That is the only explanation I can think of for suggesting I read this piece by Megan McArdle. The man wants my head to explode.
McArdle’s essay is about her thyroid, which is on the fritz. Unfortunately, her endocrinologist did not treat her thyroid problems aggressively enough based upon a reliance on outmoded guidelines. This, apparently, means there’s something wrong with evidence-based medicine as currently understood in America.
Please excuse me while I put on my old rantin’ hat.
I assume Megan McArdle is a decent writer when it comes to economics. I will confess to sufficient ignorance on the subject that I don’t know if what she writes is generally sound or not. But this essay is just so painfully poorly thought-out, it should never have been published.
Things start to go south around about here:
As a result, as thyroid patients have now been complaining for years, doctors stopped paying attention to the symptoms. They treated the number instead of the disease. As long as your levels were under 5, they told you that you were fine and there was nothing more they could do.
Changing the guidelines has helped some, but most primary care doctors just target the numbers they learned in school, and older endocrinologists have also often been resistant to change. I know several people whose weight gain was clearly entirely due to their thyroid, but whose doctors stopped well short of a dose that would make them lose the weight; when they pointed out that they still had a lot of symptoms (weight gain, but also the baldness and the fatigue and so forth) they were told that they were just drug-seeking for vanity purposes and they should stop eating twinkies and get their fat asses to the gym.
(Okay, I paraphrase. But that was reportedly the clear implication.)
Really? “Most” primary care doctors do? That’s a pretty broad brush you got there, Megan. Seems to me you’re taking what some people you know said, and leaping to an unsupported conclusion. And I loooooooove that snarky paraphrase, based on the “clear implication” reported to you by someone else. That’s a fantastic basis for criticizing the medical community as a whole!
Let me step back for a moment and concur that doctors should treat their patients, rather than some number on a lab slip. That’s certainly what I was trained to do. In fact, I try to order lab tests only when I think they will help answer a question that my clinical skills can’t on their own. So McArdle is right to be peeved that her first endocrinologist failed to do so. If she’d decided to write about how glad she was to find a new one, I wouldn’t have anything to complain about. But no…
I think about this a lot when people start talking about “evidence based medicine” and how it can help us control health care costs. The question is, “evidence of what?” Evidence based medicine works best on things that are very easy to measure, like blood levels. It is very easy to tell whether a statin reduces blood cholesterol levels. It took a lot longer to tell whether it actually reduced heart attacks.
Um… you know what else is very easy to measure? Thyroid hormone levels!!!!! Which is why, as McArdle notes early in her essay, guidelines for thyroid hormone monitoring were revised downward to better treat patient symptomatology. It’s not that evidence-based medicine failed, it’s that she was seeing a provider who failed to keep up with it.
And the “evidence,” which obviously varies based upon the subject under investigation, is generally some kind of outcome that can be verifiably linked to some kind of intervention. When newer, better evidence comes around, clinical practices sometimes change. Like, say, with regard to the management of thyroid problems as is noted in paragraph five of McArdle’s addle-brained essay, or whether statins should be prescribed. Sadly, “evidence” is not the same thing as “gospel,” which means from time to time it needs to be revisited. If McArdle has a suggestion for a better system, I’m all ears. Perhaps divination?
But this next bit had me attacking my computer with a blunt instrument:
The more we rely on a central board to make decisions for huge numbers of people, the more tempted we are going to be to rely on metrics which can be collected, aggregated, and mined for data. Where does that leave the thyroid patient with “normal” blood levels . . . and a collection of vague, frustrating–but nonetheless very real–symptoms?
Oh, I know. That central medical decision-making board is awful. Whenever I get a new directive from Dr. Strawman, it makes my blood just boil.
