Pioneer or Profiteer? Daniel Amen and the use of SPECT imaging in clinical psychiatry
August 11, 2012 12:58 PM   Subscribe

Daniel Amen and the use of SPECT imaging in clinical psychiatry. Daniel Amen's clinics grossed $20 million last year, using SPECT imaging to tailor psychiatric treatments to individuals. The psychiatric establishment is skeptical: "'In my opinion, what he's doing is the modern equivalent of phrenology,' says Jeffrey Lieberman, APA president-elect, author of the textbook “Psychiatry” and chairman of Psychiatry at Columbia University College of Physicians and Surgeons."
posted by OmieWise (42 comments total) 10 users marked this as a favorite

 
Register Now

If you don't have a registered e-mail address with The Washington Post, create an account now. It's free and easy.

posted by availablelight at 1:25 PM on August 11, 2012


My local PBS station keeps running some awful special with this guy. I assumed he was a crank basically for this reason alone...
posted by BungaDunga at 1:27 PM on August 11, 2012


(That obstacle aside, this is an interesting topic--my impression has been that Dr Amen sells hope and certainty to desperate people in a medical field where both of those things can be hard to come by.)
posted by availablelight at 1:27 PM on August 11, 2012


This article seems to ask you to register after the second page; I suggest you click straight through to the print edition.
posted by koeselitz at 1:29 PM on August 11, 2012 [1 favorite]


Also, this is very interesting. It mentions an article by Daniel Carlat in Wired Magazine a few years ago - here it is: "Brain Scan As Mind Readers? Don't Believe The Hype."
posted by koeselitz at 1:29 PM on August 11, 2012


My local PBS station keeps running some awful special with this guy.

The FPP article links to a Salon article about it:

Four years ago, Robert Burton, the author and former associate chief of the department of neurosciences at the University of California at Mount Zion Hospital, wrote a harsh article on Salon.com about Amen’s work. The headline was “Brain Scam.” When recently told that Amen was still in business and grossed $20 million last year, Burton asked for the dollar figure to be repeated.

“Oh, my God,” he said. “Just oh, my God. At some point this gets to be obscene — that’s just my bias — but oh, my God.”

posted by availablelight at 1:30 PM on August 11, 2012


While I agree Amen oversells SPECT, this has all the marks of a turf war. APA stands by the DSM process, which we all (including the chief editor of the DSM-IV) know is broken, while Amen is dissecting entities that the DSM-III to 5 have always considered to be unitary. For example, Amen argues for six clusters of SPECT results in ADHD, corresponding de facto to six different kinds of ADHD -- where it gets messy is when he writes a popular book proclaiming behavioral correlates of those clusters/proto-diagnoses as diagnostic criteria proper.

I used to know a bipolar-autistic-epileptic-whoknowswhatelse that was being treated pro bono in the Amen Clinic in SF; the DSM process and APA is simply silent on those people, and very protective of traditional medical research (which in turn falls short of that platonic "scientific method").

In the larger scale, this is what you need to know. The APA is The Establishment, and while up to the DSM-IV The Establishment did great good, they've been behind the gross overselling of patented atypical antipsychotics over older drugs that work just as well without the potential irreversible side effects, and they've botched DSM-5 beyond repair. Amen is The Maverick. He's not an idiot, he's an MD and (unlike most research MDs, hung up on 'naturalistic observaton' or repetitive, futile 'evidence-based' methodologies) understands the impact of extensive, almost infinite data in medicine. Amen could be wrong in a number of scientic claims, possibly most of them, he doesn't even have peers to do the peer-review boogie with.
posted by syntaxfree at 1:30 PM on August 11, 2012 [7 favorites]


...Also, beyond the placebo effect discussed in the FPP article, there's the fact that frankly medicalizing still-stigmatized diagnoses (psychiatric) probably makes folks more comfortable and compliant with treatment (i.e. treating a "brain imbalance" rather than a "mood disorder" or "learning/conduct disability").
posted by availablelight at 1:36 PM on August 11, 2012


He scanned... boys who dated his daughters longer than four months.
posted by BungaDunga at 1:37 PM on August 11, 2012


syntaxfree: "He's not an idiot, he's an MD and (unlike most research MDs, hung up on 'naturalistic observaton' or repetitive, futile 'evidence-based' methodologies) understands the impact of extensive, almost infinite data in medicine."

