Intended for healthcare professionals

Feature Psychiatry

DSM-5: a fatal diagnosis?

BMJ 2013; 346 doi: (Published 22 May 2013) Cite this as: BMJ 2013;346:f3256
  1. Jonathan Gornall, freelance journalist
  1. 1Suffolk, UK
  1. jgornall{at}

Committee walk-outs and unprecedented levels of scrutiny and criticism have marked the run-up to this week’s launch of DSM-5. Jonathan Gornall interviews key players on both sides of the debate and asks whether this mighty US psychiatric handbook has finally over-reached itself

It isn’t every day that a new medical handbook attracts as much media attention as the latest blockbuster novel by a world famous author. But the controversy surrounding the publication on 22 May of the long awaited fifth edition of the Diagnostic and Statistical Manual of Mental Disorders has generated international coverage on a scale to rival that afforded last week’s release of the latest sin and symbols thriller from the bestselling author Dan Brown.

And, though unlikely literary bedfellows, DSM-5 and Inferno—a romp through clues plucked from Dante Alighieri’s 14th century poem Divine Comedy—have something else in common: the authors of both have been going through hell at the hands of their critics.

The American Psychiatric Association produced the first DSM in 1952 and in the intervening 60 years there have been six subsequent iterations. The last was a text revision of 1994’s DSM-IV, published as DSM-IV-TR in 2000, which means that DSM-5 (the Roman numerals have been abandoned for this latest edition in deference to the demands of the digital age) has been 13 hard fought years in the making, during which its makers have come under unprecedented scrutiny and attack.

For the first few decades of its existence, the DSM had a fairly low public profile, although one of the seeds of discontent that has now reached full bloom—the suggestion that the World Health Organization’s International Classification of Diseases is the only manual the world of psychiatry and psychotherapy now needs—was in fact sown at the birth of DSM in 1952.

In 1948 WHO produced the sixth edition of the International Classification of Diseases, which for the first time included a category for mental disorders. Although this drew heavily on the experience of the US Veterans Administration and the classifications of mental disorders it had developed while working with second world war military personnel, the American Psychiatric Association committee on nomenclature and statistics decided to produce its own manual.1 2

DSM first came to the attention of the popular press in 1973, when members of the American Psychiatric Association embarking on the revision process that would lead to DSM-III began deliberating whether homosexuality should be “stricken from the association’s official catalog of mental disorders,” in which it was still listed as a “sexual deviation,” alongside sadism and masochism.3

Among those who agreed it should go was Robert Spitzer, the man who, as chair of the task force that produced DSM-III in 1980, is widely credited with having been the architect of the modern DSM and its classification of disorders. Much later, however, Spitzer seemed to have second thoughts about the approach to psychiatry that DSM-III had crystallised—and the reservations he expressed six years ago are being echoed loudly today by critics of DSM-5.

Medicalising ordinary experiences

“We made estimates of the prevalence of medical disorders totally descriptively,” he told a BBC documentary in 2007, “without considering that many of these conditions might be normal reactions which are not really disorders, because we were not looking at the context in which those conditions developed.”4 They had, he agreed, “to some extent” medicalised much of “ordinary human sadness, fear, ordinary experiences.”

This is one of the key criticisms of DSM-55—and few have presented it as eloquently as Frank Farley, the former president of the American Psychological Association, who takes personal exception to the widely condemned decision to remove from DSM-5 an exclusion clause that prevents bereavement being diagnosed as a major depressive disorder for up to two months after the death of a loved one.

“Grieving is normal,” he says, “and if you grieve for more than two months should you be labelled with some kind of mental illness—and that’s the terminology that the public will pick up on? The answer is clearly no. My first wife died three decades ago; I don’t grieve in the usual way that I did for a long time after she passed away, but I still think of her and I still have emotional memories. Nothing wrong with that—it’s human nature.”

Farley, like many critics, is concerned about DSM-5’s contribution to what he calls “the relentless production of disorders and pathologising of normal extremes.” It isn’t a new charge. In 1999 DSM-IV came under fire for facilitating “the creation of mental illness.”6

But it remains true, says Farley, that psychiatry and psychology in general need to be “better connected with the real world. I worry that we have too much of the monastic science in us; too many laboratory studies. People don’t live in laboratories, they don’t grieve in laboratories, they don’t feel pain in laboratories.”

The American Psychiatric Association, which has invited criticism and contributions throughout the long revision process, and since 2010 has received an “unprecedented” 15 000 comments,7 has always been vigorous in defending its position. But its defence of the removal of the “bereavement exclusion,” written by a member of the DSM-5 mood disorder work group, seems to confirm wider fears that the association’s committee led system for selecting disorders for inclusion is a recipe for disorder escalation.

