The Defeat Depression campaign was a five-year national programme launched in January 1992 by the Royal College of Psychiatrists in association with the Royal College of General Practitioners. The aim of the campaign was to educate health professionals and the public about depression and to reduce the stigma of mental illness.

A door-to-door survey of public opinion was undertaken to obtain baseline data before the campaign started (Priest, et al, 1996). Most of the people questioned in the sample, that is 78%, thought that antidepressants were addictive. This finding caused some consternation amongst those running the campaign, because, as far as they were concerned, the public was misinformed on this issue. Part of the education programme, therefore, was to teach doctors that patients should be told clearly when antidepressants are first prescribed that discontinuing treatment in due course will not be a problem.

In retrospect, this guidance may seem surprising. The British National Formulary, which doctors use for reference about medication, has given a warning since 1990 that symptoms may occur if an antidepressant is stopped suddenly after regular administration for 8 weeks or more. In fact, case reports of discontinuation reactions have appeared since antidepressants were first introduced (Mann & MacPherson, 1959).

Tricyclic antidepressants began to be prescribed in the mid-1950s. A new class of antidepressants called serotonin specific reuptake inhibitors (SSRIs) were introduced onto the UK market in 1987. Discontinuation symptoms were only recognised in this new group of drugs after the SSRIs had been in widespread clinical use for several years.

The drug companies were concerned about these discontinuation problems and a consensus meeting of experts, sponsored by Eli Lilly, the manufacturers of fluoxetine, or Prozac™, as it is known by its trade name, was held in Phoenix, Arizona, at the end of 1996. This led to an editorial in the British Medical Journal in 1998 acknowledging that antidepressant discontinuation problems existed (Haddad, et al, 1998). It suggested, though, that they were both preventable and simple to treat.

The BMJ editorial seems to have understated the seriousness of the problem. The same authors only two years later acknowledged that discontinuation symptoms are common in a letter to the Lancet (Young & Haddad, 2000). There was also confusion about the meaning of terms like dependence and addiction. Eventually, the pharmaceutical company GlaxoSmithKline dropped its insistence that paroxetine, its SSRI drug, is not addictive (Boseley, 2003). Even though there may be little evidence of physical addiction, in the sense that the body gets addicted to SSRIs, commonsense understanding of the word also includes psychological dependence, and despite what the Defeat Depression campaign said, the public knew, even if doctors did not, that taking antidepressants can become a habit.

Modern guidelines, therefore, such as the one for depression from the National Institute for Health and Clinical Excellence (NICE, 2004), actually state the opposite of what was recommended by the Defeat Depression campaign. This is that all patients prescribed antidepressants should be informed that discontinuation/withdrawal symptoms may occur on stopping, missing doses or, occasionally, on reducing the dose of the drug.

Antidepressant discontinuation reactions are now established. However, there is still no accepted definition of an antidepressant discontinuation syndrome. Many of the reported symptoms associated with SSRI withdrawal are physical rather than psychological. Schatzberg, et al (1997) divided the somatic symptoms into five clusters: (1) disequilibrium (eg. dizziness, vertigo, ataxia) (2) gastrointestinal symptoms (eg. nausea, vomiting) (3) flu-like symptoms (eg. fatigue, lethargy, myalgia, chills) (4)　sensory disturbances (eg. paraesthesias, sensations of electric shock), and (5) sleep disturbances (eg. insomnia, vivid dreams). As well as the somatic symptoms, several core psychological symptoms – anxiety/agitation, crying spells, and irritability – are associated with SSRI discontinuation.

However, I want to concentrate on wider aspects of antidepressant discontinuation and, in particular, psychological dependence. In its broadest sense, dependence means a negative affect experienced in the absence of the drug (Russell, 1976). Taking antidepressants can become a habit. People may form attachments to their medications more because of what they mean to them than because of what they do. Patients often stay on medications, maybe several at once, even though the actual benefit is questionable. It can be more of a problem than it is worth to stop medication. Any change threatens an equilibrium related to a complex set of meanings that the medications have acquired. For example, patients may think, or may be unsure, that the drug is preventing relapse, so they do not want to take the risk of stopping it.