Maybe you can explain exactly what is futule about evidence-based medicine. I can agree that the DSM is currently problematic, but it's arguable that that's because it's not evidence-based enough.

Also, the rest of your comment doesn't really respond to the heart of the linked article. Amen is an APA fellow. Most of the critics mentioned in the article don't seem to be connected to the APA at all.

posted by koeselitz at 1:41 PM on August 11, 2012 [4 favorites]


Maybe you can explain exactly what is futule about evidence-based medicine.


Two points, far from exhaustive of the problem. One, aggregating studies with very different methodologies, themselves only as good as the match between underlying construct and measurement instruments, generally outcome diaries or questionnaire-based inventories. T

Two, NNT. NNT is pretty much the only thing you can do to intelligibly use many studies as evidence of one proposition, but if you've ever looked at confidence intervals for NNT or done Monte Carlo simulations on the theoretical distributions of underlying parameters, it's all a rage comic. Your point estimate may be 8, with a CI from 2 to 64.

The ultimate problem with psychiatry -- and I'm a big defender -- is construct validity. Construct validity is taken to be a good marker for stratifying samples in first place, so all sample-based inference is funky, let alone aggregate sumaries.
posted by syntaxfree at 1:49 PM on August 11, 2012 [2 favorites]


The APA is The Establishment, and while up to the DSM-IV The Establishment did great good, they've been behind the gross overselling of patented atypical antipsychotics over older drugs that work just as well without the potential irreversible side effects, and they've botched DSM-5 beyond repair.

Even if the DSM is an over-glorified Dungeons & Dragons sourcebook, and the APA a trade association for risperidone wholesalers, that's not really a positive argument that Amen is doing anything more than taking people's horoscopes with a nuclear medicine study.
posted by strangely stunted trees at 2:11 PM on August 11, 2012 [13 favorites]


Even if the DSM is an over-glorified Dungeons & Dragons sourcebook

HA!
posted by shivohum at 2:13 PM on August 11, 2012 [1 favorite]


I was touched by the comments underneath the article, asserting about all the good that Dr. Amen has done for each commenter personally -- and yet, even so, none of the anecdotes prove anything about the validity of his treatment, which makes them all the more heartbreaking.
posted by Countess Elena at 2:25 PM on August 11, 2012 [1 favorite]


repetitive, futile 'evidence-based' methodologies

I will grant you that in psychology many constructs are not as valid as one may wish them to be but I'm wondering what methodologies you propose the medical community take up as an alternative to evidence-based ones? Anecdotal based?



he doesn't even have peers to do the peer-review boogie with.

This


The Brain Imaging Council, in fact, was intrigued by — if skeptical of — Amen’s work in the mid-1990s, remembers former president Michael Devous. He says the group offered to let Amen analyze a large series of SPECT scans of psychiatric patients and a control group.

“We promised that whatever the results we would publish them with Dr. Amen as a co-author,” Devous wrote in an e-mail. “If this was successful, it would provide him with a level of legitimacy he had not had. The offer was made over the phone, by me, and he declined.


Seems to suggest that there certainly are medical professionals who are willing to consider his technique in a peer reviewed context.
posted by sendai sleep master at 2:27 PM on August 11, 2012 [3 favorites]


The title of the post is Pioneer or Profiteer?

There is a possibility he is both.

I have one of Daniel Amen's books. Making a Good Brain Great. His claims are very bold but the sucker has got like 40 footnotes per chapter. Alas, most published research findings are false. But if you want to know what is the documentation for, say, ginko biloba as a cognitive enhancement, Amen's book is a place where you can find out what to look up in the library. It is the only book I own with this feature.

The most interesting thing he says in his book is that in one of his largest clinical studies (if I recall correct N was 500 or something like that) they had a questionnaire first screening: no brain disease, no brain injury, no drug addiction, no psychiatric diagnosis, &c. Then after that they had a SPECT screening to find "normal" brains for their study. Only 15% of the second population had "normal" enough brains to include in their study.

