Last week the newspaper headlines screamed and got agitated about mood disorders – it was a rare occasion when mental health was kind of dominating the news agenda all across the media for perhaps the right reasons – because of scientific research into what works and what doesn’t – rather than the more usual prompt – because of a rare and unrepresentative example of a psychiatric patient endangering a member of the public.
So those in the mental health field should be grateful – shouldn’t we?
However the headlines on closer inspection were rather worrying – ‘Prozac, used by 40m people, does not work say scientists’ was one example from a broadsheet newspaper.
The story picked up so surprisingly massively by the media was based on a paper in an otherwise rather obscure academic journal ‘PLoS Medicine’ which published an analysis that had for the first time combined the results of 47 trials on some anti-depressant drugs, including Prozac, and found only minimal benefits over placebo, except for the most depressed patients.
The significance of the analysis was not that it was comprehensive (in fact it didn’t look at any of the research done over the most recent years) – instead this was the data set that US regulators were supposed to have been paying attention to when they decided to license newer drugs for the treatment of depression back in the 1990’s.
However it appeared that some of this vital data might have been less accessible than others, and furthermore it may have been that the less flattering studies on the benefits of antidepressants may have been actively suppressed by Pharmaceutical companies. Therefore there was the suggestion of a distortion in the scientific data on which doctors and regulators are supposed to make treatment decisions, introduced by an industry which has a vested interest in getting as much medication as possible prescribed.
Unfortunately the media missed several vital points, if the truth of the treatment of mood disorders is to be properly communicated to the public, patients, therapists and doctors.
Firstly no one who knows anything about psychiatric illness could ever have suggested that prescribing a pill was ever going to be the definitive treatment of anything in the mental health field.
Psychological problems are complex and usually require a sophisticated approach that combines treatment modalities and tailors them to the specific needs of a particular patient. So no one should be at all surprised at the idea that studies which merely looked at whether anti-depressants were superior to placebos found the medication didn’t do that well. That’s because the treatment of mood problems is always about much more than just taking a pill. It should involve a range of interventions including psychotherapeutic, behavioural, family, social, stress management etc etc
Given these facts what we really need are studies that compared patients having psychological therapies combined with medication against patients have only psychological approaches and patients having just medication.
Just such a study was published in the New England Journal of Medicine by Martin Keller and colleagues at the Department of Psychiatry and Human Behavior, Brown University in the USA. The startling results are that among the 519 subjects who completed the study, the rates of successful response to treatment were 55 percent in the Nefazodone (anti-depressant) group and 52 percent in the psychotherapy group, as compared with 85 percent in the combined-treatment group – which was a highly statistically significant finding.
We need more research like this into the effect of combining treatments compared to just having one treatment element alone – but this kind of investigation is much more difficult to undertake, and as a consequence we don’t have nearly as many studies of this kind as we need to come to a definitive conclusion about what really works with mood disorders.
The second point is that medication in these studies is often prescribed at a standard dose in order to facilitate comparison with other medications also being trialled at the same time. These doses are rarely the effective dose that most clinicians end up using.
The prescribing of a tablet requires expert supervision and the way a medication is prescribed, what dose and how side-effects are monitored and dealt with – how the particular medication is selected and how it might be changed depending on patient response – these are all delicately handled issues which dramatically alter the impact of the pill.
This issue of physician expertise is not adequately examined in most randomised controlled trials where one standard dose or a limited range of doses tends to be deployed. For this reason anti-depressants are not likely to fare well in the laboratory and benefit from an interaction effect – the right drug in the right hands for the right patient is likely to have a much more massive effect than the wrong drug, prescribed by someone who doesn’t know what they are doing, to a patient who may not suit it.
So the treatment of depression and allied conditions is much more complex than can be adequately and definitively examined by just looking at a small number of selective studies into the area.
For example a recent study published by Jørgen Bramness and colleagues at the Division of Epidemiology, Norwegian Institute of Public Health, Oslo, Norway found that the recent dramatic fall in suicide rates in Norway and its constituent counties was related to the increased sales of newer anti-depressants of the particular variety so criticised by the media last week.
