A pill to take away your sadness
I am a clinical psychologist. I believe I have a good working knowledge of the evidence-base for the effectiveness of pharmacological interventions and I think I have some understanding of the hypothesised biological mechanisms. But I do rely on the expertise of my colleagues with more specialist skills, experience, knowledge and qualifications. One unequivocal expert is Dr Joanna Moncrieff, to whom I turn for systematic and intelligent reviews of the literature. In her books 'The Myth of the Chemical Cure' and 'The Bitterest Pills',5 and online6 Jo reviews a great deal of the evidence about the extent to which psychiatric drugs are effective in helping people, and discusses the related issue of whether they are 'treating illnesses' in any meaningful sense.
Because antidepressants are prescribed so frequently, it seems likely that at least some people think they 'work'. And - at least in the UK and most other post-industrial democracies - clinicians are helped to make appropriate decisions about prescription through complicated processes that involve expert panels making recommendations on the basis of systematic reviews of published scientific evidence. Those processes lead to such useful tools as the clinical guidelines issued by the National Institute for Health and Care Excellence or NICE. These guidelines - and therefore the systematic reviews on which they are based - conclude that antidepressants 'work'. In Joanna Moncrieff's words: 'the general feeling seems to be that although they are being overused and may have some unpleasant side-effects, they certainly "work", at least in some people'.7
Perhaps the first thing to say - because often this striking point is not mentioned - is that we know that a huge range of chemicals affect our perceptions, moods and thoughts. Although it might be unwise, there are vanishingly few people who never ingest psychoactive substances. Across the world, in most cultures, people drink alcohol (which affects their brain chemistry), chew khat (ditto), smoke nicotine-delivering cigarettes, and drink coffee (which has well-recognised effects on the brain, as well as on the digestive system). Laws and social rules differ across the world, but there are enormous markets for illegal drugs - cannabis, ecstasy, heroin, cocaine, LSD etc. I don't advocate taking such street or recreational drugs (although I'm not very keen on legal prohibition or the 'war on drugs', either). The essential point is that many chemicals of many kinds - foods, illegal and legal drugs - change our thoughts, perceptions and moods. The three groups of drugs we are discussing here - recreational street drugs, commonplace drugs such as alcohol and caffeine, and doctor-prescribed medication - all follow the basic laws of biochemistry. The medication prescribed in response to mental distress is specifically designed to alter our mood. So, in that context, it would be foolish to expect 'antidepressants', 'antipsychotics' or other psychiatric drugs to have no discernible effects. Of course they'll affect our moods. That's not surprising. But - and this is the important point - that's rather different from thinking that 'antidepressants' specifically target underlying biological abnormalities responsible for an illness - depression - and therefore can be regarded as treatments for that condition.
The conventional approach to testing the effectiveness of medication and other therapies is to conduct a 'randomised controlled trial' or RCT. I've been involved in a few. Clinical recommendations - such as NICE guidelines - are typically based on systematic reviews of many such RCTs. And systematic reviews of the effectiveness of antidepres- sants8 suggest that, on average, people taking antidepressants see their scores on a mood rating scale improve a little more than people who are taking a placebo or nothing. The difference in improvement - the degree to which antidepressants out-perform placebos - is enough to establish beyond statistical chance that the medication has some kind of effect. It's worth saying that this degree of change is typically about the same as that offered by psychological therapies such as CBT (cognitive behavioural therapy). And it's also important to note that, typically, both the people given the antidepressants (or CBT) and the people given the placebo both see their low mood ease over time. There are many reasons for this, including the fact that people typically seek help when their mood is low, and therefore their mood is likely to improve over time - fortunately not many of us remain depressed for long periods.
More importantly, we need to understand what this means in practice. Again, all these caveats apply equally to trials of psychological therapies. But what counts as 'effective' in the language of such clinical trials may not be quite what we expect. First, it only means that the degree to which antidepressants out-perform the placebo is 'statistically significant'. That is, the difference is relatively unlikely to be due to chance. That doesn't mean it's necessarily a large difference. NICE guidelines recommend the use of antidepressants on the basis that many studies have found such differences, rather than because antidepressants are dramatically effective. In addition, many scientists and academics are concerned9 that drugs companies and researchers routinely fail to publish negative results (meaning that only the positive results see the light of day).
In nearly all such trials, outcome is assessed on the basis of scores on questionnaire measures of low mood. In many senses, that's reasonable. Since problems such as low mood lie on continua - some people feel desperately low, some people are quite miserable, others are broadly happy - it is better to measure the degree of somebody's low mood than to ask whether they are or are not experiencing 'depression'. And since only the person themselves can really say what they are experiencing, 'self-report' questionnaires are appropriate ways to measure these things. But there is an issue when we look at what counts as 'improvement'. Randomised controlled trials are powerful ways of discovering differences between two groups, but whilst there might be an average difference between the groups, there is usually still a very great deal of overlap. In a typical trial, it is possible for there to be a difference between two groups (for instance, those people taking antidepressants versus those taking placebos) of only two points on a questionnaire with a maximum score of 54. And the questionnaires used to assess low mood cover quite a wide range of problems. Typically, people will be asked about their mood, their motivation, their ability to enjoy life, their sleeping, their appetite, and their self-esteem. Although obviously any improvement in scores is good, improvements of two or three points on these scales could well reflect improved motivation, better sleep or better appetite rather than an improvement in mood itself. These are effects that could well be related to medication. But they don't necessarily represent a 'cure' for 'depression'.
A few years ago I was involved in a small study that looked at the effects of different clothing on our mood.10 It was a pretty straightforward study; we asked people to wear outfits they liked and also ones they didn't and asked them about their moods - using the same kinds of questionnaires used in measuring the outcome of trials of antidepressants. Unsurprisingly, the people felt better - happier and more confident - wearing outfits that they preferred. Unsurprisingly, these differences weren't huge . .. but they were about the magnitude of the differences usually reported for antidepressants. When we've had good news, our mood lifts a little. When we receive a compliment, our mood lifts a little. When we wear clothes we like, our mood lifts a little. When I have a cup of coffee (I'm not going to speak for anybody else here), I feel my mood lift a little. And when people take antidepressants, on average, their mood appears to lift ... a little.