The second thing I remember is he said anybody who dates his daughter for more than three months he has to look at their SPECT. He did not mention if this requirement applies to his son, although I would infer it does not.
posted by bukvich at 2:31 PM on August 11, 2012


He's not an idiot, he's an MD...

As an MD (and perhaps an idiot) I can assure you the two terms are not mutually exclusive.
posted by TedW at 2:55 PM on August 11, 2012 [14 favorites]


Even if the DSM is an over-glorified Dungeons & Dragons sourcebook

Technically I believe it's the appendix to Clanbook: Malkavian.
posted by Pope Guilty at 3:10 PM on August 11, 2012 [5 favorites]


One other thing . . . the book I have by Amen is very well written and I have read it cover to cover at least three times. The last time I did so after I found a listing for him on Quackwatch (link) and I read it very closely on the lookout for fraud and quackery and what not. He hedges and qualifies carefully enough that I couldn't find anything explicit I could pin him down on. I would be very interested if a neurologist or psychiatrist could post a chapter and verse with the notation "this part is clearly messed up."
posted by bukvich at 3:20 PM on August 11, 2012


Anyone who knows anything about biological psychiatry knows that Amen is claiming to have found the holy grail. There is not a single biological psychiatrist who would not climb out of the grave to have valid and reliable biological markers for mental illness, and if those markers could be located in new patients with a simple brain scan, that would be fucking fabulous. There is nothing patented (by Amen) or unusual about the process he uses, and yet other psychiatrists say SPECT scans won't work as he says they will. That's pretty much all you need to know about whether or not his scans tell him anything useful for clinical treatment. The argument isn't about whether he's ahead of the game, it's about whether this thing that other psychiatrists desperately want, and could have if what Amen says is true, is really there for the taking.
posted by OmieWise at 4:03 PM on August 11, 2012 [8 favorites]


bukvich: “He hedges and qualifies carefully enough that I couldn't find anything explicit I could pin him down on. I would be very interested if a neurologist or psychiatrist could post a chapter and verse with the notation ‘this part is clearly messed up.’”

Well, I haven't read the book, so I guess I can't go chapter and verse, but – isn't it pretty clear where they're saying he messes up? Particularly the Wired article gives this explanation: it's that Amen seems to be assuming that results that have been shown with some general statistical validity can be used to diagnose individual patients.

I don't know crap about medicine, really, but I've had some discussions with a good friend who's a doctor. She has said before that she'd never, never diagnose a patient based on general statistical tendencies that are still not pinned down. You wouldn't unqualifiedly conclude that someone had a cancerous growth, for example, just because they had a certain rash that 70% of cancer patients have been shown to have.

I'm sure Amen hedges up and down, but when it comes to it mostly you have to hedge by saying things like 'studies show patients with X and Y brain patterns tend to be ADD,' or 'patients who suffer from depression tend to show less brain activity in this node or that node.' The fallacy is in the reliance on a tendency, which really isn't a basis for any claim of superior diagnostic power, and which might actually have no validity for any single patient at all.
posted by koeselitz at 4:16 PM on August 11, 2012


One reason why his claims are seen as bogus is because there is nothing in his process that suggests causality. He doesn't do controlled experiments, its all based on correlational evidence. The SPECT-derived neurophysiological criteria he says is a marker of say, ADHD, could be entirely epiphenomenal. If Amen sees certain brain areas as being implicated in a certain disease state, there is no way for him to know if dysfunction in some other network/area has some influence on the areas he is using for the diagnosis. This could lead to treatment of epiphenomena and completely ignore the source (treating symptoms, not the cause).

Also, there is huge debate in the field about the degree to which any measure of blood flow in the brain (fMRI, SPECT, PET, ASL, etc) is evidence of neuronal activity as opposed to other aspects of brain function (astrocytes, etc). Its also my understanding that Amen does a lot of resting state type scans, as opposed to task related scans. Analysis of resting state data is still on the cutting edge and its way too early for anyone to be able to draw reliable medical conclusions from it at the individual subject/patient level.

