Russ Roberts

Jonathan Skinner on Health Care Costs, Technology, and Rising Mortality

EconTalk Episode with Jonathan Skinner
Hosted by Russ Roberts
Yuval Levin on The Fractured R... Medicare Part P...

AdobeStock_109509551%20%283%29.jpeg Technology and innovation usually mean higher quality and lower prices. Is health care different? Jonathan Skinner of Dartmouth College talks with EconTalk host Russ Roberts about how technology and innovation affect the cost and efficacy of health care. The conversation concludes with a discussion of the rise in mortality among middle-age white males--a surprising reversal of trend--that has been linked to use of opioid painkillers.

Size:29.0 MB
Right-click or Option-click, and select "Save Link/Target As MP3.

Readings and Links related to this podcast episode

Related Readings
This week's guest: This week's focus: Additional ideas and people mentioned in this podcast episode: A few more readings and background resources: A few more EconTalk podcast episodes:


Podcast Episode Highlights
0:33Intro. [Recording date: June 9, 2016.] Russ: Jonathan Skinner... has written extensively on health care, which is our topic for today. Let's start with a paradox of health care technology. In most, maybe all areas, technology lowers costs, improves quality. But in health care, it seems to raise costs with uncertain effects on quality. What do we know about that phenomenon? Guest: That is a great question. There is a clear difference in, I think, a lot of health care technology. When, for example, health information technology came in--that is, all these electronic record-keeping to keep track of whether the physician has actually done the right thing for their patient and asked them all the right questions--great potential for improving health care outcomes. But the problem is that it takes time for the doctor to actually fill out all of these forms and click on different boxes. And so, that's actually reduced the amount of time that the doctor has to spend talking to the patient. Some institutions have actually adjusted to that by hiring scribes to follow the doctors around. And so, what you end up with is a new technology that costs a lot of money, that actually leads to a boost in health care jobs; and with uncertain effects on outcomes. So, that's a little bit of a quick paradigm of health care technology. Russ: Well, my General Practitioner sits with a laptop--not a laptop; it looks like a Magic iPad; I don't know if it's a customized tablet or whether it's just an off-the-shelf iPad, it was some software running. But he sits there dutifully acting like a scribe, noting my responses, checking off boxes, etc. It's pretty smooth. I want to go deeper into this question of technology, but in this particular case, there was a lot of enthusiasm for electronic record-keeping. What is the expected return on health care outcomes if that goes smoothly? Of course, there are mistakes. You know, you can mis-enter things. Things can be forgotten to be entered. But assuming that we really do keep good electronic records, what is the promise of that technology, in particular for outcomes? Guest: There is, actually, a tremendous potential for very good outcomes. In fact, it's hard to imagine in 20 years a health care system that wouldn't be highly dependent on their health care technology--the record-keeping. If somebody gets sick and they end up seeing a different doctor, there's really no substitute for that other physician being able to just tap on their iPad and find out all the records. It's also invaluable for monitoring what physicians do. That is, are they giving these necessarily, say, tests for diabetic patients? Are they doing the Hemoglobin A1c test, which kind of checks to make sure that their blood sugar is under control? So, I think the problem, in practice, as these things usually end up being, is the implementation of something that's sort of new and difficult to use. I think the technology is going to get better. But just like, you know, flying airplanes, that I think in the future that the electronic part of health care is going to be indispensible. Russ: Yeah. I mean, in theory, you could use smart technology, smart machines to do the monitoring you talked about, to make sure that certain procedures were done. Of course, you can also check the box and not do the procedure. That's a different kind of mis-entering. You can lie, or be negligent, or you can decide, 'Oh, I didn't think that was important,' and just check the box anyway. And we've had episodes on EconTalk where those kinds of issues arise. But in general, you'd think that that kind of collection of data would be very helpful, as you say. And right now, I had a shoulder issue; I went and got an MRI (Magnetic resonance imaging). My doctor never saw the MRI. He got the summary of the findings by the MRI person in writing about what the nature of the problem was. But I found that kind of remarkable, that that wasn't just uploaded into the Cloud. And then the other part that's remarkable about it--and we'll get into this later as well--is: Who owns that? You know, that should--arguably that should be my record; and I eagerly should share that with my doctor. Instead, it gets written onto a disk, that I happen to have forgotten that day; and he said, 'Oh, don't worry about it. I'll just look at the...'. And it's just weird that in 2016 we're still using what is essentially a paper or a physical-based system. Guest: Yes. It is remarkable. It's probably one of the few industries that still use FAX (Facsimile) machines to send things around, records, for example. And I think that the idea of putting one's data up into the Cloud where an individual patient can access it is also something that's coming along very rapidly. But the key thing, also, that I see for a computer is that, you know, there are probably 30,000 randomized trials that are done every year. And no human being can keep track of all of this new knowledge. And so, I think, computers are good at finding patterns; and they are good at sort of noticing things in people's records that can at least prompt, or at least buzz the physician to say, 'You know, the patterns that we see for this patient are consistent with such-and-such disease.' Now, it may be so uncommon a disease that, you know, the physician may have run into it in medical school 20 years ago but has never seen such a pattern before. But even though it is fairly uncommon, it does happen. And so that's, again, I think the promise of electronic prompts or nudges that can really sort of help the physician do a better job of diagnosis.
7:14Russ: The other issue, though--and this--electronic health records are just one kind of technology that is part of the process. The real issue, seems to me, and you've written about this, is the use of innovations in diagnosis; innovations in treatment; various devices--that are incredibly expensive, incredibly cool, amazing. But where the bang for the buck is not always there. So, talk about some of those--the variation across technologies and innovation. Because we think of ourselves in the United States as the most innovative health care provider in the world. And I think that's true. But it's not always worth it for people to accept. Do I have that described correctly? Guest: Exactly. Yes. My co-author, Amitabh Chandra, and I tried to think about different types of technologies, and maybe I'll just kind of lay out the way we kind of split things up here. It's very simple. One set of technologies are things where they are no-brainers. They are just obviously good. They save lives. They are marvelous. They are wonderful. And they are limited to a fairly narrow group of patients. So, for example, the anti-retrovirals as treatments for HIV (human immunodeficiency virus). If you didn't have HIV or weren't close to getting it, you would never take these drugs. So it was sort of a limited population who took it. And the results were life-saving. And those, we can all agree, even if they are expensive--for example, the treatments for Hepatitis C that have come out recently that have gotten so much attention--yeah, they are expensive, but boy do those treatments work. So, those are great. The second category are things, are treatments that are great for some people or not so great or maybe we don't really know much about, for the, for the larger group of people who might be candidates. I want to come back to that, because I think that's the most intriguing part of health care technology. And I think what worries me most with regard to future health care cost. An the third group are treatments where we really don't have good evidence that they are useful or not, but they are used anyway. And, for example, there are a variety of approaches to treating prostate cancer. And some of them cost twice or three times as much as kind of the standard open prostatectomy, but with no evidence that they are any better than the alternative. And I think those are things, again, that we can sort of look at and kind of wonder why we are paying so much money for those treatments. Some will argue for this third or latter category, 'Yeah, well, we haven't figured out how to use the technology yet, but if you give us a little more time, we can figure out really how to make this work.' And that holds, for example, for proponents of what these huge machines called proton beam, that do proton beam therapy. But let's set that aside. I want to go back to this Category 2, where, this happens quite a lot--where a randomized trial comes out, which shows huge benefit for a very specific group of patients. They are patients that are chosen to not have complications, who would be the ones who would be most likely to benefit. And the procedures are done by the very best doctors in the very best medical centers. And they find a positive benefit. And doctors read the New England Journal of Medicine, JAMA(Journal of the American Medical Association); they say, 'Hey, this thing works.' And all of a sudden you see this tremendous growth, this diffusion of this new technology. It's quite uneven across areas of the country, by the way. And there you start wondering, 'Hmmm. How much benefit are we really getting for that marginal patient?'
11:09Russ: Yeah, that's the challenge because, obviously, the protein B[?] manufacturer or the doctor who gets really good at it has a natural incentive to think it's the greatest thing since sliced bread. And we had an episode with Adam Cifu on this question of how do you reliably know what you know about what works? And of course one trial, one study often is not adequate for figuring it out. And you pointed out a couple of really interesting reasons for why that might not be true. The doctors doing it might be especially talented; or the patients receiving it might be especially prone to find it helpful. And of course then it gets used more widely. But ultimately the question is going to be--and this is, I think the biggest problem we have in American health care is, 'Well, why--as I'm sure some listeners have been thinking--why would you ever use something that doesn't work? That costs three times as much? Or why would you ever use something that might not be any better that costs three times as much?' And that gets at what's peculiar about how we adopt health care innovation as opposed to innovation in other industries. So, what do you think is going on there? Guest: Well, I think that's right. I think that the FDA (Food and Drug Administration), and in particular Medicare, who often makes these decisions of whether to pay or not, they are not in the position legally of deciding whether something is worth certifying or qualifying to make available for people to receive and have it covered under their insurance. Instead, the nature of the certification process, for example to be able to bill Medicare for providing this particular service--that nature is whether the treatment is medically effective. And for example, proton beams, it's as effective as other treatments but they are not asked to actually compare whether this one is any better. Sometimes there will be randomized trials where they don't compare the new drug against the next best drug--the opportunity cost, if you'd like. But instead they compare it against the placebo. And they show that it's effective against the placebo. And so the drug is certified for use. And so I don't think that the regulatory structure in the United States is set up to make these kinds of comparisons, to say, 'Yes, this is cost effective, and this isn't.' Russ: So, that's crazy. I mean, I try very hard to be even-keeled on this program, as the host. But right now I'm trying to fight off the urge to scream and say, 'Are you crazy? Isn't that nuts? Why would you do it that way? Is that bizarre?' And so, when we look at rising health care costs--I look at, the single factor I'm drawn to both as an economist and as an economist who likes the free market--is that out-of-pocket spending as a proportion of total spending has slowly declined in the United States over the last 50 years; and that correlates with a nice increase in how much we spend as a nation. And I think there's a relationship there that's not just correlation but causation. We can debate how important it is, how big it is; but surely when you are spending other people's money, you don't spend it as carefully. You are likely to buy