Beware Mind Drugs

One in eight Americans take prescribed mind drugs, which probably hurt on average (vs. talking cures):

I first took a close look at treatments for mental illness 15 years ago while researching an article for Scientific American. At the time, sales of a new class of antidepressants, … SSRI’s, were booming. … Clinical trials told a different story. SSRI’s are no more effective than two older classes of antidepressants. … Antidepressants as a whole were not more effective than so-called talking cures. … According to some investigators, treatments for depression and other common ailments work—if they do work—by harnessing the placebo effect. …

In retrospect, my critique of modern psychiatry was probably too mild. According to Anatomy of an Epidemic … by … Robert Whitaker, psychiatry has not only failed to progress but may now be harming many of those it purports to help. …

As recently as the 1950s, Whitaker contends, the four major mental disorders—depression, anxiety disorder, bipolar disorder, and schizophrenia—often manifested as episodic and “self limiting”; that is, most people simply got better over time. Severe, chronic mental illness was viewed as relatively rare. But over the past few decades the proportion of Americans diagnosed with mental illness has skyrocketed. … One in eight Americans, including children and even toddlers, is now taking a psychotropic medication. …

Whitaker compiles anecdotal and clinical evidence that when patients stop taking SSRI’s, they often experience depression more severe than what drove them to seek treatment. A multination report by the World Health Organization in 1998 associated long-term antidepressant usage with a higher rather than a lower risk of long-term depression. …

Before the introduction of Thorazine in the 1950s, Whitaker asserts, almost two-thirds of the patients hospitalized for an initial episode of schizophrenia were released within a year, and most of this group did not require subsequent hospitalization. Over the past half-century, the rate of schizophrenia-related disability has grown by a factor of four, and schizophrenia has come to be seen as a largely chronic, degenerative disease. A decades-long study by the World Health Organization found that schizophrenic patients fared better in poor nations, such as Nigeria and India, where antipsychotics are sparingly prescribed. …

Beginning in the 1970s, Harrow tracked a group of 64 newly diagnosed schizophrenics. Forty percent of the nonmedicated patients recovered—meaning that they could become self-supporting—versus 5 percent of those who were medicated. … Electroconvulsive therapy … fell out of favor in the 1970s, in part because of its negative portrayal in the 1975 film One Flew Over the Cuckoo’s Nest, and yet about 100,000 Americans a year still receive ECT. … virtually everyone who receives electroconvulsive therapy relapses within a year without further treatment. (more)

Added 1p: I’ve blogged before on antidepressants as placebos.

Added 21Sept: Yvian has convinced me to doubt the claim above I had found most interesting, that schizophrenia changed from a temporary to a chronic condition. So now I doubt this author in general as a source.

GD Star Rating
loading...
Tagged as:
Trackback URL:
  • Yvain

    You know, every time you’ve posted something about medicine, it’s had glaring errors or biases or other things that make it worse than useless. Doesn’t the Outside View suggest by now that no matter how shiny the newest pop sci article on medicine might seem, maybe it also has glaring errors and biases?

    *sigh* Here goes.

    “SSRIs are no more effective than two older classes of antidepressants”

    No one says they’re more effective. What people say is that they have fewer side effects. Tricyclics can stop your heart and MAO inhibitors can cause you to go into potentially fatal hypertensive crisis for silly reasons like eating too much cheese. SSRIs have very few serious side effects, which is the only reason they are preferred over the other drugs. This guy is attacking a straw man.

    “Antidepressants as a whole were not more effective than so-called talking cures, whether cognitive behavioral therapy or even old-fashioned Freudian psychoanalysis.”

    It’s true that antidepressants don’t work better than some psychotherapies (they have been shown to work better than others). Antidepressants are often prescribed in conjunction with psychotherapy, because therapy+medicines have been proven to work better than either alone. They’re also prescribed to people who don’t want therapy, since eg Freudian therapy can require sessions two to three times per week for years to decades costing hundreds of dollars each session. It’s easy to criticize people for not being “pure” and pursuing non-pharmaceutical solutions first, but many of the depressed are low-income people who may not have the opportunity to pursue therapy, or who may think it carries a stigma. Most good psychiatrists will offer people a choice between drugs and therapy, but many people do choose the drugs.