I apologize for the following unavoidable use of the caps lock key — THERE IS NO CENTRAL BOARD THAT MAKES DECISIONS FOR HUGE NUMBERS OF PEOPLE, NOR AM I AWARE THAT ANY SUCH BOARD HAS BEEN PROPOSED BY ANYONE. Yes, there are bodies within the scientific medical community that every now and then issue new guidelines for management or screening for various illnesses. Not infrequently their recommendations are controversial, and it takes some time for the broader medical community to reach a consensus. This controversy often stems from physicians’ clinical experience, which may not seem to square with the newer evidence at hand. These guidelines are nothing more than that, and individual providers are free to practice how they see fit.
Now, obviously the government (which I infer is the boogeyman responsible for this ominous “central board”) sets guidelines for what it is willing to pay for through Medicare reimbursements, based upon the best evidence currently available. Again, I’d love to know that system McArdle thinks would be a fit replacement. You know who else does this? Private insurance companies. While it’s nice to have congruity between what I’ve prescribed and what insurers will pay for (government or otherwise), the former is not necessarily dictated by the latter.
I honestly have no idea what point McArdle thinks she’s making with this sloppy, sloppy piece. Her symptoms weren’t well-c0ntrolled, but her endocrinologist didn’t pay attention to her complaints. She found a new one, who is taking better care of her. Hurrah! What this has to do with evidence-based medicine is completely beyond me. I hope she’s better on economics than she is on healthcare policy.
Update: I’d like to begin by thanking Megan McArdle for taking the time to comment at such length in response to this post, which I think demonstrates a great deal of graciousness and class, particularly when considering the tone I took. As I say in the comments, I now wish I had used more temperate language. In the future I’ll try to keep this experience in mind when I choose my language in describing what other writers have said.
From what I understand after reading her comments, her main concern is that evidence-based medicine will be used as a rigid set of rules for determining how medicine should be practiced, and what services will be reimbursed. I certainly agree that EBM works best as a starting-point for medical decision-making with plenty of room for well-reasoned deviation, and should not be used as a strict metric for setting regulations. I would hope that anyone crafting healthcare policy would treat it this way, and I am more optimistic than she that the ACA will do so. However, after having seen how EBM has been used to deny patients of mine treatments that were recommended and appropriate for them, I can understand her concern.
Fold up the laptop and walk slowly away from the computer, doc. As I did not too long ago, you have at least two choices available for stress relief and being mad at Megan McArdle has never done anyone any good. Choose wisely.
I am planning on asking certain parental figures for a nice bottle of high-quality scotch for Christmas. (Powerful household entities would likely veto such purchases at this time, citing budgetary concerns.) In the meantime, ranting does seem oddly therapeutic.
For some reason Megan McArdle seems almost uniquely able to inspire responses like Doc’s, for reasons that I can only partially explain or make sense of.
I assume Megan McArdle is a decent writer when it comes to economics.
Why, after reading a sloppily reasoned piece with a badly-hidden agenda, would you assume that?
Misplaced charity, perhaps?
I would have guessed that The Atlantic wouldn’t pay someone who didn’t know what she was talking about to write for them. It seems I would have been mistaken.
But she does know what she’s talking about. It seems you would have been mistaken.
In this particular area, Koz, while I sincerely appreciate that she has taken the time to reply at length, I still have qualms with her viewpoint.
Yeah, I understand, and forgive me for projecting all the failures of lib advocacy onto you. I don’t care that much about the subject matter. But the advocacy angle is important, and it comes up a lot, and nowhere more often than when McMegan is involved.
“Megan McArdle wrote A, B, C about EBM. I’m a doctor and have some clinical experience in that, and in the main I think she’s wrong because of X, Y, Z.”
But no, that’s not what they write. There always, ad hominems, cheapshots and allcap detritus thrown in, and the audience likes to delude themselves that all the extraneous bullshtt in the “smackdown” means more logical strength (see Schilling above).
In rereading what I wrote, particularly since Ms. McArdle has taken the time to reply at such length, I wish I had been more temperate in some of the language I chose.
That said, I have endeavored to the best of my ability to explain why I still have points of disagreement.
