'First of all, we need to get clear on what the placebo effect is. There are a lot of assumptions about what it is - even among medical doctors - but not much knowledge. So let me help to put the record straight.
The placebo effect is a genuine therapeutic mind-body effect. It can be triggered by a range of different phenomena in medicine and healthcare.
The branding, the expensiveness, the colouration of pills, the number of times you have to take them per day, even experiencing side effects to medication can increase the placebo effect.
Some people incorrectly think that the placebo effect is bogus. Or they wrongly assume that only fake medications (such as sugar pills) can trigger the placebo effect. Both of these ideas are common but incorrect.
The placebo effect can be triggered by dummy pills, but it can also play a role in routine medical care.
So, the next time you take some painkillers, part of the explanation for why you start to feel better will be owed to the pain-relieving properties of the pills, but a large part of it will be owed to the placebo effect.
When researchers want to find out how effective a medication is, they test drugs against placebos.
This is how it works: patients who agree to take part in these trials are randomly allocated to one of three different wings - one-third of participants will receive the drug, another third will receive sugar pills (so-called placebos), and the rest will be placed on a waitlist (they will receive no treatment).
The idea is to see how much better patients do when provided with the drug than when they are given a placebo or are simply on a waitlist (and may start to feel better anyway).
Such trials aim to ask: how much better is the drug than the placebo? If there is a significant effect on patients' symptoms then we can say that the drug works better than a placebo. But this hides the fact that the placebo effect may still be a big effect.
The Harvard psychologist Irving Kirsch, who is one of the world's leading authorities on the placebo effect, has done a lot of research into the effectiveness of antidepressants.
Kirsch and his colleagues discovered that antidepressants are not more effective than placebos for mild to moderate depression.
They did find some research that antidepressants outperformed placebos for individuals with severe depression (in these cases, there is a small but clinically significant difference).
But it is worth remembering that the vast majority of people who are prescribed antidepressants suffer mild to moderate depression.
It's really important that we proceed with caution here because many health professionals misunderstand or dismiss this research.
It deserves to be taken seriously. First, even if antidepressants do not outperform placebos this doesn't show that antidepressants are not effective: they are. We know that the placebo effect is a sizeable therapeutic effect.
Does this mean that antidepressants are just placebos? It's hard to know.
Even if antidepressants don't work better than placebos, they may work by some other means. There is a lot of debate and controversy over how antidepressants work within biomedical research.
History shows us that the livelier the debate in science, the less mature the field. The fact of the matter is: psychiatry is an immature branch of medicine.
It is also worth emphasising that Kirsch was not interested in antidepressants in particular - he had no axe to grind or no conflicts of interest.
He stumbled on this clinical research (and around 40 per cent of these drug trials were never published). He had to use the Freedom of Information Act in the US to gain access to this data. This is because the drug companies who fund this research get to decide what gets published. Kirsch's interest was in the size of the placebo effect for depression - that's all.
His research is a bitter pill for the pharmaceutical industry (and for many doctors) to swallow. Drug trials tell us if medication is effective but not why they are effective.
There is a risk of oversimplifying depression and of harming patients with misinformation.
The research I described has caused a lot of controversy among psychiatrists and it is often simply ignored by GPs who have been trained in particular theories of depression and mental illness. For example, if you attend your GP feeling depressed, they may say that (among other things) that they would like to prescribe an antidepressant that will "help to restore low serotonin in your brain".
But this chemical imbalance theory of depression is a gross over-simplification of the causes of depression.
As one professor of psychiatry memorably told me: "Trying to explain depression as low serotonin is like trying to describe Shakespeare with the letter 'p'."
Some doctors may think it is irrelevant to provide patients with the right information - so long as the prescriptions work, who cares if we oversimplify?
Actually, it matters a lot. How people understand their illnesses has implications for their health behaviour.
Research shows that people who believe their depression is just caused by a 'chemical imbalance' are more likely to perceive themselves as "essentially depressed" and in need of long-term treatment with antidepressants.
This is a risk because we don't fully know the long-term health effects of antidepressants but we do know that long-term use is associated with increased risk of type 2 diabetes.
Patients who believe this are also less likely to make lifestyle changes or changes in their life that may help to prevent future relapses of depression.
It is vitally important that we find progressive ways to treat people with depression: it is an enormously debilitating illness.
However, I am not advocating that people don't take these drugs - they are very effective.
If you are on antidepressants, please do not stop taking them without your doctor's advice."