So, in summary, doctors do not always know best. It took some time for them to recognise the nature of antidepressant discontinuation problems. They did not use their common sense to realise that discontinuing a drug that is thought to improve mood may cause problems. Antidepressants are likely to be habit forming, so however much the medical profession may declare that they are not primarily reinforcing like psychostimulants, the public has always understood that there may be difficulties in discontinuing antidepressants. The general public might reasonably have expected that psychiatrists, who are supposed to be specialists in disorders of the mind, would recognise psychological dependence, base their advice on clinical experience, and use their common sense.

I want to suggest that there are at least three reasons for doctors’ neglect of the significance of antidepressant discontinuation problems:

These factors help to explain why there was so much delay in doctors recognising that antidepressants can cause discontinuation problems. In general, doctors are advocates of antidepressant treatment and this has led to them overlooking what should have been obvious about the risk of discontinuation problems. I will consider each of the reasons I have listed in turn.

Medication is often prescribed in life crises and serves to reinforce defensive mechanisms against overwhelming anxiety. When people are desperate they will accept almost anything that is proposed to help them. The power of the placebo needs to be recognised (Shapiro & Shapiro, 1997; Kirsch, this volume). Medicines may provide relief just because patients and their doctors believe in them.

It is not just doctors that hope for a ‘quick fix’. Ideally, we may all want a simple, quick, cheap, painless, and complete cure. In fact, in the real situation, motives may be more complex than this. In surveys, people express a reluctance to take drugs, but an inability to be free of them (Townsend, et al, 2003). Nonetheless the wish-fulfilling nature of both patient and doctor expectations is a driving factor in outcome.

There can be disastrous consequences for patients of investing their faith in the omnipotence of doctors. The combination of impressionable patients and misinformed doctors is a particularly powerful mix. As noted by Oliver Wendell Holmes as long ago as 1842:

Modern medicine likes to think it has moved on from such quackery. It may believe itself to be ethically beyond reproach, glossing over, for example, how practitioners still accept considerable amounts of hospitality, gifts and other freebies from the pharmaceutical industry and commercial medical device manufacturers. Rather than just going along with any intervention that may seem to make a difference, it tries to assess the evidence for treatments scientifically by performing randomised controlled trials. It, therefore, attempts to justify its interventions as evidence-based.

However, clinical trials tend to be tests of short-term interventions. The results do not always generalise to real-life clinical practice. For example, the average duration of trials in the NICE analysis of SSRIs vs placebo is only 6³/₄ weeks (National Institute for Clinical Excellence, 2004). This emphasis on the short-term and the episodic nature of depression helps to create the impression that occurrences of depression are easy to treat.

The resources required for longer-term studies can be prohibitive. Moreover, there are methodological problems due to attrition of subjects, leading to potential bias in the results. When longer-term treatment is studied, this tends to be done by looking at what happens when medication is withdrawn, rather than comparing outcome from the start of treatment. Patients do not do well when antidepressants are stopped. Discontinuation trials of antidepressants have a substantial relapse rate, with estimates from 36% (Klerman, et al, 1974) to as high as 92% (Prien, et al, 1974).

At face value, the evidence of discontinuation studies is that people should maintain their antidepressant treatment. But then why, if people should continue antidepressants, is the outcome for the treatment of depression over the long-term so poor? The introduction of antidepressants onto the market should have been expected to improve the long-term prognosis for depression. However, only ¹/₅^th of people make a complete recovery over a period of 15 years (Andrews, 2001). Such poor long-term results contrast with the impression created over the short-term that antidepressants are effective.

Withdrawal studies may also not be the best way to assess the value of antidepressants over the long-term. In clinical practice, distinguishing discontinuation reactions from true relapse is not always clear-cut. Maybe what is being detected in withdrawal trials is more due to discontinuation effects than true relapse. After all, we recognise the placebo effect of antidepressants when they are started. Expectations are as likely to play a role in discontinuing medication, producing a negative placebo effect, known as a nocebo reaction.