Of course. And please don’t think I’m being excessively antagonistic toward you. It’s the overall pattern I’m concerned about that obviously predates today by a lot.
And about the pattern, there’s one thing that’s worth mentioning. Megan isn’t as cursory as her critics seem to think. Nowhere at the Atlantic post does she say EBM can’t work or that it has no value at all. It still doesn’t mean she’s right on the merits of course but she usually has thought about more than her critics give her credit for.
McArdle’s neither great on economics nor as bad as liberals believe. She has a general understanding of economics that’s sounder than what most liberals have, but she has a sad tendency to write too glibly rather than delve down into the real logic of situations she discusses (as demonstrated here).
She may be wrong, but she’s rarely if ever ridiculous. Once you get that, it’s pretty easy to keep you emotional balance while reading McArdle. Unfortunately, that’s quite a lot to ask of some libs.
You’re right, Mr. Saunders; I didn’t take a representative survey of every single doctor. What I did take was a survey of all the thyroid patients I know, who with the exception of one, report the same experience. It also dominates the thyroid patient discussion forums, though of course, that could be selection bias. And the overwhelming consensus of all of this is that, with some exceptions, the problem is worst among primary care physicians (who are generally unaware of the ACCE recommendations and treat to whatever they learned in med school, which for the majority of practicing primary care physicians, is the old 5.5 TSH); second worst among older endocrinologists; and thankfully fading among younger endocrinologists. Young residents, who are under the guidance of old doctors, are the worst; you can’t reason with them, and you can’t talk to the doctor who is actually making the decision.
Patients shouldn’t have to doctor-shop the way the majority of thyroid patients seem to in order to get someone who will look up from the TSH levels and pay attention to the fact that they still have symptoms.
The article was not an argument against evidence-based medicine; it was a warning to pay attention to what evidence is used, and how it is used. There’s a bit of the “No True Scotsman” fallacy in some of the people who have given me over-the-top responses. It’s totally true that really well-designed and implemented EBM would not deliver these results. It is, however, unfortunately not necessarily the case that all EBM programs are going to be well-designed and implemented.
As for the central board, you’re surely aware that we have a whole new health care board being created by ObamaCare to make reimbursement decisions based on EBM among other criteria. In the case of hypothyroid it probably doesn’t matter–levothyroxine is too cheap to care about. On the other hand, lab tests and follow-up visits aren’t. What if we decide, down the road, that the clinical guideline is 3, and we don’t reimburse for the follow-up labs and doctor’s visits necessary to start someone like me on thyroid hormone? This is not some dystopian fantasy; it’s exactly the way the law is supposed to control costs.
I would say that my condition is probably too small for IPAB to touch . . . but there are a lot of conditions out there that aren’t cheap, and real concerns about how this is going to affect health care decisions. I am very much in favor of evidence-based-medicine. But I am very worried that it will be implemented to further bureaucratic, not medical, priorities, and that a lot of health care wonks are expecting it to deliver far more than it actually can. Humans already have a strong tendency to fixate on numbers because they’re available, not because they’re useful. It seems quite plausible to me that the way we’re going now (at both insurers and the government) will exacerbate this tendency in an . . . unhealthy way.
First of all, it was very courteous of you to take the time to reply, Ms. McArdle. I sincerely appreciate your doing so.
There are a few things I would say in response. Firstly, all I can do is restate that the problem thyroid patients seem to be experiencing isn’t a flaw in EBM, but in older practitioners or providers less familiar with the standard of care deviating from best practices. I can certainly understand why this would be very frustrating, but this isn’t a flaw with using EBM to guide management. If anything, it’s the opposite.
And yes, obviously under the ACA there will have to be some kind of entity that determines what is paid for, and how often. Again, this isn’t really any different from the status quo, wherein insurers public and private have to use some set of criteria to determine what they’ll pay for. Having personally had to hector insurance company personnel to get “unsupported” treatments approved, I am all too familiar with how that goes. I don’t know if there is a basis for fearing it will get worse under ACA.