The grief exclusion criteria, wrote Kenneth Kendler of the Virginia Institute for Psychiatric and Behavioural Genetics, in a document posted on the APA website in 2011, had been added to DSM-III “largely on the basis of the work of one of the DSM-III task force members who was then studying grief and was carried forward with little modification into DSM-IV.”8

Pet interests

Peter Tyrer, professor of community psychiatry at Imperial College London and editor of the British Journal of Psychiatry, is chair of the personality disorder group for the revision of ICD-11, due out in 2015, and says his field offers “a very good example of what I think is wrong with DSM. A lot of clever people sit around a table and say ‘I’ve done work on this and I want to have narcissistic personality disorder included,’ ‘I want to have dissocial personality disorder,’ ‘I want to have avoidant personality disorder.’”

In fact, he says, “these are categories of personality disorder that actually have no scientific basis behind them and yet here are worthy people sitting in committees all agreeing that this is important and we’ve got to include them.” The result, he says, is that “you are in danger of medicalising people unnecessarily.”

Although he concedes that for a long time the under-resourced ICD “followed in the wake” of DSM, Tyrer is not alone in believing that, while it has “been a stimulus to interest in classification,” DSM has over-reached itself and, in the face of a strengthened, streamlined ICD, is now “on the way out.”

“I think there are disorders out there that are pets of particular doctors or groups of doctors,” says Gary Greenberg, a psychotherapist in Connecticut and author of the newly published The Book of Woe—The DSM and the Unmaking of Psychiatry.9

“They have an agenda to get [a disorder] into the DSM and they succeed partly because they are the people on the committee.”

The best example of this in DSM-5, he says, is disruptive mood disregulation disorder, one of several new depressive disorders. It has been introduced, says the APA, “to address concerns about potential overdiagnosis and overtreatment of bipolar disorder in children.”10

Others newcomers include excoriation disorder, for which there is “strong evidence for its diagnostic validity and clinical utility,” and hoarding disorder, introduced in the light of “evidence for the diagnostic validity and clinical utility of a separate diagnosis . . . which reflects persistent difficulty discarding or parting with possessions due to a perceived need to save the items and distress associated with discarding them.”10

Curiously, one of the fiercest critics of the DSM-5 process has been Allen Frances, the chair of the DSM-IV task force. Since 2010, his less than friendly fire from the high ground of a column in Psychology Today (DSM-5 in Distress) has sniped at everything from the “price gouging” of DSM’s $199 (£130; €155) cover price—which he attributes partly to limited demand created by “all the mistrust that has been generated by DSM-5’s flawed process and reckless product”—to the suggestion that the new manual “will give drug companies running room to continue their disease mongering of female sexual disorders.”11 12

He joins a diverse band of opponents, including the Society for Humanistic Psychology, which in January 2012 began a public exchange by sending an open letter to the DSM-5 task force asking it to submit its proposals to an independent group of scientists. The letter was backed by an online petition that attracted the signatures of more than 14 000 individuals and 53 organisations, including the British Psychological Society.13 14

“It spoke of a sort of unease, to say the least,” says Farley, “and basically we got a thanks, but no thanks. They said they had the science well covered in terms of expertise.”

Internal dissent

But not everyone on the task force agreed. In April last year, two members of the DSM-5 personality orders work group walked off the job, claiming the members had thrown away “an important opportunity to advance the study of personality disorder by developing an evidence-based classification with greater clinical utility than DSM-IV.” Instead, with a “truly stunning disregard for evidence,” they had advanced a “seriously flawed” proposal which was “unnecessarily complex, incoherent, and inconsistent.”

Important aspects of the proposal, wrote Roel Verheul and John Livesley, in an email made public by Frances, lacked “any reasonable evidential support of reliability and validity.” Not surprisingly, they added, “the proposal has received widespread criticism to which the work group seems impervious.”

Verheul, professor of personality disorders at the Viersprong Institute in Amsterdam, and Livesley, former head of psychiatry at the University of British Columbia, were the only international members of the group, which Frances dismissed as a “small group of cloistered DSM 5 ‘experts’ stubbornly ignoring the sharp criticism from within their own group and the near universal rejection of their proposals by everyone else in the field.”15

Competing interests

One of the most serious accusations levelled against DSM is that leading psychiatrists who have worked on it have been influenced by funding relationships with the drug industry,16 but the APA has taken it in its stride. While such speculation was “bound to occur,” said David Kupfer, professor of psychiatry at the University of Pittsburgh and chair of the DSM-5 task force, in a press release in January, “we think it is important to stay focused on the fact that APA has gone to great lengths to ensure that DSM-5 and APA’s clinical practice guidelines are free from bias.”