Discontinuation reactions may be particularly likely to bias the results of withdrawal studies if patients are unblinded in the clinical trials. To try and eliminate the effect of expectancies, both patients and doctors are masked from knowing whether active or placebo medication is prescribed. However, ‘double-blind’ designs are not truly double-blind (Oxtoby, et al, 1989). Both patients and raters are cued into whether active or placebo medication are being prescribed by a variety of means, including the recognition of the side-effects of the active medication.

Although not specifically examined for withdrawal studies, there is evidence from short-term antidepressant trials that blindness can be breached. When patients are asked whether they have been put onto a placebo or an active antidepressant, it is found that they can guess correctly better than would be predicted by chance (Even, et al, 2000). This "unblinding" effect is likely to be at least as strong in withdrawal studies, as patients become accustomed to the effect of taking their antidepressant and will notice when it has been discontinued.

There is also evidence of a loss of benefit emerging with long-term treatment and also on re-treatment after discontinuation of antidepressants (Baldessarini, et al, 2002). In clinical practice, it is not uncommon to find that the effect that a patient obtains from medication may seem to reduce over time. Similarly, medication may not work as well after the first time if it needs to be reintroduced. These findings reinforce the notion that the original reaction was a placebo effect.

In fact, there is some naturalistic evidence that people treated without antidepressants may do better over the long term. Using antidepressants may actually increase recurrences (Fava, 2003). The possibility that taking antidepressants may, therefore, create a vulnerability to relapse needs to be taken seriously. Maybe people who work through their problems without medication actually do better (Whitaker, 2010). This may require hard work, not a quick fix.

Evidence to support this view comes from the finding - in trials comparing psychotherapy and medication - that psychotherapy shows a significant advantage over medication at follow-up (Imel, et al, 2008). The longer the follow up the greater the difference between drug and psychotherapy. Furthermore, patients previously exposed to cognitive therapy are significantly less likely to relapse following treatment termination than patients withdrawn from medication (Dobson, et al, 2008). This indicates that antidepressant discontinuation could well be hindering patients doing better over the long-term.

Whether antidepressants actually create a vulnerability to relapse should be a major research question. However, besides the methodological difficulties of testing this hypothesis, there is also an ideological barrier to considering it, as doctors want to believe in their prescriptions. However, proper learning from the lesson of the history of the resistance to the recognition of antidepressant discontinuation reactions should help to create a more open attitude to examination of this important issue.

It is commonly believed that mental illness is due to a chemical imbalance in the brain. In fact, doctors frequently tell patients that their problems are due to a ‘chemical imbalance’. However, there is no rigorous corroboration of any chemical imbalance theory, such as the serotonin theory of depression (Lacasse & Leo, 2005). In fact, there is a significant body of contrary evidence to the simple notion that depression is due to a deficiency of serotonin in the brain. The serotonin theory of depression is no more than a theory, and most of the evidence is against it.

Psychopharmacologists trying to understand the mechanism of action of antidepressants long ago abandoned the theory. It is practising doctors who continue the myth in their everyday work. They have faith in the theory and influence their patients to believe as well. Doctors generally assume there is widespread evidence for mental illnesses being proven biological diseases of the brain. To question this presumption is almost heretical.

This conviction extends to professional medical bodies that represent doctors. For example, the American Psychiatric Association (2003), in a statement on the diagnosis and treatment of mental disorders, properly acknowledged that there are: "no discernible pathological lesions … that in or of themselves serve as reliable or predictive markers of mental disorder". However, it then went on to say that: "mental disorders will eventually be proven to represent disorders of intercellular communication or disrupted neural circuitry". Psychiatry has always had the belief that the answer to mental illness will eventually be found in the brain. It will not give up this wishful thinking, even if the evidence is against it. To be a psychiatrist seems almost to demand this step of faith.