The paper was recently published in the prestigious Journal of Affective Disorders and examined the sales of antidepressants and suicide rates in Norway and its counties from 1980–2004. The authors of the study pointed out that the marked recent decline in suicide rates in many countries beyond Norway has indeed appeared to be linked with the increased prescribing of newer anti-depressants. They argue that one reason for this is that the introduction of newer antidepressants, and especially the selective serotonin reuptake inhibitors (SSRIs), at the start of the 1990s accelerated the diagnosis and treatment of depression.
Studies that have examined this phenomenon across many countries have concluded that an increase of antidepressant prescribing of approximately 10% is associated with a corresponding on average 2.5 percent reduction in suicide rates.
Another possible reason for this however is more indirect – perhaps the introduction of the newer treatments lead to an increased awareness and more treatment optimism among family doctors who are burdened with an impossibly short time in which to wrestle with the complexities of depression, and who now believed they had a tool which was relatively easy to deploy compared to older treatments that appeared to be more problematic.
Kenneth Liu and colleagues at Merck Research Laboratories in Philadelphia, USA are about to publish a fascinating paper in the Journal of Psychiatric Research entitled ‘Is bigger better for depression trials?’ which found a significant effect on the size of a trial on its ability to detect a benefit for antidepressants. Basically the bigger the study in terms of number of patients recruited, the less likely it was to find antidepressants superior to placebos. The authors of the study came up with a fascinating theory which is that when a study starts the first patients recruited into it tend to be most severely effected, as these are the easiest to identify, but as the pool of readily identifiable patients becomes depleted, the later patients tend to the ones probably not just much less ill, but actually probably much less likely to actually fit the bill of even having the diagnosis!
This is because those conducting the study are anxious to hit their recruitment targets and probably become much less fussy about who they get into the trial, as they find it tougher to meet the numbers required to finish the study, towards the end of the recruitment drive.
The force of this point is that randomised controlled trials in psychiatry might be more prone to be less representative of the real world than such studies done in the rest of medicine, and more prone to various biases, and therefore have to be carefully evaluated on their individual merits rather than just lumped together – as was done in the PLoS study which got all the media attention last week.
For example, there remains the thorny issue of diagnosis which was completely ignored by the media last week. There remains no blood test or x-ray or brain scan by which we can be confident those conducting a study on mood disorders are investigating the exact same diagnosis, as those performing another study elsewhere.
When it comes to diagnoses like depression the situation is particularly complicated because of the large overlap between this particular diagnosis and others – let’s just put aside the tricky issue of how confident you can be that you are diagnosing bipolar as opposed to uni-polar depression. A recent editorial in the British Journal of Psychiatry by James Cole and colleagues at the Institute of Psychiatry, attacked the notion that there has been any progress at all in the last few decades on improving how to make an accurate diagnosis of depression.
The editorial points out that recent research has found 72% of individuals with generalised anxiety disorder (GAD) were diagnosed with concurrent depression and 48% of individuals with depression also had GAD! Many in the field are now coming to the conclusion that the word depression might be particularly unhelpful and instead suggest the umbrella term of `distress disorders’ should now be deployed to cover both syndromes of depression and anxiety much more scientifically.
What all this data means is that one paper in a relatively obscure journal no matter the media hype it attracts, cannot be the definitive pronouncement on whether anyone taking an antidepressant, or considering taking one, or any other treatments for mood problems, should be browbeaten into a hasty decision.
Psychological difficulties require careful thought and consideration, basically a good therapeutic alliance between patient and doctor which allows both to openly consider a range of treatment options and to apply what we know from research to the specific instance of this particular patient with this specific disorder.
Therefore the blanket and bald decision the media made that antidepressants don’t work was so off kilter that it appeared to many of us who work in the field to be just as dreadful as the usual low standard of coverage mental health gets.