Is DSM-5 defensible?

As we move closer to the release of the fifth revision of the Diagnostic and Statistical Manual of Mental Disorders (DSM-5) of the American Psychiatric Association (APA), the clamour of its critics appears to be growing louder (Anonymous, 2012; Watts, 2012a). The critiques are not restricted to scientific journals, but are freely offered in the lay press (Frances, 2012) and on the World Wide Web (Society for Humanistic Psychology, 2012), and are emanating from psychiatrists, psychologists, philosophers, theologians, other professionals, lay commentators, and people diagnosed with mental illness. While the tone and emphasis of the criticisms vary, they can be broadly categorised as follows: (a) the APA, comprising exclusively of American psychiatrists, has no mandate to delineate a normal from abnormal mental condition, a determination that should involve many other disciplines; (b) psychiatric knowledge and neuroscience have not advanced sufficiently in the last two decades to warrant a major revision of DSM-IV, and DSM-5 is little more than an attempt to capitalise on a best-seller; (c) DSM-5 is lowering the thresholds of various diagnoses, resulting in the medicalisation of normal human experience and the creation of spurious ‘epidemics’ of mental illness; (d) DSM-5 is creating new mental disorder diagnoses for the benefit of the profession and/or the pharmaceutical industry; (e) DSM-5 work group members are compromised by their declared and undeclared conflicts of interest; (f) DSM-5 continues to inappropriately impose categorical constructs on dimensional mental states or conditions; and (g) DSM-5 is merely perpetuating and exacerbating the deficiencies of previous psychiatric classifications without working towards a truly scientific classification. The examples commonly used to support the criticisms are the proposed criteria for attention deficit hyperactivity disorder (ADHD), disrupted mood regulation disorder, attenuated psychosis syndrome, mixed anxiety depression, substance use disorder, the proposed removal of bereavement as an exclusion for the diagnosis of depression, and the overhaul of personality disorders as published on the DSM-5 website (www.dsm5.org) in early 2010.

As a member of the Neurocognitive Disorders Work Group of DSM-5, I am familiar with some of the procedures involved in bringing DSM-5 into shape, although being an overseas member practising in Australia places me, I believe, at some objective distance from the process. From this vantage point, some of the criticisms are plainly wrong. It is, I feel, quite appropriate for psychiatrists to draft a classification of mental disorders, just as oncologists would be expected to classify cancers and haematologists blood disorders. The APA, as a representative national body of psychiatrists, has enough intellectual and material resources to be able to do this, as it has done on four previous occasions. It has arguably done this very well, as evidenced by the de facto adoption of DSM-III and DSM-IV by international psychiatry. It is not true that other disciplines have not been involved; our own work group had a neurologist and a neuropsychologist, and external experts were very often non-psychiatrists. It has been frequently mentioned that neuroscientific knowledge is exploding exponentially; to revisit the classification of mental disorders after two decades should not be viewed as an indulgence. The process of selecting work group members in relation to conflicts of interest was the most rigorous I have ever encountered. While it may be impossible to find a group of leading experts in a field of medicine with no links to the drug industry, this group of experts was arguably one of the cleanest one could muster. I agree that the DSM ‘enterprise’ has been very lucrative for the APA, but the $25 million (USD) already spent on the fifth revision process (American Psychiatric Society 2012) does not appear to be a great investment if book royalties were the primary objective. The process of the revision is extremely rigorous, and any proposal for a new disorder or a major revision of existing criteria needs to come on the back of strong scientific evidence. The number of disorders in DSM-5 has not increased relative to DSM-IV, and in fact the field trials suggest that the overall rates of mental disorders with DSM-5 criteria are lower than that for DSM-IV criteria (Society for Humanistic Psychology, 2012).