    “According to some investigators, treatments for depression and other common ailments work—if they do work—by harnessing the placebo effect, the tendency of a patient’s expectation of improvement to become self-fulfilling.”

    The beloved “according to some…”. According to some biologists, evolution is false! More seriously, yes, there was a big metaanalysis a while back that everyone of this ilk latched onto because it showed no difference in people with mild depression (it showed significant difference in people with severe depression, which this guy conveniently forgets to mention). Although this provides some reason to reconsider antidepressant use in people who are only mildly depressed, there are other factors that complicate matters. First, some studies have shown the same for strong opiates like OxyContin – ie no better than placebo – which suggests it’s more of a matter of large metaanalyses failing to show effects in cases where people get different drug responses. Second, all studies show antidepressant/placebo being a huge benefit over nothing at all, and the legal system doesn’t currently support ability to prescribe placebo.

    Even some pop science can get this pretty much right. To quote TIME: “In practice, only about one-third of patients with depression will be helped by the first drug they try. Many people improve on their own or with talk therapy. About 8% to 14% patients will respond badly to medication, experiencing increased thoughts of suicide after taking it, a response that appears with equal frequency in placebo takers. So the question becomes not whether antidepressants work better than placebo, but how they will work, in whom and under what conditions.”

    “But over the past few decades the proportion of Americans diagnosed with mental illness has skyrocketed. Since 1987, the percentage of the population receiving federal disability payments for mental illness has more than doubled; among children under the age of 18, the percentage has grown by a factor of 35.”

    And cardiovascular disease has grown at about the same rate. It must be those evil cardiologists, out to delude others!

    There are a couple of plausible explanations. First is incentives; a whole lot of things can affect federal disability payments for a disease beyond the prevalence of the disease. Second is lifestyle; we know that urbanization, drug use, stress, broken families, unemployment, and poor diet all increase mental disease. Third is cost-benefit analysis. If the only option for schizophrenics is sticking them in a horrible asylum and giving them older-generation antipsychotics that would leave them gibbering wrecks, psychiatrists are going to be extremely reluctant to diagnose people with schizophrenia. Now that schizophrenia can be managed on an outpatient basis by relatively successful drugs, they’re going to be more likely to diagnose it. We see the same thing with non-psychiatric illnesses – no one used to talk about hypertension, because the only known antihypertensives had side effects worse than the disease. Now that we have better antihypertensives, we’re diagnosing anyone with even mildly raised blood pressure because it’s so easy and useful to treat.

    “This epidemic has coincided, paradoxically, with a surge in prescriptions for psychiatric drugs.”

    It’s paradoxical that when an illness becomes more common, prescriptions of medication for the illness also become more common?

    “Whitaker acknowledges that antidepressants and other psychiatric medications often provide short-term relief, which explains why so many physicians and patients believe so fervently in the drugs’ benefits.”

    I love how one journalist with no relevant education and an axe to grind is quoting another journalist with no relevant education and an axe to grind as his source. And now Robin quotes the second journalist! The circle is now complete!

    “Whitaker compiles anecdotal and clinical evidence that when patients stop taking SSRI’s, they often experience depression more severe than what drove them to seek treatment.”

    He acts like this is some amazing discovery Whitaker “compiled”, but he could have just looked up “rebound effect” on Wikipedia. All sorts of drugs from all sorts of areas of medicine work this way, as anyone who’s ever had an “analgesia headache” from going off painkillers too quickly knows. Psychiatrists try to minimize medication rebound by tapering people off medications slowly, by not withdrawing people off antidepressants until they’re sufficiently recovered that the rebound effect won’t push them back into another depressive episode, and, in cases where the person has a long history of depression that’s not going away, just keeping them on the SSRIs for life.

    “A multination report by the World Health Organization in 1998 associated long-term antidepressant usage with a higher rather than a lower risk of long-term depression.”

    I’m pretty sure a hospital audit would associate long-term chemotherapy use with a higher, rather than a lower, risk of cancer. Asking why would be a good question for the high school statistics class that this author clearly never took.

    “SSRI’s cause a wide range of side effects, including insomnia, sexual dysfunction, apathy, suicidal impulses, and mania.”

    Placebos also cause many of these same side effects if patients are told they’re “antidepressants”. The point is that studies have shown that suicidal impulses, apathy, etc. go down ON NET on antidepressants.

