Intended for healthcare professionals

Analysis Too Much Medicine

Medicalising unhappiness: new classification of depression risks more patients being put on drug treatment from which they will not benefit

BMJ 2013; 347 doi: (Published 09 December 2013) Cite this as: BMJ 2013;347:f7140
  1. Christopher Dowrick, professor of primary medical care1,
  2. Allen Frances, emeritus professor of psychiatry 2
  1. 1University of Liverpool, Liverpool L69 3GL, UK
  2. 2Duke University Medical Center, Durham, North Carolina, USA
  1. Correspondence to: C Dowrick cfd{at}

Summary box

  • Clinical contextDiagnoses of major depressive disorder and treatment with antidepressant drugs are increasing

  • Diagnostic change—DSM-III homogenised the diagnosis of depression and the new DSM-5 classification broadens the definition further, allowing the diagnosis of major depressive disorder just two weeks after bereavement

  • Rationale for change—To provide more patients with access to effective treatments

  • Leap of faith—Accurate diagnosis of mild depression is possible; treatment is necessary and leads to better outcomes

  • Increase in disease—Although community prevalence of major depressive disorder has remained static, diagnoses doubled among Medicare recipients in the US between 1992-95 and 2002-05

  • Evidence of overdiagnosisDepression is now more likely to be overdiagnosed than underdiagnosed in primary care. Rates of prescribing of antidepressant medication doubled in the UK between 1998 and 2010 and in the US 11% of the population aged over 11 now takes an antidepressant. People without evidence of major depressive disorder are being prescribed drug treatment

  • Harms from overdiagnosisTurning grief and other life stresses into mental disorders represents medical intrusion on personal emotions. It adds unnecessary medication and costs, and distracts attention and resources from those who really need them

  • LimitationsWe do not know whether clinicians will follow the DSM-5 proposals

  • ConclusionsPatients with mild depression or uncomplicated grief reaction usually have a good prognosis and don’t need drug treatment. Clinicians should focus on identifying people with moderate to severe depressions and sufficient impairment to require treatment.

Many patients report sadness or distress during consultations with primary care doctors. Such emotions may be related to grief and other life stresses, including the stress of physical illness. Sometimes sadness appears out of the blue, without obvious relation to external causes. Over recent decades there has been an increasing tendency, especially in primary care, to diagnose depression (commonly major depressive disorder) in patients presenting with sadness or distress and offer them antidepressant medication.1 2 3

In this paper we offer a critical review of the diagnosis of major depressive disorder, show how and why this broad diagnostic label has resulted in overdiagnosis and overtreatment, and suggest how the approach to diagnosis and management of depression should change to reduce stigmatising the sad and provide better help for those who most need vigilant care and medical treatment.

Evolving views of what constitutes depression

Descriptions of depression can be found in the Bible and Shakespeare, but no formal definition existed until the third version of the American Psychiatric Association’s classification systems for mental disorders was published in 1980 (DSM-III). The manual set out clear operational criteria to aid clinicians in diagnosing mental disorders (see box 1) and introduced the term major depressive disorder.

Box 1: Diagnostic classification of mental disorders

Diagnostic and Statistical Manual of Mental Disorders (DSM) III (1980) introduced a unitary diagnosis of major depressive disorder (MDD) with nine symptoms (mood, interest, activity, fatigue, weight/appetite, sleep, guilt, concentration, and suicidality).

The main emphasis of DSM-III was on severity ratings:

  • Mild MDD was defined as five or six symptoms of mild severity, including either low mood or loss of interest

  • Moderate MDD was defined as seven to eight symptoms with moderate impairment

  • Severe MDD was defined as six or more symptoms with severe impairment or psychotic features and strong suicidal intent

DSM III also included depressive subtypes:

  • Secondary depression arose from a clear external cause

  • Psychotic depression had associated psychotic symptoms

  • Melancholic depression involved lack of pleasure or lack of mood reactivity plus three of the following: subjective mood qualitatively different from grief or loss; severe loss of appetite or weight; psychomotor agitation or retardation; early morning waking, excessive guilt; and mood worse in the morning.