How, then, does one explain the chorus of criticism? In my opinion, it arises from a number of misunderstandings about the purpose of classifying mental disorders, and in particular the objectives of DSM-5. The primary purpose is to enable physicians (or other clinicians) to ‘diagnose’ a condition in an individual who seeks clinical attention. The diagnosis ostensibly represents a ‘disorder’ – a disease or a malfunction – but only insofar as it can be done reliably and that it is valid to do so. The failing of psychiatric neuroscience is that there is, with the rare exception (as in neurocognitive disorders), no other way of truly ‘knowing’ whether a disease really exists. This, however, is not a failing of DSM-5 or its predecessors; it is a limitation the work groups have had to struggle with. The experts fully appreciate that what they encounter in the psychiatric clinic is often a more severe form of a phenomenon that pervades society, and their role is to identify if it is severe or deviant enough to warrant attention. DSM-5 is intended to help them make that decision. The threshold they apply to make a ‘diagnosis’ is not a culturally influenced whim, but is based on the distress or dysfunction related to the symptoms, the long-term prognosis based on empirical evidence and the likely impact of intervention. Culture does influence the decision, as in ADHD or bereavement, as psychiatrists are indeed embedded in their respective cultures, but their training forces them to examine the appropriateness of the diagnosis with the lens of culture as flattened as possible.

It is inevitable that a group of scientific experts when asked to examine a problem will attempt to reduce it to its component parts in minute detail. The result is the multitude of mental disorders that comprise the DSM-5. When attempting to apply a classification after a ‘7-minute’ consultation, a general physician may find the exuberance in categories and criteria frustrating and impractical. The problem is that there is only one DSM-5 to cater for the needs of general adult, child and old-age psychiatrists, as well as general physicians, psychologists, forensic specialists and insurance companies, all of whom have different needs and expectations. Yet, a classification developed primarily by psychiatrists cannot but have a practising psychiatrist as its foremost client. Another unintended consequence of the large number of disorders is that patients are very likely to be given more than one diagnosis, what is erroneously referred to as ‘comorbidity’ when it is uncertain and unlikely that this reflects multiple diseases.

DSM-5, and any other classification of mental disorders, is not an attempt to define what is normal. Being normal is not the same as ‘not having a DSM-5 diagnosis’, and having such a diagnosis is not the same as being ‘insane’, as some have wrongly argued about the DSM (Caplan, 1995). Insanity is in fact a legal term, and ‘mad’ or ‘crazy’ are stigmatising lay terms that do not apply to the vast majority of people with a DSM-5 diagnosis, and should not in fact be used for anybody. Many individuals, including physicians (Spence, 2012), find it difficult to accept that mental illness, not unlike physical illness, is common and most of it is not madness or insanity, and it is appropriate for citizens of advanced societies to identify and deal with milder forms of it.

A DSM-5 diagnosis does not, in the same vein, suggest that the individual needs the expert care of a psychiatrist in preference to a psychologist, a social worker, a counsellor, a priest or a well-informed family member or friend (Kirk and Kutchins, 1992); or that the most appropriate intervention is drug treatment. The charge of ‘medicalising normality’ (Spence, 2012; Watts, 2012b) is therefore unfounded. Whether a physician chooses to prescribe a drug in preference to more time-intensive but arguably better alternatives, or the drug industry exploits an opportunity to sell its wares, depends mostly on society’s expectations of medical science and the economic realities of health care and medical practice.

A challenge for DSM-5, and generally for psychiatry, is to provide ‘categorical’ diagnoses for dimensional constructs such as anxiety, depression, personality disorder, cognitive impairment and substance misuse. The Neurocognitive Disorders Work Group dealt with this by formalising criteria for mild neurocognitive disorder, at the risk of being attacked from both sides – for giving a diagnosis to some who are ageing normally and will never decline into dementia, and for mislabelling others who show the early signs of Alzheimer’s disease or other dementias. It is clear that this debate will continue as our profession ponders what is worth treating and society delineates what is worth helping. DSM-5 must simply be regarded as psychiatry’s next faltering step. It is not the bible for the next millennium, and unlike the scriptures, it is neither above criticism nor mandates faith in its followers.