    “Whitaker’s analysis of treatments for schizophrenia is especially disturbing. Antipsychotics, from Thorazine to successors like Zyprexa, cause weight gain, physical tremors (called tardive dyskinesia) and, according to some studies, cognitive decline and brain shrinkage. Before the introduction of Thorazine in the 1950s, Whitaker asserts, almost two-thirds of the patients hospitalized for an initial episode of schizophrenia were released within a year, and most of this group did not require subsequent hospitalization.”

    Thorazine is no longer widely used. The link between new drugs that completely change the standards for hospitalization and re-hospitalization rates would also be a fun exercise for a high school statistics class to look into. And studies show that the longer a schizophrenic goes without receiving psychiatric treatment, the worse the long-term prognosis.

    “A decades-long study by the World Health Organization found that schizophrenic patients fared better in poor nations, such as Nigeria and India, where antipsychotics are sparingly prescribed, than in wealthier regions such as the United States and Europe.”

    And, of course, there’s no difference between India and America besides rates of antipsychotic prescriptions. Clearly, antipsychotics also decrease rates of Hinduism and of breaking out into large musical numbers.

    “Harrow theorized that those who were heavily medicated were sicker to begin with, but Whitaker suggests that the medications may be making some patients sicker.”

    Study that contradicts common sense and all other studies. Obvious potential bias. Person who conducted the study says it shouldn’t be misinterpreted because of obvious bias. Whitaker says “No, no, please ignore him and misinterpret it”.

    Baaaasically, all I see in this article is one huge case of “cum hoc ergo propter hoc”. Psychiatric disease is increasing, and psychiatric medication is increasing, so clearly psychiatric medication causes psychiatric disease! There are a few straw men and a few blatant misunderstandings of what psychiatry is supposed to do, but mostly it’s just correlation = causation all the way down.

    • http://hanson.gmu.edu Robin Hanson

      The most interesting parts to me were the cross-national comparisons, which you dismiss, and:

      Before the introduction of Thorazine in the 1950s, Whitaker asserts, almost two-thirds of the patients hospitalized for an initial episode of schizophrenia were released within a year, and most of this group did not require subsequent hospitalization. Over the past half-century, … schizophrenia has come to be seen as a largely chronic, degenerative disease.

      • http://jonathan.graehl.org Jonathan Graehl

        I agree that those are both interesting suggestive evidence (perhaps Thorazine is already accepted as harmful and no longer prescribed, as Yvain claims; still, it’s weak evidence to suspect today’s psych-med practices). Yvain’s reply to those two bits of evidence is just the generic “maybe some other difference is causing it” – agreed; it’s only weak evidence.

        I had the same doubt as Yvain about the 1998 WHO study that may have just been making the point that (long-term) antidepressant use is correlated with depression – it’s too close in wording to a ridiculously vacuous claim (as Yvain lampoons it) for the writer to not clarify how it was non-trivial.

        For some reason I trusted too much the rest of the quoted material (perhaps because I wrongly assume Robin is endorsing the claims he quotes) in spite of this clear warning. I appreciate Yvain’s deeper examination that revealed other problems.

      • Dave

        The part about Africans doing better is old information and may have been true even before anti- psychotics. Africans were are said to act crazier than Western schizophrenics but were not locked up and were thought of a being kind of magic.
        The point about reduced hospitalization seems askance,as most of the insane asylums were shut down because of the efficacy of anti-psychotic drugs. I wouldn’t jump on one guy’s band wagon just because he is counter cultural. Interesting, but will it go anywhere.Most of these kind of things don’t.

      • Yvain

        Cross-national comparisons: your article talks of an “epidemic” of eg schizophrenia in the developed world which it blames on psych medications, but all my sources (see for example Eaton and Chen 2007) disagree. Incidence of schizophrenia per 1000 people per year is .1 in Honolulu (the only US city that made it onto the study), .15 in Manchester (representative British city), .05 in Vancouver (representative Canadian city), compared to .35 in rural Chandigarh, India, .6 in Madras, and .45 in Brazil. Here the incidence of schizophrenia appears to be up to three times higher in the developing world than in the developed.