Further, using EBM as a basis for these decisions seems eminently reasonable. I don’t know what better paradigm might be substituted. Certainly, should patients require treatments that deviate from the established schedule or recommendations, I would hope there was an efficient system to get such treatments evaluated and approved, but I don’t see why using EBM as a starting point would be an issue of concern.
Of course, no matter what happens with ACA (or any other kind of healthcare reform package that may emerge to replace it), some kind of painful decision-making will doubtless be required. We can’t control burgeoning costs without making choices that will be hard for patients and providers alike. Frankly, while I don’t have any numbers in front of me to support this, I suspect our healthcare morass wouldn’t be nearly so bad if better adherence to EBM had been the norm up until now.
Once again, I sincerely appreciate your taking the time to reply.
I think you have missed the point of that post, which is not that Evidence Based Medicine is a bad idea that will kill us. If you missed it, this was probably because I was unclear, for which I apologize.
The point I was trying to make was that we need to be careful about what constitutes “evidence”, and how that evidence is used. I wasn’t blaming the problem of thyroid patients on EBM–that would be moronic, since it’s not like EBM has taken the practices and hospitals of America by storm.
Rather, I was pointing out that guidelines often get misused, and offering an example. This example was meant to guide us as we think about EBM–how it will be implemented, and the realistic limits. It was not meant to illustrate how terrible EBM is, but how terrible blind reliance on networks, and to warn that there is a real danger that EBM in practice will encourage blind reliance on metrics. I believe I said in the piece that this was not supposed to be an argument against EBM; I’m sorry that this was insufficiently prominent or clear.
I completely agree with you that this is a problem of insurers as well as the government. However, insurers do have to compete for patients–yes, imperfectly, etc, etc, but at least somewhat. The government doesn’t at all, and the pressures it is subject to from activist groups are not necessarily a very good substitute, for all that I really respect patient action groups.
A few things to say in response:
1) Any doctor who weighs a lab value more than a patient’s symptoms when deciding on treatment measures is simply not very good. Sadly, there are lots and lots and lots of not very good doctors out there. Even more sadly, there is no particularly effective remedy for this.
2) EBM, like just about any decision-making mechanism in any field, is imperfect. I’m certainly not arguing that EBM won’t create some problems, or that it will always result in healthier patients more cheaply. Lord knows all you have to do is read the medical literature to see that every few years there’s some new massive revelation that completely upends what medical providers are supposed to be doing. I’ve already gotten somewhat jaded about it, and I’m still quite young in my career. But EBM is by far the most reliable method we have of delivering the best care to the most people.
3) Sure, insurance companies have to compete for customers. However, this competition hasn’t stopped them from hiking their rates pretty much every year, much to the dismay of their paying customers. Nor have I witnessed patients finding great deals by comparing insurers. Most offer about the same level of coverage for the same conditions, at least when you’re looking at plans of comparable price. Elected officials at least have to face angry voters who, if things are genuinely bad enough, have the option of firing them. Not so much with the CEOs of Harvard Pilgrim or Aetna.
4) I don’t think anyone can plausibly argue that the problem of exploding healthcare costs will be solved in such a way that people won’t see a cut in the services they receive. Perhaps I am mistaken in this, but I doubt it. Accepting as a given that not everyone will be able to get all the care they want all the time, I see no better mechanism for determining which services merit third-party payment than EBM.
The concern is not that a *doctor* will look at a lab result and say “Your number’s less than five so your obvious weight gain and hair loss must be due to some other cause.” The concern is that someone who is *not* a doctor is going to believe that they’re just as good at making medical decisions as a doctor because they have a copy of the guidelines–and that this person, instead of a doctor, will be the one making decisions.
1) True, but there are a large number of policy wonks who do seem to think that EBM can cure this, and fail to consider that it might actually make it worse
2) Again, we are not arguing about the value of good EBM. I’m just offering a caution against what ordinary people and policy makers should expect from it, and for care in implementation.
