Here’s the abstract of an article by Martin, Rice, & Smith in the current issue of the Journal of Health Economics (generally regarded as the top journal in the field): Empirical evidence has hitherto been inconclusive about the strength of the link between health care spending and health outcomes. This paper uses programme budgeting data prepared by 295 English Primary Care Trusts to model the link for two specific programmes of care: cancer and circulatory diseases. A theoretical model is developed in which decision-makers must allocate a fixed budget across programmes of care so as to maximize social welfare, in the light of a health production function for each programme. This yields an expenditure equation and a health outcomes equation for each programme. These are estimated for the two programmes of care using instrumental variables methods. All the equations prove to be well specified. They suggest that the cost of a life year saved in cancer is about £13,100, and in circulation about £8000. These results challenge the widely held view that health care has little marginal impact on health. From a policy perspective, they can help set priorities by informing resource allocation across programmes of care. They can also help health technology agencies decide whether their cost-effectiveness thresholds for accepting new technologies are set at the right level.
Thank you for your reply, but you still did not mention why the broad category of Bayesian investors is a poor one for evaluation, yet the broad category of US medical care is not. It seems to me to be a goose/gander situation. Having been put in the position of having to accept that my profession has done as much harm as good (and I have accepted that), it is only fair to expect that those same economists shed light on the margin of other aspects of the US lifestyle, such as Bayesian investing or teaching economics, so that better strategies can be understood and pursued.
retired urologist, I assumed you were asking how to go about evaluating investment strategies. A detailed explanation of how I compare investment strategy evaluations to medical evaluations is not worth the effort it would take.If you're looking for strategies that are better than average (which sure looks like what you were asking for), I don't recommend evaluating broad categories such as Bayesian investors. Evaluating broad categories may give you insights into what produces poor results.
I Googled "Mark Hulburt" and could not find any references at all addressing your questioned statement: evaluating investment strategies is harder than evaluating medicine. Why didn't you just say, "that is my opinion or bias, but I know of no evidence" ?
I don't recommend evaluating broad categories such as Bayesian investors.
Yet, somehow you recommend evaluating such broad categories as doctors. Enquiring minds want to know, what is the distinction?
retired urologist, you could look at how Mark Hulburt evaluates individual people.If you want to evaluate a specific algorithm, there are problems with knowing whether it was selected to fit the specific data you're testing it on.I don't recommend evaluating broad categories such as Bayesian investors.
Retired, the point of my "generic traditional economist" bit was that even in an imaginary world with health insurance but without any of the biases and such that we are fond of talking about here on OB, and even where everyone *knows* the marginal health value of each unit of care (this was an implicit assumption that I should have made explicit), one would still predict a small (though still positive) health value of the marginal unit of care. Simply put, people use too much of things that are made artificially cheap, and health insurance makes health care artificially cheap in the sense that the user pays less for it than it costs to provide. Just like everyone eats "too much" at an all-you-can-eat buffet: each additional serving is free to you, so you keep eating until the marginal value (to you) of one more serving is just above zero, which is way below the actual cost of producing the food. In other words, plain vanilla economics gets us a good part of the way, but not all the way, to the result of a zero marginal value of medicine. (Is this part clear now? Let me know if it isn't and I'll take another crack at it)
As discussed above, all the biases and whatnot flying around the world of medicine seem to have resulted in a situation where the marginal value of medicine is zero (or at least that was so back in the RAND experiment days, this may no longer be true). As you point out, merely realizing this fact is not enough; what you really want is to know *which* treatments are effective and which ones aren't. Unfortunately, I have no special expertise in answering this question. I keep one eye open for studies like the one I linked on this page, and I like to think that the general tools that I've developed from places like OB enable me to ask better questions than I would be able to otherwise and maybe identify overconfidence or whatever when I see it.* And it is probably a good rule of thumb that if your doctor says that it is a close call between doing some procedure and not doing it, that you're better off not doing it.**
I know this is a little unsatisfying, and I'm not happy about it either. Of course there are people who know much more than I do about this, and maybe somebody has actually taken a crack at putting together a comprehensive list of treatments whose effectiveness have been well-established with high-quality evidence. If you find such a thing, please do let us know. The good news is that the recent past has seen a very large amount of growth in evidence-based medicine whose purpose is to address exactly these kinds of problems (one example linked below), and I expect that this will continue.
