For instance if you do a study to figure out which antidepressant works the best, that's part of (1). But when you have a guideline like "anyone who scores above X on such-and-such depression questionnaire should be given antidepressants", then changing the value of X is a change in (2).
It seems like a coherent viewpoint to say both that:
(i) The system we already have does (1) pretty well; we have a pretty good sense of which treatments are most worthwhile (e.g. vaccines are at the high end of the range, back pain surgeries are maybe at the lower end)
(ii) But the threshold in (2) is set way too low; rather than doing everything above a 50 we should only be doing things above 80.
So, the answer to the question of "which specific treatments should you cut" is "the ones that seem iffy based on our current state of knowledge - i.e. if it seems iffy then it's probably actually bad, if it seems mildly good then it's probably actually neutral, if it seems incredible then it's probably good."
===========
I think that in general, a lot of economics arguments work this way - i.e. you can tell if a parameter is too high or too low even if you don't know the specifics. For instance if it were the case that:
- the government massively subsidized widget manufacturers due to lobbying from the widget industry
- and there were no positive externalities, network effects, or anything else that make those subsidies economically beneficial,
then that is sufficient to tell you that probably too many widgets are being produced, even if you don't know anything about which specific brands of widgets are good or bad.
But you're right that it's unclear what Hanson's model actually is. It doesn't seem like his model is **just** that government subsidies are messing things up, it seems like he thinks that **individuals** are making bad decisions. One thing that is confusing to me is that he seems to think that:
(a) people are spending too much money on medicine, and not enough on preventive measures (diet, exercise, etc.)
(b) this is explained by status-seeking and conspicuous consumption.
While (a) seems very likely to be true, I don't see at all how (b) is supposed to help explain (a). Most preventive measures are a lot more conspicuous than medicine! We eat with other people all the time, and whether you're overweight, muscular, etc. is easy for other people to see. But our culture seems to me to treat medical decisions as a particularly private domain, and inquiring about other people's medical decisions is usually perceived as intrusive. This seems the opposite of what you would expect if people were using medical care as a status-seeking method. It seems like the "status-seeking" hypothesis would predict that people would spend too much on highly visible preventive care, rather than too little.
A possible steelman of the trichotomy would be the following:
- There is a small amount of interventions (like vaccines) that are net positive.
- However, we don't have a good way to distinguish between mildly positive, neutral, and harmful medical interventions.
- The mildly-positive-to-harmful group is net neutral on average.
- Therefore, we should discard everything that isn't obviously positive; trying to move the line in order to include a little more will start including some negative interventions as well, and therefore is a monkey trap.
Does it look like a consistent position to you? I think I could see Hanson holding it.
I think the charitable interpretation of Hanson’s position would be:
(Due to how healthcare is set up) the default action is to give medication. It affects nearly all types of treatments equally (antibiotics are good, but they get prescribed for every person with a running nose, whether they have bacterial infection or not) so figuring out which medicine is ‘good’ is irrelevant. Knowing this, a person should realize that for cases like this they shouldn’t take antibiotics.
I disagree with this last part, as the average person is unlikely to make good medical decisions on which treatments to ignore, but the general point seems mostly correct.
That is a simplification for rhetorical effect. Here is the actual concept gestured at. If you are maximising a smooth function (for example, a utility function) that has a maximum in the interior of its parameter space, then as we all know from first year calculus, its gradient there with respect to all of the parameters is zero. At the maximum, the marginal effect of changing any parameter up or down is zero. Small finite changes to parameters will have only extra-small effects, perhaps too small to see.
Thus, "at the optimum, everything is useless at the margin."
This has no implications for whether anything is just useless.
The 250,000 medical error deaths is crude, but it is far more sober than the alternative studies which are obviously wrong. Counting only recorded errors is an obvious bias. Studies which claim that there are fewer iatrogenic hospital deaths than central line infection deaths are drunk.
"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 wonder if part of cutting "frivolous spending" could be something like, if you have some pain that doesn't seem serious wait 48h and eat well, sleep well before going to the doctor. "that doesn't seem serious" is doing lots of heavy lifting here, but also it seems doable to have some kind of master flowchart to orient people, that's refined and improved from time to time.
On the other hand I don't know enough about american healthcare to know if people getting medicine when they don't need it is a thing that really happens. My impression is that due to the price people seems to be getting less/avoid getting treated, but there's a good chance it's my bubble.
There may be also cases where medicine is not exactly the right answer. For example, maybe X% of back problems would be better treated by physical therapy/getting more muscles than by oxycodon/surgery, and since those in that case are useless and can get pricey and make people worse (addiction) all medicine looks worse as a result.
But this is all a case of "let's look very carefully at which parts are good, which parts are bad and remove them" which doesn't seem to be Hanson's position.
Now time for the woo stuff, I feel like sometimes we have an oversupply of "we can find out what's the good and the bad part and remove the bad" but I have a vague feeling that it's sometimes not possible. It applies more to stuff like "this guy writes really well, but he would be better if his politics aligned with mine and in fact I think he could and should exist like that" but maybe Hanson's position is that it applies to the medical system, that the good and the bad can't be split.
Edit: I commented too early and ended up repeating some of Scott's points, especially about the back pain. I'd love a thing where the window for editing a comment is to the right of the article so I can edit as I scroll down the article.
So... there's a rather obvious way to resolve your two positions from my perspective.
You mention Great Britain as opposed to the USA - so far as I know, the utilization rate of medical care in both countries is more or less equal, but the PRICES paid per unit of care are dramatically lower in the UK.
It is theoretically possible to "cut half of medicine" by simply cutting the prices paid for care by half. (I am not advocating for this - I've experienced this as a small business pharmacist at the hands of health insurers and their PBMs, and it is NOT pleasant for anyone involved).
You are focusing on utilization cuts, when I'm pretty sure Robin's unit of measurement isn't care units delivered, but dollars spent on care.
The US spends dramatically more per unit of care than any other country. And the results of the insurance studies are consistent with this trend - a decent fraction of care delivered to uninsured people ends up being uncompensated care, i.e. the dollar flow for that care delivery is exactly $0. By shifting folks onto medicaid from uninsured status, you move all of the previously uncompensated care into (lower reimbursing) medicaid-compensated care, and that is, from a societal perspective (and maybe Robin's?) marginal care. Paying physicians and hospitals for care that they would have already delivered for free (bad debt writeoffs at the end) is pretty obviously near-zero marginal value to health status of the individuals - they got the medical care that they needed for free without medicaid, and the state paid for it with medicaid.
So I think that the way to read the Oregon study in favor of both of your positions is that 1) Medicaid enrollment induces some amount of marginal care delivery, which results in additional diagnoses (good!) 2) Medicaid enrollment means that hospitals (especially hospitals) get paid for a lot of care that they were previously delivering for free, which is, economically speaking, marginal deadweight loss. Taking both of those effects together, marginal healthcare SPENDING is net neutral. (disclaimer: stablizing hospital finances is not necessarily a bad goal, and is definitely more FAIR than expecting hospitals to deliver care for free, but it's not the primary purpose of medicaid).
I, of course, agree with the general principle that a fair amount of medicine is not helpful (take, for example, most use of proton pump inhibitors). I will add that a bunch of SPENDING is on schemes to game the third party insurers' reimbursement mechanisms (take, for example, Zegerid, a combination of omeprazole with baking soda, which costs approximately 70x omeprazole alone - that marginal spending is for BAKING SODA). I assume that such schemes are not unique to pharmacy, and are, in fact, widespread across reimbursement systems for medical care in general (surgeries made slightly more complex or delivered via robot when they could be done more simply and with equal effectiveness without the added complexity and/or robot, which serve the primary purpose of getting insurers to pay more money).
You keep referring to antibiotics. On average they are beneficial, but at the margins they are questionable. Every time I’ve been prescribed a z-pak, I probably works have been fine without it, but ins covered it so I took it. All the ear infections my kids got, sometimes they needed the antibiotic. But sometimes we’d get an rx and the doctor would say wait a couple days. Sometimes we waited, but sometimes we just filled it.
The point is, antibiotics are over prescribed. We should use less. But I don’t think doctors like to be the bad guy, so they overprescribe.
I think we could say the same about cancer. Treatment can save lives on average. But people also stay on treatment when it is obviously not helpful.
Back pain. I know some people that have gotten help from medicine. And others that have surgery or PT and still have pain. There are a lot of pains we just can’t do much for.
So it seems clear that average med is beneficial. But marginal is questionable. It is sometimes clear what extra med isn’t helpful. Sometimes it is not clear. But that applies to many things I purchase. So letting individuals make the choice seems right. Exposing individuals to more of the cost will incentives them to make the choices.
Has Robin Hanson considered the ethical consequences if his suggestions would be followed? It's one thing to cut expenses for seemingly ineffective treatments that supposedly result in statistics he thinks show medicine is useless. And another to cut expenses for medicine which is purely palliative. Maybe he hasn't considered what's going to feel like when insurance will cut his access to palliative interventions because he'd already used his 5 interventions that year? Why should anyone reimburse him for expensive morphine for example? It's ineffective anyway. He'll die anyway.
Is the "monkey trap" thing even real? Where the monkeys self entrap because they can't let go of a snack, to the point where they can be caught by a predator?
I've heard this a bunch of times and I've seen a video (but like, how long did the monkey actually remain stuck, I don't recall) and never was particularly skeptical. But somehow hearing it from an economist ("burly coolies", "cobra effect") makes me suspicious. This guy (https://samknight.com/?p=924) says he could verify none of the monkey trap stories.
Also, pre-1960 this kind of thing was referred to as a [derogatory term] trap. It's a fable included in the Talmud and Aseop's collection, among other sources.
There's this video, I think from some Disney documentary, but I don't think they were above fakery.. https://www.youtube.com/watch?v=e0VI_BXlFI0
got a link to the fable in Aesop or Talmud? apparently, you can call it an apologue if it's exaggerated/fake
Is it this?
https://en.m.wikipedia.org/wiki/The_Boy_and_the_Filberts
Thanks. I recommend it, and a lot of the point of the analysis is not that it's a serious trap, it's a thing that young people fall for, and they learn to do better.
Bateson presented it as a challenge-- what *general* advice would be useful for the monkey?
I'm guessing that whatever [derogatory term] was, it wasn't used in the Talmud or by Aesop.
Probably not the same derogatory term, but not at all out of the question that many of these sources use different derogatory terms for different groups.
That sounds plausible, but I can't find a source to back that up. The only thing I found is Urban Dictionary's "a bucket of fried chicken placed under a net." (Unless I'm substituting in the wrong "derogatory term." Is it the delightfully rhyming "[East Asian] trap" instead? That gets me Pacific War propaganda posters.)
"I even contacted a zoo and was assured that, while a monkey may fuss with the prize for a minute or two, it is most certainly smart enough to let go."
Maybe this distracts the monkey long enough to sneak up on it and kill it though.
Yeah, and frogs get out of water if the temperature goes too high, even if it's increased slowly.
This is why pot lids were invented.
Ironically, the boiling pot metaphor is exactly backwards in real life. If you slowly increase the temperature of a pot, the frog will jump out when the temperature gets too high. If you drop a frog in boiling water, it will die too quickly to jump out.
Like Desertopa has said, there's no "even" here. The slow increase is the only way to get the frog to jump out.
The point about the Idea of the frog being boiled slowly is that humans aren't likely to notice gradually worsening conditions. The contrast was between conditions that get worse very gradually compared to conditions that get bad quickly, but not so fast as to be immediately deadly. Of course the frog jumps out of somewhat hot water, but maybe a slow increase would fool the frog.
> The point about the Idea of the frog being boiled slowly is that humans aren't likely to notice gradually worsening conditions.
I don't think it works in even the most direct sense. Imagine you, a human, are in a room. It's getting hotter and hotter. Will you ever try to leave the room? Would it matter how slow the increase in temperature was?
Why imagine that frogs are fundamentally defective in this way, compared to humans?
This isn't about temperature. If you country is gradually getting worse for you, when do you decide to leave?
Cold-blooded animals have to be able to discern slow increases of temperature in order to survive. Otherwise the rising heat of noon would kill every cold-blooded animal not already in shelter.
Biologists correct me if I'm wrong, but it's my understanding that every vertebrate animal experiences physical sensation, and can tell when a temperature isn't optimal for it.
I'd go further: even if you put a frog in a pot of water and don't turn on the heat, the frog will jump out. Frogs are pretty skittish, they don't wanna be in this weird metal cylinder.
Further, you catch the most flies with a mix of honey and vinegar.
I read about it in Gregory Bateson. It was described as a method humans used for catching monkeys.
Now I'm left wondering about the trap for humans described in one of the early NLP books. Scientists tested rats (mice?) on their ability to run a maze for a reward, and then on how long it took the rodents to stop running the maze if the reward was taken away. The rodents quit trying after some reasonable amount of time.
They tested the result using grad students and a giant maze baited with $5 bills in a basement. The grad students are still breaking into the basement in case there's a $5 bill.
A good allegory about lost purposes. This was from the 80s, so people weren't as cynical about grad school then. The story reliably gets a laugh, speaking of incentives. But is there any truth at all?
Rats don't have access to religious purpose, born from tradition of a forgotten origin.
The comments on whether monkey traps actually work reminds me of a proverb I've heard, "You catch more flies with honey than with vinegar."
This was intended as a metaphor, of course, but I found it amusing, because vinegar is in fact an excellent fly trap.
I think I've heard that one less often than "you catch more flies with honey but you catch more honeys being fly"
I was going to say I tried testing this myself in college, and neither worked very well, but I suppose it depends on what kind of fly you're targeting. Vinegar works quite well on fruit flies, but appeared not to with whatever kind of flies the ones I tested it on then were.
Phorid flies don't seem to care about vinegar, but look a lot like fruit flies.
(shudders) The only thing that worked was hanging fly paper *everywhere*. And of course I had to repeatedly explain to the property manager that no, they weren't fruit flies, so vinegar traps weren't going to work.
I recently catched a lot of fruit flies using traps with wine and vinegar under perforated plastic foil. A mix of wine and vinegar seemed to work best. Did not try honey, though.
I also tried this at one point out of curiosity, setting out bowls of water, honey-water, and vinegar-water. The vinegar worked best, although I admit I have no idea what it was catching.
Relevant xkcd: https://xkcd.com/357/
I suspect this is a case where an essential element of a tool that evolved in different times in the hands of different people has vanished from the telling. Imagine a world in which foot traps in their historical form have been so long vanished that children only know about live traps. The metaphor about a wolf chewing its own foot off to escape no longer makes sense, so they change it to a story about a wolf who ignores his hunger and does not go into a live trap. Still later someone in a discussion about incentives where this metaphor comes up quibbles about whether or not animals ever get hungry enough to go into live traps.
The monkey trap probably works great if its original form was anything like the DIY coon traps in rural America a century ago: it's a can with sharpened nails driven through at an angle so that the little paw fits through empty but not full. The goal is to cause the animal pain upon initial attempt at extraction, causing it to panic and disable its paw to where it can't let go. When you find it hours later its mangled forelimb is hooked on at least one nail and killing it with a club is an act of mercy.
The "Where The Red Fern Grows" raccoon trap, that is - the protagonist trapped one raccoon (dishonorably) to fund the purchase of his hounds.
Ah yes, I'd forgotten that. Is there a drawing of the trap in the book? I remember seeing a schematic for it in an old book. Perhaps one of the Foxfire series, but I doubt it (Foxfire would have snares). I think the same book had an illustrated how-to on butchering snapping turtles.
It's not. Just like the boiling frog experiment is fake and the "5 monkeys and a banana on a ladder" experiment are also fake.
Crabs in buckets too, probably.
This one is, from experience, very true : crabs will kill each other if you put a bunch of them in a bucket. They will also kill each other if a small crab enters into the hole of a bigger one by the way.
The monkey trap sounds about as likely as Buridan's Ass starving to death.
"...I guess I must be misrepresenting him, and I apologize."
Without reading to the end, I predict that this is the most important line in this post. Polite and human disagreement is possible and valuable!
I'm going to unironically say that this is the *worst* line in the post.
Robin Hanson is using a motte and bailey here where the motte, of course, is "medical care isn't good at the margins" and the bailey is "medical care is useless".
Scott's being a quokka here. He doesn't seem to ever be able to recognize that someone's acting in bad faith, so when he's presented with a motte and bailey, all he can do is mutter about how he "apologizes" for pointing to the bailey and "misrepresenting" it as the motte instead of calling out for what it is.
Robin's actual proposal is to cut medical spending in half. He was explicit about that from his opening essay at Cato Unbound, it wasn't a position he retreated to after his interlocutors pushed back. If other people agreed with him on that then he would no longer qualify as a contrarian on that issue!
So, does he have any strong opinions on *which* half ought to be cut? If not, I think this still resolves as "Hanson thinks all medicine is equally bad."
He has also explicitly said that some treatments work, but he doesn't propose to select which treatments work himself. He's an economist and his proposal is instead to cut spending, and other people would make downstream decisions on which things to cut. There aren't people arguing "We should spend money on useless treatments", so merely saying we should get rid of them doesn't move us from the status quo (similar to people saying they will lower the budget by getting rid of "waste, fraud & abuse").
But, at least according to some of the quotes in this article, he disagrees with people who think we should expend any effort figuring out which half to cut. And the first bit of this article is pointing out that (at least as Scott sees it), Hanson has made individual statements that seem to disagree with _any_ particular way of figuring out which half to cut.
If medicine is taxed rather than subsidized, people will simply make choices about which things is cut and I don't think Robin would object then. If people insist on first figuring it out before cutting, then I think Robin won't expect it to get cut.
>But, at least according to some of the quotes in this article, he disagrees with people who think we should expend any effort figuring out which half to cut.
Doesn't
>eg Britain spends only half as much per person as the US on healthcare, and Brits have _approximately_ as good health outcomes
imply that British NHS decision makers _already_ know where to cut?
And, presumably we (writing from the usa) could either
- copy their policies
- have them remotely make these decisions for us (presumably at a price)
- kidnap some of them and install them in our health care system
and similarly for individual decisions such as
>So my second question for Robin is: how do you recommend I proceed? Do I avoid going to the doctor for some specific subset of these categories, like 5-10?
Admittedly, this isn't _exactly_ right. The UK and USA populations are a bit different, and e.g. the fraction of medical care that goes to treating e.g. gunshot wounds probably doesn't match across the pond... Still, the British experience seems to imply that _most_ of the right choices are already known.
People keep saying that. He was very explicit, e.g. end subsidy, read Cochrane reviews. He didn't personally review every extant medical intervention, is that what you mean?
Did you intend to reply to someone else? I agree he wants to end subsidies.
In countries with public healthcare, the government decides what to spend it in. The free market has nothing to do with it. In countries with private healthcare, courts decide what you can sue your insurance company for not paying for. The free market has nothing to do with it. So where exactly is he planning on implementing this?
I think you're right about the motte and bailey but wrong in your judgment of Scott's reaction. Assuming good faith is often an effective play in this sort of situation. If it turns out they are actually arguing in good faith and just communicating badly, it avoids you being a dick for no reason. And more importantly, even if they are acting in bad faith, it allows you to avoid *looking* like a dick when you politely press them over the "confusion," which is exactly what Scott does here.
Assuming good faith makes you look epistemically virtuous, but sometimes at the expense of causing a slight deterioration in collective epistemics. People end up thinking a view is more plausible, or a thinker more honest, than they should, and would if they had instead been exposed to a critic more willing to call a spade a spade.
If you assume good faith and they're doing it in bad faith, they're going to say "I'm sorry you're confused but..." and keep running with the scam.
Hanson has already responded to Scott's response, and his re-response is basically "here's some more motte that I've posted in the past, see, I have actual quotes"--he's doubling down on the motte and completely failing to address the *inconsistency* between his past motte-posts and his past bailey-posts.
It is possible to read the whole "oh poor me I must be mistaken I so sorry" line as vitriolic sarcasm. I suggest trying it.
Scott's past writings don't support that interpretation. He couldn't even admit that *Cade Metz* was malicious all along.
I'm a fan of how Scott handles this sort of thing! If he wrote Robin off as acting in bad faith then I would have very little interest in reading this post, since Scott would have no reason to address any of Robin's arguments outside of pointing out contradictions. I think if as a writer you don't think the person you're addressing has a real position that can be understood, it would be a waste of words to write about them in any detail.
When you politely assume a person is smart, and you may just be misunderstanding them, you're leaving the door open for them to clarify themselves. In the best case you might understand their position if you restate it and they correct you, but in the worst case they make a more specific contradiction or weaken their argument and either way you prove your point.
You might think there's a downside to being a quokka (never heard that one before lol) but from where I stand it's the best way to address an opponent, and also selfishly I would not be very interested in reading Scott just dunk on someone's contradictory position without really getting into the weeds of what they believe and making an effort to steelman them.
If someone is not sincere, "the weeds of what they believe" is a mirage. It may not be possible to get what they believe from what they're saying because their words and their beliefs *aren't connected*.
Oh totally, but it's very hard to be 100% sure someone is insincere, and doesn't cost you much to assume (or pretend to assume) that they are acting in good faith.
The alternative to responding to someone as if they were sincere shouldn't be calling them out for being insincere, it should be ignoring them. If you think you possess overwhelming evidence that someone won't engage in good faith, a big breakdown of their position like this followed by a "you must be a contrarian" call out won't change their mind anyway. And if Scott is making this post to try to convince the audience that this is a nonsensical position, he might as well give the benefit of the doubt, to look to the audience like a fair and even minded evaluator instead of someone trying to find extra reasons to dunk on a bad argument by claiming that anyone making it must be arguing for the sake of arguing.
Agreed. I hadn’t heard of Hanson before now, and for me to be able to confidently dismiss Hanson as a kook or a troll, I need to be able to tell that Hansen genuinely has nothing useful to say. A discussion that left open the possibility that Scott and Hansen were simply talking past each other would be too ambiguous for me to conclude much of anything without reading more of Hansen’s writings. Attempting to seriously engage a troll may be unrewarding work, but it provides a record that allows other people to determine with relatively little work on their part that the troll is indeed a troll or a kook.
Based on the exchange so far, I’ve provisionally placed Hansen in the troll/kook category, but I need to see Hanson’s response to this latest post by Scott before before I can pass judgement with confidence.
Here's Hanson's response. It's not good.
https://www.overcomingbias.com/p/second-response-to-alexander-on-medicine
Prof. Robin Hanson is no troll. Though known for surprising ideas. Same GMU-faculty as are Caplan, Cowen, Tabarrok. And very active and well known in EA/rationalist circles. A friend. Who is now surprised Scott did not just call him to clear stuff up beforehand.
Quokkas rule, therefore Scott does, too. Your meme metaphor has unwittingly played straight into my hands!
... Don't quokkas throw their children at predators in order to save themselves?
I have teenagers. If this is true, it just makes me love them more.
being a good faith arguer regardless of who you're arguing with is in fact virtuous and valuable.
Too many people think that once you learn about a fallacy, the only thing you should do is diagnose the fallacy.
It's usually much better to use your knowledge of the fallacy to press the real question, and leave the diagnosis of the fallacy implicit. I know it *feels* like you're letting the terrorists win when you don't call them out. But calling out isn't the goal - understanding is.
I read that line as basically "Robin Hanson says I'm misrepresenting his opinion, and Robin Hanson is an honorable man."
Indeed. I've rarely seen anything as low key vicious as, eg, "So this was exactly what I said in my post, except that Robin takes out my explanation and quotes only half of the section, so that I look like a moron who didn’t read the paper."
