Despite considerable advances in scientific knowledge of the brain, mental illness and its treatment still suffer from a lack of public understanding and, hence, sympathy. The tendency to shrug off conditions such as depression - “it's just in her head, she should pull herself together” - has the pernicious consequence that, even today, many sufferers do not seek treatment. This will remain a heavily stigmatised area of medicine without more, and more enlightened, public discussion of causes, consequences and available treatments. Yet, precisely because taboos persist and because this area of research is particularly difficult for laymen to follow, the subject does not lend itself easily to public debate.

Depression is a potentially serious condition of hopelessness that can wreck or even take lives - it is responsible for 70 per cent of recorded suicides in Britain. Yet few people understand how radically it differs from just feeling down, or anxious about your job, or even from the piercing sadness of a personal tragedy. Antidepressants are designed to normalise levels of the brain chemical serotonin, which affects mood.

Brain-altering drugs have revolutionised the treatment of mental illness since the 1960s, making it possible, for example, for people with manic depression and schizophrenia, who tended previously to be compulsorily institutionalised, to live nearly normal or at least tolerable lives in the community. The biggest single difficulty with this more humane form of treatment is that, because some of these medicines have unpleasant side-effects, inadequately monitored patients stop taking the drugs and then relapse.

This explains the keen interest that greeted the development of a new generation of antidepressants, known as selective serotonin reuptake inhibitors (SSRIs), which had few known side-effects. These drugs, the best known of which is Prozac, have given new hope to millions. A University of Hull team has come to the unsettling conclusion, based on its meta-analysis of unpublished as well as published clinical trials of six widely prescribed SSRIs in the 1980s and 1990s, that placebos may work almost as well as the real thing for all but the most severely depressed patients. This finding merits careful study by government and the profession, comparing it with clinical evidence and other trials, such as the largely positive studies of SSRI efficacy in relapse prevention. The debate centres on the margin of difference between treatment with these drugs and with placebos. But placebos are not the same as no treatment. The patient may be responding to otherwise identical elements of care; in mental illness, placebo responses are complex.

What is not warranted is a rush to judgment that these drugs are no good: policymakers and doctors should have sharply in mind that the people most likely to say that pills are pointless are the people who most need them. Nor should it be concluded that the drug companies concerned, which submitted both sets of trial reports to the licensing authorities in this country and the US, pulled a fast one. Trials do not necessarily go unpublished because the results do not “fit” with the hopes of the drug developer; the common reason is that, if results are abnormal, it quite often indicates that the sample was flawed. A study conflating both sets may also be flawed.

The armoury against mental illness is still small. Each weapon in it contains the priceless salve of hope. The need is not to jettison what exists, but to intensify the search for better cures.