*Tiny anecdote. I have asthma, which has occasionally caused me to visit a pulmonologist, who would routinely give me a breathing test with a fancy machine. For a while that same guy was also my GP, and when I came in for a regular checkup, he wanted to give me a breathing test, despite the fact that I had had one a short while before when I came to see him as a pulmonologist. It was clear to me that this was just bill padding, and I declined it despite the fact that it was free to me and the test doesn't cause me any particular trouble or discomfort. It just seemed like a waste.
**Even this is not always right. I have a friend who is a very healthy athletic guy who hurt his knee. It was clear to him that he wanted surgery to get it taken care of once and for all, and a lot of surgeons were telling him to do therapy or whatever because surgery is invasive. He was convinced that the surgery was very low risk, and he was well willing to endure the risk and pain rather than to have a nagging, sports-limiting injury for the rest of his life.
@ D J Balan:A person with no insurance would start off buying the medical services that had the greatest health benefit per dollar, and would continue to buy services of successively diminishing health value until reaching a point where the marginal value of one more dollar's worth of services provides an increment of utility no greater than the increment of utility that the person could gain by spending that dollar on their next best alternative.
How would he "start off" knowing which services had the greatest benefit per dollar? Dr. Hanson mentions there are probably some effective treatments, but he doesn't say what they are. Is there a list somewhere that we can check?
Of course, if a person is really sick (e.g., needs multiple expensive operations)
How would he know if he needs expensive operations, unless he learns it form the medical community that his studies have already shown to have no marginal worth?
a person will not buy that unit of care simply because they can't afford it, not because it doesn't help.
How does he know that he will not be out both his money and his health? What if his son offers to mortgage his house to pay for it? Should he do that? Unscientific anecdote: I was on a team that performed a related-donor renal transplant from a super-type-A air traffic controller to his brother. The donor made no bones about the fact that he was unwilling, but had been shamed into donation by the family because he was the best allele match. In spite of the match, the recipient experienced acute graft rejection, and the organ was removed the same week it was transplanted. The donor developed a pneumothorax, then pneumonia, and was out of work for 2 months. Retrospectively, how should the various decisions have been altered, according to economics theories?
Since they have insurance that pays for most of the costs of their care, they will consume care until the marginal health value of one more unit is equal to their *private* marginal cost (i.e., their co-pay and whatever inconvenience/pain they must endure, which is "too much" care in the sense that it is more than would be provided if the services were reckoned at their "social" marginal cost, which is the full value of the resources used to provide that care, regardless of who pays for it.
Patients complain that they don't understand what their doctors tell them. Imagine you have written this statement to a patient (anyone of any educational background). What is the math-thingy probability that he will understand what you are saying? I don't.
it would be a good thing to cut way back on care.
You have a 100% convert here. Which specific things does the evidence confirm must be eliminated?
he doesn't mean don't accept your second insured life-saving brain surgery even if you couldn't have been able to afford to pay for it out of pocket.
Why did he accept the first one? Did a doctor tell him to do that? Did he know because of a margin calculation? Since he needs another operation, how does this impact on the value of the first one?
@P McCluskey;evaluating investment strategies is harder than evaluating medicine..
I'm interested in reading more about this. Could you give me the best evidence-based reference you know about it?
I did hope that they would give you some idea of how little of an answer any of us have to your questions.
That point I do understand.
The Internet is so mysterious when it comes to determining tone or emotion; I hope this set does not seem confrontational. I'm trying to learn how to make practical, intelligent decisions based on your math.
retired urologist, your puzzling "unconscionable" comment was the only thing I was directly responding to.The other paragraphs in my comment were not a response to any one person (sorry for not making that clear), but I did hope that they would give you some idea of how little of an answer any of us have to your questions. And evaluating investment strategies is harder than evaluating medicine.