Specifically, the sentence seems subservient, but it's much more devastating than a direct "Robin is lying here". Politeness and respect can hide devastatingly sharp blades.
Yeah, that was the vibe I got.
I endorse this in principle, but in practice it feels kind of weird because the rest of the essay goes on to say "...but actually, Robin Hanson *did* say what I'm accusing him of, here are the quotes."
I don't read it as an accusation so much as an invitation to communal debugging:
- Robin Hanson claims he doesn't say what I thought he did.
- This is a highly detailed paper trail of why I came to believe that he said that.
- Help me find the contradiction.
This way of dealing with the views of others is sometimes referred to as the golden rule of hermeneutics.
That is, if you present the views of someone you disagree with so that their views seem obviously stupid and/or evil, you are probably misrepresenting what they believe. Since very few want to come across in the eyes of others as stupid and/or evil.
Instead, present your opponent's views so that they may possibly make sense to you.
That said, sometimes some people are really stupid and/or evil. (As Kurt Vonnegut remarks in The Sirens of Titan: The trouble with stupid people is that they are too stupid to understand that not everybody are stupid.) But if so, let that "show itself" through further communication. Do not have that as an ...eh.. prior. Since that will block communication.
A lot of medicine is useless, and fear of death makes people overspend on medicine, but a lot of other economic activity is also useless and motivated by arguably irrational reasons. Is it inherently more problematic that people overspend on medicine than on luxury cars for instance? If people did follow up studies on other kinds of consumption there would be a lot of results showing that it’s useless or worse at improving people’s lives.
This is still useful to know for my private life. If a doctor suggests surgery for back pain, I'll know that it's useless and won't do it.
You won’t know that about any specific case of back surgery.
Yeah, I want to be really careful not to make anything that looks like a medical recommendation here. There's a lot of overtreatment in back pain, but I bet there are specific kinds of back pain where surgery is effective and important. I would recommend starting off from a position of mild skepticism, but doing some research and asking your doctors frank questions and maybe getting a second opinion, and letting those things push you to yes if they all seem plausible.
Or just find a specialist who thinks that there's a lot of over-treatment in back pain, and go from there.
Going to a back surgeon with a herniated disk and seeing on the intake form, " Is this your first, second, third, fourth, fifth.....fifteenth.... back surgery," was very eye opening.
I apparently got pretty lucky in my back pain adventure. I have (had? I was told it might eventually heal and I haven't had a check in a long time) a partially herniated disk that was, for a while, causing quite significant (although not debilitating) back pain. I saw a specialist who recommended physical therapy and, even though he was a doctor who performed both steroid injections and surgeries, told me that the level of pain I had was not enough to warrant something that invasive.
He was absolutely correct, because now, several years later, I am _mostly_ pain free (I'm unsure how much the physical therapy helped vs. it just got better with time), but apparently the fact that I wasn't pushed quite strongly towards a surgery or at least a steroid injection is somewhat surprising.
I do think that the MRI that my non-specialist doctor ordered was probably a waste. He should have just sent me straight to physical therapy and saved me the couple of thousand dollars on imaging that didn't (from my perspective) seem to impact my care very much.
Good point. I know of one person in particular whose back surgery got him out of a wheel chair and walking and swimming normally.
This is not a "gotcha" - but isn't this statement functionally equivalent, from an economist perspective, to what Hanson is saying? You both agree there's a continuum of cost vs benefits on back pain treatments, and you both agree that actual dollar costs go up the more you lean on surgery/medicine vs non-medical interventions, you both agree that the efficacy of those costlier medical interventions varies a bunch, and that it is often hard to tell.
Hanson says "therefore, spend less on medical treatments of back pain" and you say "therefore, approach the use of medical treatments of back pain with mild skepticism, get second opinions and only accept treatments that seem plausible."
But these things seem very similar, in both the near and far way? "Medical interventions may be more costly/less effective than you think!" and "you should approach claims of medical intervention efficacy with mild skepticism, and demand a higher level of plausibility" result in the same actual actions by patients when confronted by back pain.
In both cases, in an economics context, you are both saying "when doing a pro/con on whether to get a medical treatment, you should put less weight on the pro side." Hanson isn't medical, so he can't provide *medical* recommendations on whether a given treatment passes the cost benefit test, (whereas you can) but he can look at aggregated data on outcomes, and make *economic* recommendations (i.e. based on the patient's desires).
At a high level, I think RH is thinking on the margin of the economic incentives: that if we convinced Americans to cut their aggregate medical spending by 100m, say by reducing medicare reimbursement rates across the board by a %, this would change every single one of their cost/benefit analyses for treatments. This would have a marginal impact on decisions, but would only actually change patient's decisions to seek medical treatment in a tiny % of cases: those where the patient's perceived net benefit from the treatment, minus costs, was greater than the small % change: and this (by definition) would be the (allowing for notoriously bad patient assessment) least beneficial cases: i.e. the guy saying "yeah my back hurts, but I've decided to lift weights and do PT from youtube instead of the now-more-costly-surgery". RH is saying that cost cut would, on the margin, likely make every patient better off (and also the payers, who have cut costs).
Obviously, yes, there would be some patients who go "100 bucks more to pay to treat my gunshot wound? No thanks!" and then die but RH is betting that those cases are rare enough that the benefit from the other vast majority of cases where patients do properly weight their cost-benefit will outweigh that.
Many more people have enough scientific literacy to investigate medical questions for themselves than actually do. If you are able to read literature and understand it well enough to have a meaningful back and forth with your doctor (to ask a question like "a systematic review found only low-quality evidence that people with my condition benefit from this surgery more than they would from physical therapy, why do you think it's recommended in my case") I feel like you should do that--especially for something as intensive and risky as surgery--rather than take a hearsay observation, even from someone you respect who is also a doctor, and run with it. This has got to be one of the most fruitful uses for a scientific education. I don't mean to impugn you specifically, especially not knowing your background, but I see people reasoning about their own medical care by lurching from one hasty generalization to the next all the time (I have surely done it too) and I'm really curious about what reasoning actually governs these decisions and why, besides laziness and time constraints, we rely on it anyway.
A close relative of mine got back pain surgery that helped, so careful with following this kind of blanket statements.
I think we ought not to treat back pain as a specific disease, because we'd get in the same trap as treating headaches as a disease. That's a symptom and not a cause. The cause is what's relevant. Try to under-treat a herniated disk(take a severe case), then we'll see. Don't make general statements about surgery. It might save your life some day.
Luxury cars are not generally paid for with other people's money. The enormous amount of medical spending that is paid for in really indirect ways makes it much more susceptible to perverse incentives.
I believe it's part of Robin's point, but we don't subsidize luxury cars and definitely don't guarantee that everyone can get a luxury car. Some people argue that medical care is a human right, and almost everyone even in the US agrees that certain care should happen regardless of ability to pay. I think if we allowed people to get whatever medical care they can afford (as they can with luxury cars), that he would be just as okay with that as someone cutting half of all medical expenditures.
That is a good point but in the Fear of Death and Muddled Thinking article he seems to concentrate most on how this is all a problem of human rationality, not one of subsidies.
The countries where health is subsidized more by the government are more keen on cutting down on costs. So there’s a certain paradox there.
The US might have one of the worst overall systems when it comes to cost incentives. I think the cost incentives in countries with universal systems is bad, but not nearly as bad as in the US. In universal systems there's at least pressure on the payer (government) to limit the care given.
Big picture, the US separates the people paying for procedures from those deciding when/how/who for those procedures. Employers pay for employees to get coverage, but can't decide how it's spent. The government mandates certain levels of coverage, but puts in few to no meaningful limits on that spending. Insurance companies pay for the care, but pass along costs to others. Hospitals convince patients they need care (right or wrong) while the patients only pay a small portion of the overall costs.
It's like if you were buying a car for your own use but someone else was paying for it. You might decide to be frugal and careful to get a car, but there's no compelling reason to do so. Some people would get exactly what they needed, but most people would instead fit into one of two buckets - 1) they would buy a car they otherwise would not have gotten, whether because they couldn't afford it or because they didn't need it, or 2) they would get more car than was needed for their purposes, at least getting a few upgrades to features but maybe even a higher level of car entirely.
Well, yes. But there's a difference -- I think -- to taking what's essentially a placebo, or most likely to be, or at best "just in case, won't harm" or a vehicle for improving one's feeling of control and agency (I take vitamins + couple other supplements purely on this basis) and getting treatment that we believe will DIRECTLY improve physical state of the body in the desired direction. One reason for that being that many (maybe most) effective treatments have significant costs in effort, pain, time, potential side effects and if you live in the place like the US, also money.
So I can pop a 4x DRA of D3 daily because the cost is minimal and it gives me a good feeling and who knows, maybe leaving at a cloudy 57N I need it. But I'll not take a SSRI for moderate trauma induced anxiety because the cost (potential side effects and the dangers of long term use in this case) are not worth the possible benefits. IF I knew that that very same Sertraline was guaranteed to glue my broken brain back to where it was 6 years ago, I'd take it for sure.
I'll have massage because it feels good to be touched by another adult human, and it's relaxing but I'll cancel the appointment if I don't have money for it. Had I believed it was HEALING my hip arthritis I'd prioritise this spend. Etc etc.
Rarely, if ever, have I seen someone go to such lengths to prove something so obvious.
Huge respect for Robin Hanson, but this is an issue where he seems to have fallen into the contrarian trap of defending a stupid premise purely to show off how cleverly he can defend it. "Some ideas are so absurd that only an intellectual could believe them."
That's the impression I'm getting as well, particularly from the apparent internal inconsistencies and refusal to engage with Scott's main points.
I think Hanson is a worthwhile thinker in the Rule Thinkers In, Not Out sense, but this kind of incautious contrarianism (and lack of clarity about his own central points) isn't at all atypical of him.
Yes. It's the unfortunately-familiar libertarian game of intellectual whack-a-mole. I'm frankly not seeing here any reasons to have huge respect for Hanson (which is ironic since I'm personally a cautious/skeptical health care consumer to the point of regularly exasperating my spouse on the topic).
This newest post clarifies that Scott's frustrated and almost frantic re-engagement with Hanson's childish intellectual bullying on this topic is being fueled by the situation with Scott's son. Every veteran parent grasps how that is feeling for Scott and his wife right now. Hanson obviously does not and never will.
The misshapen head link Scott supplied mentioned that it is more common with twins. Not speaking to his situation at all, but in general the significant rise in multiple births and hence the problems associated with them, could potentially be considered one of those ways in which medicine both gives and takes.
My understanding is that the primary cause of increased flat-headedness is a success of the movement to put kids to sleep on their back rather than their stomach in order to avoid SIDS.
"Childish intellectual bullying" is what you recall Hanson responding to Scott's critique?
> Every veteran parent grasps how that is feeling for Scott and his wife right now. Hanson obviously does not and never will.
He has multiple offspring.
I really don't think that labelling this kind of thing with _any_ particular political ideology is either A) true (in the sense that only one particular ideology uses this kind of strategy) or B) useful/helpful/elevating to the conversation.
His proposal to municipalize medicine was an attempt to "pull the rope sideways" rather than arguing in favor of private over public medicine:
http://www.overcomingbias.com/2009/06/municipalize-medicine.html
I was going to say the same thing. To add, at an extreme of this phenomenon you get flat earthers: use a complex edifice of clever arguments to arrive at an obviously wrong conclusion, then make that conclusion a key part of your identity, and defend this identity at all costs.
I think that's a slightly different phenomenon; there are no flat-earth arguments that are "clever", at best they are confusing.
Modern medicine is sufficiently complicated that a logical case can, with effort, be constructed against it (albeit one that is pretty easy to demolish by someone of equal intelligence to the edifice's designer). The earth's roundness, on the other hand, has been settled science for thousands of years.
Broadly agreed, and that’s why I assigned flat-earth arguments to the extreme end of the spectrum.
Sorry, can’t help this one: their arguments can be both confusingly clever and cleverly confusing. :)
> Maybe we should skip it - I never see any adults with obviously mis-shapen heads out there
For anecdotal evidence, I had a friend in middle school with this issue. It's less noticable as an adult (possibly he learned to hide it better with haircuts or something? Or maybe it just got milder naturally), so you might not notice it passing him on the street but you probably would if you were sitting down for a conversation with him. I don't think the helmets were available back then.
I also know some people like this. It's not a big deal in the grand scheme of things, but I might have done double-takes and stared a little on first encounters.
Thanks. I'm terrible at noticing things about people, so if other people can detect this then I'm happy to defer to their expertise.
Our son also had plagiocephaly, vociferously rejected attempts to sleep on his other side and ultimately we decided to go the helmet route
It worked fine, his head is not perfectly symmetrical but definitely not something you'd notice without calipers. It requires an extra dozen or so doctor appointments but there are so many appointments those first years I hardly noticed. Also need to give him regular baths and wash his helmet and hair so it doesn't smell terrible, but again, kids smell terrible a lot of the time anyway
Meanwhile one of my coworkers said she had a flathead and was never treated for it. I honestly hadn't noticed but she seemed kind of self conscious about it, so avoiding that also seem worth the $$$
I had a family member who was born with the same condition, got the helmet, and it made a significant visual difference at that age. There's no visible sign of the condition at this point, though he's a teenager now.
I don't know how well these findings have survived the reproducibility crisis, but there are definitely studies that show that facial asymmetry is perceived as less attractive even if it isn't at the point where it's something you'd consciously notice. For this reason, I too would buy the helmet.
I had a friend in varsity with this issue as well - his skull looked a bit like a muffin that had risen out of the pan. It had no functional effects that I'm aware of, and only become really noticeable if he cut his hair short.
I hope he doesn't go bald.
My daughter (now 3 years old) had a misshapen head, probably from always sleeping with her head turned the same way. The doctor said "you can get a helmet, but I wouldn't". So we didn't, and now she looks fine.
He's a old school country doctor (obstetrician/pediatrician/geriatrician all in one, maybe 75yo) and as far as I can tell he generally would never treat himself for anything. When he does recommend treatment, I take it very seriously.
To be honest you can’t take an economist seriously on anything.
Minor warning, fewer comments like this please.
Fewer criticisms of economists or more weight to the criticism of economics? It would derail this thread to point out the flaws in most economic models, it might even derail open threads. So I’ll leave it at that.
I think the issue was with broad-brush attacks, regardless who you're targeting.
I agree. Specifically, denouncing a whole profession as unreliable, though it might be interesting to look at reliability of economists in general. It would be a hard thing to test.
>though it might be interesting to look at reliability of economists in general. It would be a hard thing to test.
Agreed on both counts.
<mild snark>
Though if one wanted to look at a profession where reliability would be _really_ hard to test, consider what a double-blinded RCT to test the reliability of clergy would require... :-)
</mild snark>
You would have to know something about economics to be able to point out flaws in the models. I have suggested this to you before, and will do so again, learn some economics. It will enrich your intellectual life. Try marginal revolution university's microeconomics 101 course
I would prefer to see fewer criticisms of anything or anyone that take the form of a single sentence saying in essence, "you can't trust *those* types", without elaboration or justification.
I'm surprised there aren't more prediction markets on economic takes.
This is the closest thing I could find on either Metaculus or Manifold:
https://www.metaculus.com/questions/3778/will-a-democratic-nation-whose-economy-transitions-from-capitalist-to-market-socialist-sustain-adequate-economic-growth/
Wouldn't what happens if we cut health care spending in half be an easy market to make?
I think the problem here is neither of you has framed this in terms of computational complexity. You seem to be objecting to Hanson’s lack of specific actionable advice at the macro or individual level. But this doesn’t make his fundamental claim wrong!
Health care is computationally opaque. The effects of anyone one purchase aren’t obvious or legible; they are subtle and interact combinatorially. The placebo effect plus the finance model of insurance mean there’s some psychological boost for _any_ intervention, and a while supply of people getting paid for you to keep using it.
Suppose you model health interventions as being probabilistic in nature - what works for some, doesn’t work for others. You can’t know what will work in advance. If you find yourself with problems, you go to a doctor, and the doctor says “well, intervention X works for 80% of people who try it with this condition.” So you try it, and it seems like things are a little better… maybe? Plus insurance covers it so you keep doing it. If most people just blindly follow doctors advice, you should expect that roughly 20% of the people taking that intervention aren’t benefiting from it.
My own experience has been, after a decade of intense focus on my health, picking up and dropping numerous approaches and interventions, I’ve finally found the right diet that works for me, and habits that on net make me better off. My wife did none of this and has overall about the same health as me.
As I type this I’m using lamisil to treat a rash I got on my foot from spraining an ankle and then wearing a boot, the sprain came from exercises I was doing to try and strengthen muscles I thought were too weak. I wish I could say this was a one off case but this general pattern has been the story of the last ten years of my life, when I focused immensely on trying to improve my health.
What I’ve discovered only in the last year or so, is that what I pay attention to and how good I feel about the present moment seem to be the single biggest lever I have for improving my health outcomes. Regular prayer and gratitude and trust in my body to manage itself seem to be working better for me than attempting to tinker in a complex system I don’t understand, and then incurring the anxiety of knowing I can’t control this very important thing. No amount of dietary fidgeting is as effective as increasing my sense that my body is a gift from God so maybe don’t stuff it with junk food.
So, yeah, if there is an answer here, I think it’s leaning into the placebo effect and intentionally trusting reality itself to be sufficiently, without knowing the specific details of how that will be. Google SRE has this motto, “hope is not a strategy,” to which I respond, bullshit, hope is the optimal strategy for navigating a complex system when your navigation mechanism (ie brain) tries to reduce the error in your valence assessment of that system. Mental health is, imo, a necessary precursor to good physical health.
I think Hanson's specific advice on the macro level is to remove all subsidies for medicine and perhaps tax it instead.
I believe that in some sense it's simpler and perhaps easier to solve a problem than to improve a system that already functions pretty well.
> frame[d] this in terms of computational complexity
Username was not chosen at random I see :)
I don't think Hanson is unaware of the computational complexity angle. The solution he most favors, according to his latest response, is to parallelize the computation by incentivizing individual insurance companies to "compute" what spending should be cut (and he ostensibly achieves *that* by mandating that medical insurance be coupled with life insurance).
Let me try to steelman the anti-healthcare position here.
I have chronic heart disease. Based on some early symptoms and family history, I was put on several medications (beta blockers, ACE inhibitors). Sounds great, right? However, I have developed chronic fatigue in the years since, which is 10x worse than my (mild) heart problems.
Years into this, I find out that the heart medications I've been taking have extremely strong system-wide effects: reduced activation of the sympathetic nervous system (beta blockers), increased inflammatory response and lowered pain threshold (ACE inhibitors), and even reduction in the efficiency of respiration (apparently, lowered heart function, which my meds induce, can lead to slight respiratory alkalosis even as oxygen saturation read as "healthy"). My cardiologists never mentioned any of these effects to me. These don't correspond to the typical image we have of side effects (rare, acute complications, e.g. "rashes in less than 5% of patients"). Rather, this is cumulative damage to half a dozen vital subsystems, throwing homeostasis way out of whack.
My point is that some outcomes are relatively easy to measure and correlate (early start with heart medications reduces incidence of heart failure), while diffuse downstream effects that sap your vitality and make your life worse are extremely easy to miss, in all but the most egregious cases. If we assume that such things are systematically happening in the treatment of many diseases that aren't immediately life-threatening, we can end up with a picture where lots of people walk around saying that modern medicine has saved their lives many times over (I used to say such things in my younger days), while simultaneously the health of the population gets mysteriously worse, in ways that are easy to dismiss with a "better screening" pseudo-explanation.
Yeah, I see this kind of thing in my practice too.
I think we really need to spread the announcement "If you take a medication and feel much worse, then unless this is part of the plan (eg everyone knows they'll feel worse on chemo, but it's worth it), tell your doctor and unless they have some great counterargument consider stopping the medication."
Did you eventually stop the medication, and did your fatigue improve?
A cardiologist eventually took me off the beta blocker. I'm considering stopping with the sacubitril/valsartan (similar to an ACE inhibitor). Based on everything you've written on predictive coding etc, I freak out 10x more about things that mess with homeostasis than I used to.
Note that the heart medications arguably *are* in the category of "feeling worse is part of the plan". Reduce cardiac output, get low blood pressure and a more sluggish body in exchange for protecting the heart. The really bad fatigue problems showed up a decade down the line.
Also, the cause-effect here is not clear cut at all. I elided a bunch of stuff (reduced excercise due to covid lockdown + some injuries, stressful job, fibromyalgia).
ETA: my fatigue is improving now! I tried many things, what seemed to work is a combination of breathing and relaxation exercises, improving posture, deep meditative states to better feel where the body is tense, learning a lot about anatomy and becoming better at loosening specific muscles.
Lo ok at it this way: If a heart condition makes you 2x as likely to die of a heart attack in any given year. What does that mean?
When you’re young instead of a 1 in a 100000 chance of dying of a heart attack per year, it’d be 1/50000. And when you get old it’d be more like 2/3 instead of 1/3.
The heart disease takes about one year off life expectancy, but has a HUGE impact on the proximate cause of death.
So if the pills cured the heart problems perfectly they’d only a year to LE, and if they aged you even a little it’d easily wipe out the gains.
Do you see? Everyone dies of aging, aging is five orders of magnitude more consequential than your little heart disease, and all the diseases (and medicine) in the world just shuffle cause of death around.
Except for the few really bad diseases that can easily kill a young person.
But the rest of it is obviously just shuffling around cause of death.
Just think for one second! You’re heart disease won’t kill you when you’re young, it’ll kill you sometime after you start to age. And how likely it is to kill you is a function of your aging.
Which is to say the heart disease sin’t changing when you die, but what the proximate cause is!
No one has died of old age in the United States since 1951, when it was removed as an official possible cause of death. http://www.hdncms.org/images/editorial%20archive/sattler/It%27s%20Illegal%20to%20Die%20of%20Old%20Age.pdf
Note that the article, in addition to citing this, actually disagrees with the position, kind of like HIV doesn't actually kill people, just allows something else to kill people which otherwise wouldn't. Which is kind of like saying guns don't kill people, but bullets do.
Aging is almost exactly like HIV (before they had the medications they have now). It's a total death sentence.
We take it for granted, but it is responsible for >99.9% of the lost life expectancy. You might expect to live to 100 000 without it.
If an 82 year old dies of cancer, the aging was the cause of the cancer in the first place, and the aging was likely to kill them shortly anyway.
If you allowed aging as a cause of death it would be the primary CoD for nearly every single person 65+, with their proximate CoD being at most a small contributing factor.
Maybe it makes sense not to bother listing it.
Something else I learned from the article is about cellular apoptosis, a normal way for cells to die. We have much more to learn about aging, but I agree with the decision not to include "old age" as a cause of death. It may make things more complicated in some cases, but, as far as I see it, old age merely makes more and more maladies more and more likely as time goes on. Maybe a combination of them can kill someone whereas taken individually they are survivable.
But putting "old age" on a death certificate would be as informative as listing "suicide", as that isn't the actual manner of death. HIV, however, does seem like a reasonable cause of death to use, even though another disease actually caused the death.
It may be that old age itself can be conquered. Redwood trees can live for thousands of years (https://en.wikipedia.org/wiki/Sequoioideae). What is the difference in how they live?
But the point is that the benefits of the medications are bounded and the potential harms are (nearly) unbounded.
The best case for the benefits are: return this patient to being on track to dying of old age, and the harms are are: this might accidentally be aging the person.
So as long as the dangers of the heart disease are much smaller than the harms of aging (how on Earth could they not be, aging is 10x worse than AIDS) then the patient has FAR more to lose than to gain.
