The Empire of Humbug: Not So Bad Pharma

April, 23, 2013 | 20 Comments


  1. At long last, in this series of posts, we have someone who does both clinical and research medicine and who is prepared to say, loud and clear, what the limitations of the RCT are. Furthermore, David is telling us what the implications of the current infatuation for this tool have for the future of medicine.

    So far, I think I have listed from David’s posts, the following consequences of over-use of the RCT as follows (in no particular order):

    1. Failure to discover and record adverse events adequately.
    2. Intrinsic weaknesses in their formal approach which allows “gaming” by those with a vested interest in the outcome.
    3. A restraining or even blocking influence on future research for new possibilities; partly by the gobbling up of financial resources and partly by the inversion of hypothesis driven research.
    4. Failure to tell the clinician what they actually want to know.

    These are all things which Goldacre does not point out in his book, but rather, goes on to advocate more of the same. His publicity machine is so efficient, and the hype so loud and ubiquitous, that it is begining to look as if the medical professions are going to render themselves redundant in favour of menu-driven software to diagnose and treat the patient.

  2. I have forgotten:

    5. What David calls “disease mongering” which (I think) is the search for any significant result emerging from multiple outcomes (often proxies for real conditions) and the marketing of a drug on the basis of having found significance. There is a tendency to do something similar by the description of “new diseases”, giving them new names and then finding a drug which will treat this new disease, but I suspect this latter practice is less prevalent than the variant David has highlighted.


    Who were those seedy men from GSK and Roche and what was that woman from the EMA talking about. Is she blind to trials transparency prior to 2004 and why is she saying it is all so complicated.

    It is not complicated at all.

    It is easy, if they all get their heads out of the sand.
    One point to Ben, zero for the committee…..
    Do any of these people actually consider lives lost, or is this ongoing committee talk destined to allow GSK and Roche spokespersons to sit in their high office of corruption and sit in these sorts of committees and give out their usual talk the talk and not give a damn about lives lost.
    Even Ben, with his tough talk, forgets about lives lost……..
    Back to the drawing board…..
    These sort of meetings where they all talk the talk is so crass, that I do not wonder that, us, at the bottom end of their talk are left in sheer bewildering bafflement that talk like this is uttered……
    Congratulations to Professor Healy for a goodly article in the Mail Today about Rxisk. Well done, Jerome Burne, and Anne-Marie for just a little bit about us….
    Committees. Good try. Fail. No-one came out of that particular meeting very well.

  4. Thank you for this.

    Just as some feed-back from a reader, I would find it helpful to have links/citations for some of the claims being made, (although I realise this can make blogging more work). I’m not familiar with a lot of the issues being discussed, so it can be difficult to know where to start with some of the claims, so if you do have links easily available when you’re typing something up, I’d vote for you including them!

    Also, while I need to read more of your earlier blogs (I have read some), I’m not sure why there’s any necessary problem with RCTs. RCTs can be spun and manipulated, and it’s foolish to think that just because results are from and RCT they can be trusted… but RCTs should allow for sub-grouping, etc to find out which drug is likely to be best for Mr Jones. Is the argument that the cost of doing RCTs well is such that they tend to be done by those with the money to do them and a financial incentive for a certain result? If one had access to sufficient funding for an RCT, I do not see how this could be less useful than any other way of testing a treatment. I don’t see how the justified criticism of RCTs leads to the implication that there are better ways of testing treatments. Thanks.

    • RCTs are a good way of casting doubt on claims a treatment is effective. If there is an overlap between an effect of treatment good or bad and an the effects of the illness, then RCTs are not a good way to work out whether the drug has benefitted or harmed particular individuals – that is much more likely to come from N of 1 trials

      • Ta.

        So the limitations of RCTs and our understanding as to how to meaningfully sub-group patients are such that RCTs showing that the best treatment for patients with the characteristics X, Y, Z should not allow confidence that the treatment is helpful for a patient with characteristics X, Y, Z. The use of the treatment should be seen as a cautious trial.

        (I may have imposed my own thinking on what you said there). In a lot of ways, this sort of medical scepticism and caution about would seem likely to lead to quite a dramatic change in the traditional relationship between doctor and patient. It sounds like it would be an improvement. Thank you.

        • We have known for a long time that imipramine and some older antidepressants are good treatments for melancholia. We don’t know this from RCTs – we know it from clinical experience. In RCTs imipramine often fails to beat placebo if given to a group of patients united only by having depressive symptoms. You could go back and do an RCT in melancholia and then it would beat placebo easily (SSRIs wouldn’t). But companies would never do this because this would likely restrict their market.

          Having said all this, imipramine has a number of therapeutic principles in it and a good doctor can put these to good use in a variety of patients – bearing in mind the risks of agitation and suicide, dependence and withdrawal that come with it

          • I see, ta. While I appreciate your criticisms of RCTs and the trust some place in them, I have even less faith in clinical experience as a guide to treatment. There is a serious problem with money and power skewing medical research (and other areas of research too!), but couldn’t those clinicians who believe imipramine is a preferable treatment for melancholia perform some sort of RCT of their own? I realise that there are limitations on what can be done with limited funding and time, but honestly done RCTs do have many advantages over a series of N of 1 trials. There is a long history of ineffective treatments being supported on the basis of clinical experience, and so many potential biases which can effect people’s perceptions of results, so I think RCTs are still the best of a bad lot. Thanks.