It is worth adding that there would also be a moral argument in favor of foregoing some care even if it had a small positive marginal value and it was "free" in the sense that you didn't have to pay for it out of pocket. It's the same argument in favor of not abusing a rented car; it is a good thing not to waste resources even if someone else is paying the bill.
A generic traditional economist who didn't know anything about medicine or about the evidence regarding its effectiveness would probably say something like the following.
A person with no insurance would start off buying the medical services that had the greatest health benefit per dollar, and would continue to buy services of successively diminishing health value until reaching a point where the marginal value of one more dollar's worth of services provides an increment of utility no greater than the increment of utility that the person could gain by spending that dollar on their next best alternative. Of course, if a person is really sick (e.g., needs multiple expensive operations), it may be the case that there is a large marginal health value of an additional dollar of medicine even at a high number of dollars (e.g., the $200,001th dollar of care might still make the difference between living and dying). In a case like this, a person will not buy that unit of care simply because they can't afford it, not because it doesn't help. For this reason, people buy insurance against becoming ill.* Since they have insurance that pays for most of the costs of their care, they will consume care until the marginal health value of one more unit is equal to their *private* marginal cost (i.e., their co-pay and whatever inconvenience/pain they must endure, which is "too much" care in the sense that it is more than would be provided if the services were reckoned at their "social" marginal cost, which is the full value of the resources used to provide that care, regardless of who pays for it.
So our generic economist would predict that the health value of the marginal unit of medicine will be smaller than would be the case if only the "socially optimal" amount of care was provided.** And in those cases where co-pays and inconvenience plus pain costs are small, the economist would predict a marginal value of medicine that is also small. This standard stuff alone is maybe enough to sustain an argument that it would be a good thing to cut way back on care. The results that are often discussed on this blog take things a step beyond where our generic economist would go. They say that the marginal benefit of medicine may be zero, and to explain *that* you need to enter the world of biases and whatnot; it doesn't make sense for the patient to pay a cost, even a small one, for care that doesn't help at all.
So when Robin says that you shouldn't accept a unit of care that someone else is willing to pay for if you wouldn't pay for it yourself, I'm pretty sure he doesn't mean don't accept your second insured life-saving brain surgery even if you couldn't have been able to afford to pay for it out of pocket. I'm pretty sure he means that we are living in a weird world where the care that someone else is willing to pay for likely doesn't help, and things that don't help are a bad deal even if they're free.
*Today in the U.S. most health insurance is simply an employment benefit that pays for health care, and not "insurance" in the sense described above. There's some stuff to be said about that, but it's a whole 'nother issue.
**There are also distributional issues at work here: there is evidence that there is a health harm to not having insurance at all, and there are a lot of people who fear that illness will break their finances even if they do end up getting needed care. But that's another 'nother issue.
Peter McCluskey: it may surprise you to learn that I agree with what you have said, regardng the medical aggregate margin findings that Hanson and Balan have discussed. But it shouldn't, because I stated that in the opening of my comment. However, other than the word "unconscionable", your comments do not address anything I wrote, good or bad. Re-read it, taking the word "unconscionable" in the tongue-in-cheek fashion it was used, and see if you (or Dr. Hanson) can actually answer my question (an infrequent occurrence on this blog, unless the answer itself is a question). And yes, from an evidence-based stand, I do not think that Hanson or anyone else can prove that he would be willing to pay for a service unless he actually does so.
I'm sincerely trying to learn something here, from people who possess tools that I have not been taught to use. My question, restated, is that if one agrees that the aggregate is no better off seeking medical care, how does that affect one's decisions in individual circumstances?