I personally would love for someone or some group to create better diagnostic procedures than the APA uses...their DSM is partially designed so that as many psychiatric drugs as possible can be prescribed. It was partially written by former big shots of pharmaceutical companies. BUT Amen's way is just as much, if not more, BS-ridden than APA's. You can profit from cultivating a maverick reputation, from offering some alternative to the status quo, even if your solution is not actually superior to the status quo. Based on what I know, Amen's clinic is basically selling snake oil. But unfortunately, in his industry, so is most everyone else.
posted by captain cosine at 5:08 PM on August 11, 2012 [4 favorites]


The APA is The Establishment, and while up to the DSM-IV The Establishment did great good, they've been behind the gross overselling of patented atypical antipsychotics over older drugs that work just as well without the potential irreversible side effects,

There's potential irreversible side effects with both the typical and atypical classes of antipsychotic. The extrapyramidal side effects (tremor, parkinsonism) are more marked with the older drugs, metabolic disturbance (diabetes) with the newer. Saying they work 'just as well' doesn't mean much if people stop taking their drugs after developing tremors.
posted by chiquitita at 5:59 PM on August 11, 2012


Also, there is huge debate in the field about the degree to which any measure of blood flow in the brain (fMRI, SPECT, PET, ASL, etc) is evidence of neuronal activity as opposed to other aspects of brain function (astrocytes, etc). Its also my understanding that Amen does a lot of resting state type scans, as opposed to task related scans. Analysis of resting state data is still on the cutting edge and its way too early for anyone to be able to draw reliable medical conclusions from it at the individual subject/patient level.

ASL is one kind of image acquisition procedure that falls under the fMRI umbrella. ASL is often contrasted with imaging that uses BOLD signal. The debate about whether the changes in the brain that are observable using fMRI represent actual activity (as opposed to, say, recording from implantable electrodes, etc.) are difficult to resolve, but I think they've safely moved into the background, and most researchers feel confident about this assumption. I also think it's safe to take resting state activation from the "cutting edge" box by now. Recording the brain's activity while it's not engaged in any explicit tasks is not a difficult idea. If you're going to look for correlates between brain activity and mental disorders, you might as well look at resting state activation.

Finally, how much does it really matter if the markers he's purporting to have identified are epiphenomenal or not? A runny nose is not the cause of a cold, but it's a reliable sign of one. If there was a pattern of activation one could point to and say, "There! A sign of clinical depression!", that would still be enormously valuable.

But yeah, validation and peer review are crucial steps, and it's a huge red flag that this guy has chosen making money hand over fist instead of letting his method speak for itself in the scientific arena.
posted by Nomyte at 6:08 PM on August 11, 2012


Finally, how much does it really matter if the markers he's purporting to have identified are epiphenomenal or not? A runny nose is not the cause of a cold, but it's a reliable sign of one.

It matters because you would not want to just get rid of the mucus if you had a runny nose. I mean, decongestants are great temporary relief, but what you really want is for your white blood cells to fight the infection. In many cases, the appearance of symptoms is evidence that your immune system is actually working properly. In terms of the brain, if you find an epiphenomenal "sign" of clinical depression, and you specifically treat that sign, who is to say if you will even come close to treating the causes of that sign?

I also think it's safe to take resting state activation from the "cutting edge" box by now

What makes you think this? I follow the default mode network and resting state literature fairly closely and people can't even agree on how to analyze the data (global signal regression? influence of head motion on measures of functional connectivity?), much less how to interpret its nuances in individual subjects. There is indeed a lot of great research being done in this area, but the most reliable conclusions are from group averages from massive datasets, not individuals.
posted by captain cosine at 6:27 PM on August 11, 2012 [1 favorite]


Aside from the complete lack of published, reviewed data, the problem is that he's saying "I see something in the brain" therefore "It's not your fault." One does not follow the other. Your brain *must* change in order for you to learn or to have any experience at all— and so seeing some difference in the brain doesn't mean that a) the difference causes your different behavior and it's not your fault or b) the difference means "damage" or "pathology."

I imagine he gets his pumped up success rates (if they are actually real and not just the usual "not counting the people who drop out" gambit used in addiction treatment) because people with psychiatric problems (particularly addictions) are constantly being blamed for their problems and being told "it's your fault," and anyone who tells them "it's not all in your head" will automatically be seen as wonderfully understanding and compassionate. And, indeed, it's true that addictions and other psychiatric disorders *can* impair will in ways that are distressing. But you don't need pretty brain pictures to see that.