DSM-IV (1994) used a similar classification system.4 DSM-5 (2013) now allows grief reaction to be classified as major depressive disorder (MDD) after two weeks.

The International Statistical Classification of Diseases and Related Health Problems (ICD), produced by the World Health Organization, also includes criteria for mental disorders. This is used more widely in Europe and other parts of the world. The coding systems of DSM and ICD are designed to correspond with each other

Since then major depressive disorder has received more research attention than any other diagnosis in psychiatry but has created many problems. The criteria, which have not changed since 1980, capture too heterogeneous a population for research studies and are so loose that, in everyday clinical practice, ordinary sadness can be easily confused with clinical depression.5

Unhelpful classifications of mental disorders

Under DSM-III the term major depressive disorder combined what had formerly been described as “melancholia”—characterised by severe, disabling, and sometimes life threatening depression, often coming out of the blue and characterised by marked diurnal variation, suicidal thoughts, and somatic symptoms—with “reactive depression.” Reactive depression contrasted in almost every way with melancholia, with onset closely linked to a definable life event and with symptoms that were milder and typically including sadness, loss of interest, and feelings of guilt and unworthiness. Somatic changes, including difficulty sleeping and loss of appetite, were less profound and enduring in reactive depression than in melancholia. Those affected retained the capacity to feel pleasure. Symptoms were usually least troubling in the morning and patients tended to get better over time and respond well to placebo and psychotherapy. Those with melancholia, by contrast, were more likely to have disturbed sleep and abnormal dexamethasone suppression test results6 7 and to respond to drug treatment or electroconvulsive therapy.8

Although the DSM-III definition of major depressive disorder was meant to provide simplicity and increase reliability of diagnosis, from the beginning it was recognised that it would capture a heterogeneous population of patients. The definition therefore provided severity ratings and different subtypes (box 1). Unfortunately, however, the valuable distinctions offered by severity and subtype ratings were generally ignored in both clinical practice and research.9 Major depressive disorder became homogenised to include “mild” major depression—arguably a contradiction in terms for it is not major, nor really depressive or a disorder.

DSM-IIIR and DSM-IV carried forward the DSM-III definition, and the recently published DSM-5 broadens the diagnosis of major depressive disorder still further. It allows major depressive disorder to be diagnosed just two weeks after a bereavement. The change in the diagnostic status of grief from bereavement (not a mental illness) to depressive episode (a mental illness) introduced by DSM-5 was designed to provide more patients with access to effective treatments.10 This is particularly relevant in insurance based health systems such as the US, where a specific diagnosis is needed before funders will agree to pay the costs of treatment. It has, however, provoked both controversy and concern focused on the medicalisation of the normal human experiences of loss.5 11 12

Homogenisation of depression has been a mistake

People with uncomplicated episodes of major depressive disorder (lasting no longer than two months and not including suicidal ideation, psychotic ideation, psychomotor retardation, or feelings of worthlessness) are hardly more likely to have a further episode within 12 months than people with no history of major depressive disorder (3.7% v 3.0%).13 These episodes, along with mild and non-melancholic episodes, may be better understood as normal intense sadness.13 An Australian primary care study of 789 patients with depressive disorders found four different trajectories: most patients (n=532) had a mild and static symptom trajectory, very different from the experience of the small minority of people (n=69) with severe persistent depression, who had high levels of disadvantage, abuse, morbidity, and disability.14

Including people, as the DSM- 5 classification does, who are experiencing grief only two weeks after the loss of a loved one is a mistake. Bereaved people follow a course very different from those with recurrent major depressive disorder. A study of over 30 000 US citizens found that single bereavement related brief depressive episodes have distinct demographic and symptom profiles that differ from those of other types of depressive episodes and are not associated with increased risk of future depression.15 Uncomplicated bereavement is not associated with an increase in suicidality.16

Increase in diagnosis of depression and antidepressant drug prescriptions

The prevalence of depressive disorders in the community is stable. In the United States two national comorbidity surveys a decade apart found prevalences of major depressive disorder of 6.1% and 6.6%.17 18 In England the one week prevalence of depressive episodes among adults decreased from 2.6% in 2000 to 2.3% in 2007.19 20

