Does it really take several weeks for antidepressants to work?

In my previous post, I have gone on about the errors about the treatment of depression in the latest blog entry on the APA's Healthy Minds. Health Lives. There's another mistake about how long it takes for antidepressants to work. Patients are commonly told, as they are in the blog entry, that antidepressants can take up to 4 to 6 weeks to have an effect.

What the blogger should have said is that it commonly takes 4-6 weeks for a statistically significant difference between active and placebo treatment to be detected in clinical trials. But this is an artefact of the way in which statistical significance is measured. Larger size clinical trials will detect a statistical difference earlier than trials with smaller numbers of subjects.

Actually, the largest improvement per unit time produced by antidepressants occurs within the first 2 weeks of treatment (Mitchell, 2006). Recovery from depression fits an exponential model (Priest et al, 1996). The false 'delay' hypothesis has been used to call for research into better antidepressants that act more quickly. Maybe it would be better just to be honest and accurate with patients about the limitations of antidepressants.

How easy is it to treat depression?

The latest blog entry from APA's Healthy. Healthy Lives., besides making the misleading statement that antidepressants work by increasing the amount of serotonin between nerve cells, also gives the wrong impression that depression can be readily treated. The blog makes out that there are lots of options to try, so if you just keep making changes of medication everything will be alright.

I don't want to appear pessimistic about the outcome of treatment for depression. However, the reality is not as simple as the blogger makes out. Some people fail to respond to treatment and some relapse after responding. Over 6 months, maybe, about half of people do quite well, a third have a fluctuating course and 1 in 9 remain unwell (Mulder et al, 2006). Of those who are doing quite well at 6 months, maybe about half relapse in the following year (but a third of those depressed at 6 months recover) (Mulder et al, 2009).

Over the long-term, recurrence is high. Figures in studies vary from 40-85%. If about a quarter of people are improved by treatment, there's a quarter of people who do not have good outcomes (Hughes & Cohen, 2009).  Even if there's a clinical improvement, this does not necessarily  mean there's been a social recovery (Kennedy et al, 2007). Many patients still report residual symptoms despite apparently successful treatment (Fava et al, 2007).

Outcomes for non-drug treated samples are not necessarily any worse over the long-term (Hughes & Cohen, 2009). Doctors do not generally tell patients about the small effect size and substantial non-response rate of antidepressants for fear of undermining the effectiveness of medication. The serotonin imbalance theory is used as a means of encouraging patients to take their medication, which is why the Healthy Minds. Health Lives. blog mentions it.

The role of psychiatry is to give hope to depressed people. It is also to be honest with them about the cause of their problems and the appropriate treatment. Patients are able to understand that the 'chemical imbalance theory' is only a theory. What they find more difficult to appreciate is why they are told that this theory has been proven, when this is clearly not the case. They may also struggle when antidepressants may not give the simple and easy answer they have been led to expect.

Are pharmaceutical companies really moving away from psychiatric drug development?

In my previous post, I mentioned that Thomas Insel, Director of NIMH, had noted that pharmaceutical companies are moving away from psychiatric research. He elaborated on this further in a previous post on his blog, when he expressed concern about GlaxoSmithKline and AstraZeneca apparently terminating their psychiatric medication development programmes. Jeffrey Lieberman discussed this issue further in a Medscape broadcast.

However, PhRMA reported earlier in the year that a record 313 new medicines to treat patients suffering from mental health disorders are being developed by America’s pharmaceutical research and biotechnology companies (see press release with link to full report). Doesn't sound like a lack of investment to me, even if there are no drugs working by new mechanisms. As has always been the case, the motivation for research is to fulfill the wish for medication that will provide the cure we've been hoping for.

A report from NIMH takes the issue of cure  further to prevention and emphasises the importance of so-called personalization and preemption as the foundations for new treatments. As the report itself notes, psychiatric genetics has not yielded a single validated target for any mental disorder. The causes of mental disorders and their mechanisms are unknown making the development of so-called personalized psychiatry a risk for the pharmaceutical industry, which it understandably may not wish to take (see News and Notes from Psychiatric Services).

Mental health breakthroughs in 2010

Thomas Insel, Director of NIMH, who I have mentioned in a previous post, has listed his top 10 research events and advances of 2010 on his blog. I'm not quite as convinced as he is that these represent advances for psychiatry and I think we do need to question whether we are really getting value for money from NIMH (see previous post).