However, an interesting study also showed that if you add neurobabble about imaging to virtually any explanation, people find it much more credible. So, add these two together and you get a powerful recipe for a strong placebo effect that may indeed help people without having anything to do with him knowing anything about anything.

Moreover, just because the man can write a well-referenced book doesn't mean that the references actually support his claims. The fact that he's making causal claims about imaging shows that the doesn't understand the references or is at least deliberately ignoring the fact that it's actually impossible to draw causal influences from them.

What's also missing is the perspective of people who receive the news that their addiction has "damaged" their brain (like he provides on Dr. Drew) and then go out and say "It's too late; may as well keep getting high."

There are plenty of severe problems with mainstream psychiatry but you can oppose mainstream psychiatry and still be wildly wrong and highly profitable (see under: Scientology).
posted by Maias at 6:28 PM on August 11, 2012 [3 favorites]


Also, where's he getting his normative data from? I imagine very few people show up to his clinic without any problems and if he's just using himself and his friends as "normal," well, you can see the problem there that isn't just small sample size.
posted by Maias at 6:30 PM on August 11, 2012 [1 favorite]


Why is he using SPECT rather than fMRI?

Why hasn't he published (or partnered with other researchers to publish) his findings in a way that other researchers might replicate?
posted by anotherpanacea at 6:45 PM on August 11, 2012


“SPECT can specifically help people with ADHD [attention deficit hyperactivity disorder]. ... SPECT can specifically help people with anxiety and depression. ... SPECT can specifically help people overcome marital conflict. ... SPECT can specifically help people age better. ... SPECT can specifically help people with weight issues. ...”

a scientology registrar once said almost that exact same thing to me.
posted by quonsar II: smock fishpants and the temple of foon at 6:54 PM on August 11, 2012 [2 favorites]


As Quackwach notes, he has not established that his patients have derived any benefit from the scan-based therapy, compared to patients treated without it. That seems pretty devastating to any claim that this stuff is clinically established. Quackwatch are pretty circumspect in their language, actually; it seems there was some legal to-and-fro. They do say that that, at minimum, he should tell people that his treatment is experimental.
posted by thelonius at 7:18 PM on August 11, 2012


syntaxfree: APA stands by the DSM process, which we all (including the chief editor of the DSM-IV) know is broken

I am a psychiatrist. Like many successful things that have stuck around for a while, the DSM is broken in places, but it works well enough in others. Yes, it proceeds from nosology rather than etiology, but this approach is at the core of a whole lot of medicine. Many of the common classic psychiatric DSM diagnoses with large sample sizes exhibit inter-rater congruency with sensitivities and specificities in the 80-90% range. Which is surprisingly about as much as you can count on for other some other diagnostic modalities, such as CT or MRI imaging for abdominal, intra-cranial or intra-thoracic syndromes.

SPECT for psych diagnoses? Not so much.
posted by meehawl at 11:48 PM on August 11, 2012 [1 favorite]


I think psychiatry is incidental to the Amen controversy. I had mentioned the rotten DSM-5 process and the problems with current methodology for context, but the main point seems to deserve more elaboration.

There has never been one golden scientific method. There isn't one to be found in Tha Bible or in the Epic of Gilgamesh. Renaissance astronomers seem to have willed that there were seven planets, because, well, seven is such a pretty number. Later, science became strongly inductive and verificationist; quantum mechanics arose from the black-body radiation problem, which in turn arose from the problem of inductively find a law for EM radiation curves or whatever.

The problem is -- at the same time biology and physics and, yes, economics, were making great strides, so were phrenology, psychoanalysis and various methods of runic or spirit divination. This became such an issue that several pseudoscientific theories were held in stride until the mid-20th century, such as psychoanalysis and darwinian evolution. (Note well: darwinian evolution was made scientific in the mid-20th century by the combination of big loads of evidence and good theories for mutation and selection).