Meanwhile rates of diagnosis have increased considerably. Although community surveys that use lay interviewers have shown little change in prevalence over time, diagnoses of depression among Medicare beneficiaries doubled between 1992-95 and 2002-05.1 This is not because primary care doctors are getting better at identifying major depressive disorder; overdiagnosis is now more common than underdiagnosis. A meta-analysis of 41 studies, including 50 371 patients, estimated that for every 100 unselected cases seen in primary care, there were more false positive cases (n=15) than either missed (n=10) or identified cases (n=10) of depressive disorder as judged by standard diagnostic criteria.2 In a study of 5639 participants from the 2009-2010 US National Survey of Drug Use and Health, clinician identified depression was compared with assessments for major depressive episodes using a structured interview. Only 38% of adults (including only 9% of those aged 65 and over) with clinician identified depression met diagnostic criteria for depression during the previous year; nevertheless, most participants were taking psychiatric drugs.21 The trend to overdiagnosis may increase as DSM-5 diagnostic criteria loosen.

Rates of prescribing of antidepressants to patients having no evidence of major depressive disorder, or fewer symptoms than DSM would advise, are also increasing in primary care. The proportion of visits to non-psychiatrists at which antidepressants were prescribed but no psychiatric diagnoses were noted increased in the US from 60% to 73% between 1996 and 2007.22

About 11% of the US population aged over 11 now take an antidepressant, including 23% of women in their 40s and 50s.3 In England, antidepressant prescribing increased at over 10% each year between 1998 and 2010,23 a rise far greater than for any other psychiatric medication. This is explained mainly by an increase in long term prescriptions.24 Similar rises have been described in other Western nations including Australia,25 Canada,26 and Denmark.27

Drivers of overdiagnosis

The homogenisation of major depressive disorder has been in part a consequence of heavy drug company marketing and an overstrong focus among many psychiatrists on the biological correlates of psychiatric symptoms rather than the psychological, social, and cultural.28 The rate of diagnosis of depression has increased substantially since the development and marketing of selective serotonin reuptake inhibitors,29 a trend assisted by drug companies’ financial support for prominent academic psychiatric units12 and direct to consumer advertising in the US.

General practitioners and the public are complicit in this. For GPs a diagnosis of depression may be an attractive instrument for managing uncertainty in the consulting room,30 especially as its commonest treatment comes in the form of a once daily pill and is encouraged by clinical guidelines and indicators. Patients often request treatment for symptoms of sadness, and doctors and patients can feel obliged to offer and accept a diagnosis of major depressive disorder.31

In addition, there is a trend in Western societies to expect the right to happiness and a need to restrict the range of negative emotions that are considered “acceptable and normal.”32 Pharmaceutical companies and psychiatric nosologists derive their positions on depression diagnoses from a set of common but implicit value judgments.33

What the evidence shows

The weight of evidence from meta-analyses of placebo controlled trials shows that antidepressant drugs have little or no effect in mild depression.34 35 36 37 Although there is some evidence that the benefits of treatment compared with placebo are not related to baseline severity,38 there are continued concerns about publication bias in data provided by drug companies.39

The placebo effect of antidepressant drugs is substantial and increasing, partly because less severely depressed people now take part in drug trials.40 The role of regression to the mean in assessing the effects of antidepressants is also important since many people with reactive depression get better with time, regardless of treatment. Watchful waiting can have a stronger effect than antidepressants.41

There is no substantive evidence that people with uncomplicated bereavement benefit from antidepressants, and a dearth of clinical trial evidence of response to medication in those with complicated grief reactions.13 Many conditions currently diagnosed as major depressive disorder, especially those related to other forms of loss, are better understood within a model of grief that does not assume drug treatment.42

Harms from overdiagnosis

Turning grief and other responses to loss into a mental disorder is a medical intrusion into private emotions.43 It substitutes a superficial medical ritual for deep and time honoured cultural ones and stigmatises the experience.33 It leads people to act under the description of a psychiatric diagnosis, believing themselves to be and behaving as if they are someone with a mental illness and compromising their sense of agency.33 44 By putting a simplistic time frame on recovery from grief, DSM-5 is stepping further away from the personal interaction that should be the basis of healthcare.

