Call for an accounting


I have my doubts about the efficacy of the growing welter of petitions for every possible cause, but at least we can express our concerns. There is now a petition to nvestigate psychiatric research misconduct at the University of Minnesota, my place of work, so I feel an obligation to bring it up.

Dan Markingson was mentally ill and committed suicide almost ten years ago. He was in bad shape and was committed for psychiatric care (he was suffering from violent delusions) when he was recruited into a study by AstraZeneca of anti-psychotic drugs. This was not a good idea; running an experimental clinical trial on patients at serious risk of doing physical harm to themselves and others is not recommended.

In fact, the CAFE study also contained a serious oversight that, if corrected, would have prevented patients like Dan from being enrolled. Like other patients with schizophrenia, patients experiencing their first psychotic episode are at higher risk of killing themselves or other people. For this reason, most studies of antipsychotic drugs specifically bar researchers from recruiting patients at risk of violence or suicide, for fear that they might kill themselves or someone else during the study. Conveniently, however, the CAFE study only prohibited patients at risk of suicide, not homicide. This meant that Dan—who had threatened to slit his mother’s throat, but had not threatened to harm himself—was a legitimate target for recruitment.

He was also signed on by his own consent: the consent of a committed mental patient with serious concerns about his competence, and against the wishes of his mother.

The petition does not assign blame, but only asks for an objective examination by external reviewers. You know there’s a problem when the university’s own bioethicists are supporting the call for investigation.

In 2009, the Minnesota state legislature passed “Dan’s Law,” which prohibits researchers from recruiting a patient under an involuntary commitment order into a psychiatric drug study. Media outlets such as Mother Jones, the St. Paul Pioneer Press, City Pages and Scientific American have published accounts of Dan’s story. His story was also featured in the documentary film, Off Label. In 2010, AstraZeneca, the sponsor of the study in which Dan died, settled federal fraud charges for $520 million, and a University of Minnesota psychiatrist was implicated. Last year, the Minnesota Board of Social Work found serious wrongdoing by the study coordinator for the research study in which Dan died.

More recently, evidence of fraud and serious privacy violations in psychiatric studies at the university have emerged. It is possible that other research subjects have died or suffered serious injuries, or that they have been mistreated in other ways. Bioethicists at the University of Minnesota itself have called for an external investigation, yet the university still refuses.

I signed it. I’d like to recommend that more people sign it.

Another scary aspect of this is the possible abuse of science. There’s some funny stuff going on in clinical trials…

A 2006 study in The American Journal of Psychiatry, which looked at 32 head-to-head trials of atypicals, found that 90 percent of them came out positively for whichever company had designed and financed the trial. This startling result was not a matter of selective publication. The companies had simply designed the studies in a way that virtually ensured their own drugs would come out ahead—for instance, by dosing the competing drugs too low to be effective, or so high that they would produce damaging side effects. Much of this manipulation came from biased statistical analyses and rigged trial designs of such complexity that outside reviewers were unable to spot them. As Dr. Richard Smith, the former editor of the British Medical Journal, has pointed out, “The companies seem to get the results they want not by fiddling the results, which would be far too crude and possibly detectable by peer review, but rather by asking the ‘right’ questions.

I’ve been reading Goldacre’s Bad Pharma: How Drug Companies Mislead Doctors and Harm Patients. There really are grounds for more global concern.

Comments

  1. neuroturtle says

    Would you happen to have a citation for that 2006 study referenced in the last quote? I’ve been poking around PubMed, but don’t have good enough search terms to find the paper. I teach statistics, and I want to see if this is too technical for my class or if I can use it to demonstrate that stats isn’t just a boring required class, but has real-world implications that can be very good or extremely damaging. I may buy Bad Pharma and Bad Science too.

  2. Blattafrax says

    #2

    Why olanzapine beats risperidone, risperidone beats quetiapine, and quetiapine beats olanzapine: an exploratory analysis of head-to-head comparison studies of second-generation antipsychotics.