But no one in science was bothered bu inductivism, because runic divination was a nuisance limited to early nazis, theosophy was stationed far in India and everyone had the common sense and cartesian skepticism to distinguish their Freud from their Blavatsky. But what's this, this common sense? I could digress into Kant's sensus communis logicus, but it's not especially helpful, and actually shows the limits of letting people's understanding and intuition to steer them away from quackery.

One man was bothered by Freud, though. This man saw that psychoanalysis and theosophy weren't all that different. This was Karl Popper, and he brought an entirely different concept of "scientific method" built not around verification of theories, but falsification of hypotheses. And of course this was just philosophy and not taken immediately by scientists (c.f. pre-Samuelson economics) but it's the concept to which we refer now when we speak of the scientific method.

Here's how falsificationism works. You have a theory, and find a few propositions that, if false, disprove your theory. Then you attempt to statistically show that it's unlikely that they're false. The problem starts when you start digging on "propositions". What's their extensional domain? Are we matching well the extensional domain of propositions with the sample designs in statistical analysis? How do you do this when the very existence of these domains is only indirectly known?

You can do two things: you can built a Great Big System that can crumble on the falsification of one hypothesis, or you can use construct validity. The Standard Model in physics is almost a Great Big System for the fundamental forces, except at the end it turned to be verificationist (i.e. see if the expected Higgs boson shows up). Psychiatry uses construct validity methodologies, some contrived and suspect, some simpler and quite ingenious, such as inter-rater reliability.

So some construct is taken to be valid -- even though it's been clear to "kantian" common sense since naturalistic observations by Kraepelin and contemporaries -- and statistical tests are made on top of it. But how are you really sure that the extensional domain of your studies matches the extensional domain of true constructs? There are clearly bipolars, cleared from epilepsies, who respond to Topamax, and bipolars who don't. Does Topamax make a dent in the whole idea of "bipolar"? I'm inclined to say yes, but the current process doesn't accomodate that.

On the other hand, there's the data deluge in science. Philosophers and scientists occasionally make into the Wireds and Gizmodos of this world to speak of the theory-free science of the future. While this hasn't been clearly articulated yet, Big Data changes everything. And response to Topamax is a datum. The plural of anecdote isn't data, but the plural of datum is.

So dr. Amen is in a position to use massive amounts of data to check for clusters on his own, unshackled by tradition and to treat patients in a highly individualized way (try to find something that works for an autistic-bipolar-epileptic coming off a toxoplasma-induced bout of agoraphobia). This is by all means an excellent, if speculative and "unscientific" by popperian standards, endeavor. This is how research should be done. The autistic-bipolar-epileptic dude I know founded a forum and met others like him, dubbed this "uberspazz syndome" and would remain lost in morgellons territory if there weren't "fluid" corners of the medical establishment like the Amen Clinic.

I don't like it when Amen goes pop. I don't like it when he writes popular books and references his clustering studies as new diagnostics and gives behavioral correlates as diagnostic criteria. It gives hope to people -- people who aren't even acknowledged by the glacial (and misguided) pace of construct validity -- but not all hope is good. I think he overreaches his grasp and overmarkets. In that I can't disagree. But it's one thing to denounce the excesses of marketing-driven medicine, another to validate the established method and negate the potential of emerging data-driven methodologies like the Amen approach.
posted by syntaxfree at 9:22 AM on August 12, 2012 [1 favorite]


> So dr. Amen is in a position to use massive amounts of data to check for clusters on his own, unshackled by tradition and to treat patients in a highly individualized way (try to find something that works for an autistic-bipolar-epileptic coming off a toxoplasma-induced bout of agoraphobia). This is by all means an excellent, if speculative and "unscientific" by popperian standards, endeavor.

While I am glad that he is doing this and sharing his results and ideas in accessible books that retail for $13.95, your "by all means" description is way over the top. It would be a lot better for all of us if competent careful research is performed by researchers and clinicians have efficient reliable processes to get the latest-and-greatest treatments out to the people who need them. The problem with psychiatry is they are mostly agents of the drug companies. If you go to the American Psychiatric Association annual meeting, it is little else but a pharmaceutical trade show. The quackwatch guy asserting that he has objective science on his side and Doctor Amen is a quack is just not credible. Pharmaceutical corporations employ scientists, but they do science as part of their overhead, not as a core competency. What they do mainly is make money.
posted by bukvich at 10:20 AM on August 12, 2012


Maybe I should add is that there's nothing about "maverick" that says "he's right, he's the Spirit of History riding on a horse", as Hegel is said to have said of Napoleon. All I'm saying, the DSM process is rotten, "evidence based' medicine is poetry with numbers and ultimately we're relying on naturalistic observation and inter-rater reliability, both of which are highly subject to sociological biases. Amen is bringing on something new -- he's awake to the meaning and potential of the data deluge.