These problems are greater in cross-cultural consultations with patients for whom depression may be an alien concept. Recent asylum seekers, for example, have often experienced severe and traumatic losses. Subsuming the consequent distress within a diagnosis of depressive disorder replaces loss with illness and individualises previously social problems.45

Bringing grief within the category of major depressive disorder adds unnecessary medication, with its inevitable side effects and carries added risks including increased suicidal thinking in children and young people.46 and risks of interaction with drugs prescribed for other health problems.47 Unjustified use of antidepressants increases the costs of healthcare. The excess cost of care associated with prescribing antidepressants to older Canadians without depression was estimated as $C1800 (£1000; €1250; $1700) per person.48 An expanded focus on grief will also affect existing psychiatric diagnosis, distracting attention and resources from those who have severe mental health problems.

How to do better

Diagnostic criteria should be tightened. Milder symptoms must be persistent throughout the day, be present for at least a month or two and cause significant distress or impairment before a diagnosis of mild major depression is made. For moderate and severe major depression, existing diagnostic criteria should be accurately applied, so that diagnoses are made only in the presence of substantive symptoms and clear associated impairment. Patients presenting with milder or loss related symptoms should not be dismissed, but more attention should be given to benefits of time, support, advice, social networks, and psychological interventions.49 There are opportunities to avoid the mistakes of DSM-5 in ICD-11, which is now in preparation.

GPs should focus on identifying patients with severe depression50 and provide them with better access to adequate evidence based care.51 These include two main groups of patients: those with symptoms of melancholia and those with severe persistent symptoms associated with socioeconomic disadvantage and disability (box 1).14

Drug companies should be stopped from marketing antidepressant medication to physicians and the public and from supporting professional organisations and consumer groups.

What to discuss with patients

High rates of placebo response account for much of the seeming beneficial effects of medication and this should be discussed sensitively with patients, who also need to be made aware of the side effects, risks and costs associated with antidepressants. Informing them of the way that drug companies have acted to boost sales of their drugs may also be appropriate. There is still a widely held view that all depression is “brain disease” caused by chemical imbalance which can be “corrected” by pills, and countering it is important by noting the relevance of life circumstances.

Patients can be helped by listening carefully to their story, promoting the value of time as a healer and encouraging them to build resilience through exercise, support, and (where possible) making changes to their circumstances in dealing with everyday life problems.52 A diagnosis of depression may not be necessary (box 2). A shared approach to decision making is essential. Patients should also be encouraged to share experiences and learn from others through organisations such as Healthtalkonline ( Watchful waiting over multiple visits can enable doctors to see if the problems will resolve without intervention, an approach that plays to the strengths of experienced primary care doctors.

Box 2: Is my patient really depressed?

  • Before diagnosing depression, listen carefully to the patient’s story and consider the context:

    • Has the patient experienced grief or other life problems?

    • Are symptoms mild and recent?

    • Is this a first episode?

  • Mild symptoms, or symptoms related to grief or other life problems, usually do not become more severe over time and a diagnosis of depression should be avoided whenever possible

  • Patients with mild depressive symptoms don’t need antidepressants—any benefit is likely to be due to the high placebo effect

  • Mild symptoms, or symptoms related to grief or life problems, are likely to resolve with time, psychological support, and environmental manipulation


Cite this as: BMJ 2013;347:f7140


  • doi:10.1136/bmj.f7225

  • This article is part of a series on overdiagnosis looking at the risks and harms to patients of expanding definitions of disease and increasing use of new diagnostic technologies

  • We thank Tessa Richards, Josie Billington, Bernard Carroll, Jane Gunn, Iona Heath, Tony Kendrick, Peter Lucassen, Don Nease, Gordon Parker, Greg Simon, Don Stein, Brett Thombs, Jerry Wakefield, Chris van Weel, and Evelyn van Weel-Baumgarten for advice and comment on earlier versions of this paper:

  • Competing interests: We have read and understood the BMJ policy on declaration of interests and declare the following interests: AF was chair of the DSM-IV Task Force and has written two books related to DSM-5.

  • Provenance and peer review: Commissioned; externally peer reviewed.


View Abstract