things that are not as valuable as you would if you were paying for it out of your own pocket. That, I understand. But the idea then that you would say, 'Okay! Here's your menu of free stuff; and we're going to add to it some stuff that's really expensive because, you don't care, and you don't pay for it, and the doctor doesn't care.' So, what kind of a crazy, nutso system is that? Guest: I agree. The question is coming up with a system that's less crazy and nutso. I think it's possible. But I think there's a lot of debate as to whether any of the proposed reforms would improve things or not. I think it's useful--it's really useful to look at Europe. Now, when people talk about European health care, it is not a homogeneous system. I mean, the Netherlands is way more like we are than they are like England. England is kind of a classic national health service. They actually have a panel called NICE (National Institute for Health and Care Excellence)--it stands for--I can't remember what it stands for. But-- Russ: NICE? Guest: Yes. It's like National Institute for Comparative Effectiveness. Russ: Uh, huh. Guest: And they actually do the task that you are suggesting that we should do, which is, they look at each treatment and they use a cost-effectiveness ration, oh, about £30,000 per life-year. And if the drug is priced above that, oftentimes the drug company will readjust the price so they come in under the line. So that they qualify. And if the treatment is not sufficiently cost-effective, they will often say, 'No, we're not going to provide this. We are not going to pay for this.' And in the National Health Service (NHS, British healthcare system), if the NHS doesn't provide it then you are probably not going to get it except perhaps through private sources in England. And it's tremendously controversial. Because some drug will come along that doesn't quite make that guideline or the hurdle, and all of a sudden it's this problem where, 'How can you let these people die?' Russ: And of course the answer to that is: 'Because we are not going to let other people die, because we are not going to overspend on health care and we are not going to lower the quality of life.' But I certainly agree that it's equally crazy to have a committee that formally decides who should live and who should die and under what circumstances. I get that. That's not a fun committee to be on. Guest: no. Russ: But it sure beats, it probably is better--probably, it's not clear--but it's probably than saying, 'Well, if it's a little bit better, regardless of the cost, we'll have it; we'll pay for it. Yeah, pile it on the plate. We're hungry.' Guest: Yeah. Yeah. And so, sometimes what people will say is, 'Look, the U.S. government should only pay for things that are cost effective. If you want to pay for it yourself out of pocket, that's fine.' Russ: Sure. Help yourself. Guest: And that's another way to do it. If you want to pay for proton beam therapy, fine. You just--it's sometimes called reference pricing, where the insurance company or the government will pay for sort of the thing which is least cost but which is quite effective; and if you want to kind of wander off the path then you pay the differential between what the government was willing to pay for this low cost treatment versus this higher cost treatment. And that's certainly one way to go. Although, you would introduce something that, in the United States is still somewhat foreign, which is, especially in Medicare, which is: Rich people buying things that poor people can't afford. Russ: Which doesn't bother us so much when it's a Maserati. But when it's, say, anesthesia, it would. The question is: How much is it like anesthesia and how much is it like a Maserati? And with health care we tend to have an emotional reaction, it's more like anesthesia every time.
18:58Russ: Let me ask a question related to--let me give you a tangible example. I mentioned my shoulder; so, I had a cortisone shot for my shoulder. And when my doctor put that shot into my shoulder, she was watching the needle go in on a real-time imaging thing that let her see where that needle was relative to the bones and muscles. It was really extraordinary. Whereas in the old days, I guess they just--and I'm sure in some places it still happens--they have a pretty good idea of where to put it, and they put it in and hope it gets close. So, that piece of technology which let them look and see exactly where it was going, that's a beautiful thing. And it must cost a fortune. And--who decides whether that gets in that office? Does the doctor say, 'Well, we're going to raise the cost of our cortisone shot treatments because we have this technology' and will Medicare or my health plan pay for it? Who--there are other gatekeepers besides Medicare. I'm not on Medicare. So I assume it's a question then of whether my health care provider, my insurance provider, decided that's a legitimate expense. But, of course, no matter what, I've never been asked, 'Do you want to just have this done by feel, like we used to do it, and it will be cheaper?' They just assume--everybody just assumes, just give him the max. Give him the luxury. Give him the Maserati. Do you know what's happening there? Because--same question comes up with a hospital: how many MRIs (Magnetic resonance imagings) should they have? Should they adopt the new, improved PET scan (positron emission tomography scan), whatever it is. I'm sure it gets better every year. Who is making those decisions? Guest: That is an excellent question, and I think as you might expect the decisions are made but not probably in a rational or thoughtful way, necessarily. In the case of your imaging machine to make sure that the shot goes where it should, I'm like totally in favor of that. I think, in fact, the question should be asked about, how often do they give shots, for example, sometimes for back pain? They'll give these epidurals, like women get at childbirth, only they do it to kind of relax your back. So far, there's no evidence that this works. But it's not hard to get one. You've got somebody in--you're a doctor, you've got somebody in pain; they are suffering, and they say, 'Look, Doc, you've got to do something for me.' And sometimes, you know, you get some relief for a little while; maybe it goes away. But the physician goes into the business in order to help their patients and not to, like, protect the budget. So I think those kinds of questions are hard. I'm sort of willing to pay for the best, for very good technology, assuming it really does yield better outcomes-- [?]-- Russ: But doesn't it depend on the cost? Supposing it's a billion dollars? Guest: Well, a billion dollars, yeah, no, no, no. I agree. But this is electronics. It's like once you buy one of these things the marginal cost of using it is pretty low. And you've got to think that after a while, like for you know, ball point pens, that the price is going to come down and it's going to be just sort of off-the-shelf thing. At least we hope it is. Russ: You'd think that. But that goes against the question we started with. Which is, that prices don't come down. And the reason, I think, is obvious: It's because--they come down everywhere else because people have an incentive to spend their money carefully. Here, they just make a better one, that's more expensive. There's no incentive to make a cheaper one. That's the problem. Guest: Yeah. Except at some point you can make them that--no, that's right. There is actually quite a bit of debate about like some of this 3-D imaging for breast cancer, for mammography. And now it's that they are actually getting to be so accurate that there's a lot of concern that they are picking up too much. That they are picking up benign, little tiny abnormalities that are not predictive at all of breast cancer; and they are identifying people as being sick. And my colleague, Gil Welch, has written extensively about what he calls over-diagnosis: that is, you walk in, in the morning, and you feel like you are a healthy, productive member of society. And you walk out of the doctor's office--you feel fine, but you've just been told you have this illness or this disease which in fact may or may not lead to something more serious in the long term. Russ: We've done a bunch of episodes on this problem, and it's a huge--it's fascinating. Because, I probably have prostate cancer, right now; and yet here I am and healthfully and only in my own mind moving along with my life. But that prostate cancer might not kill me for 50 years; and something else might get me in the meanwhile. Guest: That's right. Russ: I really don't want to spend money finding out that I have that inside me. In fact, I think every man who lives to be--if a man could live to be 100 he'd have prostate cancer, often benign and often slow-growing, and so it's not relevant. We don't like to know that necessarily. But the problem is that if we diagnose it and then we feel we have to do something, it can often be worse than being in ignorance. Guest: Right. Once you get the PSA (Prostate-specific antigen) test, and if the PSA test is sufficiently high, then you are just going to worry about it all the time. Gil Welch has this very nice way of explaining screening for these kinds of diseases: we want to set up a fence to keep rabbits inside--and that's kind of the idea of like screening for disease so we can find that cancer and contain it and keep it within. But, either you have a tortoise, in which case you don't need to build a gate anyway because it's not going to go anywhere; or you have birds, in which case you put up the fence and the birds just fly off anyway. And so, to the extent that you have diseases that are characteristic of either tortoises or birds, along with rabbits, then screening for the disease is going to be limited in effectiveness. Which doesn't mean you shouldn't do it. In many cases of course you should. Russ: Of course. Guest: But you don't get the benefits that you might think you would from a widespread screening program. Russ: Yeah. We don't give medical advice on this program. But when people tell me, 'I have to get test A,' whatever test A might be, and I say, 'Well, I don't think so,' they look at me like I'm crazy. Or, the alternative I hear is, a friend says, 'This surgery saved my life,' and I always say, 'How do you know?' To myself. I don't-- Guest: Yeah. What's a counterfactual? Russ: Yeah. Exactly. Because you don't know. But there's an emotional, psychological part to this, of course.
26:14Russ: Let me ask you about--do you know anything about Lasik? There's this folklore among free market folks like myself that Lasik, the eye surgery, has fallen over time, probably because it's not covered by insurance and people are paying out of their own pocket; and the incentives which normally produce good results with technology in other areas apply there, whereas they don't in other parts of the medical area. Do you know anything about that? Guest: I can't speak with authority; but I have heard that there are places, in particular in Canada, where that's all they do and they specialize in that. And they get very good outcomes. But I haven't kept up with the trends. Russ: Yeah, but my understanding is it's gotten cheaper over time and more reliable. Guest: Yeah. Russ: Which is not true of anything else that the health care system--or very few things that the health care system touches. Guest: I--but, let me go back to the example of Lasiks. That's something where the diagnosis is clear. And the treatment is well-specified. Like, there's not a lot of question about whether you should go ahead and do something about it. If the patient wants to, they are well equipped to decide. And so I think that's kind of a special case. When people ask me, 'But isn't it true that health insurance has really contributed to health cost growth?' yeah, that's certainly true. You need to provide some funds in order for a provider, health care providers, to do what they do. But then you look at pets' health care. Russ: That's another good example. Guest: And Medicare, Part P hasn't yet been implemented for pets. Russ: Any day now. Guest: Any day. Any day. Russ: Could be in February. Guest: The next Administration. Yeah. Russ: Depends who it is, but it could be coming. Guest: And we will get all the cat people on board with Medicare Part P. But it's been remarkable how much growth there has been in health care spending on pets. And that's why I often think of technology growth, the fact that when you bring your pet in, the doctor will say, 'Well, we can put in, implement a medical device into the pet, into this dog,' and you are in the funny position where, if you say, 'No'--as an economist you might say, 'Well, my dog is 12 years old, it's like we've had a good life,' and your kids are looking at you like you're Simon Legree. So, it's a tough one. Russ: But that's true of everything. It's like when--to take out the emotional part, which is perhaps not appropriate but I think it's still illuminating: We go shopping for a car and my sons want a Porsche, say, and I want a Honda Accord. And they say, 'Aw, Dad, we really want a Porsche. It's going to be so cool.' And I say it's just not worth it. And there we say, that's life. And yet, the Honda Accord today is such a better car than it was 25 years ago. Ten years ago. It's an amazingly great car--as are all of it's competitors in the class. Because to be a competitor, they have to pretty much match the features. And the reliability. And the gadgets and the gizmos and the gas mileage, etc. So, the real question is, when we say that health care expenditure on pets is rising: Sure. We're a rich nation. I think, despite the data, I think we're a richer nation very much across the board than we were 20, 30, 40 years ago. And I would expect that we'd probably want to make that decision to save Lassie, even though she's 14, or 12, more often than we did in 1940 when we said, take the dog out back and kill it. And that's great that we're more able and willing to do that, because we have the resources to do it. The question is, does competition among vets help reduce the rise in that demand [cost--Econlib Ed.] by introducing innovation and other cost-saving things? Now, one of the reasons it's going to get more expensive is that it's a labor-intensive activity. And all labor intensive activities have gotten more expensive. But if there are gadgets used to work with dogs and cats, I wonder if they have got--and that don't work on humans--because that contaminates--there's a cross-effect there--do those get cheaper over time? And easier? Maybe it's not even an answerable question. I don't know. Guest: That's a great question. And again, I'm not an expert in the pet health care industry, but I'm thinking that might be a good place to go into, if--you know, I do have tenure so people can laugh at me-- Russ: No, I think it's interesting-- Guest: But I think it's an alternative world where there's no such thing as health insurance--and to kind of trace that out would be really interesting. Russ: I think it's understudied.