Insel concentrates on genetic research. I'm not saying this research doesn't need to be done, but would question its value for psychiatry. For example, whole genome sequencing has demonstrated the surprising number of variants in normals - as he says each child shows "50 – 100 new mutations not present in his or her parents". But it's speculation to correlate rare “structural” variations in the genome with autism, schizophrenia and other neurodevelopmental disorders. Parental imprinting is an interesting phenomenon and we need to understand it further, but it's unlikely to change the way we approach mental disorders. Epigenomics does need to be developed as a basic science but it's unlikely to provide us with a new way of understanding mental illness.

I don't think funding for genetic research should be obtained on the back of what is provided for psychiatry. Nor can I see induced pluripotent stem cells (iPSCs) cell technology, despite its interest and potential, contributing to psychiatry. Similarly, the Human Connectome Project may well produce basic scientific advances in the understanding of patterns of brain-function connections. It's the belief that this will uncover abnormal brain circuits that worries me. HIV/AIDS is also an important condition for research but so-called potential progress in its prevention isn't really going to help psychiatry. There was a thought that we could all take medication to prevent us developing mental illness, but this seems to have dropped off the horizon for the moment. I think we should also be sceptical about the value of expensive antiretroviral chemoprophylaxis for HIV/AIDS - shouldn't we be spending money more on prevention through condoms?

I'm not exactly sure why pharmaceutical companies have moved away from the development of psychiatric medication. Insel makes out that there are various basic science developments that they could pursue. In terms of the basic science hypotheses, I think the ones he mentions are just as likely to end up on the failed heap of neurobiological hypotheses of the basis of mental illness. Nor have we really made the advances this year in the understanding of the autistic brain that he implies. And surely it's scraping the barrel to screen for chemicals capable of enhancing neurone formation in the hippocampus of adult mice to develop them as antidepressants.

I hadn't realised that Nature had produced a whole issue on schizophrenia this year, which Insel classes as one of the events for psychiatry of 2010. I'll have to look at it further.

Is money well spent on mental health research?

A New York Times article suggests that an entire psychiatric textbook was ghostwritten by a writing company funded by a drug company. Perhaps we shouldn't be too surprised by this. As another New York Times article points out, ghostwriting has not been that uncommon in medical journal articles.

Doctors are not always neutral agents in the marketing of pharmaceutical drugs. Understandably, maybe, they want to find effective medications for their patients. Their promotion of these medications may well be biased. However, patients do look to their doctors to provide a balanced assessment of the effectiveness of medication, even if they may wish for a simple, quick and complete cure.

Academic psychiatrists may see their presentation of material in a textbook as scientific knowledge. From their point of view, it therefore doesn't matter too much who writes the chapters. After all, they sign off the final copy. They accept responsibility for what has been written. A press release from the American Psychiatric Association (APA) admits that editorial assistance from a writing company has not been that uncommon and insists this isn't ghostwriting by a drug company. From their point of view, it's merely compiling and checking facts. The problem is that "facts" in psychopharmacology are usually open to interpretation. I suppose it depends how much of the "compiling" has been done by the writing company as to whether it should be seen as "ghostwriting". Actually, perhaps what the APA is more objecting to is that technically it wasn't the drug company doing the ghostwriting - which is what it says the original NYT article implied (it's been amended since) - it was a writing company paid by the drug company.

What the authors of the book don't mention is that they have been paid handsomely to put their name to such a book. The NYT article actually implies that they didn't tell their publisher about the writing company (but anyway, according to the APA press release, the publisher wouldn't have been too bothered if it had known). Nor is the book likely to have very high quality scientific content, in the sense of critically and independently examined and reviewed. It's these researchers that obtain large research grants, and have been shown up before for not disclosing their interests to their University (eg. see another NYT article).

The Project On Government Oversight (POGO) takes a keen interest in strengthening the integrity of federally funded science and has written to the National Institutes of Health (NIH)  I think NIH are the right target here. The funding they put into mental health research and medical research in general is very significant. Such vested interests do encourage a biomedical bias (eg see my article) within psychiatry. Challenging the myth that a biological basis of mental illness will be elucidated by further research undermines the basis for these large NIH grants. Losing research funding is what biomedical psychiatry finds difficult to accept. Academic standing to obtain research grants can be improved by writing a textbook, and it's even easier if a writing company does it for you, and money can be made out of it.


(With thanks to posting on Mad in America blog)