It may also surprise you to learn that I link to your site on my blog, although I'm not aware of any evidence-based studies that show a positive margin for Bayesian speculative stock investors. In your rush to condemn my comment because I'm a doctor, you forgot to notice that I learned my current position from my your "side". Lighten up and try to help me understand how "day-to-day" life works within the depressing statistics of the aggregate.
retired urologist, what's unconscionable about health insurance? It appears to reduce one's financial risk as long as there are some treatments one is willing to pay for, and I don't see evidence that the average person's health is harmed by any additional treatment it results in.
Would it be possible to get a much better than random guess about the effectiveness of various treatments by modeling how doctors' biases are likely to affect their judgment?It's fairly clear that they are biased to overdo things that seem to show off their skills, and underdo tasks that require them to act like mindless robots. So I'd expect them to underestimate the risks of surgery (by underestimating the risk that they'll kill the patient by failing to wash their hands, overestimating how steady their hands are, etc.).Prescribing drugs typically seems to show off less skill. If the risks of the patient getting a harmful drug (due, say, to illegibly written prescriptions) is no greater than the equivalent risks in surgery, I'd expect somewhat less overoptimism here leading to drugs being better on average than surgery. Plus if my cynicism leads me to check for mistakes more than the average person does, my ability to find the information needed to detect drug errors seems greater than my ability to detect mistakes such as failing to wash hands.
@ Robin Hanson: I'm happy to admit there are probably some effective treatments
Not exactly a resounding endorsement, but one takes what one can get. You have won me over with the power of statistics versus the weakness of my anecdotal experiences (and the selectiveness of my memories). With no math skills, I'm unclear about the value of individual instances (as opposed to the aggregate margin). You have stated, "if you would not pay for medicine out of your own pocket, then don't bother to go when others offer to pay." You and Dr. Balan have opened my eyes about the wisdom of this.
How do you determine if you're *actually* willing to pay for it, unless you *actually do* pay for it? I realize it would be unconscionable for you to be in possession of health insurance (because this would not allow you to prove that you are willing to pay for the service out-of-pocket), so when your crushing chest pain occurs and you feel that you are about to lose consciousness, are you going to pay for an interventional cardiologist to *try* to save your life, or not? If yes, that must be one of the treatments you would put in the "probably some effective treatments" groups category(and perhaps you could mention some of the others). If no, would it be because you think it is ineffective, or because it's not worth the cost?
Instrumental variables estimates are famously sensitive to model assumptions. But I'm happy to admit there are probably some effective treatments, and England does seem to try hard to allocate medical resources according to studies of effectiveness.
P. McClusky:
Thank you for your reply, but you still did not mention why the broad category of Bayesian investors is a poor one for evaluation, yet the broad category of US medical care is not. It seems to me to be a goose/gander situation. Having been put in the position of having to accept that my profession has done as much harm as good (and I have accepted that), it is only fair to expect that those same economists shed light on the margin of other aspects of the US lifestyle, such as Bayesian investing or teaching economics, so that better strategies can be understood and pursued.
retired urologist, I assumed you were asking how to go about evaluating investment strategies. A detailed explanation of how I compare investment strategy evaluations to medical evaluations is not worth the effort it would take.If you're looking for strategies that are better than average (which sure looks like what you were asking for), I don't recommend evaluating broad categories such as Bayesian investors. Evaluating broad categories may give you insights into what produces poor results.
@P. McClusky
I Googled "Mark Hulburt" and could not find any references at all addressing your questioned statement: evaluating investment strategies is harder than evaluating medicine. Why didn't you just say, "that is my opinion or bias, but I know of no evidence" ?
I don't recommend evaluating broad categories such as Bayesian investors.
Yet, somehow you recommend evaluating such broad categories as doctors. Enquiring minds want to know, what is the distinction?
retired urologist, you could look at how Mark Hulburt evaluates individual people.If you want to evaluate a specific algorithm, there are problems with knowing whether it was selected to fit the specific data you're testing it on.I don't recommend evaluating broad categories such as Bayesian investors.