31:37Russ: Let's move to something that's heavily studied; you've worked a lot on it. Which is, a topic which has gotten a lot of attention, which is variation in both price and usage of various medical techniques geographically. So, it's surprising--give us an idea of what the range is--it's surprising how much variation there is, again, compared to, say, other markets where prices tend to be somewhat similar. Guest: Well, yeah. I should start out with--the person who really got me interested in this topic, and lots of people interested in this topic, was a real pioneer, a guy named Jack Wennberg who was driving around in Vermont in the 1960s. And he was worried that there were communities in Vermont that were being underserved because they didn't have access to hospitals, or to doctors. And so he started actually collecting all of this data from all of these towns in Vermont, and found remarkable variation across little towns. In fact, in the community where he lived, divided by a street. So, in one school district none of the kids had their tonsils. And across the street, in the other district, most of the kids had their tonsils. And it was because the doctor in one school district believed in taking tonsils out, and the doctor in the other school district thought that was kind of a dumb idea. So, he compiled all of this data, and he sent the paper to all of the major medical journals; and they all rejected the paper. And he finally, one of his friends said, 'Why don't you send it to Science?'--which is the highest prestige journal in the business? And they took it. And they published it in 1973. And ever since then we've been sort of following on this article in Science. And more recently Zack Cooper and his associates have started looking not just at differences in utilization across regions--that is, how many back surgeries per thousand or how many doctor visits or admissions--but has also looked at variation in prices in the under-65 population. And, there is way more variation across regions than even I thought possible. First of all, there's a lot of variation in the Medicare population. And, remember that Medicare is a Federal insurance program, so prices are pretty much fixed across regions. There's some adjustments for--you know, New York is more expensive than Oklahoma. But what Zack and his colleagues found was that there was just as much variation and utilization in the under-65 private insurance programs. But there was even more price variation--that is, prices in these small communities that were actually doing pretty well with Medicare spending, were astronomical compared to more competitive regions that had lower prices. So, my takeaway is like, everywhere you look, you see variation. Now, I should say that it's not just health care. That, you know, people in Philadelphia are more likely to consume Philadelphia cheese steaks; people in California are going to eat salads. And I remember looking at this map of, like, chicken consumption. And it showed just as much variation across the country in chicken consumption. So-- Russ: But not in price. Not in price. I guarantee you. I'd bet large sums of money. Guest: Well, exactly. Yeah. There is not that kind of variation there in price. I think that people are beginning to realize that health care is often a natural monopoly. There is one dominant hospital in the area. Where I live, there's Dartmouth-Hitchcock Hospital. And-- Russ: You wouldn't be caught dead going anywhere else. Sorry, I couldn't help myself there. Guest: Absolutely. But it's also a long drive to go anyplace else. And so those hospitals, you know, not surprisingly exploit their natural monopoly. Because it's easier to do that than it is to figure out how to cut costs. Russ: But to come back to my earlier example, if I need an MRI for my shoulder and I called 5 MRIs within 10 miles--and there's a lot here; I'm in suburban Maryland, and for better or for worse, there are quite a few. One could call that a great thing. I'm not sure. But there are quite a few. So, I call around--they are not close to the same. They are all over the place in what they charge. And it's obvious to me why, and that it's a bad thing--and it's because nobody's paying attention. Well, that's not fair. I'm not paying attention, for sure. The only entity paying attention is going to be my insurance company. They have some range, I assume. But they don't dictate the price--unlike Medicare. And yet, the incentive for a provider of any service, whether it's back surgery or MRI or whatever it is, it's just not like any other market. There's no market forces--there's limited market forces working there. Guest: Oh, I agree. I think insurance companies are trying to figure this out. In other words, some are trying to introduce this reference pricing we talked about before, where you would--the insurance company would sort of find a legitimate high-quality provider that charges not very much and say, 'This is how much we'll reimburse you for your colonoscopy.' If you want to go someplace else, that's fine but you'll have to pay for it. That kind of policy leads to very rapid response from the companies, from the hospitals providing colonoscopy. But I want to give you--I'll give you a personal example about this issue of MRI pricing. I got a phone call--I was scheduled for an MRI and I got a phone call from my insurance company and they said, 'You know, there's this other place over here which is fairly close, and if you go there, you'll save on your co-payment; and you can save like maybe $100, $150.' And so I checked with my doctor--and this was a fairly involved MRI--and the doctor said, 'Well, first of all, that's not quite as good an MRI. Second of all, we don't get the results as quickly, because inside the hospital you have electronic records. And so the only risk is that you might have to come back and do it again.' Russ: It's horrifying. Guest: And so, which--right. Which is not much fun. And furthermore, the insurance company was only promising me a reduction in the co-pay. It was not a highly incentivized program, because they would get 70% of the gain or 80% of the gain and I'd get 20% of the gain. So, it wasn't clear even then that they had sort of figured out that if they are really serious about getting people to go that they would use a more high-powered incentive. But I think that's where people are going now, that's where insurance companies are going. Because that seems like an obvious one to do. Russ: I just think it's so bizarre. Guest: Yeah. Russ: It's just business as usual. I mean, no one even thinks twice about this, the way it is. The idea that your insurance company would make a suggestion to help you save a few bucks--in this case maybe it wasn't worth it--but the idea that they would call you and notify you about it is comical, in a tragicomic kind of way. Guest: Well, it's a start. It's a start. Russ: It's true.
39:46Russ: Let's shift gears. I want to talk about something that was in the news recently, and you've written some on it, which is, a recent study, Angus Deaton and Anne Case, about rising mortality rates among, I think it was middle aged white men. And that's shocking because mortality falls steadily for all groups, year in and year out in the United States over the last 50 years, maybe longer; and it's one thing to say it's not falling as quickly for some groups, which of course can be--by random variation it's not going to fall the same for every group--but that it's rising for one group is kind of shocking. And it's been speculated--I don't know how reliable this speculation is--that this is due to an increase in prescribed opioids for pain that have either been abused or misused. So, what do we know about that? Guest: So, that is a very remarkable finding. Recently, the last few days, a new result came out. Some preliminary data from the CDC (Center for Disease Control) suggests not just that mortality is rising for these white, non-Hispanic, kind of older middle aged folks, but that mortality is rising for the entire United States in 2016. Russ: Yeah, I saw that. Well, that's in the aggregate. But is it rising--which could be because just for that one group, the rise is large enough to offset the declines elsewhere. Guest: Well, the rise actually wasn't that big. So, there was a small increase in mortality, and the increase itself could be explained largely by opioid deaths. The real puzzle is why didn't mortality drop for this older group as rapidly as mortality in other groups? But the fact that the aggregates are going up--and this is in combination with the elderly, who have always been reliably living longer every year--and of course a lot of the mortality rate is driven by older people, because unfortunately they are the ones who are more likely to die--that was a shocker to me, because you'd think that there would be certain subgroups of the population that are doing worse. For example, again, whites in the 25-34 group are also particularly badly hit by this opioid epidemic. But to find that overall mortality rates are coming back up--that is a surprise. Now, to me that suggests that we may be devoting our resources rather than--perhaps in the wrong place. Russ: Well, we know that's true, John. [?] we've been talking about that. Guest: By definition. Russ: That's a given. The question is: we've overcome that before. Guest: Right. But that we're spending lots and lots of money for these new treatments: personalized medicine is coming around the corner and all of these wonderful drugs, $100,000 drug regimens that extend lifespan by a median of 6 months. And we're kind of totally missing the boat on all of these folks that are having serious problems with opioids. And the deaths suggest the tip of the iceberg. I think it is true that there are more and more people who are on opioids. And are genuinely in pain. And it is a--it's not necessarily addiction, but it is true that once people start taking opioids, that to go off of opioids is really, really difficult. Because if they were in pain initially, kind of the dorsal horn is opened up and sending these pain impulses directly to the brain, and so that when you stop taking opioids, you are in real trouble. And so I think that is certainly contributing to it. And also just the falling apart of jobs and community structure. And again, not just jobs of individuals, but all of the kind of public goods that used to be provided by communities that had vibrant factories and manufacturing. That all seems to be gone. The evidence seems to be that most of the mortality increases are coming in rural areas, and less in cities. Which I thought was quite interesting.