    Am J Psychiatry. 2006 Feb;163(2):185-94.

  3. says

    I read Bad Pharma as soon as it became available last month. I recommend it. (I also read about an interesting little proto-debate between Goldacre and radical economist Harry Shutt, but haven’t yet followed up on it.*)

    One problem I have with Goldacre’s presentation, although I understand the reasoning behind it, is that he doesn’t single out any specialty for particular concern. The problems he describes exist throughout medicine, of course – though the research shenanigans and conflicts of interest are especially bad in psychiatry – but psychiatry can’t be talked about in exactly the same terms as other areas of medicine, because there’s a problem of urgent concern: forced/coercive drugging. This makes the corruption of research and medical practice in this field all the more serious: it’s one thing to mislead people and doctors and corrupt clinical practice so that useless or harmful treatments are prescribed or chosen, but it’s quite another to mislead and corrupt in a a system that includes coercive drugging. It’s a human rights question, and the major issue of the psych rights movement.

    *The responses I’ve read so far from Goldacre don’t inspire confidence. in his grasp of the arguments.

  4. eric says

    I’m thinking of a general rule of thumb, “use the oldest scientifically-shown efficacious medicine available.” If you’ve got a choice between aspirin and clinically-tested-in-2013 1,2 diooxynewpharmacopaeia, and both are reasonably expected to work, choose the aspirin. Not only will it likely be cheaper, but the longer its been out and used, the more likely its actual performance matches its clinically claimed performance.

  5. dianne says

    running an experimental clinical trial on patients at serious risk of doing physical harm to themselves and others is not recommended.

    This was the wrong trial design, the wrong drug, the wrong consent process, and wrong in a number of other ways. However, is a blanket denial of clinical trial access to people who are suicidal a good idea? What if they’ve failed all available treatments and are ready to kill themselves if they don’t get any relief? We allow people dying of cancer to be on clinical trials, why not people dying of depression?

  6. dianne says

    If you’ve got a choice between aspirin and clinically-tested-in-2013 1,2 diooxynewpharmacopaeia, and both are reasonably expected to work, choose the aspirin.

    Excuse my fussiness, but a proposed addendum: “…unless aspirin has side effects that are unsafe or unacceptable to the patient and 1,2 diooxynewpharmacopaeia does not.” Even then, you might try tylenol or lastyearsnewpharmacopaia first.For the reasons you’ve outlined. People who take newly approved drugs are, by the nature of the process, beta testers. Some side effects simply won’t show up until the drug’s been tried on a few thousand people.

  7. dianne says

    The companies had simply designed the studies in a way that virtually ensured their own drugs would come out ahead—for instance, by dosing the competing drugs too low to be effective, or so high that they would produce damaging side effects.

    Or they compare to placebo when an active drug is available and accepted in clinical practice so you don’t know anything about which is more effective. Or they simply refuse to try an even slightly risky trial, even when they have a large potential market if it’s positive. For example, animal data suggests that the anti-factor II drugs might be powerful anti-cancer agents, possibly effective in a wide variety of cancer. The pharma companies making them have zero interest in trying them in this context. Because they can’t be certain that the product would come out looking good and they don’t want to waste money.

  8. says

    @dianne #8 – “However, is a blanket denial of clinical trial access to people who are suicidal a good idea?”

    The dilemma comes up that studies — well, reliable studies — use a protocol where at least one cohort of volunteers receives a placebo rather than the drug being tested. The protocol also requires that the study be double blind, meaning that neither the volunteer nor the people who interact with the volunteer know which cohort the volunteer is in: the idea is to validate that it was the drug, and not the volunteer’s faith in the drug, that actually had an effect. For example, a vaccine study I was in from 1998-1999 had three cohorts: one the got the engineered vaccine, one that got the unmodified carrier virus, and one that got plain saline.