However, my experience is that patients are able to understand that the "chemical imbalance theory" is only a theory. After all, people believe all sorts of things. What patients find more difficult to appreciate is why they are told that this theory has been proven, when this is not the case. Their doctors rarely tell them, for example, that there is evidence against the serotonin theory of depression.

So why does the myth persist? One reason is that the chemical imbalance theory is used as a means of persuading patients to take their medication. Doctors do not want to give up the theory, because they want patients to take their prescriptions. If people believe that medication corrects a chemical imbalance, this provides a rationale that they can understand and follow by complying with doctors’ orders.

The way doctors think about illness and treatment may satisfy a variety of needs, including their own professional security needs. The chemical imbalance theory helps to protect doctors’ roles and their income, prestige, and power. Modern psychiatry is no different in this respect to its 19^th century origins in society’s acceptance of the need to ‘care for’ the mentally ill by building the asylums. The chemical imbalance theory is just a modern variant of the belief that mental illness is due to brain pathology.

Reductionistic beliefs, reducing mind to brain, have always been dominant in psychiatry and still persist. It is obviously attractive to believe that the phenomena of human experience can be understood in exclusively biological terms. This viewpoint seems to give some certainty, perhaps particularly in the field of madness and mental illness, which may be difficult to understand. However, it is legitimate to question whether an explanation of human nature can take the same form as the laws of natural science. Human beings do not behave like machines.

The implication of taking the step of faith and accepting the biomedical hypothesis is that psychosocial approaches tend to be disparaged (Clark, 1981). The insistence on somatic explanations of madness produces a resistance to psychological and social interpretations. For example, Henry Maudsley, in the latter half of the nineteenth century, broadened psychiatry by promoting treatment and research rather than just confinement and ‘asylum’. As editor of the Journal of Mental Science, he expanded its scope to include psychology and philosophy. However, he still believed that, "the explanation, when it comes, will not come from the mental, but from the physical side" (Maudsley, 1874)). Maudsley’s position was that the only sound psychological science was founded on physiology.

The biomedical hypothesis is so fundamental to the edifice of psychiatry that the "chemical imbalance theory" is still believed despite contrary evidence. However, psychodynamic thinking and attempting to integrate such understanding should not be avoided because it seems difficult and requires effort to make sense of people’s feelings and actions. Honest assessment should prevail rather than professional security needs taking precedence.

To summarise, the problem for biomedical psychiatry is that, at its most extreme, it reduces people to brains that need their biology cured. There is insufficient focus on the person. Objectification of people by reducing them to abnormalities of their brains may have ethical implications for the doctor-patient relationship in assessment and treatment. Psychiatric assessment needs to be explicit that it is about understanding of the person. Shared decision-making with patients needs to be encouraged in treatment.

Furthermore, suggesting that mental illness has a physical basis serves as the justification for psychiatric interventions and institutions. The biomedical hypothesis functions as an apologia for psychiatric practice. It needs to be defended and promoted because it appears to provide the foundation for and legitimises psychiatric intervention, such as antidepressant prescribing. To deny that mental illness is a physical disease may therefore be seen as hazardous as it seems to undermine orthodox practice.

However, the apparent consensus about psychiatric conditions and their treatments, as, for example, expressed in clinical guidelines and other apparently authoritative sources, may merely represent the opinion of a dominant group (Stone, 1990; Moncrieff & Timimi, submitted). The inadequate scientific basis of psychiatry allows for widely varying interpretations and the inevitable clash of different opinions. The current authoritarian control of practice by biomedical psychiatry, based on its alleged firmer foundation, needs to be rebuffed.

Over recent years mainstream psychiatry, in fact, has become more biomedical in emphasis. There was a time when it was more pluralistic. For example, Karl Menninger’s (1963) The Vital Balance represented a broadly conceived psychosocial theory of psychopathology. The perceived need to create explicit diagnostic criteria, as in DSM-III (American Psychiatric Association, 1980), ushered in a new emphasis on biomedical aspects of psychiatry. This approach has been called neo-Kraepelinian (Klerman, 1978), as it promotes many of the ideas associated with the views of Emil Kraepelin, often considered to be the founder of modern psychiatry.