          • Well here’s a question then. RCTs of beta-blockers will unquestionably show that these drugs slow heart rate. If i put you on a beta-blocker and your heart rate goes up dramatically. What would you want me to believe – the RCT evidence or the evidence of our joint pairs of eyes? The useful role of RCTs is to eliminate from the market treatments that don’t work but they are now being used to all but forcibly medicate people with treatments that don’t work for them.

          • I’d certainly agree that some have come to place too much faith in the results from RCTs, and that doing so can be harmful to patients. We should not assume that, because RCTs show that a drug tends to have a certain effect, it will do so in all individual patients, so should be cautious and wary of side-effects, like beta blockers potentially increasing some patient’s blood pressure. However, even if a doctor were to have a series of patients for whom beta-blockers did serve to increase blood pressure, then this clinical experience does not mean that they should believe beta-blockers are a drug which tends to increase patient’s blood pressure!

            I’m not currently confident that RCTs are even able to eliminate from the market treatments which do not work, particularly for talking/cognitive/behavioural treatments where blinding is difficult and control groups often inappropriate. In unblinded trials with an inappropriate control, homeopathy can be shown to improve self-reported symptoms – I don’t think that this shows it to be effective in any meaningful sense. Also, RCTs should, if enough of the right sort of information is gathered and they are done honestly, be able to help test for sub-groups of patients who respond in different ways (although this is also open to abuse), as well as providing information about the levels of improvement that tend to follow different interventions. While the problems with RCTs and the environment in which medical research takes place are currently such that I think it is best to remain cautious and sceptical about the results being presented from RCTs, if I was to conduct research examining the efficacy of particular treatments, I would want to run an RCT. Thanks.

          • There are two points here – one is testing for efficacy – where RCTS are often not needed. In testing for effectiveness they are needed and are mainly helpful when they exclude ineffective treatment. When treating you the evidence of your and my eyes sd be paramount – supplemented but not replaced by a belief system drawn from good studies which give us some grounds to believe that doing what we are doing might be helpful. Studies here can include but does not mean RCTs. Pretty well 100% of new treatments are discovered because of studies that are not RCTs.

          • I agree about the limited use of RCTs and the danger of a misguided faith in the results presented from them, but I’m afraid that I still don’t understand why non-RCTs would be seen as a preferable way of gathering data. re the imipramine for melancholia example, I don’t see why a series of N = 1 trials would be preferable to an appropriately designed RCT. Is it just that the added cost of RCT’s means that only certain interest groups tend to be able to perform them? Other than that, RCTs seem a good way of accounting for certain biases. Sorry if I’m missing your point here. Ta.

          • RCTs derive from agriculture. They account for the biases that might come into play with fertilizers. The biases of sunlight and drainage etc. But if both the drug and disease can produce the same thing – either a benefit or a harm – RCTs are lost. And when in fact I treat you how are we going to tell whether what is happening to you is good or not – do we go by the Evidence base and say well the evidence shows the drug works and has no harms therefore even if it doesn’t look like you’re getting better you must be?

          • I agree that some can place too much faith in RCTs, and that patients and doctors should be committed to looking for evidence that an individual patient is not responding to treatment in the way which data from RCTs indicated that they would be likely to, but don’t see how this leads to the belief that a series of N= 1 trials are preferable to an RCT. When the symptoms of an illness and possible side effects can overlap, working out which is which is difficult, but randomisation and the use of an appropriate control group is still surely useful.

            While I agree with your specific claims about the problems with RCTs and the way in which they are viewed, I don’t see why this leads you to believe that non-RCTs are preferable. I’ve thought about ways of assessing the value of imipramine as a treatment for melancholia, and haven’t been able to think of a reason why an RCT would be less useful than a series of N=1 trials. Ta.

  5. The Caronia case you mention is a complicated one, but to me the issues seem awfully similar to the lamentable Citizens United case allowing corporations unrestricted campaign spending in the name of “free speech.” Are corporations people, and is money speech? A growing majority of us say no – if millions of dollars from anonymous private interests can drown out the voices of local voters, it brings us way too close to a system of one dollar, one vote.

    The Caronia decision goes farther in expanding freedom of commercial speech to corporations marketing a potentially hazardous product: In order to sanction any claim they push, the burden is on the state to “prove” the claim is false. Even if we had a government agency with the money, power and motivation, how could that be done? To a 95% certainty? In a randomized trial? Short of that, if the corporation has one “peer reviewed” journal article to stand on, they could be home free. And if the FDA has no power to limit Pharma’s claims, forget about the unions and consumer groups who have sued them for fraud in recent years. Most of those suits, which have at least rocked the boat and publicized the abuses, will now be thrown out of court on the first bounce.