        Your article’s point on greater recovery rates in the developing world is well-taken and does have real data in support of it, but it’s still controversial and the latest studies suggest it may be an artifact. In particular, if the worst schizophrenics die in the developing world, but survive in the developed world, then we will find that on average, living schizophrenics in the developing world have milder disease. Studies that control for this fail to find the same effect, or get mixed results. See Cohen, Patel, Thara, and Gureje: “Questioning an Axiom: Better Prognosis for Schizophrenia in the Developing World?”

        There are also other major differences in schizophrenia between developed and developing nations: catatonic schizophrenia is ten times more common in developing nations, hebephrenic is three times more common, and acute onset is twice as common (all as percent of total cases). This means that there are serious epidemiological differences between developing- and developed- world schizophrenia from the onset of the disease, as would be expected in a psychiatric disease across two wildly different cultures.

        As I said in my previous post, there are many, many diseases which differ between developed and developing countries in currently mysterious ways. The multiple sclerosis incidence rate is two hundred times higher in Canada than in Zimbabwe. Type 1 diabetes (the type supposedly not related to diet) is forty times higher in Finland than in Tanzania. Diseases that can be directly related to diet, like heart disease or stroke, are even more dramatic.

        In the course of all this, to take one disease that *may* differ between developed and developing countries, and blame it on medication without any evidence, is wildly jumping the gun.

        A more reasonable suggestion would be to look at different populations within a country, which brings us to your quoted portion, about Harrow and Thorazine. You say:

        “Before the introduction of Thorazine in the 1950s, Whitaker asserts, almost two-thirds of the patients hospitalized for an initial episode of schizophrenia were released within a year, and most of this group did not require subsequent hospitalization. Over the past half-century, … schizophrenia has come to be seen as a largely chronic, degenerative disease. ”

        But the numbers tell a different story. From the Wikipedia article on chlorpromazine (the chemical name for thorazine):

        “The development and use of antipsychotic drugs like chlorpromazine was one of the forces that propelled deinstitutionalization, the systematic removal of people with severe mental illness from institutions like psychiatric hospitals and their reintegration into the community[citation needed]. In 1955 there were 558,922 resident patients in American state and county psychiatric hospitals. By 1970, the number dropped to 337,619; by 1980 to 150,000, and by 1990 between 110,000 and 120,000 patients.”

        Then there’s the Harrow study cited about the differential recovery rates of schizophrenics. This study is not randomized (!!!!!) That means that they just found schizophrenics who already were (or weren’t) on antipsychotics, and included them in the study. Why would schizophrenics not be on medication? The obvious reason would obviously be because their schizophrenia isn’t very bad and they don’t need it. In this case, the obvious reason is the correct reason. According to the text of Harrow’s study, which I have in front of me:

        “The results suggest that the subgroup of schizophrenia patients not on medications was different in terms of being a self-selected group having better earlier prognostic and developmental potential.”

        The only reason not to shrug and forget all about this study is that there are very few good long-term studies on schizophrenia, so someone should probably replicate it with better selection criteria. It does not seem impossible to me that there might be a long-term detrimental effect of “typical antipsychotics” like the ones in Harrow’s study, although it should be noted that modern “atypical antipsychotics” are completely different drugs with a completely different mode of action and it would be surprising if by a strange coincidence they ALSO made the disease worse. However, uncertainty of long-term effects aside, there are lots of good shorter-term studies, which show (quoting Dixon, Lehman, and Levine):

        “The Baldessarini et al. (1990) and Davis et al. (1993) reviews find overwhelming evidence that conventional antipsychotic agents reduce the risk of relapse of the positive symptoms of schizophrenia. Based on 44 placebo-controlled studies of antipsychotics with a total of 3,939 subjects and an average followup period of 9.8 months, Baldessarini et al. (1990) found that the rate of symptom exacerbation on placebo was 55 percent versus 14 percent on active medication (mean daily dose = 397 CPZ equivalents), a 3.9-fold overall sparing of relapse attributable to medication.”

        All of this stuff like looking for selection bias and so on are things any epidemiology student would learn in a first-year course, which is one reason why I hate it when people with no epidemiological training read a couple of studies and think they have discovered a massive conspiracy in the medical world. As an economist, imagine the grating feeling you get when a sociologist or philosopher uses the broken window fallacy or says raising tariffs would increase economic growth and so determines that all economists must be part of a massive conspiracy because they disagree. Then please try to have a little more respect for doctors, epidemiologists, biostatisticians, and so on.