    If a person is suicidal, or is otherwise showing signs of extreme psychological disturbance, they need to go on actual medication. It would be unethical, and potentially quite dangerous, for them to end up in a placebo cohort. Worse, a number of psychological drugs can have an initial effect of enhancing depression, suicidal thoughts or psychosis as the brain rebalances its chemistry, and otherwise minor problems can very quickly become overwhelming.

    What worries me is that aside from the ethical questions of possibly denying effective meds to someone who needs them is the fact that chemically unbalancing someone during a drug trial will — ok, should — end up skewing the data against the manufacturer: the FDA will look askance at an anti-psychotic drug that seems to make the people who take it psychotic. Ethics aside, it is in the manufacturer’s best interest to take steps that the study volunteers are as stable as possible before they get recruited. That they do not seem to care can only mean that the results are already so corrupt that the actual results don’t matter one way or the other.

  9. dianne says

    @Gregory: Not every trial is or should be placebo controlled. Some use best standard treatment as their control. If, for example, you’re unfortunate enough to get pancreatic cancer and you elect to go on a clinical trial, it would probably be a trial of gemcitabine (or, recently, foloxfiri), versus an experimental treatment. So there’s no reason why a clinical trial of anti-psychotics or anti-depressants couldn’t be standard treatment (at effective and non-toxic doses) versus experimental treatment. Apart from, of course, drug companies not wanting to use an effective dose of their competitor that might make them look bad.

    In any case, the situation I was postulating was a patient with intractable depression who has been through all known effective or semi-effective treatment options (drugs, talk therapy, even ECT) without improvement. If every known drug is acting like a placebo, how much higher risk is there in enrolling in a placebo controlled trial versus continued known ineffective treatment? Alternately, s/he could go on a phase I/II open label study of a newer drug, i.e. one not placebo controlled. They aren’t as definitive scientifically but can be useful for giving an option to a patient who is otherwise about to die.

  10. says

    A 2006 study in The American Journal of Psychiatry, which looked at 32 head-to-head trials of atypicals, found that 90 percent of them came out positively for whichever company had designed and financed the trial.

    Why the living fuck are companies allowed to design or run trials for their own fucking medications? Is that not the definition of a conflict of interests? I’m seriously asking this, because it makes no sense at all to me here.

  11. Doug Hudson says

    Dalillama,

    Because drug development and testing is enormously expensive. Only the drug companies have enough money and interest to fund the testing.

    The government could theoretically test drugs, but that would require truly massive tax increases and a huge expansion of the FDA, neither of which is going to happen.

  12. Doug Hudson says

    I should also note that, in light of problems such as this one, some people have suggested doing away with Phase III trials altogether. Just test the drugs in the general population. It’s not as crazy as it sounds–it’s basically what’s been happening anyway, so we might as well be upfront about it.

    Reminds me of a recurring complaint about software–these days, customers are essentially “paying money to do beta testing.” The same could be said of pharmaceuticals.

  13. says

    @dianne #13 – A lot depends on how far along in the research the drug is. Typically, drug studies are done in three phases.

    Phase 1 is where the drug is being used in humans for the first time and typically involve a few tens of volunteers meeting a lengthy list of qualification criteria. The purpose of a phase 1 study is to verify that the drug is safe, to look for side effects and to measure whether or not the drug has the desired pharmacological effect on body chemistry.

    If a drug meets the thresholds set by the phase 1 studies, it moves to phase 2. These studies usually have a few score, maybe a few hundred, volunteers who meet slightly less rigid qualification criteria. Phase 2 studies continue the investigations of the previous studies (is it safe, what side effects occur, does it look like it would work) with the added focus of adjusting the dosage to find the proper balance between side effects and desired effects.

    If the drug does well in phase 2, it goes to phase 3. While still tracking the items of phase 1 and phase 2, a phase 3 study is when they start looking at efficacy. These studies typically have a few thousand participants, with fairly minimal qualification criteria to better test the drug across a larger variety of people.

