Intended for healthcare professionals


How scientifically valid is the knowledge base of global mental health?

BMJ 2008; 336 doi: (Published 01 May 2008) Cite this as: BMJ 2008;336:992
  1. Derek Summerfield, honorary senior lecturer
  1. 1Institute of Psychiatry, King’s College, London SE5 8AF
  1. derek.summerfield{at}
  • Accepted 26 January 2008

Mental health is a construct that cannot be seen as independent of culture, society and situation. Derek Summerfield argues that Western definitions and solutions cannot be routinely applied to people in developing countries

Global mental health now has its own academic units, literature, study, and training courses and the World Health Organization is a major articulator of this work. Last September, a series on global mental health in the Lancet asserted that mental disorders now represent a substantial “though largely hidden” proportion of the world’s overall disease burden, that every year up to 30% of the global population would develop some form of disorder, and that there was strong evidence for scaling up mental health services worldwide.1 2 In this article I examine the evidence for these claims and challenge the assumption that Western frameworks can generate a universally valid knowledge base.

Classification and diagnosis

Psychiatric research and practice rest on empiricist convention rather than on timeless discovery. The principal classification systems, the International Classification of Diseases (ICD) and the Diagnostic and Statistical Manual of Mental Disorders (DSM), are shaped by contemporary notions about what constitutes a real disorder, what counts as scientific evidence, and how research should be conducted. They are Western cultural documents par excellence.3

The diagnostic categories within these classifications are essentially conceptual devices emerging from committee decisions. The authors of both of these classifications are careful to point out that “there is no assumption that each category of mental disorder is a completely discrete entity with absolute boundaries dividing it from other mental disorders or from no disorder.”4 Despite this, in everyday practice these categories have been effectively accepted as if they were unequivocal diseases like, say, tuberculosis.

Claims for the universality of a particular psychiatric category would be compelling if a straightforward biological cause had been established. But this is the case for only a few categories in the Western psychiatric canon. Official psychiatric categories evolve, disappear, or appear, reflecting Western social and cultural trends as much as medical thinking. For example, DSM-III, published in 1980, expunged homosexuality as a classified mental illness and installed post-traumatic stress disorder for the first time. It would be extraordinary if these classification systems made sense across the great diversity of societies and situations worldwide.

Validity in mental health research

The psychiatric literature contains large numbers of studies of non-Western subjects, and publications on refugees and populations affected by disaster rose exponentially during the 1990s.5 However, in a review of 183 published studies on the mental health of refugees, four fifths of the studies relied exclusively on measures of psychopathology developed with Western populations.6 Although a few researchers have sought to garner culture specific information, they too have retained Western psychiatric categories as their basic template.7 8

The relevance of much of this published literature is therefore questionable because it fails the fundamental test of scientific validity. Validity is a concept meant to assess “the nature of reality” for the people being studied. Methods for research into the mind cannot rely on the merely technical and quantifiable, since “the nature of reality” is bound up with local forms of knowledge and philosophy. To ignore the test that follows – whose reality, whose knowledge, whose philosophy – is to risk pseudodiagnoses and interventions that are an unwanted distraction in the hard pressed lives of non-Western subjects.

It is telling that published studies of non-Western populations often refer to participants’ “limited knowledge of mental disorders,” their lack of “mental health literacy,” or the need to “teach” health workers and the people they serve about mental health. Here Western psychological discourse is setting out to instruct, regulate, and modernise, presenting as definitive the contemporary Western way of being a person. It is unclear why this should be good for mental health in Africa or Asia. This is medical imperialism, similar to the marginalisation of indigenous knowledge systems in the colonial era, and is generally to the disadvantage of local populations.9 10

Cultural variability of mental disorder

Many ethnomedical systems have categories that range across the physical, supernatural, and moral realms and do not conceive of illness as situated in body or mind alone.11 Distress is commonly understood and expressed in terms of disruptions to the social and moral order. As Kleinman and Good put it, “Cultural worlds may differ so dramatically that translation of emotional terms means more than finding semantic equivalents. Describing how it feels to be aggrieved or melancholic in another society leads directly into an analysis of a radically different way of being a person.”12

Since the 1970s many ethnographic studies have shown that the presentation, attribution, classification, prevalence, and prognosis of mental disorders varies greatly between cultures. Journals such as Social Science and Medicine, Transcultural Psychiatry, Anthropology and Medicine, and Culture, Medicine and Psychiatry are replete with this work. However, it is under-represented in the mainstream psychiatric literature, which tends to overemphasise cross-cultural similarities and minimise the differences.13

Depression as an example

WHO has stated that “depression” is a worldwide epidemic that within a decade will be second only to cardiovascular disease in terms of global disease burden. But it is a basic error of validity to assume that because Western mental phenomena can be identified in non-Western settings, they mean the same as they do in the Western world. Some central and southern African societies recognise a local condition - often translated as “thinking too much” – which has some physiological overlap with Western depression, but the understandings, attributions, and remedies applied to it are very different.