That’s the underlying reason people have been having this debate since the dawn of time.
I wonder how much it's consistently possible to tell this. I'm not questioning your experience, just wondering how generally applicable it is. I've been on beta blockers for many years, but I have no idea if it's sapped my general vitality. I am slower and more tired than I was years ago, but not cripplingly so, and I'm also a lot older.
For what it's worth, I finally stopped reading Overcoming Bias after some anti-medicine post where I tried to find a charitable reading that didn't imply medicine didn't work at all and didn't find a reasonable one.
Which one was that?
I really ought to have booked marked it, but it was a while ago.
We used the helmet for child #1's mildly flat head. Not sure she wouldn't have normalized naturally, but feeling like we had some agency and doing 'what we could' did certainly bring us peace of mind. We have no regrets.
The current version of the helmet dilemma for us is atropine eye drops for slowing myopia progression. Useless treatment at the margin? Or maybe will it make a meaningful impact on our daughters QoL...
"We used the helmet for child #1's mildly flat head. Not sure she wouldn't have normalized naturally, but feeling like we had some agency and doing 'what we could' did certainly bring us peace of mind. We have no regrets. "
NOOOoooooo, don't say this, now you're proving Robin's signaling theory right!
I don’t want to over-react to a sly wry joke, but if Robin were saying medicine is 100% signaling, and you were saying it is 0% signaling, my guess is that you’d both be wrong.
I don't think anyone's argued that 0% of medical expenses are signalling; certainly Scott has previously discussed cases in which he fully agrees that the requested care was pure signalling ( https://slatestarcodex.com/2013/07/17/who-by-very-slow-decay/ )
You're probably already aware, but dropping a comment in case you hadn't heard about the ongoing, promising research linking lack of bright light exposure in infants and children to the development of myopia. A search for "early light exposure myopia" or similar will bring up a host of studies, or a more casual overview from Wired: https://www.wired.com/story/taiwan-epicenter-of-world-myopia-epidemic/
Curious: did your child naturally refuse to sleep on their stomach? The issue (AFAIK) doesn't happen if babies are allowed to sleep on their stomach.
No, we followed the "back is best" advice. Another intervention with questionable evidence base, but fear of death is an even stronger motivator when it comes to your children.
Thanks, that confirms my suspicion.
Re: sleeping on back, best evidence shows that it reduces the odds of SIDS by 1/1000 over the childs first year, at best. In other words, if nobody followed the back-to-sleep advice, we'd have (at most) 0.1% of children die from SIDS compared to the counterfactual.
With the caveat that thinking about very small probabilities is very hard...:
Isn't lowering risk of death for a newborn in their first year of life by 1 per 1000 kinda huge? Thankfully, few babies die of preventable causes anymore in Western countries - the SSA says the prob of death under 1 is about 0.5-0.6 per 1k. In reality, a large fraction of these deaths are probably due to differences in baby "type" than anything parents can change by behavior/action - babies born with conditions such that they survive only a short time after birth, or have conditions such that the first time they catch anything, they die. If this is half of deaths under 1, then a sleep on back vs front SIDS margin of 1 per 1k would amount to a 1/3 to 1/2 increase in *aggregate* death risk under 1 for 99% of people. That's pretty large.
Or if you want it in expected life year terms, a 1 per 1k risk of losing 80 years of life is 0.08 life years per person. By comparison, we had a society wide freakout over a pandemic a few years ago that killed something like 4 people per 100k , people whom on average would otherwise have had something like <10 years left to live. (And I don't think ELY is quite the right metric: 1 baby dying at a few months old from SIDS or some other preventable cause is devastating in a way that 10 elderly people dying at 78 rather than 82 is not).
1/1000 is a *very* high end estimate. In reality once you control for smoking, mattress firmness and the presence of choking hazards in the infants bed it’s likely to be closer to 1/3000 for prone sleep alone. Note that Back to Sleep focused on multiple things, not just prone sleep alone.
Is a 1/3000 chance of dying due to prone sleep worth the 16-45% chance of developing a head deformity? Most head deformities go away on their own but they do cause stress for parents and potential stress for the child if it’s not cured.
I was just using the numbers you stated! (You did say "at best" and "at most", but still, its the number you gave!)
Anyway, I think 1/3k is still pretty non-trivial.* This is still a very sizeable chunk of preventable child deaths. "Is a 1/3000 chance of dying due to prone sleep worth the 16-45% chance of developing a head deformity?" 16-45% causal effect??? Since WAYYYY fewer than 16-45% of young kids are walking around with one of those head helmets on, either the vast vast majority go away on their own / they are pretty minor to begin with, or the vast majority of kids are walking around with them untreated and few people seem to mind/care. Since I don't notice many kids with hugely misshapen heads I'm guessing "correct not to notice/mind/care" seems quite justified.
As for back of envelope math: if VSL = $10m, 1 per 3k chance of death = $3300. So as long as prob increase of misshapen head x min(cost to fix, loss if not fixed) is below this, seems probably worth still doing.
Or to put it a different way, I am (a) completely open to the idea that most medical interventions have much smaller preventative effects than advocates often like to claim, and (b) that much medical "expert advise" is ad hoc and not really based on sound evidence (which is how it can completely flip on a given topic so often), but this conversation isn't doing anything to dissuade me from discarding this particular piece of 2020-era "expert medical advice".
(Off topic: I know about, but don't really understand the claimed effect of smoking on SIDS. Parent smokes, so... kid dies in their sleep. Is the claim a direct causal one - kids lungs are messed up in some way that makes situations that otherwise would be survivable, deadly? Or is it "smoking parent of young kids is probably correlated with social class which in turn correlates with a whole basket of stuff wrt parent actions that we don't observe, and these are what drives SIDS. I've always assumed its the latter, but I don't have any evidence for it beyond "social class almost always predicts this stuff which suggests the direct cause isn't properly observed/controlled for" and "the claimed effect seems underspecified").
>My wife and I recently took my four-month-old son to the pediatrician. The pediatrician said he had a mis-shapen head, and referred him to a head specialist for a second opinion. The specialist said yup, looks pretty mis-shapen, and referred us to a helmet-maker. The helmet-maker said yeah, definitely mis-shapen, and wants us to pay $300 for a helmet to correct it so my son doesn’t get stuck permanently looking like Frankenstein when he grows up.
> (there are some studies of this intervention, which are neither obviously wrong nor obviously unimpeachable. They say the helmet works, but not necessarily better than “repositioning therapy”. Our son refuses repositioning therapy, so for us it’s the helmet or nothing.)
>That helmet is probably our “marginal” health care expense, in the sense that it’s less obviously important than the other two things we’ve used healthcare for this year (childbirth, a scare with our son’s breathing). So, if we’re trying to cut the marginal health care expense, should we skip the helmet? Maybe we should skip it - I never see any adults with obviously mis-shapen heads out there, and surely they didn’t all get $300 helmets as kids. Maybe it’s all a racket.
Well, the ancient Peruvians and other ancient cultures notably altered the shape of their infants' heads by applying bindings: https://en.wikipedia.org/wiki/Artificial_cranial_deformation
So I think a helmet would likely be effective in this case. Just make sure your doctor isn't an ancient Peruvian :)
Disclaimer: I have not read Hanson's book.
What does "net positive" mean here? I expect the marginal medical spending to always be net positive, but approaching zero as spending increase. Isn't the case being made by Hanson that medical spending on the margin has quite high opportunity cost? To me, this seems like a very defensible claim.
I also expect cutting expenses would preferentially defund ineffective medical interventions over effective ones just by the system naturally "wanting" to help. I expect this to be true for both individual spending (in expectation) and for public spending.
I think it is likely that pretty much everyone has someone within a couple degrees of separation from oneself that is benefitting from the least effective half of current medical spending. In Scott's case, that is his son. Likewise, I think it is likely that pretty much everyone has someone within a couple (or three) degrees of separation from oneself that would benefit from the medical interventions that would have been made possible by doubling the current medical spending. This is a question of opportunity cost.
Rather than fiddling with the dial of how much money is allocated to healthcare, I think the more effective use of relevant political capital (in the US at least) would be to focus on sorting out price ambiguities, perverse incentives, and overwork.
Requiring that hospitals charge everyone the same price for any standard procedure, and spell out that price as part of getting consent, with no slack for unexpected complications, might motivate administrators to look for ways to actually improve efficiency rather than just gouge and obfuscate.
On the sleep-deprivation side of things, how about an overtime formula which adds 50% to hourly pay rates for all medical personnel after every 40 hours worked in a week, and the "week" doesn't officially end (resetting the combo multiplier) until off-duty for 36 consecutive hours? Budget decisions about bringing in enough staff to avoid burnout would get easier.
Is it magic that, "would preferentially defund ineffective medical interventions," or do you know of a mechanism? It seems that you are claiming that there is a mechanism that favors effective over ineffective interventions but only at an absolute spending level that is less than what we currently spend. Seems unlikely that such a mechanism would be tied to absolute funding levels.
I do not think I am claiming what you are claiming that I claim. I think there is a mechanism that (in expectation) preferentially funds effective over ineffective interventions (also) at our current level of spending:
- For individual spending "on a budget": The individual has some idea of which medical interventions have low opportunity cost.
- For public spending under spending cuts: Governments/hospitals/doctors do triaging, governments/hospitals/doctors have some idea of which medical interventions have low opportunity cost.
The case being made (at least in my comment) is that at our current level of spending (it might be the case that) even among the effective medical interventions most have high opportunity cost.
Let me know if the above clarification clears things up or not. If not, I might be misunderstanding something.
EDIT: Maybe I see where I went wrong. See my subcomment.
Hmm, I was thinking about this a bit more and it is indeed more complicated than I made it out to be. Any effective medical intervention would likely be low opportunity cost (the medical condition they fix would have an associated cost and the resources used to cure the condition would unlikely to be better used somewhere else). However, if for every effective medical intervention we defund we also defund 9 ineffective ones (since we don't know exactly but at least have some idea of which ones are effective and which ones are ineffective) then the alternative use of the money saved could (maybe) have a better use.
But isn't this just making the rather banal statement that we spend money very inefficiently (and not just on healthcare). In that case, the political issue that grabs all the headlines (spend less) becomes completely divorced. It is pretty easy to show that there are tons of people out failing to be treated for conditions we can very effectively treat. So, why don't we spend more efficiently and save more lives? Spending wisely is hard!
>I can’t figure out how these claims fit together coherently with what he’s said in the past.
IMO Hanson's sincere opinions are actually only a little outside the mainstream but he likes trolling by saying things exaggerated to the point of ridiculousness. Therefore he is not coherent overall but he apparently respects you and wants you to engage only with his sincere opinions, although he can't say so explicitly.
I recognize that saying someone is often not debating in good faith is a serious accusation, but I used to read a lot of what Hanson wrote (yes, I own his book) and it's the conclusion I came to.
A more polite way of saying that he likes trolling by saying things exaggerated to the point of ridiculousness is ton say that he uses hyperbole for rhetorical effect.
I've read Hanson's pieces on all of this for quite a bit and talked with him on my podcast here too (https://open.spotify.com/episode/0uhlVXNYeDz85io4cGNMg8?si=b8da366ffeb54dec)
If you're concerned with the "so what?" question, his clear answer is: bundle healthcare with life insurance, to give providers the incentive to keep you alive. There are a couple of kinks with this idea that he's worked out, of course there are practical challenges that we don't know yet how they will play out. But the upshot of it is: this healthcare-insurance bundle provider would have the right incentives to figure out which half of medicine is bad, and which is good (i.e. efficacious).
It seems there is a bigger question, one that Robin does not address at length, about the efficacy of clinical research. I'm not deep into the details, but the replication crisis may affect it. The incentive for pharma after a successful clinical study that leads to FDA approval is to not ever question the results and sell as much of the drug as they can. As far as I understand, they have no liability after approval. Doctors also don't seem to have the right incentives (which I can't describe from experience or with sufficient detail, but from what I understand it also has to do with liability / malpractice lawsuits).
"But the upshot of it is: this healthcare-insurance bundle provider would have the right incentives to figure out which half of medicine is bad, and which is good (i.e. efficacious)."
Ah yes: the infallible market will solve it all by magic, and there is no way it could go wrong. Mastectomy is very efficacious in reducing risk of breast cancer, so everyone (male and female) should have mastectomies!
https://www.cancer.gov/types/breast/risk-reducing-surgery-fact-sheet
I think the problem of perverse incentives is the driving force behind pseudo-therapeutic bloat. The question for the medical system isn't whether it's good at producing benefit to patients, but whether it's even designed to do so. A good example is the cancer question. Clearly some cancer treatments work, but just as clearly many are very expensive and do not work. Yet every year NEW expensive therapies that don't work are approved and pushed out to patients, proving this is an ongoing problem, not just an observed phenomenon from years past.
I recommend Vinay Prasad's book "Malignant" for a clear-eyed view of the story in cancer, which in my experience is representative of the systemic problem. It's hard not to conclude cancer research - and much of medical research in general - is more a profit-generating machine for pharmaceutical companies than it is about developing effective therapies.
Now don't get me wrong, the world's medical researchers aren't mad scientists, rubbing our hands together as we watch the money flow. We're really trying to get this thing right. And I think this is part of why the question of efficacy is so difficult to answer for an outsider who isn't involved in helping craft new interventions. Everyone is working to build a better world here, and sometimes we even succeed! But the incentives are driving the car. And the incentives all point to "generate profit from drug sales", without the stabilizing element of "... by selling therapies people want/need."
The problem, IMO, is that the consumer is decoupled from the payer, both in the US and most of Europe. There's an incentive to get drugs approved, not to produce drugs people want for prices they're willing to pay. Here's a good summary from Prasad's YouTube channel:
https://www.youtube.com/watch?v=qXZuM_BlNDY
Relevant quote: "It's a financial transfer of money from all the people being taxed to all the people who own stock in the companies. And the patient is only deriving some small sliver of that money (the cost of the product plus R&D) and may not be deriving any of the benefit. In some senses it's a financial product, with the rare adverse event of improved health outcomes."
I think this is a good description of what's really going on here. The reason it's hard to disaggregate which treatments are effective and which are bunk is because there's a financial incentive to ensure we can't tell the difference.
I feel like the "why don't you both bet on it" equivalent for these debates is: "what is the RCT that should be run?"
I mean, the best one would just be to take a population and ban half of them from acessing medical care for about a century. Obviously not feasible to do, of course.
"If I thought medical care was mostly effective and just needed to be trimmed around the margin, and my readers were posting that I thought medical care was “useless”... I would be horrified and try to clear it up as quickly as possible."
Cutting medicine by 50 percent seems quite a bit more than it being "trimmed around the margin", that when Hanson talks about marginal effects, those are pretty big margins. Moreover one of his big claims isn't merely that a large amount of medical utilization is wasteful, but rather that stress, air pollution, diet, exercise, sleep quality, smoking, excess alcohol consumption etc. seem much more important. In which case seemingly flippant comments about healthcare seem fairly reasonable, that on a societal and individual level it seems much more important to emphasize the more robust correlates of health, as opposed to medicine. Also lots of the examples you link seem misleading eg (https://www.lesswrong.com/posts/yEGZcJtPS4L4qsB4f/testing-hanson-s-hypothesis-about-the-uselessness-of-health) and the last 3 tweets, that as best as I can tell they are consistent with someone actually understanding Hanson's views, "2. We over consume healthcare in the USA" does not mean that all healthcare is waste, obviously. etc. etc.
So, relative to my health, I do consume a lot of health care, and much of it has proven to be useless.
A few years ago, my doctor noticed that I had some elevated liver enzymes. After getting this result for several consecutive lab tests with no other explanation (negative for hep, etc) he had me do a liver ultrasound. In that liver ultrasound a polyp on my gallbladder was noticed. Since then he's had me repeat that ultrasound every 6 months for the last 4 years or so, during which time there has been no change in this polyp. I feel fine. Seems like a lot of wasted healthcare to me.
Then my labs started to show an abnormally low thyroid hormone. Again, after a few consecutive labs, he ordered an ANA test lab. That came back positive, and so he did a whole (expensive!) ANA test panel. They all came back negative. So I have unexplained high liver enzymes, unexplained sub-clinical hypothyroidism, unexplained positive ANA, and I feel fine. Thousands of dollars have been spent to tell me that, "Yes, you are indeed fine."
The counter-point to this, I suppose, is that I take a statin for high cholesterol, and if I hadn't learned of that need from similar lab tests, my risk of heart attack would be much greater, so just the fact that you feel fine doesn't mean that there's no medical need for treatment. But I wonder if there's some "marginal healthcare" identification to be achieved by identifying conditions that are extremely unlikely to occur while "feeling fine" and completely ignoring that they even exist unless a person exhibits bodily distress that would indicate those conditions.
Well, in some places you would not have had the liver tests or thyroid tests done without any symptoms.
Most of these things were done to rule out cancer and the problem is that cancer often doesn’t cause symptoms before it is advanced.
+1. It's not fun when you find out all at once that you have stage IV cancer. This kind of ongoing screening ends up finding something in like 5% of people, which makes it well worth it.
I heard that finding it this way still does not allow to treat if, because aggressive cancers are hard to treat, and slow cancers will be found without excessive tests.
That was stream with criticism of over-screening
Some slow cancers, most famously pancreatic cancer, are known for producing essentially no symptoms until metastasis is well underway (stage III - IV).
Yup, that was what killed my brother-in-law.
Well, John Hopkins University says:
Is there a screening test for pancreatic cancer? There is no single diagnostic test that can tell you if you have pancreatic cancer. Definitive diagnosis requires a series of imaging scans, blood tests and biopsy—and those tests are typically only done only if you have symptoms
My understanding: that means basic screening is unlikely to help. Biopsy is invasive, and can cause side-effects for healthy people
"Feeling fine" isn't an indication that you are indeed fine. I'm very glad your health problems never developed into anything serious, but it's understandable that your doctor would be concerned about unusual results with no obvious cause.
Better to have four years of tests that turn out "you're okay" than presume "well it's probably nothing", do nothing, and then whoops turns out it's cancer.
https://www.ncbi.nlm.nih.gov/pmc/articles/PMC6969319/
"Gallbladder polyps (GBPs) are generally harmless, but the planning of diagnosis and treatment of the GBP is of clinical importance due to the high mortality risk of delays in the diagnosis of gallbladder carcinomas that show polypoid development.
...Polyp size is the most important predictor for neoplastic polyps. A diameter greater than 10 mm is a risk factor alone. Because of the possibility of malignancy of polypoid formations containing risk factors, surgery should be recommended first and follow-up should be recommended in clinically nonsymptomatic polypoid formations after the differential diagnosis. Especially in non-neoplastic and polypoid lesions below 10 mm, a “watch-and-wait strategy” should be recommended every 6-12 months with USG. "
Yes, as I wrote above it’s a matter of onus. Would you rather save the system or yourself some money and risk missing a serious but treatable condition or carefully follow up on potentially worrisome signs, symptoms and test results based on good medical judgement?
It's the Take that brings the boys to the yard.
The best Takes are immediately antagonising and difficult to precisely define.
When the Take gets challenged, the Artist starts shapeshifting around each specific argument, and "I never said x" is one of the Artist's favorite defenses.
In the end, "medicine doesn't work" is a million times more successful than "medicine doesn't work as well as people think", so that's the Take that sticks.
I think your assumption that “either we can’t distinguish between good and bad medical interventions or we can’t” is too simplified. You can get around the trilemma if there exists a class of clearly identifiable good treatments and another class of treatments where some are good and some bad, in total they have negative value and they are indistinguishable.
I think the simpler explanation is that we won't in practice distinguish between good & bad treatments that way. If Hanson is right that medicine is considered quasi-sacred and sought ought for reasons of signalling, and that people will flat out ignore information about which hospitals have better outcomes, then expecting people to cut useless medicine while keeping the good stuff is to just misunderstand people's motivation. It's all "good", so people just won't cut it.
We already do distinguish between good and bad treatments that way. Some treatments are compulsory for a hospital to provide if you show up at the hospital needing them. That is, mostly, treatments that are clearly good (and have some other characteristics like urgency). Most treatments aren't like that, in both senses.
The big one I hear about is over-medication, where medication is given, and more medication is given for side effects (sometimes recognized as side effects), and the whole medication load doesn't get evaluated for quite a while.
Does getting the medication regime evaluated count as more medical care or less? Does getting off medications count as medical care, since some of them really need to be tapered carefully?
There also the matter of aggressive end of life care. I don't know whether it's half of medical care, but it's a lot.
https://slatestarcodex.com/2013/07/17/who-by-very-slow-decay/
I'm with you about the whole medication load not getting evaluated often enough.
About drugs to treat side effects though, I want to speak up for that as not (maybe even mostly) being over-medication. Drugs prescribed to treat primary conditions often have quality-of-life reducing side effects but are necessary for one's health and so the side effect managing drugs are an essential companion for treating the primary condition. Cancer would be the prime example, but we see it with depression and a bunch of other conditions.
It does at one level seem absurd that we have to take drugs to manage the drugs we take, but at another level, it makes total sense and I think should be seen as part of the primary treatment regime. If the hormone suppressing drugs you take to prevent cancer recurrence cause insomnia, then the drugs you take for insomnia are an essential part of the treatment plan.
I don't think you probably disagree with any of this, so thank you for letting me put this in here anyway.
Ah, once again when I ctrl-f for keywords in a point I'd like to make I find the name of Nancy Lebovitz, followed by my point. This is more flattering to me than it is to you. Aren't you a Less Wrong OG?
Anyhow, I just left a similar (but uglier-worded) comment on Robin Hanson's reply to this reply. I really wish everyone could just get together and shout at both of them about the problem not being $300 helmets for newborns but the 4th open-heart surgery for an old fat man on dialysis who is already scheduled for chemotherapy next month.
Thank you. Yes, I'm a Less Wrong OG, and same for Overcoming Bias.
Putting aside your terrible gun trial analogy, it might be worthwhile to go over some of the correlational evidence that Hanson has alluded to over the years, if you are so hung up about insurance experiments. Here's something from RCA "Objectively read, i.e., studying the actual data instead of weak characterizations of it, the weight of the evidence points strongly against differences in health care access, especially via health insurance, being a significant contributor to broad differences in health outcomes. We don’t find this at an international level amongst broadly comparable countries; we don’t find it domestically between counties, states, MSAs, etc; and we don’t find this with more carefully controlled experimental evidence.
I do not mean this to suggest that health care is totally irrelevant, but rather that the entire developed world gets almost all of the bang for its buck vis-a-vis broad mortality measures in the first few dollars spent. This sort of high impact medicine is broadly similar everywhere in the developed world (antibiotics, vaccines, trauma medicine, etc) and tends to be targeted at cases where its apt to make a real difference (those with obvious need), so the differences in broad averages between places in the developed world largely come down to genetics and lifestyle. The US could probably cut spending in half with little obvious effect on overall life expectancy (not the same as specific outcome measures), but the same is almost certainly true for most of the developed world too!" (https://randomcriticalanalysis.com/2016/11/06/us-life-expectancy-is-below-naive-expectations-mostly-because-it-economically-outperforms/)
What did you think was terrible about the gun trial analogy? I thought it was pretty apt--let's take a case where we already know the effect of an 'intervention' and imagine we were measuring that effect using either the controlled intervention-specific trials used in medicine or the population-level varying incentive trials of economics, where many additional variables intervene so that statistical power is reduced and inferential distance is increased. The one point of breakdown I see is that 'gun insurance' would be analogous to, like, cardiac stent insurance or something, because medicine comprises many more and varied things than does gun ownership. But that only strengthens the case that a null result in a hypothetical gun insurance voucher trial wouldn't surprise you even if specific interventions (like getting shot with a gun) have clear and knowable effects.