44:51Russ: So, let's try to go a little deeper into this, and give us a little more background. So, opioids are painkillers, right? Guest: Yes. Russ: And do you know the names of some of these, that people might be taking that are leading to problems? Guest: Well, oxycodone is a pretty popular one, for people who become addicted. There is a new kind of, a newer drug called Suboxone which is useful for sort of maintenance of drug addiction. So, that's useful for example for getting people from heroin off. But then they are still on maintenance drugs. There is also methadone. And I think a lot of the concern is people start with, say, OxyContin or some kind of standard Percocets, some kind of standard drug; and you know they are really expensive, especially if you don't have a prescription, if you buy them on the black market they are very expensive. And so people find that heroin is actually cheaper. Russ: Do we know--how much of this is known? In the following sense: So, obviously there are death certificates that will give a reason for death. Is the reason that are given in these cases--has it been established that this particular group has a rising rate of death from drug overdose? or is it drug interaction? Do we know anything about that? Are we just guessing? Guest: The death certificates do make an effort to find what the causality every of death was. So, we do have measures. The general classification that case [?] is poisoning--death by poison. And that is rising very rapidly. But there are other problems as well that are kind of related to drug abuse, which aren't quite so obvious. So, you still may have cardiovascular reactions. Or car accidents. Or, certainly cirrhosis of the liver is something that has been on the upswing as well--from drinking. Russ: But are we saying here that people in rural communities are depressed because they can't find work--which I find to be a strange argument in a time when the national unemployment rate is about 5%. I know that labor force participation is down, so employment is not great. But we are not like--it's not the Great Depression with unemployment of 25%. Yes, it's depressing to not have work; and yes there are pockets of the country where it's much worse than that, than 5%, obviously. But it's hard to understand how that can create a national rise in the death rate of a very large part of the population. I understand that, if you are a rural coal miner and coal mine jobs are gone and you don't want to move, that life is hard, and you might turn to overuse through black market purchases? Or are you getting extra prescriptions from your friendly doctor? We don't really have a good idea of what's going on here, do we? Guest: I think that the one thing we can say with certainty is we don't have a good idea of what's going on. Russ: We're guessing to some extent. Guest: And that's why--I hesitate a little bit. Ellen Meara and I wrote a commentary on this, on the Case-Deaton paper where we did some speculation; and one possibility is that there's just a sense of futility that, especially for white non-Hispanics, that their parents did better than the grandparents and the great-grandparents and all of a sudden they kind of hit the wall and they are stuck with at best jobs that pay, you know, $9 or $10 an hour and there's no advancement up. That they may be working, but they are not working in a way that allows them to sort of support their family. I don't want to become a pop-sociologist here. So I won't speculate any further. Russ: But just to clarify the mechanism: Then what? I mean, I find that possible. But I'm a little bit surprised that it can work at the national level for that size of a group. But let's say that's possible. I don't really believe it, actually; but I'll put my disbelief aside. What do I do now? I'm depressed; life is futile; I can't get a good raise; I'm not living as well as my parents. And so what do I do? I take an overdose? Is this a suicide story, or is it an abuse story? Guest: No, no suicide. But at some point-- Russ: Numb myself? Guest: No. Your goal if you have a dead end job with a jerky boss, your goal is, or one way out is Social Security Disability Insurance (SSDI). That is, you are working someplace; you start having horrible back pain, which, by the way, is related a lot to stress. You have mental illness issues--you qualify for SSDI. That allows you to go to see the doctor. We've found evidence of just remarkable rates, levels of use--maybe at least a quarter or in some areas far more than that--of people who are on SSDI and are taking a lot of opioids. And there, it becomes more plausible that these are unhappy people who aren't looking after themselves. And they can die of other things, as well. But they are not, kind of, locked in, working hard. They don't show up on the unemployment statistics, but they are kind of there. And so that's a group that I really am concerned about. Russ: Is there any question--I just quickly googled--these are white non-Hispanic males, 45-54, in the prime of their working productivity career, etc., that have shown--it's actually peak earnings; and again I understand the peak might not be as high as it once was. But generally people in that age group experience rising wages up till then at least; start to decline a little after that. But just to clarify: Are they then--when a death is characterized as 'death by poisoning,' is it that they took much? Or they just died from the side-effects? I'm just trying to figure out what the basic phenomenon is here. Guest: So, if it's 'death by poisoning' it means that they overdosed. So, they--they'll have a--they'll have 30 days or some supply of opioids, and often what will happen is it will be mixed. So they, maybe a friend of theirs shares some Suboxone with them. And they take something else; and maybe they have a few drinks. And before you know it, the sort of interactive effects can be very, very dangerous. It's not the kind of thing you realize until it's too late. Russ: And is there a phenomenon of the drugs that you need a larger dose to have the same pain impact for people? Guest: That's true for, like, Percocets. Yeah. That there is an increasing--that people become opioid-resistant. And in fact hospitals are very much aware of that, when patients come in and they are in pain, they kind of sometimes kind of ask, 'Well, have you been taking opioids before?' because the difference they might give you 10mg pills versus if you haven't taken opioids before; but they might give you 60mg pills if you have. So, yeah, you do become more resistant. And of course that's a rising price for maintaining your opioid addiction. And that is what, that's where the problem comes: these kids in rural towns can't afford to buy Percocet from the black market, and so they turn to crime or they turn to heroin or they do both. Russ: Well, that's a very depressing story. The other thing I wonder about--I wonder if these data are accurate. But that's always a question--you know, sometimes when I see a trend reversing like this I wonder if they just changed the way that they defined death by poisoning. But I assume that's not the issue. Guest: Nah, I don't think. These are total deaths. In other words, they can break it down. But most of the puzzle was not so much that deaths by poisoning were rising but that other groups--like cardiovascular disease kind of stopped falling as much for this group, compared to, say, African Americans or Hispanics or people in other countries. Russ: That's fascinating, in a tragic, tragic way. I guess it's an issue, I'm sure we'll come back to as scholars and as EconTalk, because it's so unusual. Guest: Yeah.
54:36Russ: Let's talk about digital medicine. We've had Eric Topol on the program before talking about all the changes that are coming, and he's a big proponent of something we've talked about before, which is owning your own data and being in charge of your own data, being in control of your own data, and eventually of your treatment. Which, digitization or a lot of the applications that are coming or at least talked about have the promise of doing. What do you think the potential of that is, and regulatory issues do you think are going to have to be resolved or pushed aside, a la Uber (, before some of these technologies can really have an impact? Guest: Well, I think in terms of ownership of your ownership of your own data, I think that seems like a no-brainer. It's like, yeah, you should. It's your body. You should have access to your own data. In fact, if you want to analyze it, that's great. I think that this idea of owning your--you know, getting your DNA (deoxyribonucleic acid) sequence done, I think that's getting cheaper and cheaper. And people may want to do that. I'm not sure they are going to learn that much from their DNA unless they have some very specific DNA markers which would cause them to change, you know, to get certain treatments. My impression of the DNA revolution is that it's been much less of a payoff for medical purposes than people thought. And, I think in large part because, you know, the DNA may be the template but what's really important for health is what proteins are generated by those DNA. And even identical twins produce very different proteins--the so-called epigenetics field. And I think that's where the science is going. But we're not very far along that path, yet. Russ: Do you think this digital revolution is being oversold in terms of what its potential is for, being, say, doctor-free, that you just, I put my iPhone next to my body somewhere and it will send a message to the cloud that I've got such-and-such a thing, and it will be diagnosed by artificial intelligence and the drug will be dispensed by my 3-D printer, and all that? Do you think that's a real vision that could happen? Guest: Um, yes, I think it could happen. But I think most humans still want to talk to a human about the different options and what it means, and what it means to their life. I don't discount the possibility that in 20 or 30 years we develop enough artificial intelligence that we can put doctors out of business. But, if that's the case then the economists will be out of business, too, because, as we know, all we need is a parrot to say 'supply and demand' and we've basically covered most of our profession. So, I do think that that's a possibility. But I think, if I hear the two words 'Big Data' any more, I will like, throw up on my shoes. But this is often the case. A new technology comes along and everybody becomes enamored of it. And over time, they gradually say, 'Well, you know, it does provide some value, but maybe it's not the value we thought it would be.' Russ: Why are you tired of hearing about Big Data? Isn't it going to be amazing, when all our health care records are up in the Cloud and we can find out what works and what doesn't work? Treatment is going to get better; medicine is going to get customized. Blah, blah, blah, blah, blah. You think that's just blah, blah, blah blah? Guest: Yeah. I mean, no, there are some value. But I've done some of this analysis--one of the wonderful things about Dartmouth (Dartmouth College) is that we have terabytes and terabytes of Medicare Claims Data. And so we can put the computer to work to sort of find out interesting associations between, for example, interaction of drugs leading to worse outcomes. And so, what we've often found is that, well, somebody's taking this drug and that drug, that they're more likely to, you know, have a bone fracture. But, there's a reason when you actually start looking at the data and talking to our expert clinicians: they say, 'Well, yeah, usually somebody who is ferile[virile?] and elderly will be prescribed this drug and that drug.' And so-- Russ: Uh, uh, yeah. Guest: And so, it's not causal. It's like we are just finding that, yeah, that people who have these particular problems will be prescribed the two drugs. And they also happen to get, you know, and it's because they are likely to fall that they are prescribed these two drugs. So, that said--I mean, we have had some pretty interesting opportunities to find relationships that people didn't expect in the data. But it's still very much an art. Russ: I'm just going to mention that for my shoulder, I went and got physical therapy, and was discouraged by the progress that was being made, and stopped. And my shoulder has gotten a lot better since then. And so, if I'd continued, I would have said, 'What a great physical therapist I have!' But now I've learned that the key for shoulder pain is to get physical therapy and then stop. The key question is: When do you stop? Yeah. It's--causation is tricky. Obviously. Guest: It's--randomized trials, for all of their faults, are very good at answering questions like that.
1:00:34Russ: Yeah. Well, we're almost out of time, or we are out of time. Let's close with the following: I'm going to give you a really unpleasant question, which is: If you were Health Care Czar, what would you do? See, if I were Health Care Czar, I'd just dismantle most of the things that we've tried to mantle to make things better, that I feel have struggled to make things better. And I'd let a thousand flowers bloom; and I'd let private charity and civil society take care of people who couldn't afford the health care treatments that are going to be a lot cheaper and [?] not artificially pushing up demand and causing prices to rise. So, I have a Nirvana in my mind. What's yours? Guest: Well, I'd start with Medicare for Pets. But I think that won't go far. But seriously, I would start with something a little bit less ambitious, but I think that would yield tremendous benefit: and that is, a central processing institution for all insurance claims. So, think of the Fed Reserve is the central institution that processes checks. When you write a check, it goes to the Federal Reserve and the banks get credited with the money. That, the only way that insurance companies can reimburse you for a fee is if you fill out a common form. That would yield two benefits. First, it would save billions and billions of dollars in administrative costs. And second, is that it would provide instant monitoring and a picture of both regional variations and where is spending going, and are we ramping up too quickly on this treatment or that treatment? I think that would be a no-brainer in the sense that everybody should be in favor of it. And that it would save money and improve health care.