    A drug must successfully pass all three phases for FDA certification, and the FDA protocols require the use of placebo cohorts in all phases. It is allowed — and depressingly common — for a manufacturer to get approval for non-FDA track studies. These do not require placebo cohorts, and it with these studies that manufacturers generate the results used in marketing, such as “Clinical research has shown that our new brand X is 10% more effective than their brand Y!” Because they are not part of the FDA data, the researchers have considerably more leeway in setting up the studies, and they can usually get away with cooking the results.

    If you are in a study that makes no mention of placebo cohorts, or assures volunteers that they are going to get the investigational medication, then be advised that the study will be under a lot less scrutiny and probably carries more risks.

  14. says

    @Dalillama #14 – Because US law requires that the drugs be tested, and that the manufacturer bear the costs of performing those tests. In theory, there are agencies and boards to oversee the research, and studies that are used for FDA approval must follow specific protocols. The real problem comes up, as I said above, in studies run conducted to generate marketing data.

  15. Nerd of Redhead, Dances OM Trolls says

    Dalillama:

    Why the living fuck are companies allowed to design or run trials for their own fucking medications? Is that not the definition of a conflict of interests? I’m seriously asking this, because it makes no sense at all to me here.

    Historically this is how it was done. Also, the drug companies wanted to keep the regulatory agencies as far away as possible, So the regulatory agencies in Europe, Japan, and the US enforce regulations covering the manufacturing, clinical testing, analytical, and documentation for the regulated industry. It allows the drug companies to maintain control of their intellectual property in their eyes compared to the alternatives.

  16. Rob Grigjanis says

    Interesting talk by Allen J. Frances on the overdiagnosis of mental illness and the DSM-V here, which is at least tangential to this topic.

  17. says

    Doug Hudson #15
    I said design the trials; a reasonable system would be one where they pony up money to say, the FDA or a similar agency, which then designs and implements the trials without further input from the manufacturer.
    Nerd#19
    I suppose I already knew that, yes. REfining my initial query :Are there any valid arguments taht this should be the way that things are done. ‘Because the pharma companies want to not be regulated’ does not, incidentally, qualify as a valid argument.

  18. dianne says

    a reasonable system would be one where they pony up money to say, the FDA or a similar agency, which then designs and implements the trials without further input from the manufacturer.

    I like this idea. Now, about getting it implemented…

  19. Blattafrax says

    #21 I don’t see quite what you think is broken here. There are regulations enforced by almost every country in the world on safety in drugs. Trials are designed to ensure the products provably abide by these regulations and (strange as it may seem) it isn’t good for a company to have a reputation for killing its customers anyway. After that, the pharma industry is entitled to decide for itself who it wants to market a drug to and what the trial design should be to most effectively enable that. The industry doesn’t actually have a moral obligation to develop ever better drugs, so once safety is taken care of and provided marketing strategy is honest**, whether they make a good product or a bad one is up to the manufacturer.

    Here’s a hint though – good products make more money.

    ** Open access to all clinical trial data _should_ be compulsory – I’d agree that is broken.
    And I’d also agree that marketing in general is a parasitic cancer on our society; but don’t see the need to single out Pfizer et al any more than (for example) Unilever, Apple or Mr. Bob’s Hardware Emporium.

  20. Nerd of Redhead, Dances OM Trolls says

    Are there any valid arguments taht this should be the way that things are done.

    Should be done? Either the companies do it on their dime or the government has a department of clinical trials that does it using public money. At about half a billion per full clinical trial (cheaper if it dies early on due to toxicity/side effects), with associated liability and all the problem of enrolling folks. I don’t see a hybrid system, where the government runs the trials, but the company pays for it, working. There is no political push to have goverments run the clinical trials.

  21. Blattafrax says

    #22
    Okey dokey. First get the EMA, NICE, FDA, Peruvian, Australian, Japanese, Chinese, Indian, Georgian, Icelandic, St. Kitts & Nevis, etc, etc, etc governments to agree on a common set of regulations for approving drugs. Then decide which one of them gets to spend the billion+ dollars income from the next chronic heart failure drug in their own country. Then decide what to do if the drug company spots a mistake in the protocol you’ve implemented on their behalf and wants their billion dollars back.