Did you even read Scott's article? Scott didn't disagree with this, "The US could probably cut spending in half with little obvious effect on overall life expectancy," and he very clearly made the case that this is 'obviously not true' for all methods of deciding which "half."
If we assume that the value of healthcare is, on average across the whole spectrum of healthcare solutions, neutral (some harm, some help), then we should be able to cut it in half by the first letter of the therapy/medicine and still end up on average net neutral. We can then cut that in half again, and again, to the limit of cutting it all and still have net neutral medicine.
I think what Hanson is trying to claim is that we have data that suggests that people who have more money *do* have better outcomes (in aggregate) even after you tease out medical spending, so we should be willing to make the trade of "just remove all medicine and have people spend more of their money on other things, and we'll end up net positive in terms of health outcomes."
While one *could* do what Scott suggests and try to separate the good treatments from the bad, we have already tried this (e.g., FDA) and it doesn't seem to work (at least not today). So rather than trying to solve the same problem in the same way (decide what medicine is good and what is bad in a centralized manner), we should just scrap all of medicine, let people be healthier because they have more money, and then separately maybe experiment with some new novel systems (in controlled/small settings) that are incentive aligned for health outcomes (like his health insurance ideas).
The trouble is that biology is complicated. A common test for one patient will do no good because they really don't have anything wrong with them, but for another patient it will uncover the problem. Since you don't know which patient is which, you order tests for both patients.
Sure, people have a lot of unnecessary tests and procedures which probably don't do anything for their general state of health, but US medicine has the latest drugs and tech, the population expects a lot of intervention to sort out what ails them, and health insurance is there to pay for it (in theory).
Take gastroenteritis - most people will get better at home, but if you need the sick cert for work, you have to go to the doctor. And just maybe it's something more serious, so you should go. It would be easy to say "Okay, we should cut out all trips to the doctor just because your stomach is sick and you've been vomiting" - until it turns out that it's a symptom of something more serious.
I fully agree that some therapies are useful. I also fully agree that some therapies are useful for some people and not useful for others.
The argument here though is that the system we have for figuring out which is which (i.e., the FDA) has resulted in an aggregate outcome where you are not obviously better off blindly accepting all healthcare the system recommends than if you accept no healthcare the system recommends. Since healthcare costs are very far from free, it is possible that the opportunity cost of having a healthcare system *at all* ends up net negative for society.
An individual can certainly do better than the aggregate by doing their own research and picking out the good healthcare advice and ignoring the bad, but the vast majority of people either get "no healthcare" or "blindly get all the healthcare".
When we are talking about healthcare for all of society, we need to look at the system as a whole and its aggregate outcomes, not whether the system has produced any individual good/bad outcomes. The system we have in place isn't working, so the argument is that we should get rid of it in favor of the far cheaper system that just does nothing (and achieves a similar aggregate outcome).
One can (and many do, including Scott to some extent in these posts) argue that the current system is not net-negative/net-neutral in aggregate, and I think that is a worthwhile debate. However, even taking Scott's aggregate claims at face value I'm not sure it is worth the cost to society. The quibbling seems to be over very marginal benefits in aggregate (e.g., single digit percentages) while US healthcare spending is measured in double digits of GDP.
He is not proposing to "scrap all of medicine", but to instead cut it in half.
My reading of his argument as to why he suggests half is for two reasons:
1. Scrapping the whole thing is less palatable to humans, because humans have a bias towards "doing something" over "doing nothing", and half of something is still something.
2. If you cut it in half, there is a reasonable chance that any method of cutting *other* than random selection will result in aggregate quality of the system improving (even if only marginally). This assumes that we have some mechanism of cutting in half that isn't incentive backward-aligned, as one could certainly imagine some incentive structures that result in medical care actually getting worse with a not-random cutting. e.g., cut antibiotics and keep <insert obviously bad therapy here>.
I thought he clearly said that the RCT evidence he bases his conclusion show that variance in the amount of treatment got didn't seem to be associated with better health, so this justifies cutting within that margin. If there were studies of people prohibited from getting any healthcare, that could potentially justify cutting it all. I believe Tyler Cowen referred to that sort of thing when he brought up "studies of Christian Scientists and Amish". https://marginalrevolution.com/marginalrevolution/2024/04/sunday-assorted-links-466.html Of course, Scott himself has written that the Amish don't get zero medical care either. https://slatestarcodex.com/2020/04/20/the-amish-health-care-system/ Instead they spend 1/5 as much as the average, so Hanson's proposed 1/2 is still less of a cut.
Something interesting about the Amish is that they are spending on healthcare that is output from a different system. I don't think it is impossible to have a healthcare system that is net-positive in aggregate, I just think it is a reasonable argument (debatable whether true or not!) that the modern western system is not such a system.
It is certainly possible I'm misunderstanding Hanson's arguments, but I don't think what you have said here necessarily supports that. One can believe that "cut it all" is reasonable while simultaneously recommending "cut it in half" because that is more palatable to your target audience. Also, one can imagine ways to cut it in half that result in a significant increase in healthcare quality in aggregate, thus one can be hopeful that such a cut is achieved.
You can imagine what's going on in Hanson's head that he's not saying, but I'm saying that he explicitly lays out his 50% as an extrapolation from the variance in those RCTs. So if anyone wants to argue against cutting 100%, that's not countering an argument he actually made.
Under
>I believe Tyler Cowen referred to that sort of thing when he brought up "studies of Christian Scientists and Amish". https://marginalrevolution.com/marginalrevolution/2024/04/sunday-assorted-links-466.html
I only see
>I might add that studies of Christian Scientists and Amish also dent one’s faith in the very high value of medical care. That said, in my own life medical care has only done me good.
am I misreading and overlooking a link to a study of Christian Scientists?
Christina the StoryGirl had a horrifying description of what she saw first hand in
https://www.astralcodexten.com/p/contra-hanson-on-medical-effectiveness/comment/54680067
>potentially justify cutting it all
seems like a very unlikely conclusion to me.
Cowen didn't link to any, he just verbally referenced their existence. And one of Cowen's laws is that such literature must exist.
> seems like a very unlikely conclusion to me.
Unlikely, but I'm saying that's the sort of study which could potentially do that. The studies Hanson focuses on aren't like that.
Many Thanks!
>I'm saying that's the sort of study which could potentially do that
True, if it were done and if it came to that conclusion.
Robin Hanson says, "Scott then quotes 500 words from a 2022 post of mine,"
What is the purpose of pointing this out? Should Scott have quoted the entire post?
I think he's trying to say "Scott cherrypicked the stupid bits" but the problem is, the stupid bits were there already.
Nothing worse than having cherry-pickers drawing attention to one's stupid bits.
I think he's just explaining that he's not going to bother quoting Scott quoting himself when he can instead summarize that quote by "none of which have me saying all medicine is useless on all margins".
So, my uncle was apparently born with a misshapen head (firstborn, difficult birth, pointed head). The doctor just said, here's some oil, massage his head every day. Apparently this worked, and he had a university career and is a great-grandfather, so there don't seem to have been any significant side effects. Sample size 1, of course. I wonder if this is still considered a good treatment, or if it's just for a different kind of case. This would have been in 1934/5, so it's quite plausible that doctors today would be horrified by the idea.
Unfortunately we can no longer ask his mother how she did it.
Robin Hanson is very smart. Smarter than me. And also a very original thinker. So I'm always a bit surprised when he is somewhat regularly wildly, obviously wrong and overconfident in the way that economists seem to be.
The "medicine doesn't do anything because look at this data I cherry picked" is one such weirdly wrong thing. It reminds me of a theoretical physicist who says something like,
according to this model, whenever a particle is observed two new universes are created." I'm like, "I believe that's what the math says and I believe your match is correct. I don't have the knowledge to tell you why it's wrong, but it's simply obviously wrong." Robin is sort of like that guy, but instead it's a rape auction or something.
I think "original" does cover it. Like the old joke about the book review: "This is both original and good, but the parts that are good are not original, and the parts that are original are not good".
I don't think the particle thing is obviously wrong. You have no evidence (and never could have evidence) that it's wrong, so why not follow the math?
We can and do have evidence of *coherent* superpositions, eg. from quantum computing. But that doesn't add up to Many Worlds in the popular sense.
There is an approach to MWI based on coherent superpositions, and a version based on decoherence. These are (for all practical purposes) incompatible opposites, but are treated as interchangeable in Yudkowsky's writings. "The decoherence (a.k.a. many-worlds) version of quantum mechanics ....". Decoherent branches are large, stable, non interacting and irreversible...everything that would be intuitively expected of a "world". But there is no empirical evidence for them , nor are they obviously supported by the core mathematics of quantum mechanics, the Schrödinger equation.
Whatever is going on inside a Quantum computer isn't classical computation (using the same number of particles) as Deustch correctly states. However, it also isn't a set of *decoherent* worlds, since quantum computing works via coherent superpositon. (It also isn't a set of macroscopic, irreversible, objective , etc, worlds). So a QC is neither a single world *nor* multiple worlds, for reasonable definitions of "world". Quantum computing is evidence for quantum m superposition, but simultaneously evidence against the popular idea that irrevocable splitting into non-interacting worlds occurs at every elementary interaction or observation: that would predict that coherent superpositions don't exist, for all practical purposes,.since they instantaneously decohere...and QC therefore doesn't work.
"David Deutsch, one of the founders of quantum computing in
the 1980s, certainly thinks that it would. Though to be fair, Deutsch
thinks the impact would “merely” be psychological – since for him,
quantum mechanics has already proved the existence of parallel uni-
verses! Deutsch is fond of asking questions like the following: if Shor’s
algorithm succeeds in factoring a 3000-digit integer, then where was
the number factored? Where did the computational resources needed
to factor the number come from, if not from some sort of “multiverse”
exponentially bigger than the universe we see? To my mind, Deutsch
seems to be tacitly assuming here that factoring is not in BPP – but
no matter; for purposes of argument, we can certainly grant him that
assumption.
It should surprise no one that Deutsch’s views about this are
far from universally accepted. Many who agree about the possibil-
ity of building quantum computers, and the formalism needed to
describe them, nevertheless disagree that the formalism is best inter-
preted in terms of “parallel universes.” To Deutsch, these people are
simply intellectual wusses – like the churchmen who agreed that the
Copernican system was practically useful, so long as one remembers
that obviously the Earth doesn’t really go around the sun.
So, how do the intellectual wusses respond to the charges?
For one thing, they point out that viewing a quantum computer in
terms of “parallel universes” raises serious difficulties of its own.
In particular, there’s what those condemned to worry about such
things call the “preferred basis problem.” The problem is basically
this: how do we define a “split” between one parallel universe and
another? There are infinitely many ways you could imagine slic-
ing up a quantum state, and it’s not clear why one is better than
another!
One can push the argument further. The key thing that quan-
tum computers rely on for speedups – indeed, the thing that makes
quantum mechanics different from classical probability theory in the
first place – is interference between positive and negative amplitudes.
But to whatever extent different “branches” of the multiverse can
usefully interfere for quantum computing, to that extent they don’t
seem like separate branches at all! I mean, the whole point of inter-
ference is to mix branches together so that they lose their individual
identities. If they retain their identities, then for exactly that reason
we don’t see interference.
Of course, a many-worlder could respond that, in order to lose
their separate identities by interfering with each other, the branches
had to be there in the first place! And the argument could go on
(indeed, has gone on) for quite a while.
Rather than take sides in this fraught, fascinating, but perhaps
ultimately meaningless debate..."..Scott Aaronson , QCSD, p148
See also:-
https://www.greaterwrong.com/posts/wvGqjZEZoYnsS5xfn/any-evidence-or-reason-to-expect-a-multiverse-everett
Because the math doesn't say anything about other universes, its only job is to help you predict experimental results. All that "interpretation" stuff is philosophy, which may or may not be "correct" in some sense, but it most certainly isn't straightforwardly implied by our current math.
People who try to convince me of--let's call them Quantum World curiosities--I think are doing card tricks so well that they fool themselves. That is card I was thinking of, that is my signature in an unopened pack of playing cards. It somehow got into my jacket pocket. I am astonished but I don't believe you are doing magic.
Almost every reasonable sounding and surprising "infinite worlds" theory still ends up somewhere like this:
Me: So there's infinite worlds?
Quantum Theorist: Yes.
Me: That means a universe exactly like this one exists and nothing else is different except on election day a majority of Americans will write in my name on their Presidential ballots and I will become President despite being almost totally anonymous?
Theorist: Infinity is infinity. So every possible thing that *could* happen *will* happen. There's also a universe where Margot Robbie invents teleportation and the first thing she does is teleport into your birthday cake naked. Yes.
Me: I don't know exactly why that's wrong but I know that it's wrong.
Maybe there's some sort of quibble where there's not INFINITE universes or it's one infinitely big universe or whatever. But as far as I can tell every one of those situations leads me to a place where informed horse sense allows me to be confident that the math is incomplete. The same way I know that medicine has a positive impact without studying it. If the publication is rigorous and it says vaccines don't move the needle, I don't know how they got it wrong, but I know it's wrong.
1. The math doesn't say that. 2. There seems to be significant confusion as to how math is used in physics. 3. Considering how absurd the many worlds theory is, it is almost nitpicking to point out that it doesn't even give the experimentally observed probabilities. 4. There are theories that agree with experiment and are not absurd like many worlds or unprofessionally vague like Copenhagen.
Robin doesn't seem to be selectively cherry-picking studies that fit his conclusion, rather he's got criteria for taking studies seriously that eliminate most without further discretion on his part.
Observation is only important under the Copenhagen interpretation, whereas I believe Robin rejects that in favor of the Everett interpretation in which universes are created regardless of observation.
I think the same as you about Hanson, except that... of course he might be right about something.
I have a couple of ways of reminding myself how wrong I must be. One is religion: I'm not religious, but that means I am disagreeing with most of the brilliant thinkers in Europe that I revere over the last 2,000 years. If I am certain that Descartes was wrong about god, then it's certain that I'm wrong about much, much more.
Another is an idea from a philosopher I like called Eric Schwitzgebel, that he sometimes calls crazyism. He looks at lots of philosophical questions which remain open, things like dualism, infinite universes, the objective existence of morality, and so on. He notes that if you pick any possible answer to these questions and work out all of the implications of that answer, you always end up with at least some conclusions that are "crazy". And yet, one of these answers must be true! Whatever the truth of these unknowns, it's going to mean something crazy. Therefore, the fact that an idea is crazy, or has crazy implications, should not be seen as an argument against it.
Worst case scenario: Hanson is an interesting thinker who has spewed out a lot of interesting ideas into the world, and they're all wrong. That still makes him massively valuable, simply for generating and elaborating these ideas. We shouldn't look to public intellectuals to be *right*. Often, that's not their job. Their job is to make and explore new ideas.
The particle thing is perhaps a bad analogy, since it is widely believed around here (IMO it is wrong for non obvious reasons)
If Hanson said "modern medicine is useless" then I agree, that would be a dumb thing to say.
The problem is, Hanson has said so many dumb things in the past, it's easy to believe he *did* say this as well.
"Also, we can look today at how the death rates of individual households correlates with the water sources and sewer mechanisms used by those households. "
I would love to dump this fool down into the house I was reared in as a child, the same kind of house that other neighbours and family members were raised in, and let him see if it makes a different to have running water and sewerage. When was the last time he took clothing out of a wardrobe to find it covered in mildew, for example? Frost on the *inside* of the window pane? You get your drinking water from a pump in a field? You can't ask my parents now since they're dead, but their siblings who died young from "we just don't know what medicine is good or bad, or why some people live and some people die", is that it, Robin?
How does anyone take this man seriously? *Why* does anyone take this man seriously?
It's useful to look at the rest of the paper and not just the line you quoted:
"Most of the obvious theories have serious problems, you see. For example, exercise,
smoking, social status, and urban living appear to have large effects on individuals. But the time trends for exercise, urban living, and smoking have been in the wrong direction - those trends would predict decreasing lifespan. And since social status is usually thought to be relative to contemporaries, it is hard to see how average social status can increase with time.
Finally, while the biggest advances in nutrition, medicine, and public health seem to
have occured during the first two thirds of the twentieth century, death rates have fallen
just as fast during the last third of the twentieth century. Perhaps some new influence rose in importance just as those other influences became less important, but if so it seems a remarkable coincidence that the total rate of improvement has remained pretty steady."
Hanson's point is that if improved water/sewer infrastructure was the key to longer living, lifespan should increase after these improvements and then level out. Yet lifespan keeps increasing at a steady rate generations later.
"Hanson's point is that if improved water/sewer infrastructure was the key to longer living, lifespan should increase after these improvements and then level out. Yet lifespan keeps increasing at a steady rate generations later."
When was the last time you heard of someone dying of cholera? It used to be very prevalent:
https://www.irishtimes.com/news/ireland/irish-news/wakes-beggars-and-bad-air-when-typhus-killed-65-000-people-in-ireland-1.4203488
"Typhus was not the only disease to kill tens of thousands of people in Ireland during those decades. More than 50,000 people died from cholera during the 1830s, according to the 1841 census report, with more than 46,000 of those deaths occurring in three years of epidemic, between 1832 and 1834.
Cholera is a bacterial infection caused by water and food contamination. Acute, sudden diarrhoea can cause severe hydration and death. The World Health Organizastion primarily links cholera with poor access to clean water and proper sanitation.
The census report from 1941 [sic] notes: “Cholera prevailed most in the towns of the civic districts; and ... in 1833 it had spread throughout the country at large, and prevailed most in the rural districts.”
Dublin had many cases. A news story in the Belfast Newsletter, April 3rd, 1832, detailed a Dublin case before the central board of health, before which “reports were laid before them of the deaths of a man and two women, in the neighbourhood of Summer-hill, and of the sickness of a young woman who had been at the wake and funeral of the man”.
The census report identifies Sligo, Drogheda and Belfast as hotbeds for the disease. The 1832 outbreak in Sligo has since been associated with the works of Bram Stoker, whose mother apparently related vivid descriptions of the devastation to her son. A report in the Tralee Mercury on September 1st, 1832, also notes some "melancholy details" of the illness, which "broke out with fearful violence" in the town: "In almost every case where the sick person died in his own house, three or four of the same family have been attacked, thus paralysing every exertion and rendering them incapable of any effort to subdue the virulence of the plague."
Cholera, that decade, accounted for one in every 7.36 deaths from all causes; by comparison, influenza killed 10,575 people, or one in 112.28 deaths."
Typhus and typhoid fever are also spread due to poor sanitation. People died from those in Ireland, and from diphtheria:
https://corkfolklore.org/health/diphtheria#:~:text=Diptheria%20death%20rates%20range%20between,appetite%2C%20sore%20throat%20and%20fever.
"Diptheria is caused by a bacterium called Corynebacterium diphtheria and is transmitted in respiratory droplets. Diptheria was once a highly significant illness and a major cause of death in children. Diptheria death rates range between about 20% for children under 5 and people over 40, with a slightly lower rate of between 5-10% if the infected person is aged between 5-40. Symptoms include loss of appetite, sore throat and fever. In 1948, 521 cases of diphtheria were reported in Ireland and due to the introduction of the combination diphtheria/tetanus/pertussis (DTP) vaccine in 1952/1953, the case numbers declined progressively until none were reported in 1968. After a gap of nearly 50 years, one case was reported in 2015 and another in 2016."
Vaccines are a great boon, but it's easier to be hygienic when you have a reliable source of plentiful, clean water and sewerage system to take foul water and waste away. Contamination of water sources still kills, even in developed countries:
https://www.independent.ie/farming/rural-life/in-a-country-like-ireland-we-had-a-child-die-from-a-water-borne-infection-40pc-of-private-wells-contaminated-scientists-warn/a4556353.html
"“Water sources in Ethiopia are regulated, but private wells in Ireland are not,” said Dr Hynds.
In his first PhD, he looked at hundreds of wells across Ireland, and during nearly two years of taking a sample every month, he got to know many of the families.
“One of these wells belonged to a dairy farming family. Their grandchild got sick from water from the well. He got hemolytic uremic syndrome from a STEC infection,” he said at a recent UCD Earth Institute event on the Hidden Hazards of Unregulated Private Wells.
“He then went into renal failure and got a 400pc blood transfusion. Every drop of blood in his body was replaced four times and he died.
“So in a country like Ireland, a rich country, we had a child die from a water-borne infection, and that was not on the Six One News.
“I’m an epidemiologist, and in this day and age, this is simply not good enough.”."
Hanson is like someone claiming that "You say these Irish peasants are dying of 'famine' by 'starvation', and yet once they get fed, we see that people are still dying afterwards! So your obvious theory has a serious problem!"
"Finally, while the biggest advances in nutrition, medicine, and public health seem to have occurred during the first two thirds of the twentieth century, death rates have fallen just as fast during the last third of the twentieth century. Perhaps some new influence rose in importance just as those other influences became less important, but if so it seems a remarkable coincidence that the total rate of improvement has remained pretty steady."
That's the low-hanging fruit being picked; first we get huge boosts from solving obvious problems like poor sanitation, lack of effective medicine, and insufficient knowledge. Then, as we clear those problems out of the way, we learn more, improve more, and build on the successes of the past. Now you're much less likely to die before you reach your fifth birthday, and since there is now better medicine, more access to public health services, education, and all kinds of attempts to motivate the general public to eat healthy, exercise more, stop smoking and drinking and engaging in risky behaviour, these too pay off. Healthier people have healthier children who live in healthier conditions and are exposed to less dangerous factors at work, so there's a slowly improving increase in lifespan. We're not roofing our houses with asbestos and painting the walls with lead paint anymore, Robin.
Also, hmmm, lemme think - was there anything in particular that happened in the first two-thirds of the twentieth century that might have bumped up the death rate? Maybe even happened twice? Like some global martial phenomenon? Nope, can't recall anything of that sort!
I had the same reaction as you when I first read this. What kind of moron claims water quality doesn't affect health outcomes? Look at the work John Snow did on Cholera in England during the mid-19th century. At the broad street pump site, there were two buildings that had private wells, and the mortality rate was a huge margin lower than the surrounding buildings. The chart with the residents of two different water companies was also critically illuminating; the company which piped water upstream (clean) had Cholera rates less than the London average, yet the company that piped downstream (sewage contaminated) had rates many times higher.
After reading Hanson's reply to this post, his position seems to be that medicine works but half of it is junk and the medical establishment is incapable of separating the two. Also, I think the point of the paper quoted above is the question about why lifespan continues to increase at a continuous rate despite the factors supposedly influencing it happening at discrete intervals. My position on Hanson has gone from 'moron' to 'bad at communicating what he really means.' Although in fairness, I never heard about him before these posts by Scott, so maybe that lens makes things seem worse than they are.
As a final note, I hope the excess mortality from war is obvious enough that people exclude it from their data on life expectancy.
For some conditions, we have treatments that are really, really effective. Antibiotics for example.
I guess I could entertain a claim that doctors try to treat everything, and the distribution has quite a tail of things where they actually don't have an effective treatment.
I was once in hospital (outpatient only) where the consultant was reduced to making up a diagnosis with "idiopathic" in its name. This is like, sure it looks like there's something wrong with you, but we have just run off the edge of our diagnostic flowchart here, and are in the land of rare diseases. NB: i am not criticizing the doctor here .. i am quite in favour of being given an honest "no idea what this is" is that's what the doc is actually thinking.