      • http://hanson.gmu.edu Robin Hanson

        Well if the author got his facts dead wrong, he would of course draw incorrect conclusions. But I’m not clear on if you are disagreeing with him on the key claim that what was once mostly a temporary condition has turned into a chronic condition.

      • Yvain

        “But I’m not clear on if you are disagreeing with him on the key claim that what was once mostly a temporary condition has turned into a chronic condition.”

        I didn’t want to comment on that because I don’t know too much about the history of schizophrenia. But looking it up on Wikipedia:

        “From the outset, dementia praecox [the old name for schizophrenia] was viewed by Kraepelin [its discoverer] as a progressively deteriorating disease from which no one recovered. However, by 1913, and more explicitly by 1920, Kraepelin admitted that although there seemed to be a residual cognitive defect in most cases, the prognosis was not as uniformly dire as he had stated in the 1890s. Still, he regarded it as a specific disease concept that implied incurable, inexplicable madness……modifying his previous more gloomy prognosis in line with Bleuler’s observations, Kraepelin reported that about 26% of his patients experienced partial remission of symptoms.”

        So long before the discovery of antipsychotics, only 26% of patients have even partial remission. Although I don’t know what criteria Kraepelin used for his study or whether there have been more recent studies using the same criteria, modern psychiatry does view schizophrenia as a disease that often remits as its sufferers grow older, and one study quotes a 33% recovery rate with more people getting better but not recovering completely.

        I see no evidence that there has been any deterioration in prognosis since Kraepelin’s time, and there are staggeringly more schizophrenics who have been successfully integrated into the community.

      • http://hanson.gmu.edu Robin Hanson

        OK, you’ve persuaded me to doubt; I’ve added to the post.

    • Matthew C.

      Actually, Robin’s posts on the effectiveness of medicine are some of his best work — in fact some of the best researched and most important blog posts anywhere. The fact that he gores many sacred cows with them is not a surprise. He has put a lot of thought and research into this topic.

    • Douglas Knight

      “the legal system doesn’t currently support ability to prescribe placebo.”

      What do you mean by that? Half of US physicians surveyed report having prescribed them. Wikipedia says that half of Danes prescribe them monthly. Maybe you mean that they are limited to untreatable illness?

      In any event, there are many legal systems and an opportunity to try this in the most flexible one.

      • Yvain

        From Yale Law & Policy Review: “In November 2006, the American Medical Association (AMA) adopted a new ethics provision categorically prohibiting doctors from using placebos deceptively. In so doing, the AMA shifted the legal landscape, making it almost certain that courts will decide that placebo deception violates informed consent requirements. ”

        My father, a US primary care doctor, says that his practice has always assumed that if they tried it they would be sued and it would be an open-and-shut guilty verdict. There’s some gray area around suggesting treatments that probably won’t work for scientific reasons but would make good placebos (eg antibiotics for viral infections) but just prescribing sugar pills is a total legal disaster.

      • Douglas Knight

        thanks

    • assman35

      My view is that if you have an intervention and the underlying problem actually gets worse than the intervention is probably not that useful.

      Psychiatric interventions are just not the useful. Maybe we should be prescribing Hinduism and Bollywood movies.

  • http://silasx.blogspot.com Silas Barta

    How about if we pulled the anti-depressants’ FDA certs, and moved them to “woulda-banned stores? 😛

    Yvain: excellent post!

  • Bob Unwin

    Very good post from Yvain. I suppose at this point it would take a while for Robin to prepare an adequate response to the evidence Yvain has marshalled. But I’d be interested to see that response!

  • Laika

    As a former psychiatrist with no skin in the game, I have to say that Yvain has done a great job of shedding some light on the staggering misinformation that Whitaker has been putting out.

    You cannot understand historical psychiatric data without a knowledge of how the field has evolved, and even if you have that understanding, conclusions are difficult to draw. Whitaker ignores all that, and makes assertions that are misleading and at times embarrassing. He frequently claims to compare ‘data’ on illnesses like schizophrenia and bipolar disorder from the 1950s to the present. What he doesn’t admit to understanding is that what was diagnosed as ‘schizophrenia’ in the 1950s is substantially different that what is diagnosed as schizophrenia today.