If he has refused peer review, let him. SPECT is not patented, go and try it. Stop trying to make it about one guy, the real issue is the data-led scientific revolution in course. Far from complete or even mature -- on the contrary, in real need of expert guidance and smart people generally working from a data-centric standpoint.

Finally, I should note that the DSM process is rotten re: DSM-5. DSM-IV is flawed but saved countless lives. Contrary to what many uniformed commenters seem to think, mental illnesses of biological etiology and treatment do exist, often become severely life-impairing and need to be addressed. It's evident that pharma has some pull -- manic depressives wouldn't otherwise be started on Sapphris before trying lithium, or lamotrigine, or quetiapine, or even the valproates.

But let's not throw out the fucking baby with the bathwater here. There's a fine line before being critical or even dismissive of a process like the appearance of new diagnoses, and being dimissive of the discipline as a whole. I've once said here that lithium saves lives, and got "no it doesn't, I was like a zombie for a year and then quit it and I'm happy" in return, which isn't even an answer. If all you know about a subject is "meh, buncha quacks", you don't get an opinion. How many suicides have you indirectly caused by pooh-pooh-ing lithium?
posted by syntaxfree at 11:18 AM on August 12, 2012


Amen is bringing on something new -- he's awake to the meaning and potential of the data deluge.

If he has refused peer review, let him. SPECT is not patented, go and try it. Stop trying to make it about one guy, the real issue is the data-led scientific revolution in course.


Yeah, this is why what you've written is essentially meaningless.* Psychiatry would jettison the diagnostic criteria in the DSM for valid and reliable markers of mental illness in the time it takes me to write this comment. It's precisely because psychiatry wants what Amen claims to have discovered so very badly, and precisely because anyone can replicate his "experiments," that we know that what he's doing is bunk. All of your concern for the philosophy of medial knowledge is completely ancillary to the question of whether Amen is producing anything meaningful because it is so central to the concerns psychiatry has about its own validity. If psychiatry could embrace Amen's work it would rush to do so yet the very people who would love love love to have what Amen claims to be getting are saying that there is not there there. There is no "Big DSM" (your understanding of the relative importance of the DSM to psychiatry seems not very nuanced) that would have anything but praise for valid and reliable brain scans for mental disorders.

*Sure, there is plenty to write and think about in re the history and philosophy of medical knowledge, but it's all completely beside the point with respect to whether or not Amen is doing something meaningful and useful. You're simply wrong that there is a vested interest in psychiatry that would reject Amen's work, if valid, out of hand. The opposite is most assuredly true, and that his work is nonetheless being rejected obviates any discussions about the philosophy of science as pertinent to his claims.
posted by OmieWise at 11:38 AM on August 12, 2012 [3 favorites]


syntaxfree: “‘evidence based’ medicine is poetry with numbers”

Again, I call bullshit. Evidence-based medicine is the only kind of medicine that makes sense, that has external validity. The fact that evidence-based medicine has not thus far been possible within psychiatry does not invalidate evidence-based medicine as a whole. In fact, I note that Dr Amen's claim is that he has brought evidence-based medicine into psychiatry – this has been a dream of many for decades – and the claim of his detractors is that he has done no such thing.

“If he has refused peer review, let him. SPECT is not patented, go and try it.”

People have tried it. They found only very rough predictive validity in it. That kind of vagueness doesn't justify the fact that this guy is publicizing himself all over television and in popular books and pushing this treatment on every patient that walks through the door, charging a hefty $3000 a head. This is not valid evidence-gathering or research. It's strongly suggestive of a major ethical lapse, in fact.
posted by koeselitz at 11:42 AM on August 12, 2012



The fact that evidence-based medicine has not thus far been possible within psychiatry does not invalidate evidence-based medicine as a whole.