And as for random inexplicable stuff .. one time, I get severe pain in the extremities, like in my fingertips. it Felt like the sort of thing that deserves immediate medical attention. (I guess something with similar symptoms is the onset of diabetes and you're getting crystals in the extremities. Tests reveal it was not diabetes.) Doctor is like, "have you had a recent tetanus shot? here, have a tetanus shot right now.") Symptoms went away after a few days. No idea what it was. Shrug.
Total cost of that was presumably a tetanus vaccination, a blood glucose test, and ten minutes of a doctor's time, to (probably) no net benefit to the patient.
>That helmet is probably our “marginal” health care expense, in the sense that it’s less obviously important than the other two things we’ve used healthcare for this year (childbirth, a scare with our son’s breathing).
I think this one parenthetical here contains one of the clearest examples of a major medical expense we probably *could* cut with minimal loss. Child delivery is *much* more expensive in the United States than most countries (https://www.statista.com/statistics/312026/total-hospital-and-physician-costs-of-a-normal-delivery-by-country/). How much safety are we buying with that extra expense? Well, the CIA factbook figures on maternal mortality rank the US second-to-last out of the selected countries (https://www.cia.gov/the-world-factbook/field/maternal-mortality-ratio/country-comparison/) better than South Africa, but behind Kazakhstan, Chile or the United Arab Emirates. And it's not like poor countries simply get more bang for their buck due to lower costs of labor, costs in Germany are lower than costs in Chile.
Unfortunately, as a consumer, I don't think you have the leeway to buy the services/price package of German child delivery. But if we looked at questions like "why does American child delivery cost so much more than German child delivery?" I think we might be digging into a pretty rich seam of low value marginal dollars spent on medical care.
Maybe one of my disagreements with Robin is that I doubt the medical waste forms a clear category that patients can just opt out of. For example, during postpartum, nurses came to check up on us a bunch of times. My wife was very tired and asked them not to do that. They said they had to (I assume this was a legal liability thing).
I don't think Hanson said "We should cut spending by getting individual patients to opt out of treatment". But if, say, we taxed rather than subsidized medical care, many such practices would go away.
This was how I interpreted:
"Unfortunately, even if you believe everything that I have said, your behavior will probably not change much as a result. You will still spend nearly as much on medicine for yourself and your family, and spend much less effort on the more effective ways to increase lifespan. After all, your sick family would consider it the worst kind of betrayal if you did not “do something,” and give them all the medicine that your doctor recommends (Hanson, 2002). Alas, the problem of the fear of death muddling our thinking is so much worse than we imagined."
This is one of the places I think Hanson is really off-base in his views on medicine; I think if you cut out all signalling-based medical expenditures, it'd probably only cover a little bit of the distance between US healthcare spending, and spending in countries like, say, Kazakhstan. I think a larger portion of the price tag probably comes from CYOA procedures based to limit legal liability. You and your wife didn't deliberately seek out a hospital which would send nurses to check up on your wife repeatedly despite your expressed wishes because you wanted to signal the intensity of your commitment to your children. The hospital insisted because they wanted to minimize their risk of being sued.
Great point.
As someone who went through something similar:
I think it is fun to think through this. The patient finds this annoying, but perhaps it is net helpful. You never know when you are going to catch a furtive co-sleeper (or something).
Or perhaps this is useless with healthy patients, but it's difficult to train nurses to only check on red-flagged rooms with consistency and net positive to just say "check them all". So the healthy patient pays the price to on net help the unhealthy one.
Or perhaps this causes healthy patients to prefer to leave AMA (like we did) in order to avoid the constant intrusion, but there are a group of healthy patients where maybe something happens and it would have been better for them to stay at the hospital. So on net it goes back negative compared to "only check on red-flaggers".
Or maybe... four more levels down. Is someone going to do a RCT on this? Will that RCT be designed in the same way this hospital's training and staffing levels are designed? Will following that RCT from some other hospital during normal staffing work the same in this hospital dealing with understaffing? Or from a nurse's supervisor who just doesn't care? Or from a local wave of furtive co-sleepers?
I don't know, but fun to think about. How do you collect evidence in a world with ethics advisory boards and administrators that are just trying to survive one day to the next and maybe were never really a facts person to begin with?
> Well, the CIA factbook figures on maternal mortality rank the US second-to-last out of the selected countries...
Funnily enough, even *that* is more complicated than it first appears: https://foreignpolicy.com/2024/01/31/united-states-maternal-mortality-crisis-statistics-health/ (The U.S. Maternal Mortality Crisis Is a Statistical Illusion -- Accurate counting has produced a seemingly dire death rate.)
"However, these figures are completely wrong, and they have been known to be wrong for many years now. The U.S. National Center for Health Statistics, the branch of the Centers for Disease Control (CDC) charged with collating health and vital statistics, has published three separate reports elaborating in excruciating detail on one crucial fact about U.S. maternal mortality: It is measured in a vastly more expansive way than anywhere else in the world.
As a result, U.S. maternal mortality is overestimated by two to three times. Properly measured, the real U.S. maternal mortality rate in 2019 was 9.9 maternal deaths per 100,000 births, which would put it at 36th place—still not impressive by comparison, but somewhat better than Canada and a bit worse than Finland or the United Kingdom.
...
Historically, the United States and most countries have tracked maternal mortality using data based on the cause of death listed on death certificates. When a person dies and the cause is assessed by an examiner of some kind, certain “maternal” causes, such as “eclampsia” or “obstetric trauma,” are commonly tracked. If a woman has died due to one of these pregnancy-related causes, she is listed as a maternal death. This process is fairly straightforward and has been widely adopted across many countries.
But in 2003, the U.S. CDC decided to launch an improved death certificate form. Among the various changes proposed was the addition of a checkbox, wherein whoever filled out the paperwork would identify if the deceased had been pregnant in the last 42 days or the last year. The reason for this checkbox was that the CDC believed (correctly, as it turns out) that in only measuring “maternal causes of death,” it might be underestimating the true health hazards of pregnancy. Pregnancy might alter the course of other diseases and conditions or interact with them in important ways.
The CDC anticipated that the checkbox would increase measured maternal deaths; it did not anticipate just how much it would increase them. As it happens, the CDC’s own reporting, which I have confirmed elsewhere, shows that the addition of the checkbox approximately doubled maternal mortality rates.
You might think a sudden doubling in maternal death rates would be obviously flagged as a data issue to correct, but this turns out not to be so. Because the United States has a federal system, individual states added the checkbox in different years. While individual state maternal deaths showed sharp level shifts, the national maternal death count drifted upward gradually as states added checkboxes to their death certificates: California in 2003, Florida in 2005, Texas in 2006, Ohio in 2007, Tennessee in 2012, etc.
In 2018, further modifications were made to the data-processing protocols used by the National Center for Health Statistics for pregnancy-related checkbox deaths, leading to more thorough inclusion of them. The result was a massive but gradual artificial inflation of maternal mortality.
This doesn’t mean that the American way of measuring death is wrong. It’s just quite different from the countries that it’s being compared to...
But the U.S. case is particularly beguiling, since the United States now tracks all deaths of women who were pregnant, not only women who gave birth. Women who miscarried early or had abortions—whether officially reported or not—are also counted in the checkbox method. As a result, the United States may be the only country in the world where central vital records systems track all pregnancy-related mortality, not just maternal mortality."
This does a lot to explain why listed US outcomes are worse than countries like Kazakhstan or Chile, but it still leaves the US with similar outcomes to countries with similar standards of living (better than some, worse than many) despite spending dramatically more money on childbirth services. Even without the US doing so anomalously poorly in its actual outcomes, the figures suggest a lot of dead weight spending.
Cut all antibiotic use for ear infection in children or cut it in half. To paraphrase Robin Hanson "Ear infection treatment is not about curing ear infections".
One type of ear infection in children accounts for 10% of all antibiotic use in Iceland (From the paper 20% of antibiotic use in Iceland is consumed by children under 7 of which 50% is used to treat a type of ear infection.) Relative to other Nords Iceland has a larger problem with antibiotic resistance. - https://www.ncbi.nlm.nih.gov/pmc/articles/PMC3442319/.
Instead ear infection treatment is about -
...“the lack of stable doctor–patient relationships due to lack of continuity in medical care. Pressure from patients in a stressful society, the physician's work pressure, the physician's own personality, particularly the earnings incentive and service mentality and, last but not least, the physician's lack of confidence and uncertainty, resulting in use of antibiotic prescriptions as a coping strategy in an uncomfortable situation” Petursson P. GPs’ reasons for “non-pharmacological” prescribing of antibiotics: A phenomenological study. Scand J Prim Health Care. 2005;23:120–5.
Have you ever had an ear infection? I'm not trolling, it's a serious question.
The pain is.... I don't know what to compare it to, I've never experienced anything even close. And it was a bacterial infection, and maybe you're telling me it would have just cleared by itself, or maybe I'd go with what the doctor told me, that it was quite dangerous because of its proximity to brain, here's a prescription, you should notice an improvement withing 24-48 hrs., and I did. Take the pills and noticed an improvement.
I do not recall if I have had an ear infection. Probably. Given five out of six children have at least one ear infection before their third birthday. I may have had one but have no a memory out it.
There is a difference between your personal experience and what makes for good medical policy.
Does an ear infection cased by virus feel different than an ear infection caused by bacteria?
It my understanding that in 80% of the cases a virus is the cause of ear infection in children. To confirm bacterial infection requires a test. Did you receive a test?
Most middle ear infections resolve in 2 to 3 days.
Depending on when you went to a doctor it could be that it cleared on its own, if no bacterial test was taken it could have been a virus. If it was a virus the antibiotics were likely useless.
Did you finish the course antibiotics or did you stop taking them after you felt better?
Many people stop taking antibiotics before finishing their course of antibiotics. Failing to finish antibiotics increases the risk of antibiotic resistant strains of bacteria developing.
In Sweden the average child has about 5 courses of antibiotics in childhood while in American it is closer to 12 courses of antibiotics. Swedish children have lower levels of childhood obesity some have suggested that part of this may be due to over prescription of antibiotics in children. The conjecture relates to the effect antibiotics use has in the fattening of livestock.
Have you or a loved one died from a antibiotic resistant infection? At the margins a child given unnecessary antibiotics does not increase the likelihood. A policy that over medicates with antibiotics in aggregate does lead to increased death due to antibiotic resistant strains of bacteria.
I…. am not arguing against antibiotics overuse being the case? But you proposed “cut all antibiotics” for ear infections, and I don’t think it’s a medically defensible position. Some ear infections are viral, some are bacterial.
FWIW yes mine was bacterial, yes I finished my bottle, yes I had a friend of a friend “almost” die of a resistant infection (2-months stay in a hospital, intubated at some point).
I agree. I should have said cut all antibiotic use for ear infections in children without a test to confirm bacterial infection. Given that bacterial test takes one to five days. That most middle ear infections resolve in 2 to 3 days. That viral infection is the leading cause of ear infection in children and that getting a followup appointment can be difficult to go over test results. In practice it may amount to the same thing. Also I softened my initial assertion "or cut it in half" based on the lower occurrence of antibiotic use in Swedes during childhood compared to Americans and the two populations are about as healthy.
Even if the following were true "More people are harmed by drug resistance bacteria caused by the current level of prescription of antibiotics for child ear infection than are helped by antibiotics for ear infection" that still isn't medically defensible argument for cutting all antibiotic use for ear infection in children. Medicine concerns itself with the patient not the society. This was attempt to argue the marginal utility of medicine in this one narrow avenue to society as it practiced may be zero or near zero.
“ Medicine concerns itself with the patient not the society.”
Couldn’t have said it any better! I think you hit on the key problem with Hanson’s arguments, and why in general it’s so hard to cut medicine. We go to great lengths to reduce suffering of our loved ones, rationality be damned.
Mate, I had an ear infection that I didn't get antibiotics for, as a young adult, and I was *crying* with the pain of it and dosing myself up on over-the-counter painkillers.
Imagine a young child screaming with pain. Are you going to be the parent that goes "Just tough it out, Junior, we're not going to the doctor for something this trivial?"
Even if they're only doing it for "I can't take another day of the screaming for 24 hours solid, please make it stop doctor" reasons, I think those are valid.
Not all ear infections are that bad, and the old-fashioned treatment used to be warm up some olive oil (which was obtained from the chemist), pour a little in the ear, plug it with cotton wool and let it act that way. Presumably the heat killed the infection. You can also now buy over the counter ear drops. But that really bad ear infection? I was lancing pus out of the ear every night for a week and believe me, warm olive oil would not have cut it.
What a parent should do for a sick child is different than what good medical policy should recommend with respect to antibiotic use for ear infection.
I would not presume to know the best course of action for parents in a given circumstance.
What I do know is that about 80% of all ear infections have viral causes. In most cases doing nothing would lead to resolution, folk remedies would be just as effective as antibiotics as would placebo. It is my understanding that antibiotics do not have antiviral action.
To confirm bacterial infection requires a test. A bacterial culture test can take between one to five days. Most middle ear infections resolve in 2 to 3 days.
If one has a policy to give all children antibiotics without regard to the nature of the infection you risk doing more harm than good. The paper I linked to talks about that effect in one population. It also has a link to recommendations for when to give antibiotics by referencing a guideline. In the practice of medicine these guidelines are often ignored. https://www.nice.org.uk/guidance/ng91
I am aware that antibiotics do not have any effect on viruses. I am aware of over-prescription. But a blanket ban on "we won't prescribe antibiotics" isn't going to improve health, either, because sometimes it *is* a bacterial infection, it's bad, and it does need treatment:
https://www.cdc.gov/antibiotic-use/ear-infection.html
"The body’s immune system can often fight off middle ear infection on its own. Antibiotics are sometimes not needed for middle ear infections. However, severe middle ear infections or infections that last longer than 2–3 days need antibiotics right away."
'Parents pressure doctors to over-prescribe' is true. But so is 'the child has been sick for over 3 days and pus is oozing out of the ear'.
We're in agreement that there should be guidelines, and that there is often social pressure to prescribe antibiotics when those do no good. But there is the lovely shiny theory of "let us reduce over-prescription by issuing guidelines all doctors should stick to" and the reality of "frazzled parent with screaming child turns up in your surgery looking for something". I am not surprised a doctor might think "Well, maybe it is bacterial, writing a prescription won't hurt".
Less medicine may lead to healthier outcomes.
I would agree that a blanket ban on antibiotic prescriptions won't improve health. Which is why I did not make that assertion. I am weakly confident that some kind of prohibition, ban, tax, disincentive or policy enforcement that results in the cutting in half the average childhood prescriptions in the US from 12 to the Swedish average of 5 would not make American children less healthy. I am weakly confident it may do good on net for overall societal health and from the Icelandic paper even some children themselves .
It has been thirty years since the effects of over prescription of antibiotics have been warned about in the literature. At least two generations of doctors have now been educated knowing this to be the case. Yet the problem persists.
We are only now beginning to engage in research projects on how to stop doctors from continuing to engage in a practice, that they know most of the time does not work. Watchful waiting is a meaningless platitude once the patient parents demand a doctor do something.
This is a small issue where the problem of how medicine actually practiced is broken. Where we know it is broken and we know how to fix it have known for quite sometime but have had limited success. I do not believe this is an isolated issue. Incentives innate in medicine lead to wasteful use of medicine. Even when we know which parts are bad we can't cut them.
I don’t how much but a lot of medical spending is done to rule out worrisome conditions. Tests with negative results will dilute the medical effectiveness of the money spent testing. Eliminating these tests would be a net positive in theory except that you can’t know which cases to desist from testing in advance. The onus (justifiably) is to maximize health rather than save money.
I feel like the way to square the circle here (though I don't think this is Robin's stated position from the quotes) is that a lot of the end of life interventions we make are probably of very low value (both realized and in expectation) and are sometimes instituted by families trying to show they care. Related: Atul Gawande's "Being Mortal" does a good job covering why the way we approach death medically is likely neutral / negative for patients and very costly.
His wife is a hospice nurse, so I believe he has discussed that specifically. Aside from signalling as an explanation, I recall that he's also written that we just have an aversion to thinking about death.
There's an additional dimension not in conversation here which is where treatments for some conditions have unequivocally improved outcomes over time. Nature: "The number of people who die after a breast cancer diagnosis has decreased by two-thirds since the 1990s, a study of more than half a million women in England has shown."
I think medicine should be spending more on the kinds of lifestyle changes that make a huge difference. On cancer recurrence for instance, there's good research showing exercise and diet make a huge difference to reducing recurrence rates. We spend a quarter of a million dollars treating someone for cancer and then spend nothing on making sure they do the things that prevent more cost later and that make a huge difference to mortality rates.
But exhorting people to make these lifestyle changes doesn't work. What's it look like to take these numbers seriously and provide the kind of healthcare system support for diet and exercise that we do for getting someone through chemo? The thing is, it won't be money to pharmaceutical companies, it would be for salaries for huge teams of coaches, fitness instructors, dieticians, and recreational facilities. What's it look like to not treat these as optional add-on things but as a central part of the medical treatment?
Instead of bemoaning people's unwillingness to make the kind of health-oriented changes that make a difference, I think we should be incorporating them more into medicine as part of treatment. And then medicine would be even more effective for the people it's already spending money on.
Scott writes:
"Public health measures? Sanitation? Clean water? Robin again:
If medicine for treating individuals is not quite the miracle we have heard, does public health make up the difference? Have not we all heard how the introduction of modern water and sewer systems greatly improved our ancestors’ health? Well, a century ago the U.S. cities with the most advanced water and sewer systems had higher death rates than the other cities. Also, we can look today at how the death rates of individual households correlates with the water sources and sewer mechanisms used by those households. Even in poor countries with high death rates, once we control for a few other variables like social status we usually find that water and sewer parameters are unrelated to death rates. Well we must live longer now for some reason, right? Yes . . . but the fact is that we just do not know why we now live so much longer.
(Robin seems more bullish on sanitation elsewhere, so maybe he’s changed his mind? Like I said, I have trouble fitting all his statements into a coherent model.)"
End quote.
One of the issues in this debate (which Hanson has observed in the past) is that there doesn't seem to be a widely agreed-upon definition of "medicine". Personally, I don't really think that civil engineers and site work contractors designing and building modern sewer infrastructure counts as medicine.
This is especially true of "public health", which, we learned in the pandemic, can be defined to include basically everything that human beings do. In my region, local health inspectors (previously considered minor bureaucrats who inspected restaurant grease traps) turned out to have broad dictatorial powers to shut down businesses, churches, and schools, and to make extremely granular rules about how people could behave.
If "public health" is part of "medicine", and if it covers basically all human activity, then medicine is actually responsible for all of the gains in life expectancy that have occurred over the past few centuries.
> In my region, local health inspectors (previously considered minor bureaucrats who inspected restaurant grease traps) turned out to have broad dictatorial powers to shut down businesses, churches, and schools, and to make extremely granular rules about how people could behave.
Bret Devereaux had a series on the structure of the government of Rome, which included the fun observation that the tribune of the plebs was an important position since quite early on with well-understood powers that were used in traditional ways. This lasted for centuries.
Then one guy was elected to the position and used the powers - which everyone agreed on - in nontraditional ways. At that point, the system broke down.
"Well, a century ago the U.S. cities with the most advanced water and sewer systems had higher death rates than the other cities."
Gosh durn it, this is the kind of careless statement that makes me go "Sources!" Where is he getting these figures, does he explain?
Because if he was writing in 2002, "a century ago" would be 1902 and I'm betting the "cities with the most advanced systems" were also the largest, therefore the higher death rates. I'm just pulling that out of my backside, but since we seem to be in the land of "simply make an airy claim", then I'm as justified.
Some stats nerd in 1906, God bless 'em, did a whole "The General Death-Rate of Large American Cities, 1871-1904: [Tables for Individual Cities]". Take it away, Mr. Frederick L. Hoffman:
https://www.jstor.org/stable/2275981?seq=2
Taking the 1902 figures, we find a few samples (I could give a lot more):
Mortality Rate per 1,000
New York City (Manhattan, Bronx) 19.2
Chicago 14.3
Philadelphia 18.1
New York City (Brooklyn Borough) 17.8
St Louis 18.3
Boston 18.7
Cleveland 15.9
Buffalo 14.6
San Francisco 19.9
Cincinnati 17.3
Pittsburg 21.3
Detroit 15.6
He even gives us a handy comparison between races:
Baltimore – white population 17.5
Baltimore – black (“coloured”) population 29.5
New Orleans - white population 18.8
New Orleans – coloured population 30.3
Washington D.C.
District of Columbia – white population 16.3
District of Columbia – coloured population 30.1
Wow, just look at the difference in rates there, all due to the "coloured" people having such over-abundance of advanced water and sewer systems! Right, Robin?
So how do these rates stack up against US mortality at the time? Well, again bless the stats nerds of the US Census because they gave us Life Tables for 1890 to 1910:
https://www.cdc.gov/nchs/data/lifetables/life1890-1910.pdf
Total (reported) deaths for 1902 were 318,636 on a population estimated to be 20,408,867 in 1901 which - if my shaky maths can be believed - is a rate of 15.6 per 1,000. The cities aren't doing too bad by comparison with that, I think; some are higher, some are lower, but the figure isn't all deaths, just reported deaths, and covers the entire country where some centres of population are going to be much smaller than New York and Chicago.
Yeah, I would still like to see where Hanson is getting his comparison figures for "cities with better sewerage versus cities with worse and comparable death rates".
Digging in to the question, I'm currently skimming through "Urban Wastewater Management in the United States, Past, Present and Future":
https://www.researchgate.net/publication/253439152_Urban_Wastewater_Management_in_the_United_States_Past_Present_and_Future
Seriously, this is a fascinating account, and *may* give some colour to Hanson's argument, on the grounds that improved facilities led to demands for luxuries such as indoor lavatories, but the existing sewerage systems were inadequate to deal with the new demand (and output of waste). So yes, "better plumbing" did mean "more disease" BUT only until the city officials caught up with the necessity for "better sewerage" to go along with the better plumbing. So I can't agree with the totality of Hanson's comment there.
"Population Growth
During the nineteenth century, there was considerable urban population growth in the United States. In 1820, less than 5 percent of all Americans lived in urban areas (cities with a population larger than 8,000), but by 1860 the percentage increased to 16 percent and by 1880 had risen to 22.5 percent. From 1820 to 1880, most major cities in the United States experienced considerable growth. For example, during this time Boston’s population increased eightfold, New York City’s tenfold, Philadelphia’s thirteen fold, and Washington, D.C.’s fivefold. As a result of this increased population density in urban areas, the decentralized privy vault-cesspool wastewater management systems became overtaxed. Mitigation measures included increasing the cleaning frequency and constructing additional privy vaults and cesspools. The improvements, however, only slightly reduced the periodic overflows and development of nuisance conditions. The privy vault-cesspool system, as it existed then, was inadequate to handle the increased amount of wastewater. The centralized water-carriage sewer system, on the other hand, was being promoted as the management alternative for urban areas with increasing populations.
Public Water Supplies and Water Closets
Another major cause of the abandonment of the decentralized privy vault-cesspool system was the increased construction of piped-in water-supply systems. More and more during the middle of the nineteenth century, potable water supplies were being piped in because local water sources were contaminated, frequent disease outbreaks were occurring, and water quantities above what was available locally were needed for fire fighting and street flushing.
Water-supply systems were constructed in most of the major U.S. cities in the early to mid-nineteenth century, and by 1860, the 16 largest cities in the nation had waterworks.