    In the 1950s, ‘schizophrenia’ often included almost every possible mental illness that involved psychosis, emotional unresponsiveness, or a thought disorder. Many psychiatrists considered it a variety of diseases and therefore a garbage-can term. Some psychiatrists in the 1950s still preferred the term ‘dementia praecox.’ Since there was no effective treatment for what we now know as schizophrenia until Thorazine was introduced in 1954, the specific diagnosis simply didn’t matter (and of course Thorazine treats more than schizophrenia, so even an effective treatment didn’t clear things up completely). Basically much of what we now classify as bipolar disorder, severe depression, drug/toxin reactions, some forms of epilepsy, and so on were all diagnosed as ‘schizophrenia.’

    The diagnosis of schizophrenia have evolved dramatically over the last 60 years, and it continues to evolve. Whitaker seems to assume that the diagnosis has been utterly consistent, and any variation in the rates or typical progression of ‘schizophrenia’ must be caused by psychiatry. He has it exactly backwards. The data he claims to reveal simply doesn’t exist.

    Bipolar disorder has it’s own overlapping and complex history, making comparisons with past rates virtually impossible. Until lithium was accidentally discovered to effectively treat manic depression, many people with manic depression were misdiagnosed as schizophrenic/dementia praecox. Lithium was first reported in the literature in 1949, and of course it was years after that that it’s effective use became widespread. The term ‘bipolar disorder’ was first introduced in 1952, and wasn’t widely used for a time after that. In general ‘manic depression’ means more-or-less the same as ‘bipolar disorder,’ but until relatively recently, there was little standardization in what any diagnostic terminology meant.

    The Diagnostic and Statistical Manual was created to deal with this problem, and first published in the 1950s. If you compare the original DSM with the subsequent revisions, you will see constant fine-tuning of diagnostic criteria, as well as the creation of new diagnoses (e.g., Bipolar Disorder type 2, Bipolar Disorder NOS). What this means is that looking at the ‘data’ from a study from the 1950s, and another from the 1970s, and a third from 2008, and assuming that the terms and diagnoses are equivalent is an exercise in futility (unless one has a book to sell). Whitaker no doubt points to published studies to back up his conclusions, but he’s on a fool’s errand.

    It’s for these reasons that this grotesquely idiotic statement can apparently be made with a straight face: “Though bipolar disorder was relatively rare just a half-century ago, reported rates of it have increased more than a hundredfold, to one in 40 adults.” Of course it was rare! The diagnosis simply didn’t exist.

    Regarding the data on psychiatric diagnoses, there is also the matter of intentional misdiagnosis, which will skew much of that data. Some psychiatrists look hard to find an alternative diagnosis, for example, when dealing with a high-functioning patient with schizophrenia or major depression, since the diagnosis is so stigmatizing. On the other hand, there are financial incentives to come up with a more ‘serious’ diagnosis like ‘Bipolar Disorder,’ since this will get treatment authorized much more readily than something like an ‘Adjustment Disorder’ or a milder form of depression. Sadly, some treatment professionals will exaggerate a diagnosis (even as they perform appropriate treatment of the actual diagnosis) for reimbursement reasons. Finally, there is the fact that the majority of psychiatric diagnoses are made my non-psychiatrists (either Family Practice doctors or non-MD therapists), and their rate of accuracy isn’t always very good. A lot of this diagnostic data is collected, especially by insurance companies, but the data is almost worthless in drawing conclusions about incidence rates.

    The bottom line is that picking and choosing studies to compare from across 60 years of psychiatric literature, and attempting to compare psychiatric diagnoses, and treatment outcomes is an extraordinarily dicey proposition. We still have a pervasive misunderstanding of mental illness in our society, and Whitaker is actively contributing to that, to the detriment of patients and patient’s families. It’s a shame that both the ignorant and the intellectual elite share in the fear of the vague ‘evils’ of psychiatry, and are so quick to write off the entire field.

    And how pathetic is it that Yvain, primarily using Wikipedia (!), can poke so many holes in Whitaker’s claims. Robin, I’m glad that Yvain has made you look twice at Whitaker’s claims, but wasn’t a little fact checking in order before you wrote this? This kind of misinformation does real harm.

  • Pingback: On My Mind | John Goodman's Health Policy Blog | NCPA.org