Point taken. My beef with EBM is that dissimilar studies are aggregated and the few metrics that can be used on such heterogeneous literature are statistically very weak. Heterogeneity is more of a problem in psychiatry than anywhere else, because outcomes are measured very differently, construct homogeneity can't be assured, etc. I can't speak for whether EBM would make sense in bariatrics or something.


Evidence-based medicine is the only kind of medicine that makes sense,


... Actually, the bigger problem with EBM, and why data-led science may be the only way out, is that you can only get answers on the questions you asked. This doesn't make for scientific progress. Again, about half of the bipolar folk I know from fora respond to Topamax, and half don't (I do). Pool them together, and you get no effect. This is straight ouf of the envelope theorem, so to speak: your outcome estimate will be at the constraint estimate.

I see how EBM may be, in principle, the best way to develop first-round treatment algorithms, but science isn't a round-track bicycle race.

Maybe this is too abstract, the Topamax problem. But going back to bariatrics, maybe the problem is that you put people on a ketogenic diet and there's no response because you really can't control whether they're reaching ketosis in an outpatient setting. Maybe half of them are able to go into ketosis in N days and half takes N+k. This sensibility to study details and limitations is lost in meta-analysis. And calling meta-analysis "evidence-based medicine" is a joke -- this is basic statistics: the variance on unknown factors doesn't cancel out.

I could go full Lakatos mode on and say that studies reflect what authors want them to -- that's why they set it up in first place, to get a result, not to confirm the null. But I don't need to. The peer-reviewed, selected-publication model is still the default mode to deal with the various contingencies of science in the real world (keeping quacks out/managing reputation, sorting people for tenure, etc.) but it's not a given. It's not in Tha Bible, and we shouldn't give the literature more weight than the data itself.

I'll believe in "evidence-based medicine" when it's built on aggregated raw data, not meta-analysis.
posted by syntaxfree at 12:21 PM on August 12, 2012


Again, about half of the bipolar folk I know from fora respond to Topamax, and half don't (I do). Pool them together, and you get no effect. This is straight ouf of the envelope theorem, so to speak: your outcome estimate will be at the constraint estimate.

This sounds strikingly like anecdote, and not evidence.
posted by Pope Guilty at 2:54 PM on August 12, 2012


Time for the new film Burning Wilhelm Reich to sort all this out.
posted by Twang at 3:38 PM on August 12, 2012


The elevation of Wilhelm Reich, a quack and crank if ever there was one, by otherwise-reasonable people is sad. Was he ill-treated by the government? Yeah, I'd say so. Was he a quack peddling nonsense for money? Oh my, yes.
posted by Pope Guilty at 3:42 PM on August 12, 2012


koeselitz: Evidence-based medicine is the only kind of medicine that makes sense, that has external validity. The fact that evidence-based medicine has not thus far been possible within psychiatry

Something like this or this or even this might come as a surprise to you. EBM is nice when it's informative, and advances the care of a patient, and involves results that have not been overly skewed by various kinds of biases or interested parties. However, the key thing to remember is that medicine is not science. It uses science as one of its tools, and for many people invoking SCIENCE does activate a sense of "external validity" that can be therapeutic. However, I can find many, say, population-intervention-control-outcome studies that tell me interesting things, but in the end if something has only statistical significance without clinical significance then it's not useful to me or my patient. Many of the studies that get repeated ad nauseam about, say, cat parasites and behaviour fall into this: weakly statistically significant for some traits yet of limited positive or negative predictive value for target populations. No treatment interventions enacted. No way to evaluate NNH or NNT. Science? Yes? EBM? Sometimes. Helpful? Not really.
posted by meehawl at 4:52 PM on August 12, 2012


People are taking Popper-style falsificiation seriously? He's supposed to be a punching bag before you go onto real philosophy of science, after you make it out of the 17th century hell that is empiricism.
posted by Veritron at 4:50 AM on August 13, 2012


« Older Racetrack is a game with very simple rules which n...  |  Dog Steal Cabbage... Newer »


This thread has been archived and is closed to new comments