Piped-in water supplies influenced wastewater management in two ways. First, water-carriage waste removal required a copious supply of water, and the introduction of a piped-in water supply made water-carriage sewer systems viable. And second, the improved standard of living for urban dwellers in the nineteenth century coupled with the availability of water led to the implementation of modern plumbing fixtures and a concomitant increase in wastewater production. The water closet probably had the most significant effect on wastewater management compared to the other plumbing fixtures because it increased not only wastewater quantity, but also the quantity of fecal matter in discharges. The high level of fecal matter being discharged with the wastewater heightened the risk of disease transfer and outbreak, but this was not understood at the time.
The increased wastewater levels overwhelmed the privy vault-cesspool system, but few municipalities planned for, or constructed, additional wastewater management infrastructure. Residents had two ways of addressing the increased wastewater being produced: (1) continue to discharge to an existing privy vault or cesspool, or (2) create an illegal connection to a storm sewer or street gutter. Both choices were ineffective solutions because neither the privy vault-cesspool system nor the storm-sewer system were designed to accommodate the increased wastewater. Instead of addressing infrastructure needs, municipalities implemented ordinances to mitigate the problems created by the increased wastewater quantities. One such ordinance was instituted in Boston during 1844 that prohibited the taking of baths without a doctor’s order. Municipalities also tried to prohibit the discharge of fecal matter to the sewer system. Bans such as these were in effect in Boston until 1833, in Philadelphia until 1850, and in New York until 1854, at which time sanitary connections to sewers became required. The enforcement of imposed wastewater discharge limits and the prevention of illegal sanitary connections to the storm-sewer system was difficult for a municipality. Privy vaults and cesspools continued to overflow, while the connections to the storm-sewer system also resulted in sanitation problems. In most cases, neither the privy vault-cesspool system nor the uncoordinated sewer system were able to handle the increased quantity of wastewater. In many American cities (e.g., New York City), physicians, public health officials, and the general public demanded action to address the wastewater management problems created by the influx of piped-in water, and most supported the implementation of centralized water-carriage sewer systems.
Public Health
…In the United States, repeated cholera epidemics and other disease outbreaks gradually influenced municipalities to improve sanitation practices. Between 1832 and 1873, numerous American cities were afflicted with major outbreaks of disease, including cholera in 1832, 1849, and 1866 and typhoid in 1848. The causes of the outbreaks were attributed to a variety of reasons including unsanitary conditions and punishment from God. The experience gained from the epidemics improved the understanding of cholera and other diseases and their corresponding etiology. A cholera outbreak, following the Civil War, provided a chance to practice some of the prevention techniques based on improving sanitary conditions and disinfecting the waste products of infected individuals. The relative success of those measures indicated that the effective management of human wastes was an important component in protecting public health. The search for an effective method of protecting public health by managing human wastes invariably encouraged the construction of water-carriage sewer systems."
It feels like there are a number of different arguments one could be making here and I'm not sure which one you are making.
1) Are you just saying that Robin believes we can't distinguish useful treatments from useless ones therefore it's impossible to cut spending without cutting the good things?
Seems plausible but it doesn't entail that modern medicine is on net neutral only that benefit is maybe half of what it could be.
2) Are you saying Robin believes 1 plus a claim that the costs of harms of the non-working medicine are so large as to render the net impact of medicine zero?
3) Or are you suggesting that Robin believes that the marginal impact of extra medicine is zero or negative? This seems more plausible but like a different claim -- and again it doesn't even mean that all of it is bad only that we can't effectively sort the good stuff.
I'm not actively trying to suggest anything. I'm saying this doesn't make sense to me and I'd like clarification.
Ok, fair. My sense is that Rpbin definitely means 1 and likely means 3 but probably not 2. And I think he's unsure about how much net life extension is provided by medication even if it's net positive.
However, it can sometimes be a bit frustrating to figure out Robin's position since he likes to just toss out ideas (maybe that's something more people should do but it can be frustrating).
Part of the disagreement here is that Scott (as a psychiatrist) is a lot more focused than Robin on identifying specific actions that people should do given their views on healthcare. Robin doesn’t really give actionable advice to individuals as far as I’ve noticed. He’s more into abstract comments on society
Perhaps the amount of money “society” spends on healthcare for an individual should have an annual cap that decreases with age. Instead of what we do now, which is more or less the opposite.
I am, in fact, putting Hanson's recommendation into practice!
• My feet and legs are extremely swollen, and have been for some months
• It has recently become so bad that when I went to my monthly appt. with the head-shrinker (mostly, this is for buprenorphine; government-mandated bookkeeping more than anything else, I think), I — having started to forget it isn't normal to have feet that look like balloon-animal versions of themselves, and hence accidentally worn sandals — took quite literally a single step into the office, and one of the nurses said: "Hello, Mr. K—... look at your feet. Oh, my God. Come with me!"... and then I had to lie a bunch to all of 'em to minimize it as much as possible.
• They have demanded a bunch of stuff: tests, "I forbid you to go back to work and you need to keep your legs elevated in fact elevate them right now I'm not sure I've ever seen it this bad before if this isn't gone 100% by tomorrow you need to blah blah".
• I have done none of this. They sent me for blood and urine test right then and there, for kidney function; all normal; I think the next step was supposed to be cardiovascular checkup or something, but I just told them it went away and was probably just a bad reaction to some sketchy "pre-workout powder."
Ha ha! Fools! It is worse than ever, if only you knew!
Now, obviously, this sounds like the actions of a crazy person. No! It's not actually that bad: I think I know the reason this is occurring, and it would probably go away if I did One Simple Trick... but also, I don't care too much if I bite it, heh.
----------------------------------------
(In fact, I'd honestly been close to making that "I care a lot if I kick the bucket—as in, I will be upset if I *don't*"... but my ex-wife was very upset at the idea, and I've made her cry enough tears in this dumb life of mine... so I won't hurry it up — but I won't try to prolong it, either.¹
Unless she comes and checks on me again like she's been threatening. Then she'll see my feet and force me to stop my self-destructive lifestyle and actually go to the doctor again and a bunch of other stuff that Hanson and I are totally against.)
---------------------------------------
(...right? Robin? This IS what you'd recommend, right? I've made sure to leave a bunch of notes around saying stuff like "To My Notoriously Litigious Mother: If anything happens, please know, I am taking medical advice from Robin Hanson, so probably there was nothing to be done" and "Note to self: Hanson will be angry if you mention chest pains and arm numbness to Dr. again" and "LAWYER APPT: NEED SOON, HANSON DEMAND ONLY BENEFICIARY" etc etc.)
...I crack myself up sometimes, heh heh heh. But seriously, I do wonder what the Hansonian view here is. Is THIS still probably net-neutral in EV, to go get checked?
I mean, hey, ignoring it HAS worked so far; for a while, about half a year ago, the swelling wasn't as bad but I was having *way* worse symptoms; but they all resolved!
(...I suppose that if I suddenly stop commenting, well, y'all know how this experiment went then—)
....................................................
------------------------------------------------
~ footnotes which will inevitably become longer than the main post and which contain nothing that really even needed to be footnoted ~
---------------------------------------------
¹: Destroying your marriage to screw a pretty face turned out to be a bad idea, shockingly! You just end up stuck with someone who grates on you because she isn't your ex-wife, who was the only woman who has ever truly known you and still loved you for it; and then, considering how things started, you can hardly complain when your lover does to you what you did to your sweet innocent wife.
god damn it.
...and now I'm feeling like I can't face going to work at another stagnant and dull position, so just living off savings—but... who's going to give me a chance to change careers at this point, y'know?
"Hey guys trust me my degree is in a completely unrelated field but I would be a real good developer/data scientist/whatever, I swear! Take a chance and see!"
When I'd first graduated, I remember thinking "well, this should be easy as long as I can get someone to gimme that first look" — ...lol, dream on, moron! No one gave my résumé another glance, I expect; must have sent that damn thing out a thousand times.
Only got to such a good position as I recently had by a sheer chance meeting, in fact, not due to any application: sat next to the 'Fleet Director' of an oilfield company, on a flight back from Houston. He'd usually have been in the company jet (oilfield company: of course there's a jet!), but had let some other division charter it and just requested a ticket on a regular flight instead... luckily for me.
During the short hop over to WTX, I had joked with him about my situation, which at the time was getting pretty dire — though of course I didn't let on just how bad; said something like "someday I'll be able to get the GOOD bread, haha, the $6 kind you know" — but he's got a surprisingly soft heart, for an Important Executive™, and read between the lines.
(It seems to have preyed on his mind for the entire next week, as I later learned — he kept mentioning it to people, apparently: "Wanted to try the damn $6 bread... had to put the 'good' bread back on the shelf, and just-married too... poor fucker...")
So, though I'd left my number with him, I never expected anything to come of it — but after that week had passed, got a 9 am call from his office: "Mr. M____ said you're going to be our new Fleet Analytics guy, so when can you come up here for orientation—?"
I remember how my (then-not-ex-yet-) wife and I hugged each other and jumped around that shitty little apartment for a while, once I hung up and told her. She was so proud for/of me. "I told you! I told you you were a charmer!" "Only you think so... well, you and Mr. M___ I guess! Get your shoes on, baby — we're not having ketchup sandwiches again for lunch, not today!"
I miss her a lot. Can't expect that kind of luck again, though, heh... neither with love, nor labor.
I want to reload an old save, or else just quit.
please, get the feet treated
The majority of those tweets/screengrabs are at least broadly consistent with what Robin says he says, so it's not obvious why you'd demand he leap in to correct them. Note the first few words of the LessWrong article are "Robin Hanson has hypothesized that much health care is useless...", i.e. "much" not "all".
Even if they weren't consistent with what he said, I wouldn't expect Robin to leap in to correct random tweets.
Never having read Hanson, Is it possible that the core disagreement is around the word marginal? You seem like you’re using it for the IMO standard meaning of “least valuable” but I medicine I could kind of imagine it meaning “final”. So your sons helmet isn’t marginal because it will benefit him for years, but a lifesaving surgery on a 90 year old might be “marginal” if they die or something unrelated two weeks later.
In that regard, I could almost, if I squint at it, understand the argument “cut healthcare spending by 50% by investing dramatically less in end of life care”.
To be fair that seems like an easy thing to just say if that’s his actual belief, but it seemed like the kind of off chance thing that could be possible.
No, he absolutely means it in the first sense.
Edit: Here's an exchange that happened in the comments on the response to this post:
Commenter: 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.
Hanson: No, most med costs don't come from last six months of life.
I wonder how plausible this theory is:
1. Poor people without insurance don't get regular checkups and ignore medical conditions more often, but eventually a disease will get serious enough that they're forced to go to the doctor.
2. Therefore, chronic diseases that would eventually kill someone usually get detected and treated regardless of insurance/wealth status.
3. Therefore, although medicine "works" in the sense that early checkups catch the disease earlier, allow better and less invasive treatment, improve quality of life, etc, this doesn't show up in lifespan data because our medical system was already pretty good at making sure people don't literally *die*. The marginal medical dollar might improve healthspan but this is harder to measure.
What about the claim that the Obamacare study reported the one cohort with significance?
Generally it's pretty clear that Hanson was laying a monkey trap where unsuspecting bloggers would critique his statements that healthcare should lowkey be abolished and then he would respond by saying "haha that's super preposterous and now you must agree with my real argument which I will now reveal and pretend it was the main one all along"
However it did appear to be a real argument that the Obamacare study picked a cohort
He didn't say that healthcare should be "abolished", but instead cut in half.
Robin's monkey trap analogy was that waiting to decide which parts of healthcare to keep was an obvious mistake because it's so clear that marginal health spending is not valuable at all and perhaps impossible to demonstrate which 50% is good
Rhetorically speaking this is saying "lowkey" abolish healthcare, or "medicine doesn't work", especially since the studies supporting it don't have statistical power!
No, it's not saying abolish healthcare. If the monkey indiscriminately let some of the food out of its paw, it's paw would take up less space and could perhaps be withdrawn. If the monkey doesn't let go of anything until it's sure the food it releases is worse than average, then it's paw will remain stuck.
Oh, to really nerd out on the monkey paw analogy:
The key difference is paw open/closed. I mean, try it with your own hand: open hand has a lot of mobility, and can nicely curve to reduce its min. width. Once you close your fist, loosely or tightly, the knuckles spread out, the palm widens, and the min. width is significantly larger. So as long as the palm is curled, the exact amount of food kept is immaterial.
I feel silly just typing this. Hanson's arguing for silly things using clever silly arguments, and refuting them is like wading through mud.
Open your hand and you can still squeeze some things between your fingers.
Yeah you’re right.
Is it signaling if you do whatever (coincidentally expensive) thing the ER proposes for your kid "out of an abundance of caution" because you don't want them to call child services on you for endangering your child?
Can that actually happen?
I don't know! I don't want to find out!
It would be really surprising if the government's child abductors DIDN'T do that kind of thing as a matter of course.
"It seems to me that if we were to cut medicine in half, figuring out which half to cut would be among the most consequential decisions in history."
This seems a bit over-dramatic. Doctors pick what's most important routinely when triaging patients.
If you have lots of time and resources then that cut is to be carefully sewn or maybe you get the plastic surgeons involved if it's on their face. If you're short of time and resources? Stop the bleeding, keep it clean, treat any infections.
That list of 10 things seems pretty easy to mostly cut in half in a not-too-controversial way while retaining the biggest benefits.
If doctors lost a little bit of money (rather than gained) every time they sent a patient for an unnecessary test or prescribed a useless treatment then they'd quickly prioritise the most vital tests. They're not helpless.
> If doctors lost a little bit of money (rather than gained) every time they sent a patient for an unnecessary test or prescribed a useless treatment then they'd quickly prioritise the most vital tests. They're not helpless.
The payment-for-procedure system causes a lot of problems.
"If doctors lost a little bit of money (rather than gained) every time they sent a patient for an unnecessary test or prescribed a useless treatment then they'd quickly prioritise the most vital tests. They're not helpless."
This is how HMOs work.
Robin on HMOs:
https://www.overcomingbias.com/p/remember-the-hmo-revolutionhtml
> This is why I find this conversation so frustrating. Mention a medical treatment to Robin, even one of the “good ones” like cancer or antibiotics, and he’ll try to argue that maybe the evidence it works is being misinterpreted, and in fact it’s unclear how well it works. Then I say he thinks this stuff might not work, and he accuses me of straw-manning him.
"Cancer" isn't a medical treatment. In fact, we know that a huge amount of medical spending on cancer is deleterious. Doctors with cancer accept much less treatment than laymen do, and when you ask them why, they'll say that it's because of their experience watching what happens to cancer patients.
I've seen an article (not by a doctor) that argued in apparent seriousness that under the modern American medical system, anyone coming down with cancer dies in poverty after the system confiscates all their money, and that this is a desirable state of affairs. I take some issue with that.
When my grandmother had cancer, my mother (a doctor) specifically advised that she not accept most of the treatments she was offered. The awfulness and unhelpfulness of most cancer treatment isn't some big secret.
At the same time, survival rates keep going up. When my 𝘴𝘪𝘴𝘵𝘦𝘳 had cancer, my mother (still a doctor) was very pleased to note that a lot of progress had occurred in cancer treatment 𝗳𝗼𝗿 𝘆𝗼𝘂𝗻𝗴 𝗽𝗮𝘁𝗶𝗲𝗻𝘁𝘀.
Young patients are ideal for several reasons. Most prominently, (1) there is less wrong with them, (2) they recover faster (from everything), and (3) even stipulating that life while being treated for cancer isn't worth living, their treatment will presumably give them a lot more of the good kind of life with the cancer in remission and treatment sporadic or on hiatus.
But most cancer treatment isn't delivered to young patients for the obvious reason that there are very few young people with cancer.
Anyway, looking at what medical treatments doctors obtain for themselves and their family members will surely 𝗵𝗲𝗹𝗽 with judging which ones are likely to be harmful, in proportion to how common the problem being treated is.
There are other tells. Another essay I've read, by a doctor, pointed out that the elderly patients who get expensive long-shot (or: deleterious) medical care don't live near their children. It's the absentee children who insist on it for what seem to be reasons of guilt.
Regarding cancer, #2 on your list.
It would definitely be useful for many patients to be aware that a lot of medical treatments, particularly during end-of-life care, may have limited benefit and dramatic downsides.
It is also useful to be aware of the many psychological factors at play among physicians, patients, and patients' loved ones (including "signalling") that lead to treatments being undertaken that, if certain conversations had been had, might have been turned down.
Among other things, this may be as simple as the patient feeling comfortable to ask questions like, "Would you do this [additional course of radiotherapy] if you were in my position?".
Scott, in terms of it being a challenge to "fit claims together coherently" with what someone has said in the past: you yourself have written vividly on the horrors of excessive end-of-life care. Shouldn't the practical relevance of Robin's claims for the individual be fairly intuitive?
I recognize there are issues wrt futile care with cancer diagnoses, but in the last few years, I have seen three older relatives get diagnosed with serious diseases--two with cancer, one with another blood disorder. In two cases, they tried a treatment course, found it wasn't working, and advised ceasing all treatment except palliative care; in the third, by the time the cancer was diagnosed, the doctor advised going straight into hospice.
Yeah, you're strawmanning him; he listed a bunch of ways to decide what medicine to cut, e.g. ending subsidy as a government strategy and Cochrane reviews as a patient strategy; pretending he didn't is weird.
There's a huge "public choice" problem here. It may be true that a smart, prudent individual could distinguish which parts of medicine are "good" and which are "bad." That's does not *at all* imply that consensus could be generated sufficient to push those determinations through the political and bureaucratic process need to get them implemented at the level of the federal government. (It is a major, common rationalist fallacy to elide this distinction. "We should just do X!")
I may do good research and statistical analysis and decide that spinal fusion surgery for degenerative disks clearly falls into the "bad" category. This does not change the fact that there are many back surgeons who not only perform but will vociferously counterargue for this procedure. Rinse and repeat for many, many "bad" treatments.
I take Robin's "just cut half" argument to be partly a response to this fact. It's not that literally no one can distinguish good from bad. It's that it's hopeless to rely on those distinctions being made and accepted in advance of taking ameliorative action.
This is a good argument in favor of the decentralized/free-market approach to trimming health care, as opposed to the centralized/"medical establishment decides" approach.
I agree that simply throwing up our hands and saying that distinguishing good from bad is hopeless is not a good recommendation.
I don't know if this directly affects the way you refer to it in the argument, but it seems salient to the overall discussion that plagiocephaly is itself the product of a societal medical intervention: one of the side effects of the Back to Sleep campaign from the 90s targeting SIDS.
To me Back to Sleep is a microcosm of the issues with modern day medical science which make me personally skeptical of the value of the marginal dollar of medical spending. You have a very legible outcome targeted aggressively through public health interventions without any mechanism to weigh the benefits with any less legible costs. (Babies also seem to sleep more deeply on their stomachs - this is one of the proposed mechanisms for the success of Back to Sleep, but clearly there could also be drawbacks to babies sleeping less deeply.) And once the clarion call of the public health campaign has sounded, scientists, journals, and especially media outlets trip over themselves to frame every nuance so that it comes in line with the official orthodoxy.
Note that I claim no certainty whatsoever that the costs of back sleeping outweigh the benefits for the population at large or any slice of it. But I am confident that our system of research, public health, and primary care is not equipped to meaningfully investigate the possibility. The further question of how this relates to marginal medical spending at a patient level is complicated, but I think it's always worth considering the possibility that any particular recommended medical intervention may be myopically aimed at one legible metric.
+1, AFAIK most cases of plagiocephaly are a direct result of parents forcing their babies to not sleep on their stomachs.
I read about a third of this and shut it down due to boredom.
Modern medicine is a business and a big one at that. It’s not in existence to help or cure anything or anyone which is not sayin there are no caring physicians and nurses in existence. There are plenty of them. There are also the type that are in the medicine business for the money. Those types believe in prescribing pharmaceuticals for practically everything. Good physicians are limited by government and big business. Government tries to justify this by saying they are protecting the people from crazy quackery. That’s fine, except they take it too far. A lot of doctors’ hands are tied when it comes to how they treat their patients. This is a travesty. The government also gives a lot of criticism and objections to homeopaths…real medicine. Another travesty.
Bottom line is that Big Medicine is running the show and robbing doctors from doing not only what they’ve been trained to do, but also not allowing them what their intuition combined with their training and common sense guides them to do.
So this therefore causes much confusion and misery for doctors and especially their patients.
What’s more, this topic spills over into the ridiculous high cost that government imposes on everyone, especially the patients. I won’t call it “healthcare” because the truth is that the Big Business is not about care, only money.
Homeopathy is pure placebo. I guess the placebo effect is real, so in this sense homeopathy is real.
It most certainly is not placebo.
It has no plausible mechanism for working. Diluting an active ingredient until its presence is undetectable leaves behind expensive water.
Your comment goes to show you know nothing about it, have not studied homeopathy and so your comment is totally ignorantly based.
Sure. Water without an active ingredient nonetheless.
Ha ha ha! You damned fool, you've fallen right into our trap! Watch in awe as we eloquently explain the mechanism behind homeopathy, adduce dozens of large-scale pre-registered double-blinded studies, and conclude by rendering trivial the objection "but why isn't seawater and stuff like super potent homeopathic poison-cures?" with plausible and simple corollaries of the initially-explained physical mechanism.
Take it away, Majo! Now is when all the rigorous coursework in "Evidence, Application, & Terminology: the Study of Homeopathic Ingredients & Theory" pay off!
...Majo?
...
...look just hold tight man he's got the goods alright this is gonna blow your little establishment-drone MIND with some non-Illuminatized hyperrreality
Yes yes yes I can't wait, hopping with excitement here! :)
Like I said, you are not qualified to comment as you’re totally ignorant on the subject.
I'll comment anyway. Empty water. Good placebo.
Closed mind.
Yes, I guard my mind against junk. I duly note, however, the total lack of any counter-evidence. Anything?
Until then, empty water proffered by scam artists.
I studied homeopathy and have been using it since 1992 on myself and my family. I've attended the National Center's weekend seminar in '95 and learned much more about it, very impressed. My personal physician is a homeopath. I have in my repertoire aka., medicine cabinet, many remedies and a few helpful guidebooks written by reputable homeopaths, plus a repertory of remedies and Materia Medicas. (The remedies never expire if safely kept). I don't care what you choose to believe or think you know, nor do I wish to debate the subject. I do find your closed mindedness very sad indeed. You’re making blanket statements out of ignorance. I assume you are one of those who runs to get an RX filled at your friendly local pharmacy store as your first line of healing anything.
I sincerely recommend you reading up on how and when homeopathy was discovered. It’s a fascinating subject. The best source is NCH or National Center for Homeopathy.
I know how and when it was discovered. The theory behind it is bunk. Your assumptions about me and what I know and what I do are classic case of projection.
One minor thing - of course the "remedies" never expire, they are pure water.
This conversation back and forth is useless. I’m done and you should be too.
What about the value of this as a precautionary principle, or a bravery debate? Can't counseling something like skepticism towards the medical establishment be valuable? I'm being charitable by imagining a steelmanned version of this argument that encourages people to take a more critical look at, say, the correlation between medical spending and medical efficacy specifically. Maybe more costly treatments are not sufficiently more beneficial? Something like this seems to be at the heart of positions like Hanson's.
> What about the value of this as a precautionary principle
Easy question. Precautionary principles don't have value. Anyone invoking a "precautionary principle" is a moron who isn't capable of following simple logical chains of thought.