Comments and Sharing

TWITTER: Follow Russ Roberts @EconTalker

COMMENTS (28 to date)
Szymon Moldenhawer writes:

As a clinician working in geriatrics for the last 30 years -- I live with these medicare induced pathologies everyday but the problem of declining rate of progress in medcine is more complex, I wish Russ would invite Dr James Le Fanu author of "The Rise and Fall of Modern Medicine" he would provide additional and bigger piece to puzzle of declining progress in medical field. Dr Le Fanu opus is very Hayekian in nature.

John Garlitz writes:

Electronic Health Record software technologies and firms are limited in number, very expensive, not standard, and not portable.

McKesson recently created a joint venture with a private equity firm to sell or spin off their EHR technology. Signally a potential reduction of the number of available EHR technologies.

The most effective EHR technology is only sold to large health systems. A consequence of this market constraint is that small and medium-sized health systems give up their independence through affiliations, acquisitions, and mergers.

Imagine how the different the world would be if prior to information technology standardization and lowering of IT cost, the federal government required all organizations to immediately implement IT across all operations 'for the good of the people'.

A previous Econtalk episode on wages, non-portability of skills, and lack of standardization during the infancy of the industrial age comes to mind.

Same with Russ's favorite F.A. Hayek quote.

Craig writes:

Planet money had an episode about insurance companies who more aggressively pay patients to go with a cheaper provider.

Trent writes:

An interesting discussion, but I was surprised at the end of the podcast when Prof. Skinner said his first step to improve the system would be to create "a central processing institution for all insurance claims."

Given all the potential benefits he listed, why doesn't this institution exist - are there too many state regulatory agencies preventing this from emerging? Are there too many free rider problems where all the firms are waiting for others to invest the resources & they'll join in on the back end? Are these entities so used to only complying with federal mandates/regulators that they're not even trying to innovate - e.g. waiting until the government creates the institution/imposes it on them to react & do something? Or something else?

thomas egan md writes:

1. no mention of 10-20% increase in costs due to defensive medicine-fear of lawsuits.
2. no mention of huge increase in costs due to gov., insurance bureaucracy and layers of middle persons between MD and patient.
3.When patients are treated well, they live longer, develop new diseases, and consume more medical resources. What economic model do you use to address this?
4.the cohort of men 45-54 with increased mortality rate are not dying of "poisoning". They are self medicating their depression and killing themselves. Why? Maybe because-They are told from an early age that maleness and whiteness are bad things. They are discriminated against in jobs and college and grad school admissions and endlessly humiliated on TV shows. There are no college Men's study departments. No Men's health initiatives at the local hospital, no parades and road races against prostate cancer.
5. Hand raised: Dr. Roberts, I really do know the answer to these tough health care cost questions. Simply hire an MIT professor to fix it...say...Dr. Gruber!!!

Bob Mounger writes:

Neil Gabler wrote a piece in the Atlantic recently about how he couldn't come up with $400. If you have a $10000/annum Obamacare premium and a $15000/annum deductible you may be unable to afford medical care even though you are technically insured. Plus is there a correlation between the early dying men and their divorce status?

Madeleine writes:

Very glad opioid addiction was touched on. I remember getting knee surgery when I was 16 and having them prescribe me three different types of opioids. I am actually allergic and ended up taking ibuprofen. I've had to have two different surgeries since and have taken nothing but ibuprofen for both of them. Sure, it sucked and I was uncomfortable, but if I had to do it over and didn't have the allergy, I think I'd still take the OTC drugs and pain over the risk of addiction.

I really think a lot of people who ended up addicted just didn't have all the information. I certainly didn't when I was 16: if I hadn't broken out in hives and gotten so sick, I would have diligently swallowed the pills along wth my antibiotics, no questions asked.