Symptom checklists and formal decision-making rules for diagnoses are now established in diagnostic manuals, following the original call for such standardisation by Feighner, et al (1972). This operationalisation of diagnostic criteria was developed specifically to respond to criticisms of the basis of psychiatric classification. The attempt to make psychiatric diagnosis more reliable, combined with a return to a biomedical model of mental illness, promotes psychiatry as a scientific, medical speciality. Mentally ill patients, who require treatment, are seen as clearly demarcated from normal people. Belittling of the value of psychiatric diagnosis is discouraged.

Biological models of mental illness have been further encouraged by neuroimaging studies of the brain, genetic linkage studies and evidenced-based evaluations of the effectiveness of psychiatric treatment (Bullmore, et al, 2009). Although psychosocial approaches have always been a minority paradigm within psychiatry, biomedical attitudes are now even more dominant. This encourages an emphasis on physical treatments, such as psychotropic medication.

The increase in antidepressant precribing that has occurred over the last two decades illustrates this point clearly. Numbers of prescriptions in the UK for antidepressants increased more than twofold in the period 1975–1998 (Middleton, et al, 2001) and continued to rise by another 36% between 2000-2005 (Moore, et al, 2009). The reason for this rise may not necessarily be due so much to an increase in new cases, but because people are staying on antidepressants long-term (Moore, et al, 2009). It only needs a small number of people to stay on antidepressants long-term to increase the total number of prescriptions dramatically. This finding emphasises that it may well be fear of discontinuation problems that is the predominant explanation for the increase in antidepressant prescribing over recent years.

It is true that patients may want an antidepressant prescription, but the problem is that doctors do not always appreciate how much they may not, and doctors find it easier to continue repeat prescribing. Doctors need to be truthful about the evidence for the chemical imbalance theory of depression, and to help people make up their minds about medication, such as antidepressants. It may be obvious, but doctors should not deceive their patients. They should be interested in helping patients decide how long they need to stay on medication.

It is generally claimed that antidepressants have been proven to be effective. In fact the results of thousands of studies of antidepressants are not nearly as conclusive as they are often claimed to be. For example, about a third of the earlier published studies showed no difference between antidepressants and placebo (Morris & Beck, 1974).

These are only the studies that have been published. Negative studies of antidepressants are much less likely to be published if only because the medical journals are more interested in publishing positive findings. Even in the trials that are published, outcomes are not always reported, particularly if the findings are negative (Chan & Altman, 2005). Trials are written up – and not necessarily by the researchers who conducted them - to emphasise the positive findings.

There are other ways in which bias is introduced. For example, conclusions in trials funded by drug companies tend to be more positive than those in which a more neutral experimenter has been involved (Als-Nielson, et al, 2003). The quality of the study is also important. Better quality studies tend to find less treatment effect than the less methodically sound studies (Juni, et al, 2001).

The root problem is the belief that mental illness is a brain disorder. It needs to be recognised that psychiatry can be practised without postulating brain pathology as the basis for mental illness. Again, I do not want to be misunderstood. I am not saying that the brain and mind are separate. Perhaps a way to express what I am saying is that mental disorders must show through the brain but not always in the brain (Double, 2006).

Psychiatry has been found out on the issue of antidepressant discontinuation problems. Therapeutic zeal has led to the justification of all sorts of groundless medical interventions. Antidepressants may turn out to be yet another example. We need to be reminded of Plato’s view that appearances on the surface may be different to the way things really are. The limits of the effectiveness of antidepressants need to be recognised. More attention should be given to the fact that doctors have made so many people dependent on them.

Als-Nielsen, B., Chen, W., Gluud, C. & Kjaergard, L.L. (2003). Association of

funding and conclusions in randomized drug trials. Journal of the

American Medical Association, 290, 921-8.

American Psychiatric Association (1980). Diagnostic and Statistical Manual of

Mental Disorders (Third edition) Washington: American Psychiatric

Association.