Retired, the point of my "generic traditional economist" bit was that even in an imaginary world with health insurance but without any of the biases and such that we are fond of talking about here on OB, and even where everyone *knows* the marginal health value of each unit of care (this was an implicit assumption that I should have made explicit), one would still predict a small (though still positive) health value of the marginal unit of care. Simply put, people use too much of things that are made artificially cheap, and health insurance makes health care artificially cheap in the sense that the user pays less for it than it costs to provide. Just like everyone eats "too much" at an all-you-can-eat buffet: each additional serving is free to you, so you keep eating until the marginal value (to you) of one more serving is just above zero, which is way below the actual cost of producing the food. In other words, plain vanilla economics gets us a good part of the way, but not all the way, to the result of a zero marginal value of medicine. (Is this part clear now? Let me know if it isn't and I'll take another crack at it)
As discussed above, all the biases and whatnot flying around the world of medicine seem to have resulted in a situation where the marginal value of medicine is zero (or at least that was so back in the RAND experiment days, this may no longer be true). As you point out, merely realizing this fact is not enough; what you really want is to know *which* treatments are effective and which ones aren't. Unfortunately, I have no special expertise in answering this question. I keep one eye open for studies like the one I linked on this page, and I like to think that the general tools that I've developed from places like OB enable me to ask better questions than I would be able to otherwise and maybe identify overconfidence or whatever when I see it.* And it is probably a good rule of thumb that if your doctor says that it is a close call between doing some procedure and not doing it, that you're better off not doing it.**
I know this is a little unsatisfying, and I'm not happy about it either. Of course there are people who know much more than I do about this, and maybe somebody has actually taken a crack at putting together a comprehensive list of treatments whose effectiveness have been well-established with high-quality evidence. If you find such a thing, please do let us know. The good news is that the recent past has seen a very large amount of growth in evidence-based medicine whose purpose is to address exactly these kinds of problems (one example linked below), and I expect that this will continue.
http://psnet.ahrq.gov/publi...
*Tiny anecdote. I have asthma, which has occasionally caused me to visit a pulmonologist, who would routinely give me a breathing test with a fancy machine. For a while that same guy was also my GP, and when I came in for a regular checkup, he wanted to give me a breathing test, despite the fact that I had had one a short while before when I came to see him as a pulmonologist. It was clear to me that this was just bill padding, and I declined it despite the fact that it was free to me and the test doesn't cause me any particular trouble or discomfort. It just seemed like a waste.
**Even this is not always right. I have a friend who is a very healthy athletic guy who hurt his knee. It was clear to him that he wanted surgery to get it taken care of once and for all, and a lot of surgeons were telling him to do therapy or whatever because surgery is invasive. He was convinced that the surgery was very low risk, and he was well willing to endure the risk and pain rather than to have a nagging, sports-limiting injury for the rest of his life.
@ D J Balan:A person with no insurance would start off buying the medical services that had the greatest health benefit per dollar, and would continue to buy services of successively diminishing health value until reaching a point where the marginal value of one more dollar's worth of services provides an increment of utility no greater than the increment of utility that the person could gain by spending that dollar on their next best alternative.
How would he "start off" knowing which services had the greatest benefit per dollar? Dr. Hanson mentions there are probably some effective treatments, but he doesn't say what they are. Is there a list somewhere that we can check?
Of course, if a person is really sick (e.g., needs multiple expensive operations)
How would he know if he needs expensive operations, unless he learns it form the medical community that his studies have already shown to have no marginal worth?
a person will not buy that unit of care simply because they can't afford it, not because it doesn't help.
How does he know that he will not be out both his money and his health? What if his son offers to mortgage his house to pay for it? Should he do that? Unscientific anecdote: I was on a team that performed a related-donor renal transplant from a super-type-A air traffic controller to his brother. The donor made no bones about the fact that he was unwilling, but had been shamed into donation by the family because he was the best allele match. In spite of the match, the recipient experienced acute graft rejection, and the organ was removed the same week it was transplanted. The donor developed a pneumothorax, then pneumonia, and was out of work for 2 months. Retrospectively, how should the various decisions have been altered, according to economics theories?