I'm not saying no one needs these pills, but...yikes.

Ajit writes:

On the topic of opioids I don't think there was a sufficient causal mechanism that was explained. How does being laid off from work and being unemployed for a persistent amount of time naturally lead to opioid addiction when there was none previously? They didn't start as habitual addicts.

And as with the autor podcast, what exactly is the right policy prescription? Whether it be benign or more forceful paternalism, when and where has that ever worked?

rovingbroker writes:

Russ said, "she was watching the needle go in on a real-time imaging thing that let her see where that needle was relative to the bones and muscles. It was really extraordinary. Whereas in the old days, I guess they just--and I'm sure in some places it still happens--they have a pretty good idea of where to put it, and they put it in and hope it gets close. So, that piece of technology which let them look and see exactly where it was going, that's a beautiful thing. And it must cost a fortune."

That was likely a fluoroscope. Been around for ever.

You might find this interesting ...

Glenohumeral Joint Injection With Fluoroscopy
Test the shoulder afterward: Have the patient do Codman shoulder circles (bent-over shoulder movements without gravity) to spread the steroid/local mixture. Then have them test the shoulder against gravity to give a nice “awww” effect from the happy patient.

j mitchell writes:

Nice discussion. I am a practicing oncologist/palliative care physician and enjoyed this podcast. The one problem with the central repository for billing and a "standard" form is that it does not address one of the current enormous headaches in billing - recertification.

Most patients do not understand how this works. The doctor suggests a test or treatment (most drugs now require precertification). The doctor must then submit to the insurance company the forms (typically physician notes, labs, path reports and a form that the insurance company generates) and the insurance company reviews them before offering a determination if they feel the suggested test or treatment is acceptable. In my state they have by law 14 days to answer. If they deny it you can appeal by the physician calling one of their physicians (about 20 min of physician time for this call). This process occurs for tests/treatments that can occur in other facilities. This means the office pays for all of this infrastructure for this process and gets no reimbursement - the profitable test is done elsewhere and that entity (usually a hospital system) gets the profit. And if the insurance company does not pay, the insurance company and the entity that did the test blame the physician's office for "not submitting the right forms."

I have an illustrative anecdote. Recently a company denied an immunotherapy for a patient of mine and I had no problem with the denial - late stage disease and limited evidence of efficacy but the patient wanted to "try something" for their terminal cancer. I asked the insurance company's physician to call the patient and tell her that he was denying the treatment. He said "that's not how it is done, you have to tell the patient. We don't talk to them." I knew that was the answer because that is standard practice. But what a position to put the treating doc in. At least the UK committee NICE is upfront and you know the rules. Here the rules are fungible, post hoc and the treating doc becomes the scapegoat. This system has broken the doctor-patient relationship. Shared decision making is with the patient's insurance company and not with the doctor.

GregS writes:

Some things about those opioid poisoning deaths.

There are ~18k prescription opioid deaths in 2014, and ~10k heroin deaths, about 27k total (if you avoid double-counting). With 2.6 million deaths a year in the US, I’m having a hard time believing this one thing is a driver of overall mortality. Maybe for a young demographic, but not overall.

Please take a look at SAMHSA’s Results from the 2014 National Survey on Drug use and Health report.
See page 7: illicit use of painkillers is flat. And see page 26: substance abuse disorders involving painkillers are pretty much flat (maybe up from 0.6% to 0.7% of the population, if you believe these numbers). You see a similar trend (or non-trend) in the Monitoring the Future survey, which focuses on 8th, 10th, and 12th graders. Two separate surveys are failing to find any additional illicit use of prescription opioids, so I think the “epidemic of abuse” story is very problematic. Apparently (assuming the survey data are reliable) you can triple the number of opioid prescriptions (roughly what’s happened over the last 15 years) and not get any more addicts.

Russ asks if these are drug overdoses or drug interactions. They are overwhelmingly multi-drug interactions. With prescription opioids and even heroin, only ~30% of these drug poisonings only involve one substance. (The CDC records often list multiple substances, as they can list up to 20 causes of death on the death record.) For benzodiazepines, only about 1-2% of those poisonings *only* involve one substance. This is an under told story, and you hit right on it, Russ. We could save a lot of people if we just told them not use combine drugs, as the majority of these are multi-drug interactions. Opioids react badly with benzodiazepines and alcohol, which in combination can suppress respiration more than any of these alone. I’m not sure if these are people with serious drug problems just swallowing whatever pills they can find or if they are normal patients whose incautious doctors prescribe them a bad mix. I would think that a bigger population of problem users would show up in the drug use surveys, but that’s not what we see. There’s another kind of interaction: a normal dose of drugs interacting with an unhealthy individual. If you look at these death records, you see so many cases of “obesity”, “sleep apnea”, “asthma”, etc. Also a lot of heart conditions. It’s not always clear whether these are conditions caused by the drug use or preexisting conditions that interacted badly with drug use which wouldn’t have killed a healthy person. It seems a little misleading to lump these all together and call them drug overdoses. Here is an analysis I did on 2014’s CDC death records:

Also, the diagnosis of an overdose death is not as clear as your guest seems to imply. Pick up Steven Karch’s textbook Pathology of Drug Abuse; it talks a great deal about how hard it is to diagnose an opioid overdose. You can’t go by toxicology alone, which is how the cause-of-death is often assigned (again according to Karch). Some people have very high blood levels of opioids, because they have a strong tolerance and take a lot. If they die of a sudden heart arrhythmia or something, they’ll probably be labeled as a drug overdose even if they’re not. So we need to be very cautious about interpreting these data. I saw some pretty serious data quality issues (outlined in my post above) when I looked at these overdose deaths. You can’t just count up records with certain cause-of-death codes and say, “This is how many overdose deaths we saw this year,” which is basically what the CDC is doing. You have to issue some of these caveats. My suspicion is that the increase in deaths is probably real, but probably overstated by some undeterminable amount.

Jeff Miron is doing some really good work on this. You should have him on to talk about the opioid crisis. He’s skeptical, too, which makes me think I’m not too far off base with all this.

Kevin writes:

I apologize in advance for my long response. I work in the medical field in cancer.

All the problems I discuss below and were discussed in the podcast can probably be addressed with a Singapore-like system of radically freer health care markets. All the pricing madness, all the distortions, all the madness of third party paying. Its impossible to fix by tweaking here or there because we keep getting the worst of all worlds. Either the government takes over and we get socialized medicine like Germany or the UK with all its horribleness, or we get an American solution involving really free markets and these problems get solved with innovation and competition. Sadly, my bet is on rationed socialized medicine without innovation.

EMRs exist for billing. Almost no doctor in the world ever thought an EMR would help him take care of patients. Or if they did the first time they used one changed their mind. The fact that now, after they are being forced on doctors, people are thinking how they could improve patient care by adding more boxes for them to check is incidental to their purpose. With rare exceptions EMRs are a time wasting hindrance to patient care. When the government has to force you to adopt a technology that has existed for 20-30 years, its likely it does not fit your needs or is not cost effective, but fits the government’s needs.

Dr. Roberts, the device that allowed the doctor to see into your shoulder is probably fluoroscopy, is 50 years+ old, and is not usually necessary if the anatomy is understood by the doctor. However, for insight into the crazy way medicine works – by using fluoroscopy the billing goes up dramatically because the doctor now bills for the procedure of placing the cortisone and a second of imaging (or they might be bundled depending on how the code is set-up). You view it as amazing medicine, but very likely this is a situation where we have added cost with no benefit. In the free market the cost would be rolled in and it would be an improvement. In the socialized medical market it raises cost. If we lived in a world with private medicine, you might be willing to pay for the imaging because you may believe it has benefits.

Why does medicine use fax machines? Because medicine is resistant to technology? No, because federal laws like HIPPA are strangling how they share patient information. In medicine, like other highly regulated industries, nothing is done without deference to some law.

Prostate cancer: Should all men use the exact same method of treating their prostate cancer because it is cheaper despite lots of options at various costs that are equally effective? I would hate to live in a country with a “national policy” where every disease is treated identically despite many equivalent options. How would new technologies emerge in such an environment? Would we propose that for other parts of our lives, do we all need to drive the same government issued car? Finally, prostatectomy is not even the cheapest method of treating prostate cancer, brachytherapy is.

Protons: They are expensive. Many socialist countries have them. Japan has invested in a technology that is about 2-5x more expensive – carbon radiation (proton radiation uses a hydrogen nucleus, carbon uses a carbon nucleus). It is too long a discussion to explain why often calls to study protons are mostly going to be a waste of money. Protons are not like a new chemo, they are like having a surgeon change from brass tools to modern surgical steel tools. The promise of protons is just reduced side effects, and the reasons that side effect reduction is predicted is extremely robust. Protons for prostate – not sure if that is best use of resources since not clearly better than traditional radiation. Protons for pediatrics is close to a homerun. Is it worth the cost – very different question. It might not be except in children or a few other situations. Protons might not survive a free market, but some do pay cash for it. However, proton radiation technology is falling in price on the expected curve for prior radiation technology which means all this handwringing will result in studies being done about the time price parity is reached and no one will care.