3) If you think I’m claiming that insurance companies are exempted from this, I’m not. I think they face constraints that the government doesn’t–competition, and also, well, the government. But they absolutely do moronic things.
I think you’re overestimated the effectiveness of the response by elected officials, who are far more responsive to the noise of the lobby than the validity of their message. And underestimating the extent to which those constraints actually fall on insurers–witness the downfall of the HMO revolution, which was due to legislators who undid very effective cost control.
4) We agree! My point is that people are, in fact, frequently making sort of this argument in public–Obama did it while selling ObamaCare, as did (does) Orszag. “No, of course cost control won’t cut *services*, just worthless treatments and reimbursements to providers.” Good wonks know better, but those wonks will not necessarily get to control the final system, which will operate under heavy political constraints.
Certainly, to use EBM as some kind of panacea that will spare anyone the possibility of being denied services they think they need (or might legitimately need) is… let’s just say naive.
Again, I’ve seen how EBM can be used to deny services. My own experience has been informed by run-ins with private insurers more than public, but I have no illusion that the latter is necessarily more virtuous in this regard, though I am clearly more optimistic about the ACA than you. As I say in my update in the main post, I think I have a better understanding of what you were saying in your essay after having read your comments.
However, insurers do have to compete for patients–yes, imperfectly, etc, etc, but at least somewhat
Pah! Insurers shop for employers looking to buy insurance for their employees. The employers are looking solely at the cost and whether the benefits are satisfactory enough to their employees. This does not include any focus on whether and how EBM is being used, and it especially does not include much focus on whether the doctors covered by that insurer are good or bad, since most docs are linked with multiple insurers (out of their own self-interest, since people want to choose their doctor). Competitive pressures of insurance purchasing play no role here. None. Zip. Zilch. Nada.
There’s actually some exciting stuff being done with EBM by insurers–pharma benefit managers have started doing their own clinical research to better target expensive new patent drugs.
Also, for large group, the insurer is a pass-through; the large firms are the residual insurer, not the insurance company. And they do care about things like EBM.
Sorry for the adverse affect on your blood pressure, Doc.
I guess I read McArdle rather differently than you did (more along the lines of what she said in her clarifications here), perhaps because I have been reading her for almost a decade.
I am all about EBM and I am of the mind that bending the cost curve is going to require being more selective about treatment. And, to some extent, I viewed (and view) EBM as the best way to go about it. What McArdle’s piece got me thinking about, though, was the limits of EBM. Especially in cases like hers where the numbers aren’t matching up with what seem to be very tangible symptoms (we can accuse people of a thyroid problem of eating too much, but hair falling out)? And more to the point, those areas of medicine where there isn’t a way to get a number in the first place.
This isn’t to say that we shouldn’t rely very heavily on EBM, especially when it can provide numbers, and even to the extent of saying “No, we’re not going to pay for that.” But I was wondering what you thought, as someone with far more exposure to how EBM works than I have, about the potential limitations with it and practically what might be done about it? It seems like carving out even some “the doctor says different, so we go forward anyway,” could be very problematic because it can push patients to the doctor of greatest generosity (as often happens with painkillers now)? Or is this a false-dilemma with a more straightforward solution that’s already sort of in place to the point that we can use EBM without having to worry (too much) about cutting people off by using it as a policy tool (in addition to an MD guidepost)?
This is an area where I am very much out of my depth. So I’m kind of information-seeking.
As I’ve already alluded, any doctor who treats a patient with obviously uncontrolled symptomatology and says “Nope, paper says 5, so that’s that” is just really not very good. That’s a flaw with the provider, not EBM. EBM works well as a guideline, but there will obviously be exceptions. I don’t know any advocate for EBM who doesn’t account for that likelihood.
