Is depression overdiagnosed? YesBMJ 2007; 335 doi: http://dx.doi.org/10.1136/bmj.39268.475799.AD (Published 16 August 2007) Cite this as: BMJ 2007;335:328
- Gordon Parker, scientia professor
It is normal to feel depressed. In our cohort study of 242 teachers, the 1978 baseline questionnaire defined depression as “a significant lowering of mood, with or without feelings of guilt, hopelessness and helplessness, or a drop in one's self-esteem or self-regard.”1 Ninety five per cent of the cohort reported such feelings (with a mean of six episodes a year), indicating the ubiquitous nature of depressed mood states.
A low threshold for diagnosing clinical depression, however, risks normal human emotional states being treated as illness, challenging the model's credibility and risking inappropriate management. When the first antidepressant (imipramine) was developed, manufacturer Geigy was reluctant to market it,2 judging there were insufficient people with depression. Now, depression is all around, and antidepressant drugs have a dominant share of the drug market. Reasons for the overdiagnosis include lack of a reliable and valid diagnostic model and marketing of treatments beyond their true utility in a climate of heightened expectations.
Fifty years ago, clinical depression was either endogenous (melancholic) or reactive (neurotic). Endogenous depression was a categorical biological condition with a low lifetime prevalence (1-2%). By contrast, reactive depression was exogenous (induced by stressful events affecting a vulnerable personality style). In 1980, the American Psychiatric Association released and promoted the third revision of its diagnostic and statistical manual (DSM-III) as a reliable criterion based system. It radically divided clinical depression into major and minor disorders.
The simplicity and gravitas of the term “major depression” gave it cachet with clinicians (despite unpublished field trials suggesting that diagnostic allocation was less reliable than that in DSM-II3) and helped patients get medical insurance cover.4 Although its descriptive profile prioritised melancholic features (such as serious psychomotor disturbance or anergia), DSM-III's operational criteria were set “at the lowest order of inference.” Criterion A required a “dysphoric mood” for two weeks, including feeling “sad, blue . . . down in the dumps.” Criterion B (mandating four of eight listed items) could be met by some level of appetite change, sleep disturbance, drop in libido, and fatigue. Formal trials confirmed the low reliability of these criteria,4 and large US community studies showed variable lifetime prevalences—ranging from 6.3%5 to 17.1%6—that paradoxically fell with age. Why? Failure to recall lifetime episodes was shown to be greater for major depression than for other disorders,6 suggesting that its criteria capture less important (and forgettable) depressive states. Studies that assess cohorts at intervals (to overcome forgotten episodes) report much higher lifetime rates of major depression (such as 42% in our teachers' cohort7).
Minor DSM-III depressive disorder (dysthymia) homogenised less severe chronic conditions, requiring even fewer and less substantive symptoms (such as crying, decreased productivity, and feeling sorry for yourself). This dimensional model was subsequently extended by proposing an even less severe condition, subsyndromal or subclinical depression. Its one year prevalence in a US community database was nearly triple that of major depression, encouraging those investigators to argue for its “clinical and public health importance” and treatment.8
How low do we go?
Determining caseness for any dimensional construct requires imposing a cut-off, risking underdiagnosis of true cases or overdiagnosis of non-cases. By 1993, 79% of teachers in our cohort (in their late 30s) had already met symptom and duration criteria for lifetime major, minor, or subsyndromal depression (unpublished data). Although it was absolutely necessary to redress psychiatry's earlier weighting to melancholia, the extended dimensional model risks medicalising normal human distress and viewing any expression of depression as mandating treatment.
The reality that many people with substantive clinical depressive disorders still do not have their condition diagnosed does not, by itself, mean that depression is underdiagnosed. Such boundary concerns have multiple parallels. For example, the diagnosis of attention-deficit/hyperactivity disorder is often missed; conversely, it is often falsely diagnosed in children with other disruptive behaviours.
Does overdiagnosis matter?
Does current looseness matter if a low diagnostic threshold destigmatises depression, encouraging people to seek help and allowing clinical assessment? After all, breast screening programmes may lead to detecting more malignant lumps. However, false positives results generated by breast screening are filtered out by refined assessment, and harm rarely occurs. For false positive detection of depression, many of psychiatry's leaders mandate treatment, which for many with less severe conditions raises hopes but results in a sequence of ineffective and inappropriate treatments.
The ease of assigning a diagnosis of clinical depression, even of major depression, has rebounded on psychiatry, blunting clarification of causes and treatment specificity. As Hickie, who argues here against overdiagnosis, observed elsewhere: DSM-III defined major depression has failed “to demonstrate any coherent pattern of neurobiological changes [or] any specific pattern of treatment response outside in-patient treatment settings.”9
Meta-analyses show striking gradients favouring antidepressant drugs over placebo for melancholic depression. Yet trials in major depression show minimal differences between antidepressant drugs,10 evidence based psychotherapies,11 12 and placebo. The benefit of treatment for the minor and subsyndromal depressions is even more unclear. Extrapolating management of the more severe biological conditions to minor symptom states reflects marketing prowess rather than evidence, ignoring Nobel Laureate Richard Feynman's observation that “Things on a small scale behave nothing like things on a big scale.”
Depression is a diagnosis that will remain a non-specific “catch all” until common sense brings current confusion to order. As the American journalist Ed Murrow observed in another context: “Anyone who isn't confused doesn't really understand the situation.”
Competing interests: GP is executive director of Australia's Black Dog Institute and has served on several pharmaceutical advisory boards and spoken at many meetings convened by drug companies.