I don’t know about Lasix, but cosmetic surgery is all highly competitive and cash pay. The surgeon takes cash, the surgery center takes cash, the anesthesiologist takes cash and the procedures are dramatically cheaper and compete on price. A “cash” system would reduce costs so fast politicians would discover faith in markets. But less graft and control.

James Scheltens writes:

I’ve been listening to ECON Talk for several months now and I was particularly disappointed with the Jonathan Skinner episode. The reference to the cost reductions seen in LASIK surgery and Veterinary services missed such an obvious explanation; these are discretionary market places. If you have bad eyes you can wear glasses, or perhaps opt for LASIK surgery. If your dog gets cancer, you can opt for chemo, or palliative care, or just put the dog down. In both cases it is not un-natural to discuss costs along with treatment options. If your son gets cancer your response is entirely different. Treatment is no longer discretionary and the only acceptable treatment is the best available. There are still additional market place failures beyond the fact that you only want the best care and are willing to pay whatever you can for it. You probably don’t really know what the best treatment is, you will be in a poor emotional state to think clearly, and a profit driven system will most likely direct you to the most costly treatment. That in a nutshell is why the U.S. system has lousy cost controls. It’s not just the fact that insurance pays for it, insurance pays for automobile collision repair also.

If your base assumption is that every person should have top notch medical care (everyone gets Porsche level quality, and no one gets denied) effective market places are very hard to set up. If you allow the market place to sell a range of quality of service that are only available to those that can pay (as one sees in underdeveloped countries) then mediocre health care services can be very cheap. The market works to bring costs down, but the suffering is great. This seems to be the approach you would like to take.

The U.S. system of medical care tries to have more market place features than any other advanced nation and yet has no better outcomes and costs far more. You should be open to the reality that other countries that use more heavy handed cost controls and fewer market based approaches are delivering health care just as well as the U.S. but at lower cost. The idea that we can cram market place options into the system willy-nilly while graranteeing access to top quality health care is counterproductive and actually is driving up costs. We need to think more carefully about how market based approaches will play out in terms of incentives and costs (including emotional costs) when combined with the mandates we want to have in our system.

If you have not looked at the YouTube videos that “The Incidental Economist” health services research blog produced covering every comparable countries health care system, you should. I also think you should have Aaron Carroll and/or Austin Frakt on as a guest if you want to cover how to reform health care in the U.S. The discussion with Jonathan Skinner did not impress me as being coherent or deeply informed.

P.S. I am 64 years old, I was born with a minor birth defect and had thirteen surgeries in my childhood to correct the defect. Some were paid for by charity (Crippled Children), most were paid for by my father. I grew up knowing we were poor in part due to my surgery costs on our family. Russ’s theory of having charity cover health costs has emotional costs that I doubt he is fully aware of. My father died of cancer when I was 27, my mother had cancer in late life and died of Alzheimer’s disease. My sister had breast cancer twice, as has my wife. My wife also discovered in her fifties that she has a gene that causes intestinal cancer. She has had her large intestine removed and the rest of her intestinal track has to be monitored for cancer on frequent basis. The stress and emotional content of serious health issues is of an entirely different order than other life events. To think that the economics of serious health care are likely to play out in the same way as other economic consumption choices is simplistic thinking.

rovingbroker writes:

[Comment removed. Please consult our comment policies and check your email for explanation.--Econlib Ed.]

Greg G writes:

Great episode and many thoughtful comments, especially those by the health care professionals.

James, I thought your comment was one of the very best here but if you'd been listening to EconTalk for more than a few months I think you'd realize that Russ is well aware of market effects on prices for Lasik and veterinary services. I could be wrong but I think he assumes (probably correctly) that most EconTalk listeners already know about this.

I really appreciate the job Russ does in giving a full and fair hearing to ideas he might disagree with. I think it's great that Russ and this comment forum reliably do such a good job with that. It's something that is very rare today in policy discussions. This podcast and comments discussion has been much more focused on understanding the problems than pushing ideological solutions. Thanks to everyone who made it that way.

Andrew writes:

There was a comment in this podcast about NICE in the NHS to the effect that the body is controversial. Speaking as a casual observer in England, I don't see that. It's sometimes controversial for individuals who are being denied some treatment, but most people just accept the system.

The pernicious problem is that specialists won't tell people about treatments which are available, approved for use in the UK and would lead to better outcomes than the alternative treatments, but which aren't approved for use in the NHS by NICE.

Luke J writes:

Seems the value in EMRs, specifically the patients' charts, are realized by the practitioners not patients. In the event of a lawsuit, this is the best advocate for the doctor. I would be interested to see if EMRs have had any impact reducing payouts to patients citing negligence.

jw writes:


  • My best friend is a highly experienced and respected ICU RN. She confirms many of the stories from the commenters above. She (along with my general practitioner) is immensely frustrated with the software that was discussed. It takes weeks to learn, costs millions in installation and customization, millions more in training and millions more in maintenance and still lacks some basic functionality and is very hard to use. Besides which, the insurance company's software is also buggy and prone to problems (which always seem to favor the insurance company).
  • thomas egan md, I agree that defensive medicine is at least 20% of the cost of health care now. Also, a lot of the technology is being implemented for defensive reasons as well.
  • j mitchell, Agreed. I am currently disputing some tests rejected by my insurance company (they are covered but apparently not in my instance). Now (as evidenced here) I don't mind arguing for them and enjoy researching the journals to prove my case. But I have noticed that in both cases, their docs have 26 letter Indian names and are not locatable in US directories. It makes me wonder if they are remotely using Indian docs not trained in US standards of care. If so, what is the point of going to a US doc if every test will be challenged/second guessed by an Indian doc? And what of people with neither the time or inclination to fight?
  • My friend agrees that opiate use is out of control and that its interactions with other drugs (usually other painkillers) is a problem. But she doesn't see it being large enough to affect mortality. What about the 5x increase in diabetes?
  • One of the biggest issues for health care costs are the 50 state boards. A federal standard is needed.
  • That being said, ACA is a cost and service disaster and needs to be scrapped and replaced by direct low income subsidies (possibly Singapore-like as above). If Obama had run on the reality of "a $4K deductible and 70% coverage for $4K/yr", he would never have gotten ACA off of the ground.
  • I had to take oxy for a few months after a major operation and although I knew the weaning protocol, I decided to stop cold turkey as I would probably never get another chance to experience withdrawal. NOT recommended.
  • The Fed rarely clears checks, it is not a centralized clearer as distributed and private clearing has proven to be cheaper and more effective.
Mark Breslauer writes:

On the topic of opioid abuse and decreased mortality, a fascinating book on the subject is Dreamland by Sam Quinones. Quinones makes the case that the causes of current epidemic are (1) the decreased economic opportunities in rural America you touched on in this episode, (2) this misapprehension in the medical profession that when used for pain relief the potential for opioid addiction is minimal, and (3) the business model pursued by the black tar traffickers in Xalisco (using rural Mexican youth as employees - not dealers- to distribute heroin in the US, focusing on markets where they would not have competition from existing dealers, and providing good customer service).

Quinones would be a great guest for this EconTalk.

Todd Kreider writes:

I understand that a phrase like 'big data' might seem over used, and it was funny to hear Jonathon Skinner joke he wants to throw up when he hears that and Roberts mock it with "big data: blah, blah, blah." But we are only at the tip of the iceberg of what big data will be revealing in medicine in coming years.

Skinner said he could see doctors being replaced by enough artificial intelligence in 20 to 30 years to put doctors out of business but that he thinks "most humans still want to talk to a human about the different options and what it means and what it means to their life." But there would be no need for this consultation if the person never got sick in the first place as with many under age 40 nor would their be a need to consult if a solution to a disease is completely obvious in 2033. Recently unemployed doctor to friend in 2027: "99.9999% take X for what you have, but if you want to follow me around the links for alternatives, you may."

The more I heard Skinner discuss health care technology, etc. the less confident I was that he understood this area. I found a 15 minute presentation that he gave a couple of years ago and was not surprised to see that his first slide was of the CBO's 2007 projection of health care costs out to 2087. This is the one where three sentences were used to justify better technology has increases health care costs in the past and so assume this will continue another 80 years.
This type of extrapolation isn't as sophisticated as astrology.

Just for fun, Roberts and Skinner should both jot down what they think medicine will look like in 2025, 2035 and 2045 - (forget 2087 or 2231 since those two might be difficult...) on three index cards and stick them in an envelope. How advanced will the health pills be that reduce disease? What will be the 5 year prognosis for the stages of colon cancer, breast cancer and lung cancer? Will organs be printed out and transplanted by 2025, 2035 or 2045?

It should be interesting to see how close each got with the 2025 index card in 9 years.

Jon Skinner writes:

Thanks for all the thoughtful comments and special thanks to Russ for inviting me – it’s impressive to listen to him on his podcasts, but even more impressive to talk with him for an hour! I’m happy to follow up on a few topics raised by listeners.