Since they have insurance that pays for most of the costs of their care, they will consume care until the marginal health value of one more unit is equal to their *private* marginal cost (i.e., their co-pay and whatever inconvenience/pain they must endure, which is "too much" care in the sense that it is more than would be provided if the services were reckoned at their "social" marginal cost, which is the full value of the resources used to provide that care, regardless of who pays for it.
Patients complain that they don't understand what their doctors tell them. Imagine you have written this statement to a patient (anyone of any educational background). What is the math-thingy probability that he will understand what you are saying? I don't.
it would be a good thing to cut way back on care.
You have a 100% convert here. Which specific things does the evidence confirm must be eliminated?
he doesn't mean don't accept your second insured life-saving brain surgery even if you couldn't have been able to afford to pay for it out of pocket.
Why did he accept the first one? Did a doctor tell him to do that? Did he know because of a margin calculation? Since he needs another operation, how does this impact on the value of the first one?
@P McCluskey;evaluating investment strategies is harder than evaluating medicine..
I'm interested in reading more about this. Could you give me the best evidence-based reference you know about it?
I did hope that they would give you some idea of how little of an answer any of us have to your questions.
That point I do understand.
The Internet is so mysterious when it comes to determining tone or emotion; I hope this set does not seem confrontational. I'm trying to learn how to make practical, intelligent decisions based on your math.
retired urologist, your puzzling "unconscionable" comment was the only thing I was directly responding to.The other paragraphs in my comment were not a response to any one person (sorry for not making that clear), but I did hope that they would give you some idea of how little of an answer any of us have to your questions. And evaluating investment strategies is harder than evaluating medicine.
It is worth adding that there would also be a moral argument in favor of foregoing some care even if it had a small positive marginal value and it was "free" in the sense that you didn't have to pay for it out of pocket. It's the same argument in favor of not abusing a rented car; it is a good thing not to waste resources even if someone else is paying the bill.
A generic traditional economist who didn't know anything about medicine or about the evidence regarding its effectiveness would probably say something like the following.
A person with no insurance would start off buying the medical services that had the greatest health benefit per dollar, and would continue to buy services of successively diminishing health value until reaching a point where the marginal value of one more dollar's worth of services provides an increment of utility no greater than the increment of utility that the person could gain by spending that dollar on their next best alternative. Of course, if a person is really sick (e.g., needs multiple expensive operations), it may be the case that there is a large marginal health value of an additional dollar of medicine even at a high number of dollars (e.g., the $200,001th dollar of care might still make the difference between living and dying). In a case like this, a person will not buy that unit of care simply because they can't afford it, not because it doesn't help. For this reason, people buy insurance against becoming ill.* Since they have insurance that pays for most of the costs of their care, they will consume care until the marginal health value of one more unit is equal to their *private* marginal cost (i.e., their co-pay and whatever inconvenience/pain they must endure, which is "too much" care in the sense that it is more than would be provided if the services were reckoned at their "social" marginal cost, which is the full value of the resources used to provide that care, regardless of who pays for it.
So our generic economist would predict that the health value of the marginal unit of medicine will be smaller than would be the case if only the "socially optimal" amount of care was provided.** And in those cases where co-pays and inconvenience plus pain costs are small, the economist would predict a marginal value of medicine that is also small. This standard stuff alone is maybe enough to sustain an argument that it would be a good thing to cut way back on care. The results that are often discussed on this blog take things a step beyond where our generic economist would go. They say that the marginal benefit of medicine may be zero, and to explain *that* you need to enter the world of biases and whatnot; it doesn't make sense for the patient to pay a cost, even a small one, for care that doesn't help at all.
So when Robin says that you shouldn't accept a unit of care that someone else is willing to pay for if you wouldn't pay for it yourself, I'm pretty sure he doesn't mean don't accept your second insured life-saving brain surgery even if you couldn't have been able to afford to pay for it out of pocket. I'm pretty sure he means that we are living in a weird world where the care that someone else is willing to pay for likely doesn't help, and things that don't help are a bad deal even if they're free.