The precautionary principle answers "no" to every question of the form "should I do X?". In addition to carrying literally zero information in the entropic sense, this means that it generates irreconcilable logical conflicts as soon as anyone rephrases a question.
Okay, so you are fundamentally misunderstanding what people mean when they use the phrase 'Precautionary Principle'. But the fact that you've introduced information theory to a high-level discussion about basic heuristics leaves me feeling like it would be worthless to try addressing it. Consider this a human check and get back to me if I'm wrong or you were just having a bad day or something.
I haven't misunderstood anything. I gave you an accurate definition of the Precautionary Principle: everything is forbidden. That is both the definition people give and the meaning they intend when they invoke it.
This isn't new news or anything; you can see the problems pointed out on wikipedia. Check it out:
> As Michael Crichton wrote in his novel State of Fear: "The 'precautionary principle', properly applied, forbids the precautionary principle."
> If all that the (weak) principle states is that it is permissible to act in a precautionary manner where there is a possible risk of harm, then it constitutes a trivial truism and thus fails to be useful.
> If we formulate the principle in the stronger sense however, it looks like it rules out all courses of action, including the precautionary measures it is intended to advocate. This is because, if we stipulate that precaution is required in the face of uncertain harms, and precautionary measures also carry a risk of harm, the precautionary principle can both demand and prohibit action at the same time. The risk of a policy resulting in catastrophic harm is always possible.
> The strong version of the precautionary principle, in that "[i]t bans the very steps that it requires", thus fails to be coherent.
If anyone uses the words "precautionary principle" other than to point out how stupid it is, there are only two possibilities:
(A) That person is, themselves, extremely stupid.
(B) That person is completely unconcerned with whether the things they say are true or false.
Yeah, you're one of those people who gets the relationship between math and reality backwards aren't you?
Robin now has a response to this response:
https://www.overcomingbias.com/p/second-response-to-alexander-on-medicine
I’m very sympathetic to the concept of avoiding “low value” care. I’m a medical conservative practicing with a constrained vision (or at least striving at that).
But 2/5 Hanson prescriptions (low spending region, and small hospital, practices) simply removes “expensive” care. I would argue “expensive” does not necessarily equate to “low value”.
He also doesn’t precisely define how he would weigh those 5 steps, in order of importance (unless his number 1 to 5 is in that order or reverse order).
Finally, his suggestions can be operationalized on an individual level, but how do you apply that on a system level? The FDA already screens for “effectiveness” (admittedly doing a crap job at it) and Medicare and CMS does some attempt at covering things that are “cost effective” (also not the best at what they do). Where should it go from there?
He thinks such expensive care generally isn't valuable because there don't appear to be better health outcomes associated with higher spending regions/hospitals.
In the linked post he mentioned Cochrane reviews as one more way to determine which things to cut.
I think there's a fourth possibility that does not fit in your trilemma: antibiotics today might be more expensive than antibiotics 50 years ago without being any more effective today than those antibiotics were 50 years ago. (This is the problem of antibiotic resistance.) Perhaps some parts of medicine are an adversarial game, where you spend more and more over time to stay in place against an adversary that keeps adapting to your spending. The problem is not effectiveness but the cost of being effective.
This picture is compatible with both "increased spending is not at all correlated with better outcomes" and with "cutting anything will kill lots of people". The options here are "keep spending more and more money to stay in the equilibrium we're in" or "cap medical spending, regress to a worse equilibrium".
(Incidentally, I believe Hanson might suggest cutting out medications in order of development cost, not alphabetically.)
> Perhaps some parts of medicine are an adversarial game, where you spend more and more over time to stay in place against an adversary that keeps adapting to your spending.
"Adversarial game" doesn't mean this. You want "arms race" or, if you're feeling fancy, "Red Queen's race".
You need > 1% credence in the _combined_ proposition that both [son will have misshapen head without $300 intervention] AND [intervention will reliably fix this].
The problem is you and Hanson are (as of right now) arguing past each other. You're pointing at a study that shows if you point a gun at someone and pull the trigger they're really likely to die. Hanson is pointing at a study that says the number of guns in a society doesn't affect how many people die each year. So when you ask the question "Do guns kill people?" you answer yes and he answers no and you're both right.
You are in fact missing a plank of evidence. If a treatment is both available and effective it seems to casually follow that it has a positive effect on the population but Hanson says that looking at the end result on the population the casual chain you assume doesn't exist.
To provide yet another analogy, seatbelts are effective and available, it's easy to show that if you're crashing a car at any speed wearing a seatbelt is much better then not wearing one. But if you make them mandatory to try and apply the effects to the population and then everybody starts driving 30 mph faster you could watch more pedestrians die and even have more auto fatalities if a 40 mph crash without a seatbelt is less deadly then a 70 mph crash with a seatbelt.
Now mind you, I think you're instincts are actually the correct ones and Hanson is putting too much faith in those three studies which only kind of support his strongest statements but proving it might be difficult.
"use it to throw out every clinical study in favor of for three social science experiments that"
Grammar nit: cut out the "for"?
I mean, if people are overconsuming healthcare and this is economically inefficient presumably the thing to do is tax healthcare and reduce taxes elsewhere to offset the revenue, and the consumers will decide which healthcare options are marginal.
Is this discussion suffering from equivocation on the definition of "interventions", i.e. interventions as classes of treatment versus interventions as specific, individual medical decisions?
It is clear in some sense that antibiotics are a good class of interventions and can be extremely effective in certain circumstances.
We can hold that belief while also thinking it is possible that antibiotics are very often over-prescribed and generally used in circumstances (e.g. viral infections) where their absolute impact is zero or even negative (due to disrupting the gut biome).
It is then also credible that there are policy models available, such as increasing the cost of antibiotics for the patient, that could result in antibiotics being deployed more efficiently.
It's also credible that individual education, regarding when and where antibiotics are useful, is beneficial. It is observable that attitudes towards antibiotic use, and how likely patients are to ask for a prescription, varies substantially across countries.
Analogous comments hold for other types of medical interventions, including for cancer.
The comparison to the UK should probably mention that most healthcare in the UK is provided by the National Health Service (NHS). The NHS will (generally) only provide treatments recommended by NICE, an institute that does very explicit and rigorous cost-effectiveness estimates which form a large portion of the basis of their advice. Treatments which aren't cost-effective are therefore rare in the UK.
We can observe that the US spends more (% GDP, expenditure per capita, whatever metric) than any other first world nation, but is distinctly middling in terms of “health outcomes”. So it’s not difficult to infer that maybe the US is not getting nearly as good “bang for buck” as most comparator nations. Unless US docs are particularly inept (unlikely), or the US uses lesser versions of treatments (for identical conditions) than elsewhere (impossible), it seems to stand to reason that much/most of the money is spent without commensurate outcome benefits. Now, there is the minor issue of substrate….the average American is huge, sedentary, and eats crap…so that US dollar is spent caring for a more unhealthy person, on average, than most other comparable places. Still, even if it seems plausible that it should cost more to keep the average American from keeling over, it seemingly shouldn’t cost that much more. It’s not hard to accept the general Hanson point that there is a fair bit of wastage going on.
But Scott is asking about where the rubber meets road (and I would expect nothing less). “Cutting at the margins” is pointless hand waving unless you specify where those margins are, how you arrived at them, and what you would be cutting as a result. Then hopefully return to measure the net benefit (or harm) of what you have wrought.
And as Scott points out here, there are many devils in those details. And it’s not cut and dried. Antibiotics for bacterial infections are essential and potentially life-saving, ….and are also ridiculously overused. We need for-profit industry to innovate (and recoup investment and make profit) while ideally making sure that what they are making actually works and is only used where indicated (as opposed to usage creep and therapeutic fashion). And the best RCT only gives an average treatment effect on a population level…and an ARR of 10%/NNT of 10 would be a blockbuster in my field but still means you are wasting your time and $ on 9 out of those 10 people without any a priori ability to determine which category Mrs. Smith sitting in the office slots into…and that on an individual level, any treatment effect (on hard outcome) is categorical (ie either it works, or it doesn’t; either you’re alive, or you’re not) and not a percentage.
So what Hanson sells is not objectionable per se….but is useless drivel unless he brings some meat and details to his prescription. Otherwise it’s just pointless whinging.
> Every study on the marginal effect of medicine has some way it operationalizes “marginal medicine” for the purpose of that study. In geographic variation studies, it is the medicine done in places that spend more on medicine, but not in places that spend less. For studies that compare large to small hospitals, it is the treatments done in large but not small hospitals. For experiments that vary the price of medicine or insurance, it is the medicine chosen by subjects who faced lower prices, but not chosen by those who faced higher prices. I remember at some point also suggesting using treatments with a lower Cochrane Review rating.
https://www.overcomingbias.com/p/second-response-to-alexander-on-medicine
"Unless US docs are particularly inept (unlikely), or the US uses lesser versions of treatments (for identical conditions) than elsewhere (impossible), it seems to stand to reason that much/most of the money is spent without commensurate outcome benefits."
Or if US doctors are about as good as doctors elsewhere, but we pay them much higher salaries. Which, as I understand it, is the case.
Re: baby helmets.
I think Robin would respond that the real issue is that you've listened to doctors telling you to not allow your child to sleep on their stomach, which in turn created the need to wear a helmet to correct for the head deformation caused by forcing the baby to unnaturally sleep on its back instead of its stomach. Obviously some children _do_ naturally sleep on their back so I don't know if that's the case for your family, but in general the helmet thing is usually due to the overconsumption of medical advice that Robin Hanson warns against.
The SIDS hysteria would be a great example of medicine overconsumption in general. It had good advice about not smoking next to your child and not keeping tons of blankets (and other sources of suffocation) near the bed but then overreached and demanded that parents never share the bed with their baby, as well as never let them sleep on their backs. Meanwhile in Japan almost all mothers share the bed with their baby and their SIDS rates are very low. But the medical establishment will never admit to these Japanese statistics and keeps on repeating S-I-D-S over and over again to scare parents into submission.
Didn't the campaign to get parents to make their babies sleep on their backs correspond to a major drop in SIDS deaths in the US?
SIDS is very rare to begin with, with a prevalence rate of 1.3/1,000 at its peak in late 80s: https://www.npr.org/2011/07/15/137859024/rethinking-sids-many-deaths-no-longer-a-mystery. We've since reduced it to 0.57/1,000 through a variety of interventions. The campaign recommended lots of different things:
1. Back to Sleep
2. Use a Firm Sleep Surface
3. Room Sharing, Not Bed Sharing
4. Keep Soft Objects Away
5. Avoid Exposure to Smoke
6. Consider a Pacifier
7. Avoid Overheating
8. Regular Healthcare
9. Breastfeeding
10. Avoid Alcohol and Illicit Drug Use
The problem is that its quite hard to disentangle the effects of every individual recommendation. It makes sense to avoid putting soft objects or using a soft mattress in the baby's crib (avoids suffocation) and not sleeping next to the baby while intoxicated (risk of choking them with your body weight).
But in Japan and South Korea babies (almost) always sleep with their mother in the same bed and are not forced to sleep on their backs. Yet their SIDS rate are lower than in the West. SIDS researchers completely ignore this fact or dismiss it as some sort of a statistical fluke. But they're likewise unable to prove that _specifically_ sleeping on the back is of much use.
I don't know, I find reviews like https://www.ncbi.nlm.nih.gov/pmc/articles/PMC9284601/ , combined with the campaign effects, pretty convincing.
The article itself seems to confirm the evidence is very weak? To quote:
"Conclusions
Low- to very low-certainty evidence suggests that supine sleep position may reduce the risk of SUDI (0-1 year) and SIDS (0-1 year). Limited evidence suggests that supine sleeping probably delays short-term ‘gross motor’ development at 6 months, but the effect on long-term neurodevelopment at 18 months may be negligible."
And then:
"Of 48 observational studies, 46 were considered to be at serious risk of bias, mostly due to confounding and misclassification bias."
And then:
"An increase in positional plagiocephaly at 2-7 months of age with supine sleep position is possible (OR = 2.77, 95% CI = 2.06-3.72; 6 studies, 1774 participants)."
There's pretty good evidence of supine sleep causing positional plagiocephaly and we have a clear mechanistic explanation for it. I don't think we have good evidence for prone sleep *on a firm mattress* causing SIDS, nor is there a mechanistic explanation for it. Yes, prone sleep on a very soft mattress or on a soft blanket could cause the baby to choke during sleep, but how would this happen on a firm mattress?
Here's a research paper from Japan where very firm mattresses are most common: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC7444554/. To quote:
"Co-sleeping was recorded for 61%, and the prone position was found for 40% of cases at discovery. Mother’s smoking habit exhibited an odds ratio of 4.5 (2.9–5.8)."
It shows that smoking near babies is indeed a bad idea but there's no correlation between prone sleeping and SIDS.
I agree the evidence is weak, but it seems to be leaning that direction pretty clearly, especially when combined with the epidemiological evidence from the campaigns.
What makes you think the evidence for the plagiocephaly is better than the evidence for the SIDS? There are fewer plagiocephaly studies and they don't grade them for bias.
Do we know what percent of general infants in Japan sleep prone? I don't think you can make a conclusion until you know that.
It seems like in every country that has had a Back To Sleep campaign (many), SIDS rates have fallen dramatically; did they all campaign about other risk factors at the same time?
(and from what I could find, some other risk factors like cosleeping didn't seem to decrease during the campaigns anyway)
> What makes you think the evidence for the plagiocephaly is better than the evidence for the SIDS?
From Emily Oster: https://parentdata.org/baby-helmets/
“There is little question that the advice prevented deaths from SIDS, but it was noted as early as 1996 that there was a head-shape side effect. In the pre-Back-to-Sleep period, positional plagiocephaly was thought to occur well under 1% of the time. More recent updates suggest the current share is somewhere between 16% and 45%”
(Incidentally she recommends helmets as a secondary approach)
> Do we know what percent of general infants in Japan sleep prone?
Around half as per https://www.mdpi.com/1660-4601/21/4/471
“They also reported that more than one-half of parents did not place their infants on their backs when they noticed that the infant was sleeping prone”
A study from 1999 reported a 38% prevalence in the control group (but an 81% prevalence in the SIDS group, though note that the study size is smaller):
“The proportion of prone sleeping was 81.0% and 38.3% in the SIDS and control groups, respectively, and an odds ratio was calculated as 10.4 (99% confidence interval, 3.9 to 37.6).”
> It seems like in every country that has had a Back To Sleep campaign (many), SIDS rates have fallen dramatically; did they all campaign about other risk factors at the same time?
Here’s the original Japanese brochure from 1997: http://sids.gr.jp/english/otherrecents.html
“The pamphlet encouraged parents to do the following to lower the risk for their baby:
Sleep the baby on the back
Do not smoke near the baby or during pregnancy
Breastfeed as much as possible
Do not leave the baby alone when asleep
The pamphlet urges parents to sleep the baby in the same room with parents, to avoid leaving the baby alone, to avoid sleeping the baby in unusual places like a sofa, and to avoid soft toys and pillows in the sleep area.“
So, yes, the campaign focused on several things at once.
Counterpoint: this article claims that the Dutch campaign in 1987 focused exclusively on supine sleep and it had a good effect: https://www.researchgate.net/figure/SUDI-and-SIDS-incidence-in-the-Netherlands-from-1986-to-2019-with-addition-of-prevention_fig1_356802858
Counter-counterpoint: the Dutch data is hard to benchmark against Japanese data because Japanese beds are a lot more firm on average. Prone sleeping would also be more dangerous in the presence of pillows, toys and blankets which weren’t a focus of the initial 1987 campaign. This would also explain why Japan had a lower SIDS rate than the Netherlands before the Back to Sleep campaigns have started: there’s just less soft objects for infants to choke on in Japan.
> 3. Room Sharing, Not Bed Sharing
In California the government runs billboards with the message "if you let your child sleep with you, you are a bad parent".
I have difficulty imagining a more disgusting thing for the government to do.
> That helmet is probably our “marginal” health care expense, in the sense that it’s less obviously important than the other two things we’ve used healthcare for this year (childbirth, a scare with our son’s breathing).
Maybe I'm way off, but isn't childbirth in hospital both expensive and not that outcome-improving in normal circumstances, making it a likely candidate for marginal health care expense? Maybe not in your particular case.
Isn't "not that outcome improving in normal circumstances" an average over 99/100 times when it does nothing and 1/100 times when it saves your life?
I don't think it's randomly distributed, though (that is, for some people the expected split is 9995/5 and others it's 80/20). I double checked and of course the stats are super noisy: the midwife associations will use stats that don't control for base risk factors, so they can claim that home delivery is safer than hospital births, but attempts at controlled trials for people with low expected risk to start with (see https://www.jogc.com/article/S1701-2163(16)00089-X/fulltext ) find a 0.2/1000 increase in Ontario and 2.1/1000 in Oregon. At a statistical value of life of $9M, that means being willing to spend between $1800 to $18900 more (Though $9M is an old estimate for adults, and children tend to command a 1.5 to 3x premium, so you may want to adjust further).
We wouldn't have minded a home birth in normal circumstances, but we were told twins were a Risk Factor and we had to do hospital.
BTW, in a perfectly designed system, shouldn't we *expect* that it's net neutral on the margins? A perfect system would include all the net positive stuff, exclude all the net negative stuff, and be neutral towards the net neutral stuff, that would therefore be specifically what exists on the margins?
You expect that it is effective commesurate with its cost, net neutral with respect to the value of a dollar spent, not that it is has zero measurable effect but costs a lot of money. The claim by Hanson is the latter.
Gun analogy, yes the conclusion is that adding more marginal guns is not signficantly dangerous, and that reducing or increasing the number of guns will not change the murder numbers significantly. More murders happen without guns that with even for the population with guns; the main determinant in the murder rate is how many murderers are in the pop rather than whether they are armed. Hell maybe arming more people reduces murder rate by deterrence more than it enables murderers?
It's not that "guns can't kill you" just like Robin never said "no advanced healthcare treatment purchased at the far margin of expense has ever helped anyone ever". Complete strawmanning, are you even trying to be fair here?
To the extent that a large fraction of healthcare spending is wasted -- fails to buy extra QALYs, or does buy some QALYs but at a price wildly higher than broad social consensus could agree is reasonable -- it seems like a very large share of that wasted spending is probably measures taken in the last 2-3 years of people's lives, where everyone kind of knows that they're buying days at a time, or at best a few months, at very large expense. This is why we ought to be doing a lot more end-of-life counseling, and getting people to actually write down some kind of objective standards for deciding when they're done -- when their family (and/or subsidies from the state) should STOP spending money on trying to buy them a tiny bit more time. But of course, when that was proposed as a cost-saving measure to make the insurance subsidies in Obamacare work out better, it was caricatured as "death panels". In this sense, I do think Hanson is correct, that people feel like if they fail to spend every last dollar possible to "save" their loved one's life (for a few days), then that means they didn't really love them. But this is a far cry from his casino metaphor.
(And for the record, yes, I have sat down with my spouse and a lawyer and put medical directives in writing.)
(I originally wrote that the vast majority would be end of life care, but upon some thought decided I couldn't justify a claim quite that strong, considering stuff like the heart attack care comparison that Atul Gawande presented, some years back, where in neighboring counties you saw one of them following a standard care regimen that cost drastically more -- and profited the hospitals much more -- but did not improve outcomes. I think there is quite a bit of wasted spending that looks like that as well. Actually, my aunt wrote an entire book about that kind of wasted care in the ob/gyn field: https://www.amazon.com/Optimal-Care-Childbirth-Physiologic-Approach/dp/1598491326/ )
I think it would be interesting to know to what extent it's *actually* true that vast amounts are spent trying to prop up people who will almost certainly be dead within a year anyway.
Everyone seems to have this intuition, but is it actually true, or are we actually just talking about ~10% of total medical spending?
How does medical spending break down between these categories of people?
a) Generally-healthy non-ancient people who can be made healthy again
b) Decrepit ancient people whose lives are being artificially prolonged at massive expense?
c) Chronically ill people who may continue to live for a long time but never truly be healthy?
I feel like this third category might be the really big money sink.
Your category (c) can spend a lot of money on a per-individual basis, but those people are rare. People in category (b) aren't rare and also spend a lot of money per individual.
I think Hanson is wrong about the specifics but his main point is right. Singapore has world class healthcare and only spends 4% of it's GDP on it. So it's possible in principle to deliver the same quality with half the cost or less.
It's probably not possible in practice though. In Singapore, healthcare costs are controlled via the same mechanism of almost everything else - markets. Patients pay for healthcare out of their own healthcare savings accounts and it's up to them to decide which interventions are worth it, and it's then up to providers to convince them what interventions are necessary. Of course, this introduces the information asymmetry problem, but going off outcomes, it doesn't seem to be worse than the US system.
Westerners are used to having almost everything paid for at point of service, either from government health plans or private insurance so there is no incentive to ration. Taking that away is politically impossible.
It's not markets in the sense that all payments are voluntary , or that nothing is taxpayer funded.
"Financing of healthcare costs is done through a mixture of direct government subsidies, compulsory comprehensive savings, national healthcare insurance, and cost-sharing."
After looking up the numbers, Singapore's healthcare spending has changed more than I expected over the last decade with government spending going up and private spending going down. However, compared to most Western countries Singapore still has a much higher proportion of spending either from out of pocket or from health savings accounts (around half). That's still enough to have a large effect of how people consume healthcare. Instead of only considering the benefits when considering an intervention, patients have to consider costs and benefits.
Re the Britain point, one needs to control for the fact that people there are healthier than Americans (less diabetes etc). British people receiving US-level medical care might well become healthier.
"In the aggregate, variations in medical spending usually show no statistically significant medical effect on health"
One needs to track in each study how an "effect on health" is defined. For example, it shouldn't be necessary for healthcare to expand lifespan in order to be qualified as useful. I recently started an allergy shot treatment a few months ago. This will almost certainly have a negligible effect on my lifespan, but the increase in my quality of life has been dramatic. And insurance complicates the analysis for cases like this.
The only reason that I got this in the first place is that I am graduate student, so the insurance I get is heavily subsidized, and it covered the allergy shot treatment. Given the option between taking the cash and doing the allergy shot treatment, my past self would have chosen to take the cash. (Knowing what I do now, I would tell a parallel-universe version of myself that the price tag is worth it).
A significant fraction of medical expenses are arguably like this. Painkiller medication, palliative care, hospice care, etc. might have no effect on "physical health" but are still worth the price tag, and may be covered by insurance. The helmet for your daughter might just be cosmetic but still worth it. Most people in the United States get braces, women (and others) spend money on cosmetic products, folks get plastic surgery, people get hair transplants, and the list goes on and on.
So imagining you have some study where you give people better insurance, and you find out that people used the insurance, but it didn't increase their "health." When you look closer you find out that the types of extra treatments people received were not the type where we would expect an increase in "health" anyway. However, people in general reported that knowing what they do now, it would have been worth it to pay out of pocket for the treatment they received because of the increase in quality of life!
In this post, Scott goes "near mode" (his term: I'd say something like "lets the rubber hit the highway") by asking what Robin would have him do in actual medical situations.
One time when I posted a query about what someone giving out contradictory indications would have me do in the situation in question, I got a lot of severe replies saying it was my decision and I didn't have to do what the other person said. I KNOW that: my question was, what would he -advise- that I do? I don't know why that wasn't clear, and I wish Scott better luck with the same mode of reasoning.