    I’m not terribly worried about what a guy like Caronia says to doctors. As you have pointed out, the drug companies have learned to co-opt and use medical “key opinion leaders” to become far more effective drug reps than the drug reps could ever be. But I do worry – big time – about the “speech” Caronia’s bosses could be allowed to flood the airwaves with through virtually unregulated direct-to-consumer advertising. (Worse yet the Internet, where they can directly target the eyeballs of those seeking help, and purchase the first forty places on any Google search so that no other voice gets through.) How this decision would affect DTC ads is still hotly debated, but this Supreme Court has delivered for business beyond even business’ wildest dreams. So may as well prepare for the worst.

    I have to admit, the current patchwork regulatory system is pretty pathetic and lacks a certain intellectual consistency. That’s not to say it can’t get much worse. No doubt GSK would say that its lack of DTC advertising in most of the world including Britain is a violation of “free speech.” Should they get their “rights” worldwide, then?

    • Johanna

      As ever you have framed the issues perfectly. We need to prepare for the worst. The next two posts in the sequence are about preparing for the worst.

  6. Healy, I enjoy your articles very much. This world would be a million times better if every psychiatrist were as critical of drug companies as you are. We all need to be critical of drug companies and their meds. Here in the U.S. half the country puts our president under such scrutiny because he has the ability to control our lives in certain indirect ways, but many of these same people will take a psychiatric med and all they care to look at is the side effects sheet that comes with it and they’re satisfied this way. They will defend these meds to the very end on a forum I get on. Some even are supportive of not letting patients know of certain side effects so that they’ll be more likely to take these meds.

    Anyway, I’ve been curious about your stance on Electro Convulsive Therapy? What do you think about this treatment?

  7. Specifically I’m wondering if you think ECT can have permanent effects such as messing up someones memory or cognition?

    • Micahel – its not my role to tell you or anyone what you should or shouldn’t have. Have never wanted anything banned. The key issue if you and i are working as a team is that we have a full set of information available to us – which we don’t have for any drug and in this sense it is difficult to do pharmacotherapy with informed consent.

      When it comes to ECT there are lots of people who want it and others who would never have it. I would want it available for those who find it helpful and would champion the rights of those who do not want it to avoid having it.

      Regarding what to tell them re memory it would be something like this. When you get to the stage of contemplating ECT you almost certainly have had a bunch of treatments that affect memory. Benzodiazepines have very clear effects on autobiographical memory. Antipsychotics and mood-stabilizers have clear effects on set shifing – being able to remember someone’s phone number. ECT also certainly has effects on memory also but its hard to know precisely what. I have had people tell me about their memory problems post-ECT and describe very clearly memory problems of the type antipsychotics cause which they blame on ECT even though they are taking antipsychotics.

      Overall far more memory problems are caused by drugs than ECT and vastly more people are on involuntary drug treatment than involuntary ECT – so if there is an issue to campaign about it should be the drugs

  8. Which would you choose ssris or benzos, is the question.
    Definitely benzos because, on the whole, in my experience, there is so much publicity about benzos being highly addictive, that the message has more or less got through to some doctors.
    Ssris are a different kettle of fish entirely.
    Although having similar withdrawal, the million dollar dilemma for doctors is that they know nothing about ssris, nothing at all, and so ssris become so dangerous, so macabre in their oft sudden withdrawal that all hell breaks loose, whereas with benzos, the acute withdrawal is bad, but not so very bad.
    Depends on your definition of bad…..
    I have experienced both.
    I received copious quantities of benzos, valium, ativan, for much longer than the recommended time of two weeks to combat ssri abrupt withdrawal.
    So, not only was I was going through abrupt ssri withdrawal, but I was, then abruptedly left to my own devices, abuptedly off valium and ativan, as well.
    I would also add that valium and ativan did not do anything to help with abrupt ssri withdrawal. This was so terrible that even copious quantities of valium and ativan did not penetrate the withdrawal from the ssri.
    So, all in all, a rum do.
    I would have preferred, maybe, a two week course of a benzo, not two years of Seroxat and then plummeting to dangerous depths…….and then handed out benzos like smarties.
    People who mis-manage drugs like this, to an ordinary every day person, are a total danger to society and although, I, know better, now, and managed to come out the other end, this was largely due to pure ignorance on my part.
    The more you learn the scarier it is for the pill popper, who looks at the doctor with a new attitude of pure, unadulterated despair and desperation and, dare I say it, despise….I could now become an excellent gp as I now understand so much about it.
    I have empathy and compassion and I am sure I could work out a ‘pill plan’ not a botched and barbaric Plan B, for benzo, which was a hopeless state of affairs.
    Moreso, when patient is ‘piggy in the middle’, as I was with two opposing and different tactics, both of which failed miserably because ‘two’ failed to talk to each other and they were only a pretty, half hour drive from each other…
    I was the tennis ball ricocheting between them. Clueless doctoring.
    Benzos, without the added ssri complication, have their place.
    Ssris have yet to find their place amongst all the confusion of where to put this class of drugs in the great scheme of things pertaining to drugs helpful or drugs unhelpful…….
    Hope that was helpful.

Leave a Reply to David_Healy