Ihave been prescribing antidepressants since 1991. Like most medications, they are imperfect tools: they have side-effects and don’t work for everybody. Some patients report negative effects, or that their depression does not improve, and they may require changing to a different antidepressant. For those they do help, antidepressants undoubtedly improve depression and reduce the risk of suicide. Very rarely, in my clinical practice, do patients complain that they cannot stop their medication because of the symptoms when they try. Unpleasant physical or emotional experiences for a few days or a couple of weeks after stopping antidepressants, yes: dizziness, headache, nausea, insomnia, irritability, vivid dreams, electricity-like sensations or rapid mood swings. But patients who could not stop the antidepressant because of these symptoms? In my 33 years of clinical practice, I can recall them on the fingers of one hand. This is why I have been sceptical – along with what I believe to be most psychiatrists, psychiatry organisations and clinical guidelines – about claims in some scientific papers and the media, that “millions of people are addicted to antidepressants”. “Addiction” means that users crave the substance and cannot stop taking it compulsively, as with opioids or street drugs. However, there is no craving or compulsion for antidepressants, and our clinical experience tells us that only a small minority of people experience disabling symptoms when they stop them. The largest ever study on the topic has confirmed this. This analysis, which I took no part in, looks at 79 previous studies, encompassing more than 16,000 people stopping antidepressants, and compares them with more than 4,000 people ceasing to take a placebo. Pharmaceutical companies were not involved in this new analysis, although some of the data analysed was from trials funded by industry. The most important finding is that the proportion of people who stop antidepressants and experience severe discontinuation symptoms (which would probably necessitate restarting the antidepressant) is 1 in 30 to 35 patients: much, much smaller than the previous headline figure of about 1 in 4 patients. Even more fascinating is that about 1 in 3 patients who stop antidepressants experience some (not severe) discontinuation symptoms, but so do 1 in 6 patients who stop the placebo. This indicates that some of the antidepressant discontinuation symptoms are probably not a result of stopping antidepressants as such but rather to the attribution of some symptoms, especially now that there is an expectation that such symptoms would occur. Of course, I am not suggesting that people who stop taking antidepressants are inventing symptoms, or that the symptoms are “all in the mind” – incidentally, an unhelpful expression that serves no purpose. Rather, the symptoms are real, but may be unrelated to stopping antidepressants yet erroneously attributed to this. So, where does the discrepancy between previous alarming figures and this new study come from? Previous studies used less robust scientific research designs, because they did not include comparisons with a placebo, or they used a study design that preferentially attracted people who wanted to volunteer their experience of antidepressant discontinuation symptoms, biasing the results. For instance, online surveys are more likely to attract people who have stopped antidepressants and experienced symptoms rather than those who have stopped antidepressants with little discomfort. This previous research, albeit less robust, had the positive effect of bringing attention to the debate around antidepressant discontinuation. This new study is not perfect, and in the weeks and months to come there will be debate about the quality of the data and analysis presented. However, this work represents some of the best available research on this crucial topic. Because of that, doctors should now present these more accurate rates of discontinuation symptoms when discussing antidepressants with their patients. And people who have been advised by their doctor to start an antidepressant – indicating that they have a significant depression that is affecting their lives – should be reassured by the very low incidence of severe discontinuation symptoms. People who have been well on antidepressants for some time (six to nine months of wellbeing if it is the first time, longer if it is the second or third time) should talk to their doctors about stopping them. If they decide to do so, they should reduce it slowly over two to four months, while being aware that not all the unpleasant sensations and emotions they experience are down to stopping the medication. For the small minority who may experience severe discontinuation symptoms, a reintroduction of the antidepressant followed by an even slower reduction is needed. People will make different decisions through an informed discussion with their doctor. Some will decide that antidepressants are not for them. Some will decide that they do not want to stop the antidepressant. Many factors will influence these decisions, but at least the now-debunked myth that antidepressants are addictive will no longer be one of these factors. Carmine M Pariante is professor of biological psychiatry at King’s College London. His research funding is provided mainly by UK and EU governments and charities. Any additional research funding from industry is publicly declared in the relevant scientific papers