Along those lines, I think using EBM in determining which services merit coverage is as good an approach as is likely to be found. Yes, it’s important to both pick appropriate outcomes to measure and to reevaluate periodically. These are both in keeping with EBM as I understand it. But, as I said somewhere upthread, there must be an efficient mechanism for obtaining coverage when the clinical situation warrants it. Under our current “system,” if you can call it that, this is often an onerous, tedious process. While I don’t have much confidence that the ACA will make it better, neither do I see any basis to fear it will get worse.
I don’t disagree with this. But you get people like Peter Orszag writing op-eds suggesting that we can make a health care “Grand Bargain” where we offer tort reform–but only for doctors who follow guidelines.
Does this seem likely to improve health care? It seems mad to me–you’re going to create a system where legal safe harbor is available to anyone who slavishly follows printed guidelines, and lost if you don’t. At our current state of knowledge and experience in implementing EBM, would it really be possible to write adequate guidelines? How do you offer safe harbor to “In my clinical judgement, she was still hypothyroid, even though her numbers were within standard ranges.” From what I know about the literature on EBM and the current state of good research on many conditions, it doesn’t seem like it would be. But I assure you that in Washington, it is entirely possible, in a different congress, that some idiot would read Orszag’s editorial and push it through.
In other words, it seems to me you’re saying “But properly done EBM doesn’t have anything to do with this,” and I’m saying, “but EBM is not necessarily going to be done properly and if it isn’t, it may well look like this.” You’re imagining an idealized version of IPAB which is extremely careful about using EBM to inform payment decisions. I’m imagining the version of IPAB implemented, well, the way things usually are in Washington, which is composed of political appointees and required to deliver service cuts on a very short deadline without pissing off too many really important constituencies. (Translation: do not expect that any service will be cut for breast cancer ever, no matter how expensive or worthless. Who can we cut? Well, what about thyroid patients? It won’t kill them and they probably don’t need it anyway, and they’re not well enough organized to get congress to override us.)
There are a lot of people in Washington who really believe, not that EBM can help us determine treatment standards, but that EBM is at a state where we can improve health and control costs merely by fully implementing it right now. These are people who are familiar with things like the checklist manifesto, and not so familiar with the A1C debacle, or indeed, the difficulty of doing statistically sound studies on an actual representative sample of human beings. (Genuine question: would a broad EBM regime have stopped the A1C mistake, or made it worse?) That group includes a lot of prominent journalists, advocates, congressional and administration staff members, and as far as I can tell, POTUS. It’s very definitely informing how they think about health care policy.
So I wasn’t trying to convince the EBM movement that what they’re doing is wrong, or won’t work–I think it’s great! Rather, I’m trying to convince a general audience that EBM is not some sort of policy ginseng that will solve all problems if you take a large enough dose, and that we actually have to be careful about how we implement it, because that general audience has what seem to me to be some very unrealistic ideas, and many of them are powerful enough to put those unrealistic ideas into practice.
Incidentally, rereading this comment–written when I was just waking up–it sounds like I’m saying “you naive fool” which is not what I meant at all. More like “it’s impossible to believe how screwed up policymaking is in DC until you’ve lived here a while and had a lot of casual chitchat with with the people in charge of it”.
Yes, just to repeat my own comment: The problem is not that docs will use an EBM system based on perfect knowledge. The problem is that the federal bureaucracy will use an EBM system based on what we know right now.
And that insurance companies do not and will not.
Of course, the worst thing about PPACA is that it sets an absolute ceiling on medical care, so that anything it doesn’t cover will be illegal.
“Translation: do not expect that any service will be cut for breast cancer ever, no matter how expensive or worthless.”
OTOH, Avastin.
(though clearly that is a different mechanism)
(which itself is sort of broken)
Well, you could argue that the service itself isn’t being cut–patients can still get chemotherapy for breast cancer. They just can’t get Avastin. Which tells everyone working on breast cancer meds to not bother, because we already have one thank you.
And then everyone complains that “big pharma just works on boner pills and happy juice”.