*Today in the U.S. most health insurance is simply an employment benefit that pays for health care, and not "insurance" in the sense described above. There's some stuff to be said about that, but it's a whole 'nother issue.
**There are also distributional issues at work here: there is evidence that there is a health harm to not having insurance at all, and there are a lot of people who fear that illness will break their finances even if they do end up getting needed care. But that's another 'nother issue.
http://www.iom.edu/CMS/3809...
Peter McCluskey: it may surprise you to learn that I agree with what you have said, regardng the medical aggregate margin findings that Hanson and Balan have discussed. But it shouldn't, because I stated that in the opening of my comment. However, other than the word "unconscionable", your comments do not address anything I wrote, good or bad. Re-read it, taking the word "unconscionable" in the tongue-in-cheek fashion it was used, and see if you (or Dr. Hanson) can actually answer my question (an infrequent occurrence on this blog, unless the answer itself is a question). And yes, from an evidence-based stand, I do not think that Hanson or anyone else can prove that he would be willing to pay for a service unless he actually does so.
I'm sincerely trying to learn something here, from people who possess tools that I have not been taught to use. My question, restated, is that if one agrees that the aggregate is no better off seeking medical care, how does that affect one's decisions in individual circumstances?
It may also surprise you to learn that I link to your site on my blog, although I'm not aware of any evidence-based studies that show a positive margin for Bayesian speculative stock investors. In your rush to condemn my comment because I'm a doctor, you forgot to notice that I learned my current position from my your "side". Lighten up and try to help me understand how "day-to-day" life works within the depressing statistics of the aggregate.
retired urologist, what's unconscionable about health insurance? It appears to reduce one's financial risk as long as there are some treatments one is willing to pay for, and I don't see evidence that the average person's health is harmed by any additional treatment it results in.
Would it be possible to get a much better than random guess about the effectiveness of various treatments by modeling how doctors' biases are likely to affect their judgment?It's fairly clear that they are biased to overdo things that seem to show off their skills, and underdo tasks that require them to act like mindless robots. So I'd expect them to underestimate the risks of surgery (by underestimating the risk that they'll kill the patient by failing to wash their hands, overestimating how steady their hands are, etc.).Prescribing drugs typically seems to show off less skill. If the risks of the patient getting a harmful drug (due, say, to illegibly written prescriptions) is no greater than the equivalent risks in surgery, I'd expect somewhat less overoptimism here leading to drugs being better on average than surgery. Plus if my cynicism leads me to check for mistakes more than the average person does, my ability to find the information needed to detect drug errors seems greater than my ability to detect mistakes such as failing to wash hands.
@ Robin Hanson: I'm happy to admit there are probably some effective treatments
Not exactly a resounding endorsement, but one takes what one can get. You have won me over with the power of statistics versus the weakness of my anecdotal experiences (and the selectiveness of my memories). With no math skills, I'm unclear about the value of individual instances (as opposed to the aggregate margin). You have stated, "if you would not pay for medicine out of your own pocket, then don't bother to go when others offer to pay." You and Dr. Balan have opened my eyes about the wisdom of this.
How do you determine if you're *actually* willing to pay for it, unless you *actually do* pay for it? I realize it would be unconscionable for you to be in possession of health insurance (because this would not allow you to prove that you are willing to pay for the service out-of-pocket), so when your crushing chest pain occurs and you feel that you are about to lose consciousness, are you going to pay for an interventional cardiologist to *try* to save your life, or not? If yes, that must be one of the treatments you would put in the "probably some effective treatments" groups category(and perhaps you could mention some of the others). If no, would it be because you think it is ineffective, or because it's not worth the cost?
Instrumental variables estimates are famously sensitive to model assumptions. But I'm happy to admit there are probably some effective treatments, and England does seem to try hard to allocate medical resources according to studies of effectiveness.