    Or you can let the organisations with the knowledge and incentive to decide best for themselves how to conform to the regulations in the various parts of the world.

  22. dianne says

    If a drug meets the thresholds set by the phase 1 studies, it moves to phase 2. These studies usually have a few score, maybe a few hundred, volunteers who meet slightly less rigid qualification criteria.

    Hundreds? I do hematology. In hematology we call tests of hundreds of patients “post-marketing data”. It’s hard to find a few hundred qualified people with HIT, TTP, thalassemia, etc. This has nothing to do with anything except me looking enviously at you for being in a field where a hundred+ person study would be considered phase 2. Cardiology?

    A drug must successfully pass all three phases for FDA certification, and the FDA protocols require the use of placebo cohorts in all phases.

    Are you sure? You sound like you know more about how the FDA works than I do, but this doesn’t sound very post-Tuskegee compliant to me. And I’ve never seen an oncology trial that was placebo controlled. It’s inevitably standard of care versus experimental. I suppose occasionally standard of care is no treatment, but that’s fairly rare. How do chemotherapeutics get past the FDA?

  23. says

    @dianne #26 – For what it is worth, I am on the Community Advisory Board for the Seattle HIV Vaccine Trials Unit and have been for almost three years. The main focus of a CAB is to be advocates for study participants, working to make sure that they are informed correctly of their rights and that these rights are protected. We’ve had more than a few presentations about how the FDA approval system works and why that system exists: the Nuremburg Code, the Declaration of Helsinki, the Tuskegee and Guatamala syphilis experiments, the nuclear downwind tests and the National Research Act. I’m not saying you should take me as an authority, just that you accept that I know a thing or two on the topic of medical research.

    All research is double-blinded. That means neither the participant herself, nor the people who work directly with the participant, know which cohort the participant is in. The reason for this is pretty simple: eliminate “tells” that would eliminate a possible placebo effect. If you knew that the cancer patient you were hooking up to an IV was getting saline and not a new treatment, there is no possible way you could keep this knowledge from the person, even if you never said a word. And so, you are never told. When a person signs on to a study, they are told up front, in large print on the first page, that they might not be getting any actual treatment. There is no such ethical obligation to inform hospital personnel of this matter, and scientifically valid reasons why they should not know.

    As for numbers, the matter of “orphan diseases” is one of the reason why there are so few treatments for them: economically, there are not enough people with them to make the R&D cycle earn a return, and scientifically, there are not enough people to sufficiently test the treatment for FDA approval. Typically, treatments either remain “experimental” for decades, or an approved treatment for one thing is being used off-label for something else.

  24. Doug Hudson says

    Our current government can’t even muster the political will to keep essential services running (thank you, Republicans!) Changing the way drugs are tested and approved will only be done in fits and spurts, on those occasions when public outrage temporarily overcomes political inertia and the clout of the pharma industry.

  25. Nerd of Redhead, Dances OM Trolls says

    Changing the way drugs are tested and approved will only be done in fits and spurts, on those occasions when public outrage temporarily overcomes political inertia and the clout of the pharma industry.

    That’s been the history of the FDA, even its inception.

  26. dianne says

    Typically, treatments either remain “experimental” for decades, or an approved treatment for one thing is being used off-label for something else.

    Quite a lot of hematology is done with “off label use” of various drugs approved for something else. But then there’s hemophilia. Hemophilia is a rare disease and the treatments used really aren’t good for anything else. What else would you use factor VIII concentrate for? So the approvals must be being based off rather small numbers and probably uncontrolled studies, since there’s no ethical way to give a placebo to someone with hemophilia. I’m not sure how this passes the FDA.