He actually does explain it in the new response he wrote:
> My key point was and is that each of these operationalized definitions of “marginal medicine” offers a concrete way to avoid marginal medicine. As an individual considering various possible treatments, here are five ways:
1. Ask about a treatment’s Cochrane Review rating,
2. Ask if a treatment is done in low spending geographic regions,
3. Ask if treatments are done in small hospitals,
4. Ask your doctor how strongly they recommend a particular treatment; decline if recommendation is weak. (I’ve done this.)
5. Ask yourself and associates if you would be willing to pay for them out of your own pocket, if insurance did not cover them.
Maybe even better to ask several of these questions, and average their answers.
As a wonk considering various possible policies, you can also consider regulating or subsidizing/taxing based on these indicators. Or consider policies that make more patients face higher personal prices for treatment. When I said “most any way to implement such a cut” I had in mind these sort of options; most any should help. Though my favorite option is still creating agents who face strong direct incentives.
I completely agree with you about investigating any proposed treatment via Cochrane reviews, etc. , but I really think it's a terrible idea to ask a doctor about Cochrane reviews of something they're recommending. Doctors often *hate* it when the patient believes he's knowledgable, and hate it even more when the patient puts them on the spot. And most doctors are not going to KNOW whether there was a Cochrane review and what the upshot of it was. I have no problem being assertive with doctors, but I have stopped asking them questions like the one suggested because it almost always backfires. You don't get your question answered, and the doctor's attitude towards you is going to be less friendly and helpful forever after.
Not that I don't recognize where you're coming from in these experiences, but I think I've had better luck than this at getting the information I want out of doctors, to the extent that they have it. I also wouldn't ask these specific questions, except #4, which seems fine and I think people use variations of this pretty often ("would you do this if it were yourself/your spouse/your kid/your dog") without wasting one of the fifteen minutes that have been allocated to the visit or offending the doctor's ego. Could be luck or practice or some kind of weird doctor soft skill voodoo but now I'm wondering whether the approach can be systematized.
I dunno, I am generally pretty good at tactfully raising issues without provoking conflict. I once said to a doctor who recommended treatment X, "I've read about treatment Y is being used for this? What do you think of trying that?" And she glared at me and said "So you know everything and I know nothing!" Another muttered something about "Dr. Google." One glorious exception: a smart, cheerful dermatologist my daughter and I both see. I asked about a treatment for hair loss, which my daughter has, and the derm said "Yes! They're doing that in Philadelphia now. It looks promising. Let me see if I can remember the name of the guy heading the research." And she got on the office computer and found it for us. Agreed we should try the new thing if the present one wasn't effective enough.
The gun analogy does not seem very informative. Gun advocates in the real world do claim that guns can be used in self-defence or as a deterrent (preventing murders), and that guns only substitute for other weapons (ie murders that happen with a gun would have happened anyway with a knife).
This is why it is very reasonable to prefer to look at the society-wide, aggregate effects of gun prevalence.
The same is true for medicine: something helpful in some circumstances may be outright harmful, useless, or negative net value in others.
And this is of course only if one accepts that the evidence in favor of medical treatments in general is as compelling as a hypothetical RCT on “can a gun kill you?” would be.
On which, regarding p-values I hope it is true that standards of evidence are improving. But surely most medical treatments, and especially those that drive the bulk of medical spending, have been approved historically. Recent uplifts in evidence standards presumably then aren’t of much relevance? Or am I missing something?
Typo?
“Robin thinks it’s a “monkey trap” to try to cut the good parts of medicine but not the bad.”
This topic has made me think about my own medical care, which I generally walk around thinking has been decent, and definitely worth it, though of course not perfect. But I just sat down and did an inventory of all the medical problems and care I've had, and rated each class of care. I gave pluses to things that had undoubtedly helped me, and minuses for bad features of the procedures, including stress, expense, pain, and iatrogenic features. And when I added up all my ratings, medical care was a net negative.
Lots of caveats: I did not include immunizations in my accounting, and I regard them as an unmitigated positive. And of course I’m an n of 1, so I’m a datum not data. And I have had good health all my life, so none of my medical treatment has saved me from something that could have ruined my life or killed me. If it had, my ratings might be very different. And of course I understand there’s lots of room for disagreement about how heavily to weigh various good and bad features and consequences of medical procedures, and that my accounting doesn’t prove anything. Still, the result surprised me, and I’m curious whether anyone else using a system like mine ends up with a negative evaluation.
*Orthodontia as a kid. My teeth really did look pretty awful before I had braces, and afterwards were aligned quite neatly. On the other hand, my dentist used an old style of braces that involved placing metal bands around the teeth close to the gum line. I now have an exceptionally bad case of gum recession, which I’m told those bands contributed to. However, I do various dentist-recommended things to keep the exposed surfaces from developing cavities, and so far none has.
So overall: + -
Routine care:
*Cleanings: I believe the Dental Authorities that scraping off plaque is important, so cleanings are entirely a +
*I have had about 4 fillings and one root canal. I’m sure those 5 teeth would be gone or else hurting like crazy without the dental care, so that’s entirely a +
Overall ++
Non-dental medical stuff
* Had 6 or so warts on my hands as a kid. Doctor tried dripping some acid on the biggest one, and that hurt like crazy for a whole afternoon. A few months later all the warts disappeared on their own.
Overall. -
*Had my first UTI as a teen, did not know what it was so did not go to the doctor. Eventually I developed back pain and a fever when the infection reached a kidney and I went to the ER, where they gave me antibiotics that wiped out the infection. Seems to me that without medical care that infected kidney might have made me quite ill.
Overall +++
Birth control
Pills worked but made me gain 25 lbs in a year so I stopped them + -
Had an IUD inserted and a few days later developed a uterine infection with discharge and fever. Device removed. -
Diaphragm: It worked, but after getting it I had maybe 20 uti’s over the next couple years, and had to take many courses of antibiotics. Also visited a urologist whose exam was by far the most painful thing a doctor has ever done to me, and exam result did not find any abnormalities. Eventually I figured out that my diaphragm was a size too large, and was stretching me in a way that made it easier for debris to enter my urethra. - - -
Overall: + + for no unwanted pregnancies, - - - - for quite significant iatrogenic problems
Fertility treatment:
*Had 8 cycles of IVF, which were extremely stressful and did not work. During the first cycle the anesthesiologist made a mistake when knocking me out in order to retrieve my eggs. There are 2 drugs that I know of used in that kind of anesthesia: One that paralyzes you and one that knocks you out. He gave me some of the paralytic when I was fully conscious. My eyes went out of focus and I felt myself becoming unable to move. AND he put a mask over my nose and mouth and there was no flow of air coming through it. I have a vivid memory of trying hard to inhale, getting no air, and feeling the edges of the mask pushing deeper into my face because of the suction I was causing. I was terrified, and clumsily waving my arms around and trying to say “knock me out,” but could only produce incoherent sounds. (Afterwards I described what had happened to some anesthesiologist higher on the hierarchy than the one who’d tortured me — who was maybe a resident? — and he said he believed me, and would make sure nothing like that happened during any further surgeries I had with their group.).
Fertility treatment overall: - - - -
Some kind of post-viral syndrome after the flu
I had it for several *years*: Hypersomnia, exhaustion, exercise intolerance, pretty bad joint aches and several joints that were extremely tender, fever of 99.5 that came and went. My internist tested me for everything we could think of, and all tests were negative, then she kind of washed her hands of the matter, and began to imply there was no point in using time in my appts with her to discuss the problem. She never even told me that post-viral syndromes are a thing.
- -
Prophilactic scans and exams:
*Blood work: Does not bother me, but they never find anything except stuff whose validity seems kind of questionable to me, such as low vitamin D levels. On the other hand, I do like knowing that everything’s in a good range, even my cholesterol — and that one’s important to intervene with (I think) if abnormal.
*breast scans: Have had a bunch, about 3 of which involved some drama: 2 callbacks for further scans, one for a needle biopsy (which was aborted when the doctor found the spot in question was a cyst that popped and drained the minute his needle entered it)
*colonoscopies: Of course I loathe them, like everybody else, and they never find anything. I read recently that the pre-scope purge disrupts the gut microbiome, & that concerns me some. Also, staying up all night purging then getting knocked out and probed with a plumber’s snake screws me up so much that I don’t feel normal for a couple days afterwards. So counting the pre-scoping diet, where you can’t eat anything except shit like jello that I never eat and the post-scoping malaise, every colonoscopy takes a week of wellbeing out of my life. Oh yeah, one time I got a bill from the place I had the colonoscopy — turned out my insurance covered the procedure, but not the place where the colonoscopy was done.
*bone density: Docs push these, and I’ve had some, but now doubt they’re worth having.
*CAT scan fishing expedition: Once had a weird pain in my lower abdomen, and doc ended up sending me for a CAT scan, which found nothing in the area that hurt, but an “incidentaloma” in my upper abdomen on the opposite side: A duct leading out of my gallbladder
was dilated to an abnormal size. Doc who’d ordered the scan wanted to do the mother of all endoscopies. It had a 20 syllable name that I’ve forgotten — involved sticking a tube with a camera down your throat, but then once in the stomach they fucking *keep on going* into the gut somehow until they can get a look at the gall bladder. I looked up the procedure and found it caused pancreatitis 5% of the time. So I suggested waiting a couple months then rechecking that duct via ultrasound. Doc said OK and the duct was normal on the recheck.
So these exams earn a LOT of minus points. I’m going to give them 4 minuses, because of how goddam many of them I have had, just because I usually get whatever my internist recommends, and because of the stress, expense, discomfort and, in the case of the threatened gut probe, danger they have caused me. I’m not sure how many pluses to give the exams. I’ve read recently that both mammograms and colonoscopies save relatively few lives. Also read something pretty persuasive about how noisy the data from bone scans is, and how it’s not possible to compare readings from 2 different machines. And besides, I don’t trust the drugs given for osteoporosis. Blood work is easy, and I have the feeling that it would pick up a lot of subtle things that are taking a bad turn, so is worth done. (But I’ve never seen data about how many lives or days of serious illness are avoided by getting bloodwork, so I may be wrong.) So I’m going to give this mass of time-consuming and unpleasant prophylactic exams 2 pluses.
Overall: - - - - + +
TOTALS
Dental: +2
Non-dental: -10
Both dental and non: -8
Thanks for the detailed description!
Wow, I'm sorry you've suffered so much with the contraception and IVF! That "conscious but paralyzed" part is the stuff of nightmares. I'm going to have my first colonoscopy in a couple of years, so that's something to look forward to :)
Here's my n = 1: I've had lots of treatments in my life, mostly for small things, but I had an appendectomy at age 10, without which I almost certainly would have died a painful death of sepsis, so that right there makes my lifetime health care use a ++++++++++++++++++++++++
Also, I had scarlet fever as a baby, I have no memory of this but my mom told me I was horribly sick and had to be taken to the hospital, so there's a good chance I would have died at that point without modern medicine.
For me, the most painful medical procedure ever is a close tie between having a wart frozen off my thumb and having my stitches removed after the appendectomy (it was many years ago in Poland, and they didn't use self-dissolving stitches then). The nurse was simultaneously digging the stitches out of the scar on my abdomen with a metal instrument and pouring a disinfectant on them, and telling me to stop crying.
Also, the use of epidurals in childbirth is a +++++++ for modern medicine. Unmedicated childbirth feels like someone slowly ripping your lower abdomen apart with a dull knife, repeatedly, over the course of hours. No. Thank. You.
I'm hating on that Polish doctor.
Yes, if your life has been saved by medicine the math is different. Still, maybe it would make sense to separate out the life-saving responses from the rest. I actually do not doubt that in most US ER's someone with a common, life-threatening problem will get life-saving treatment. Seems like maybe where health care falls down is with more routine things, where there's a lot of ineffective investigation, and some treatments that cause harm. That was my experience, and I think it's likely to be a pretty common harvest from the routine stuff. What do you think of your medical care aside from the 2 emergency interventions when you were a kid, and the pain epidurals saved you? How'd the doctors do with more everyday things?
As for my bad ob-gyn experiences: The thing is, I actually had a pretty well-running reproductive system. I never had bad cramps (except for the year when I was on BC pills), or any of the other things things that are pretty common -- polycystic ovary syndrome, endometriosis, infertility at a young age (mine was just due to being in my 40's), severe PMS, etc. ALL of my bad experiences were due to medical interventions. I'm especially indignant that over 20 visits for UTI's it never occurred to any doctor to check the fit of my diaphragm when I told them that the infections invariably started the day after I'd had intercourse. I doubt that an over-large diaphragm is a rare problem, and with no specialized knowledge about fitting the damn things or about UTI's *I* eventually figured it out: Wait, I can actually feel the bottom of the loop right inside, and even protruding a tiny bit. Oh, in fact there's a stretching feeling where the diaphragm is pushing its way out. Wait a minute, could it be that . . .
I'd be long dead without modern medicine, so I have a special dislike for Hanson's elaborate arguments that - to me - amount to telling me I don't deserve to live.
Drosophilist would also be dead without modern medicine. I wonder if it's more in the "routine care" that medicine falls down. Note that the contingencies and rewards in routine care are very different from those in treatment of grave illness or injury. In the latter, there's quick feedback, visible to all, regarding whether the medical interventions succeeded. In the former there's not, and of course the medical system needs to earn its keep by doing *something*. I keep running across articles making pretty persuasive arguments that lots of routine stuff has little or no benefit: yearly physicals, various screenings and tests.
So my case is an interesting illustration why there may be a statistically sound case for routine care being less than useful, but in this case it was literally life-saving. A malignant melanoma was found during a routine-care visit, and it was just in time to still be at a stage where it didn't have a chance to metastasize - barely. The GP literally cut it out right there and sent to the lab, and I was under the knife for a wide excision a couple of days later.
Regarding the gun analogy, it seems to me that a lot of medical studies are "apply medicine, see if they get better afterwards, control for placebos, etc." like seeing how many people die after getting shot by a gun. Are there any medicines we can test in the manner of "actually looking at what a bullet does to the human body?" Like analyzing how living tissue reacts to the drug that's supposed to treat it (or the bad thing that's supposedly harming it)?
Yes, many. But it's too easy to do this (at least on a superficial level) even for things that end up not really working, so people don't take it very seriously.
One model that Hanson could have could be something like the following.
Conceptually, we can think of the process by which we decide which medical treatments to do as something like this:
1. Rate all the treatments available on a scale of how worthwhile they are, from 0 to 100.
2. Do all the treatments above some sort of threshold, say 50.
Now of course our system doesn't do that explicitly (although we might make better decisions if the system did!) but the way the system does work could be factored out that way. For instance if you do a study to figure out which antidepressant works the best, that's part of (1). But when you have a guideline like "anyone who scores above X on such-and-such depression questionnaire should be given antidepressants", then changing the value of X is a change in (2).
It seems like a coherent viewpoint to say both that:
(i) The system we already have does (1) pretty well; we have a pretty good sense of which treatments are most worthwhile (e.g. vaccines are at the high end of the range, back pain surgeries are maybe at the lower end)
(ii) But the threshold in (2) is set way too low; rather than doing everything above a 50 we should only be doing things above 80.
So, the answer to the question of "which specific treatments should you cut" is "the ones that seem iffy based on our current state of knowledge - i.e. if it seems iffy then it's probably actually bad, if it seems mildly good then it's probably actually neutral, if it seems incredible then it's probably good."
===========
I think that in general, a lot of economics arguments work this way - i.e. you can tell if a parameter is too high or too low even if you don't know the specifics. For instance if it were the case that:
- the government massively subsidized widget manufacturers due to lobbying from the widget industry
- and there were no positive externalities, network effects, or anything else that make those subsidies economically beneficial,
then that is sufficient to tell you that probably too many widgets are being produced, even if you don't know anything about which specific brands of widgets are good or bad.
But you're right that it's unclear what Hanson's model actually is. It doesn't seem like his model is **just** that government subsidies are messing things up, it seems like he thinks that **individuals** are making bad decisions. One thing that is confusing to me is that he seems to think that:
(a) people are spending too much money on medicine, and not enough on preventive measures (diet, exercise, etc.)
(b) this is explained by status-seeking and conspicuous consumption.
While (a) seems very likely to be true, I don't see at all how (b) is supposed to help explain (a). Most preventive measures are a lot more conspicuous than medicine! We eat with other people all the time, and whether you're overweight, muscular, etc. is easy for other people to see. But our culture seems to me to treat medical decisions as a particularly private domain, and inquiring about other people's medical decisions is usually perceived as intrusive. This seems the opposite of what you would expect if people were using medical care as a status-seeking method. It seems like the "status-seeking" hypothesis would predict that people would spend too much on highly visible preventive care, rather than too little.
A possible steelman of the trichotomy would be the following:
- There is a small amount of interventions (like vaccines) that are net positive.
- However, we don't have a good way to distinguish between mildly positive, neutral, and harmful medical interventions.
- The mildly-positive-to-harmful group is net neutral on average.
- Therefore, we should discard everything that isn't obviously positive; trying to move the line in order to include a little more will start including some negative interventions as well, and therefore is a monkey trap.
Does it look like a consistent position to you? I think I could see Hanson holding it.
I think the charitable interpretation of Hanson’s position would be:
(Due to how healthcare is set up) the default action is to give medication. It affects nearly all types of treatments equally (antibiotics are good, but they get prescribed for every person with a running nose, whether they have bacterial infection or not) so figuring out which medicine is ‘good’ is irrelevant. Knowing this, a person should realize that for cases like this they shouldn’t take antibiotics.
I disagree with this last part, as the average person is unlikely to make good medical decisions on which treatments to ignore, but the general point seems mostly correct.
Everything is useless at the margin.
That is a simplification for rhetorical effect. Here is the actual concept gestured at. If you are maximising a smooth function (for example, a utility function) that has a maximum in the interior of its parameter space, then as we all know from first year calculus, its gradient there with respect to all of the parameters is zero. At the maximum, the marginal effect of changing any parameter up or down is zero. Small finite changes to parameters will have only extra-small effects, perhaps too small to see.
Thus, "at the optimum, everything is useless at the margin."
This has no implications for whether anything is just useless.
The 250,000 medical error deaths is crude, but it is far more sober than the alternative studies which are obviously wrong. Counting only recorded errors is an obvious bias. Studies which claim that there are fewer iatrogenic hospital deaths than central line infection deaths are drunk.
A serious discussion of iatrogenic harm and other concerns about when it's advisable to seek less health care: https://www.youtube.com/watch?v=yr_4RoPhtu4
New paper just dropped that seems to contradict the previous insurance studies:
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 wonder if part of cutting "frivolous spending" could be something like, if you have some pain that doesn't seem serious wait 48h and eat well, sleep well before going to the doctor. "that doesn't seem serious" is doing lots of heavy lifting here, but also it seems doable to have some kind of master flowchart to orient people, that's refined and improved from time to time.
On the other hand I don't know enough about american healthcare to know if people getting medicine when they don't need it is a thing that really happens. My impression is that due to the price people seems to be getting less/avoid getting treated, but there's a good chance it's my bubble.
There may be also cases where medicine is not exactly the right answer. For example, maybe X% of back problems would be better treated by physical therapy/getting more muscles than by oxycodon/surgery, and since those in that case are useless and can get pricey and make people worse (addiction) all medicine looks worse as a result.
But this is all a case of "let's look very carefully at which parts are good, which parts are bad and remove them" which doesn't seem to be Hanson's position.
Now time for the woo stuff, I feel like sometimes we have an oversupply of "we can find out what's the good and the bad part and remove the bad" but I have a vague feeling that it's sometimes not possible. It applies more to stuff like "this guy writes really well, but he would be better if his politics aligned with mine and in fact I think he could and should exist like that" but maybe Hanson's position is that it applies to the medical system, that the good and the bad can't be split.
Edit: I commented too early and ended up repeating some of Scott's points, especially about the back pain. I'd love a thing where the window for editing a comment is to the right of the article so I can edit as I scroll down the article.
So... there's a rather obvious way to resolve your two positions from my perspective.
You mention Great Britain as opposed to the USA - so far as I know, the utilization rate of medical care in both countries is more or less equal, but the PRICES paid per unit of care are dramatically lower in the UK.
It is theoretically possible to "cut half of medicine" by simply cutting the prices paid for care by half. (I am not advocating for this - I've experienced this as a small business pharmacist at the hands of health insurers and their PBMs, and it is NOT pleasant for anyone involved).
You are focusing on utilization cuts, when I'm pretty sure Robin's unit of measurement isn't care units delivered, but dollars spent on care.
The US spends dramatically more per unit of care than any other country. And the results of the insurance studies are consistent with this trend - a decent fraction of care delivered to uninsured people ends up being uncompensated care, i.e. the dollar flow for that care delivery is exactly $0. By shifting folks onto medicaid from uninsured status, you move all of the previously uncompensated care into (lower reimbursing) medicaid-compensated care, and that is, from a societal perspective (and maybe Robin's?) marginal care. Paying physicians and hospitals for care that they would have already delivered for free (bad debt writeoffs at the end) is pretty obviously near-zero marginal value to health status of the individuals - they got the medical care that they needed for free without medicaid, and the state paid for it with medicaid.
So I think that the way to read the Oregon study in favor of both of your positions is that 1) Medicaid enrollment induces some amount of marginal care delivery, which results in additional diagnoses (good!) 2) Medicaid enrollment means that hospitals (especially hospitals) get paid for a lot of care that they were previously delivering for free, which is, economically speaking, marginal deadweight loss. Taking both of those effects together, marginal healthcare SPENDING is net neutral. (disclaimer: stablizing hospital finances is not necessarily a bad goal, and is definitely more FAIR than expecting hospitals to deliver care for free, but it's not the primary purpose of medicaid).
I, of course, agree with the general principle that a fair amount of medicine is not helpful (take, for example, most use of proton pump inhibitors). I will add that a bunch of SPENDING is on schemes to game the third party insurers' reimbursement mechanisms (take, for example, Zegerid, a combination of omeprazole with baking soda, which costs approximately 70x omeprazole alone - that marginal spending is for BAKING SODA). I assume that such schemes are not unique to pharmacy, and are, in fact, widespread across reimbursement systems for medical care in general (surgeries made slightly more complex or delivered via robot when they could be done more simply and with equal effectiveness without the added complexity and/or robot, which serve the primary purpose of getting insurers to pay more money).
You keep referring to antibiotics. On average they are beneficial, but at the margins they are questionable. Every time I’ve been prescribed a z-pak, I probably works have been fine without it, but ins covered it so I took it. All the ear infections my kids got, sometimes they needed the antibiotic. But sometimes we’d get an rx and the doctor would say wait a couple days. Sometimes we waited, but sometimes we just filled it.
The point is, antibiotics are over prescribed. We should use less. But I don’t think doctors like to be the bad guy, so they overprescribe.
I think we could say the same about cancer. Treatment can save lives on average. But people also stay on treatment when it is obviously not helpful.
Back pain. I know some people that have gotten help from medicine. And others that have surgery or PT and still have pain. There are a lot of pains we just can’t do much for.
So it seems clear that average med is beneficial. But marginal is questionable. It is sometimes clear what extra med isn’t helpful. Sometimes it is not clear. But that applies to many things I purchase. So letting individuals make the choice seems right. Exposing individuals to more of the cost will incentives them to make the choices.
Has Robin Hanson considered the ethical consequences if his suggestions would be followed? It's one thing to cut expenses for seemingly ineffective treatments that supposedly result in statistics he thinks show medicine is useless. And another to cut expenses for medicine which is purely palliative. Maybe he hasn't considered what's going to feel like when insurance will cut his access to palliative interventions because he'd already used his 5 interventions that year? Why should anyone reimburse him for expensive morphine for example? It's ineffective anyway. He'll die anyway.