As discussed above, the DSM or ICD definition of depression is merely a descriptive syndrome, highly heterogeneous and socially shaped, albeit subsuming a small subset of severely dysfunctional people who do fit the model of depression as a disease. This term simply cannot be used (as WHO uses it) to denote a universally valid mental disorder that is amenable to a standard mental health toolkit.14

A meta-analysis of 56 published studies of refugees’ mental health totalling 67 294 participants found the strongest moderating factor to be social conditions after displacement.15 Resolution of the conflict that had displaced them also had positive effects. This suggests that the mental phenomena being identified as satisfying criteria for a mental disorder (typically depression or post-traumatic stress disorder) were mostly incidental and a normal reaction to their circumstances.15 My research and clinical experience in war affected rural Zimbabwe, with rural peasants and wounded former soldiers in Nicaragua, and with HIV positive African women and other asylum seekers in the United Kingdom indicated that a diagnosis such as depression, although commonly made, had little power to explain their problems.16 17

Lastly, WHO is backing the global application of technologies like checklists to estimate population prevalences of depression and the therapeutic value of antidepressants or counselling when the strength of their evidence base even in Western societies remains controversial.18

Mental health in a broken social world

Social scientists agree that wellbeing and personal resilience are linked to social connectedness and the sense of a coherent world. Throughout the non-Western world, structural poverty and injustice, violent conflict, debt repayments, shifting weather patterns, environmental degradation, and inadequate budgets for health, education, and social welfare provide a barely viable social context for millions of people. Around 85% of Kenya’s population growth in the 1990s was absorbed into the slums of Nairobi and Mombasa, reflecting a rapid withering of traditionally self sufficient ways of life.19 Average life expectancy is falling in many countries, and one third of the world’s children are undernourished. Can psychiatric approaches honed in relatively well resourced and stable societies distinguish mental disorder from normal responses to a social world that is no longer coherent or functional? The danger of the medicalisation of everyday life is that it deflects attention from what millions of people worldwide might cite as the basis of their distress - for example, poverty and lack of rights.

Valid assessment

Until these fundamental queries about the knowledge base are answered, claims about the current state of global mental health seem risible. There needs to be agreement that Western psychiatry is but one among many ethnopsychiatries. New research ideally needs to engage with participants in a way that carries no preformed notions about what is “mental” or “health” in their world: local concepts must be the starting point for the creation of valid instruments for screening or diagnosis.

The so called common mental disorders, such as post-traumatic stress disorder and depression, have strongly featured in the research literature to date, but there are reasons to query how much the impetus for this has come from those on the ground, as opposed to interested parties from outside. WHO’s prevalence figures lack credibility and would seriously mislead health planners and providers.

Research into more severe mental disorders is on firmer ground. These disorders are much more likely to have a neuropsychiatric component and to stretch local social and healthcare resources. Based on a survey of 2739 households, Patel and colleagues estimated standardised prevalence ratios of psychosis (2.79), learning difficulties (1.48), and seizures (2.00) (seizures are commonly seen as a mental disorder in much of Africa) in rural and urban Mozambique.20 Psychosis would not of course be a homogeneous category, since it would subsume various organic cerebral states that merited attention in their own right – not least AIDS encephalopathy or dementia.

An emphasis on qualitative work would promote more grass roots ownership of the terms of reference of mental health and enable a robust and relevant knowledge base to emerge. The apparent failure of the £1.1 trillion spent on development aid in Africa over the past 50 years to change anything is partly because donors lacked knowledge of the situations on the ground in poor countries and because of over-reliance on expensive consultants from donor countries.21 Local voices were weakly represented. This poses an ethical challenge for global mental health: non-Western people can give properly informed consent only if the terms in which they are being represented, which here means candidature for psychiatric caseness, are not alien or irrelevant to their interpretations of the world.

Summary points

  • Psychiatric categories are descriptive syndromes not diseases

  • Western categories are not universally valid

  • Qualitative research into mental health is needed in non-Western settings

  • Perceived mental health problems may be a normal reaction to harsh living conditions


  • Contributors and sources: DS has studied, lectured, and published widely on health and human rights in Africa, Central America, and the Middle East, particularly in war affected populations and on asylum seekers in the UK. He is also a teaching associate at Refugee Studies Centre, University of Oxford.

  • Competing interests: None declared.

  • Provenance and peer review: Not commissioned; externally peer reviewed.


View Abstract