    Double blinding is, of course, standard and a good idea, but it’s also less certain than it might at first appear. Consider, for example, a study comparing clozaril with haldol. They have completely different side effects. There’s no way that the doctors involved wouldn’t know which drug was being given. (The patients know too if they’re in good enough state neuropsycholocially to recognize what’s going on.) So even being double blinded doesn’t completely eliminate the possibility of “tells” and different clinical decisions based on guesses about what the patient is being given. For example, quite a lot of the patients on the haldol arm of the study I mentioned crossed over into open label clozaril. Did they need to or was the decision made based on the doctors’ perception of which one worked better?

  27. dianne says

    I remember reading somewhere yesterday* that medicare part D was costing less than expected because there were few new drugs coming out and the old drugs were going off patent and becoming less expensive. That supports my claim that the worst corruption of the pharma industry is not performing clinical trials with reckless abandon, but rather refusing to perform trials and do research that is needed. They’re so scared of bad publicity and/or “wasting” money on a negative trial that they won’t conduct a trial, even if they might get a new patentable drug out of it.

    *Sorry about the poor sourcing.

  28. David Marjanović says

    the growing welter of petitions for every possible cause

    Hey, the Founding Fathers wanted it that way. B-)

  29. David Marjanović says

    Why olanzapine beats risperidone, risperidone beats quetiapine, and quetiapine beats olanzapine: an exploratory analysis of head-to-head comparison studies of second-generation antipsychotics.

    Awesome.

    Or they simply refuse to try an even slightly risky trial, even when they have a large potential market if it’s positive. For example, animal data suggests that the anti-factor II drugs might be powerful anti-cancer agents, possibly effective in a wide variety of cancer. The pharma companies making them have zero interest in trying them in this context. Because they can’t be certain that the product would come out looking good and they don’t want to waste money.

    That sounds like a great argument to socialize large clinical trials – let governments or organizations such as the EU do it, they don’t need to make a profit. I’d like a number for the “truly massive tax increases” mentioned in comment 15, especially in view of the fact that drugs couldn’t help becoming cheaper that way.

    BTW, it’s also deeply silly that drugs need to be approved separately in the US, the EU, and Japan and probably a whole bunch of other places. A billion here, a billion there, and pretty soon all this duplication adds up to real money!

    Interesting talk by Allen J. Frances on the overdiagnosis of mental illness and the DSM-V here, which is at least tangential to this topic.

    Can’t watch the video right now, but what you say strikes me as opposite to the topic.

    Besides, overdiagnosis? I wonder if light cases of several mental illnesses are massively underdiagnosed simply because they’re traditionally considered normal – even though they cause non-negligeable amounts of trouble for the people in question themselves and for everyone around them.

    Here’s a hint though – good products make more money.

    Most blatant oversimplification I’ve read in weeks.

    Emporium

    *lightbulb moment* Finally I get some kind of clue of where this most American word comes from! Is it a misspelling of imperium? :-)

    Are you sure? You sound like you know more about how the FDA works than I do, but this doesn’t sound very post-Tuskegee compliant to me. And I’ve never seen an oncology trial that was placebo controlled. It’s inevitably standard of care versus experimental. I suppose occasionally standard of care is no treatment, but that’s fairly rare. How do chemotherapeutics get past the FDA?

    Have you noticed that you made that point before, in comment 13, and that Gregory in Seattle never commented on it, not even in comment 17?

    Absolutely, the control gets the current standard of care. If the new drug isn’t better than the old one, there’s no point in making it, because you won’t be able to sell it!* Sometimes, the current standard of care is a placebo, and HIV vaccine research definitely is such a field – there is as of yet no vaccine against HIV. Chemotherapeutics are definitely not tested against placebos these days.

    * Unless, as detailed in the OP, you can fool people into believing it’s better.

  30. Esteleth, stupid fucking starchild Tolkien worshiping douche says

    Placebo is very rarely the control: more frequently it is standard-of-care. Because it isn’t enough to demonstrate that newdruggium works, it has to be better than olddruggium.


    David, “emporium” comes from the Greek ἔμπορος, “merchant.”