American Psychiatric Association (2003). Statement on Diagnosis and Treatment

of Mental Disorders. Release no 03-39, September 25th, 2003.

Andrews, G. (2001). Should depression be managed as a chronic disease?

British Medical Journal, 322, 419-22.

Blier, P. & Trimblay, P. (2006). Physiologic mechanisms underlying the

antidepressant discontinuation syndrome. Journal of Clinical Psychiatry.

67, (Supplement 4), 8–13.

Boseley, S. (2003). Seroxat maker abandons ‘no addiction’ claim. The Guardian,

3^rd May.

Bullmore, E., Fletcher, P. & Jones, P.B. (2009). Why psychiatry can’t afford to

be neurophobic. British Journal of Psychiatry, 194, 293–5.

Chan, A-W. & Altman, D.G. (2005). Identifying outcome reporting bias in

randomised trials on PubMed: Review of publications and survey of

authors. British Medical Journal, 330, 753-6.

Charney, D.S., Heninger, G.R., Sternberg, D.E. & Landis, H. (1982) Abrupt

discontinuation of tricyclic antidepressant: Evidence for noradrenergic

hyperactivity. British Journal of Psychiatry, 141, 377-86.

Clark, M. (1981). The rejection of psychological approaches to mental disorder

in late nineteenth century British psychiatry. In: Scull, A. (ed)

Madhouses, Mad-doctors and Madmen. London: Althone Press.

Dobson, K.S., Hollon, S.D., Dimidjian, S., Schmaling, K.B., Kohlenberg, R.J.,

Gallop, R.J., Rizvi, S.L., Gollan, J.K., Dunner, D.L. & Jacobson, N.S.

(2008). Randomized trial of behavioural activation, cognitive therapy,

and antidepressant medication in the prevention of relapse and

recurrence in major depression. Journal of Consulting and Clinical

Psychology, 76, 468-77.

Double, D.B. (ed) (2006). Critical psychiatry: The limits of madness. Basingstoke:

Palgrave Macmillan.

Drugs and Therapeutics Bulletin (1999) Withdrawing patients from

antidepressants. Drugs and Therapeutics Bulletin, 37, 49-52.

Even, C., Siobud-Dorocant, E. & Dardennes, R.M. (2000). Critical approach to

antidepressant trials. Blindness protection is necessary, feasible and

measurable. British Journal of Psychiatry, 177: 47-51.

course of depression? Journal of Clinical Psychatry, 64, 122-33.

Feighner, J.P., Robins, E., Guze, S.B., Woodruff, R.A., Winokur, G. & Munoz,

R. (1972). Diagnostic criteria for use in psychiatric research. Archives of

General Psychiatry, 26, 57-63.

Haddad, P., Lejoyeux, M, & Young, A. (1998). Antidepressant discontinuation

reactions. Editorial. British Medical Journal, 316, 1105-6.

Haslam, J. (1798) Observations on Insanity. London: Rivington.

Haslam, J. (1817) Considerations on the Moral Management of Insane Persons.

Imel, Z.E., Malterer, M.B., McKay, K.M. & Wampold, B.E. (2008). A meta-

dysthymia. Journal of Affective Disorders, 110, 197-206.

Jüni, P., Altman, D.G. & Egger, M. (2001). Assessing the quality of controlled

clinical trials. British Medical Journal, 323, 42-6.

Karlowski, T.R., Chalners, T.C., Frenkel, L.D., Kapikian AZ, Lewis TL &

Lynch JM. (1975). Ascorbic acid for the common cold. Journal of the

American Medical Association, 231, 1038-42.

Kirsch, I. (2010) The Emperor’s New Drugs: Exploding the antidepressant myth.

New York: Basic Books.

Kirsch, I., Moore, T.J., Scoboria, A. & Nicholls, S.S. (2002) The emperor’s new

drugs: an analysis of antidepressant medication data submitted to the

U.S. Food and Drug Administration. Prevention & Treatment, 5, Article