Opioids: The Sam Quinones book “Dreamland” is indeed amazing, and he’d be a terrific guest. GregS wondered whether opioids alone could account for the rise in mortality among 45-54 year-olds. The answer is no, as he surmises. Anne Case and (Sir) Angus Deaton show ( it’s also the increase in suicides and alcohol-related deaths, which in turn are likely related to opioid use. These combined effects are enough to explain the rise in mortality, but not enough to explain why mortality didn’t fall as rapidly as for other demographic groups – Ellen Meara and I try to tackle that puzzle here: (

Electronic records: I can understand why physicians and others are frustrated with the current incarnation of electronic health records, but it’s hard to imagine going back to paper records. Think of the dominant medical record software, Epic, as the Model T automobile. Early owners were no doubt tempted to scrap their cranky car and buy back Bessie the horse, but I don’t think many of us now would want to turn in a Tesla for even a race-horse.

Uniform insurance records: Thanks to J Mitchell for his insights on insurance recertification. The reason I suggested this reform (as opposed to other worthy reforms like malpractice) is that it seemed to be the one proposal that could get support from both sides of the aisle in Congress, and we know it would save billions of dollars in useless paperwork. Today (as it happens), CMS and the FDA called for uniform insurance records, so that (among other things) they can track medical devices ( So perhaps there is hope for this proposal.

Markets versus regulation: Not surprisingly, there was a difference of opinion about whether we should move U.S. health care more towards European systems, or towards more market-based solutions. That discussion would take at least another hour, but let me mention two modest proposals: One is to cap prices at 125% of Medicare, thus moving towards regulated prices (as discussed by Ezra Klein here The second is to free up existing regulations and laws to allow people to choose low-copay British-style NICE coverage in return for much lower premiums (

[Shortened URLs replaced with full URLs--Econlib Ed.]

Ray Peters writes:

Thank you very much for talking about what I do as a primary care provider.

The electronic health record (EHR) is great at a lot of things. Type in 3-4 letters of a pharmacy chain and the computer tells me within seconds the closest location to the patient's home address. It gives me a list of quality measures that are due like mammograms, colon cancer screening, or annual diabetic foot exams. Type in bph, pause for Benign Prostate Hypertrophy to show up in the menu, select it, and then a list of the 10 most common tests, orders, and medications related to BPH show up making it real easy to order them.

The EHR prevents mistakes cause by bad penmanship but can create mistakes from clicking the wrong box or clicking the moment the computer screen is auto-updating.

Using EHRs also involves something called Meaningful Use. Providers and health systems need to satisfy meaningful use goals in order to qualify for subsidies/avoid penalties related to EHR use. It makes sense that providers shouldn't get a subsidy for a computer that doesn't get used. So, in addition to doing what the patient wants, doing what is good for the patient, documenting enough for insurance to pay for the tests we order, documenting enough for insurance to pay us, and documenting enough to prevent or defend against a lawsuit, we need to do certain things to demonstrate that we are really using the EHR responsibly.

In the USA health care world never accept a fee schedule price as a realistic market price. Imagine what happens after 20 years of raising the fee by 5% in the hope that raising it 5% instead of 4% will convince third party payers to increase what they pay by 2% instead of 1.5%.

Regarding opioids, my state has law for the past two years that sets a limit of 60 pills for 30 days. If we prescribe more, there is required drug testing, lots more documentation, required warnings to give patients, etc. I really like the 60 pill limit.

Health care is not the only business that is largely paid for by a third party. Consider roof repairs after hail storms or college education. I think there is a real parallel between college education and health care. Both are goods that we don't think should be denied to anyone who needs them. Both are subsidized by both government and charity. Both are situations where spending now can result in future big savings/increased earnings. Both are good that have increased a lot in price over the past 30 years.

Anyway, it is a great time to be alive. Thanks to Russ for all the great podcasts.

Dallas Weaver Ph.D. writes:

The declining life expectancy (LE) is a serious issue. Life expectancy (most particularly, USEFUL life expectancy) is the goal of medicine. So why is it going down when the money and technology are going up?

One factor may be the rapid rise in disability claims. If you get tired of working, you can apply for disability to provide a steady income, but to do so is a "high band-width" series of bureaucratic challenges. Our only control mechanism to protect the disability system from free riders is to increase the hurdles people have to jump to get the brass ring (a method that cuts out the mentally marginal who should receive aid).

If you are not truly disabled and try to get in, you have to be very good at faking disability. After playing this disabled part for the years it takes to get through the system, you will start to "believe" that you are disabled (an easier way to stay "in character"). Your attornies and others will encourage you to "believe" you are disabled to win your case.

Then the Nocebo Effect starts creating its magic (the opposite of the very powerful placebo effect): you do become disabled, get your money and die early.

Meanwhile, we have lawyers telling anyone who will listen that all their problems are caused by X chemical/drug/action etc. and if they show symptoms of anything from terminal laziness to a headache they could receive money from a lawsuit. This creates a perfect nocebo effect situation where people get rewarded by "believing" they are sick and they become sick.

The nocebo effect arising from "believe" something negative can lead to real physical problems that will impact your life expectancy. This is why "witch doctors" actually work magic with their voodoo, when people believe in them. It is why sugar pills, more expensive pills and "crystals" can actually cure some people. "Belief" is very powerful, even when it is illogical nonsense.

We have a dramatic increase in the disability bureaucracy and class action lawyers who all profit from the Nocebo effect. Like modern day witch doctors with their dolls and pins, they may actually be killing people.

Todd Kreider writes:

The highest recorded life expectancy for Americans occurred in 2015. I'm not sure how one concludes a "decline in life expectancy" out of that.

Michael McEvoy writes:

Thanks Russ for another thought provoking podcast. Skinner's colleague Gil Welch is one of my primary care heroes - yet Welch seems to say little overall publicly about controlling costs. Another words , there is a wholly different approach to the problems of medicine - they have to do with a public understanding of risks and probabilities.
Many times patients tell me what they want done and I try to tell them there is no need to do those tests or they can wait - but the patient does not want to wait . I can spend yet more time trying to explain yet it is easier to opt for the unnecessary expense ( usually covered by insurance) .
Price of the proposed test has VERY little to do with the situation, which is why I doubt the market solution will really work. (Russ you may clutch hand over your heart now, gasping for a breath)

Jazi Zilber writes:

Pet comparison:

Human costs are relatively huge. So for humans the prices should have a ceiling effect.

Whereas pet healthcare is relatively minor in one's budget. EVEN for "very expensive" treatments.

Thus, if average costs per pet are 1/10 or those of human. Even exactly same growth rates will not mean that humans growth is natural. Because the cost increase in humans is so large (in utility terms, as probably increased marginal value of money spent!)

Robert Swan writes:

I'm always hearing about these centralised records and what huge benefits they'll bring. Prof. Skinner, besides the rather lame assertion that he couldn't imagine a future without them, did at least offer that they would let the bureaucracy monitor just what the practitioners are up to.

To me, that sounds like a drawback: a variant of my oft-repeated objection to treating individuals epidemiologically. Just as there is no average patient, so why strive for supposed norms in blood pressure, cholesterol, etc., there is no average doctor. A general practitioner in a popular retirement area will likely have very different patient profile from one in a college district. I suppose, over the years, the centralised system can be refined so it raises fewer false alarms, but in the end you have somebody at "head office", who has never met the patient, telling the doctor how to treat that patient. Will this do as much good as it does harm?

Centralised records systems are a boondoggle for computer IT developers. I have mentioned here before about the abandoned British NHS record system. That was 15 billion pounds for the IT people, nothing salvaged for the taxpayer.

I would have a different view if the records were kept with the patient; i.e. if your health care card had a chip in it containing your personal medical history. Health professionals could access these records as long as you remember to bring your card. If you forget it, the doctor can fall back to what they do when attending to tourists or the like.

My main hope for reduction of medical costs (besides the forlorn hope of our nations getting away from absurd epidemiological treatments) is that the Chinese will see these $1,000,000 scanners, realise they can make them for $10,000 and live cheerfully on the profits if they sell them for $100,000. Not sure how they can break into that market, but they might try starting small with hearing aids (which carry an obscene medical premium).

Finally, on word games Russ, you say you'd like to dismantle all the things that have been "mantled" over the years. You're such a moderate! Have a bit of courage and take on the things that have been "molished" too.

Robert Rosales writes:

Another excellent podcast but I think it was missing some probable causes to the increase in mortality.

The last 40 years has seen a dramatic shift in our diets to a government endorsed low fat high carb diet while seeing a dramatic shift in obesity. In 1973 our most obese state was less obese than our least obese state today.

I think more important than dietary lifestyle is the change away from the closed knit family structure as seen in the Roseto effect and the hispanic paradox.

On Lasik it has plateaued on price primarily because people are demanding quality and the latest lasers which have a per use royalty fee has created a price floor. Even though the old technology was giving great results when it comes to their eyes people tend to want the best technology.

Comments for this podcast episode have been closed
Return to top