Steps taken, Kupfer told Medscape in January, included limiting task force members to annual income from industry sources of no more than $10 000 each and to holding shares in pharmaceutical companies worth less than $50 000—limits “more stringent than requirements for staff at the National Institutes of Health, members of advisory committees for the Food and Drug Administration, and most academic departments.”17

But corruption-conspiracy theories miss the point, says Greenberg, who sees instead “a confederacy of good intentions” at work. “The DSM is created by committees, which is one of the reasons it’s such an unwieldy document, and the committees are made up of experts in the field, who tend to be people who are valued and pursued by the drug companies to do their research.

“I don’t feel there’s a huge conspiracy—it’s not like the drug companies say to a psychiatrist, ‘Look, we could really use this disorder in the DSM, here’s fifty grand.’ They don’t have to, because you’ve got an entire profession that intellectually is already predisposed to seeing mental problems as problems that should be treated with drugs.”

Nevertheless, it hasn’t helped the cause of the DSM, or US psychiatry in general, that since 2008 Senator Charles Grassley, the chair of the US Senate Committee on Finance, has doggedly unearthed a series of cases in which leading academic psychiatrists have failed to reveal large payments from drug companies.

One of the most high profile cases was that of Charles Nemeroff, who in 2008 resigned as chair of the psychiatry department at Emory University, Atlanta, after it was revealed that he had failed to report more than $1.2m of payments from GlaxoSmithKline, despite having signed an undertaking to limit payments to $10 000 a year and while acting as lead investigator on a National Institutes of Health study of the company’s drugs for depression.18 19 20 21 In 2009 Nemeroff was appointed chair of psychiatry at the University of Miami.

Challenge from biology

A bigger blow to the future of DSM, however, was landed last month by the US National Institute of Mental Health. It has launched a stinging attack on DSM, criticising its “lack of validity” and announcing it is “re-orienting its research [funding] away from DSM categories” and towards the establishment of a new classification system based on the biology as well as the symptoms of mental disorders.

The Research Domain Criteria (RDoC) project, said Thomas Insel, director of the NIMH, last month, was “nothing less than a plan to transform clinical practice by bringing a new generation of research to inform how we diagnose and treat mental disorders.”

The long term future of mental health, he said, lay in the detection of biomarkers: “Unlike our definitions of ischaemic heart disease, lymphoma, or AIDS, the DSM diagnoses are based on a consensus about clusters of clinical symptoms, not any objective laboratory measure,” and “patients with mental disorders deserve better.”22

Starting virtually from scratch, RDoC is for now a distant challenge to the supremacy of DSM. “It is the way forward, but it is nowhere near where one can use it to create a diagnostic system,” says Shitij Kapur, dean of the Institute of Psychiatry at King’s College London, who last year published a paper with Insel examining why it has taken so long for biological psychiatry to develop clinical tests.23

For now, he says, “we have a job to do and [ICD and DSM] are the tools for today.” The differences between them “are in 95% of the cases either trivial or only for the connoisseurs,” and the real question now is “does the world need a DSM and an ICD?”

In the meantime, he thinks the fuss surrounding the new DSM “will all go away, because DSM-5 is not a revolution; it is a mere sensible incremental improvement. There are people who are disappointed, because 10 years ago there was hope that we would have clear biological tests and that this DSM would radically revise all the deficiencies we have in psychiatric disorders, but that’s not the failing of DSM, because DSM cannot give you fundamentally new categories and fundamentally better tests if the field has not yet done the research and found the evidence to support it.”

He says it is not too simplistic to characterise the controversy over DSM as psychiatry versus psychotherapy, or medication versus talk therapy, and that those camps are “duking out” those tensions against the backdrop of the DSM launch. “This is not to say that those tensions are not real, or unimportant; it’s just that they didn’t happen yesterday and DSM-5 doesn’t particularly make them dramatically better or worse.”

DSM’s influence in UK

On 4 June the Institute of Psychiatry at King’s will host a two day international conference on DSM-5 and “the future of psychiatric diagnosis.” Although the UK uses the ICD, it would, says Kapur, be a mistake to assume that DSM has no effect on British psychiatry.

“It would be unwise for the rest of the world to ignore things that are happening in the US because it is undoubtedly the academic and professional powerhouse in psychiatry, generating so many concepts, ideas, new findings, and more papers than any other country almost by a factor of 10,” he says.

When it comes to scientific discourse, he adds, the DSM “trumps” the ICD: a British psychiatrist treating a patient who has schizophrenia is likely to be reading papers on the subject in which patients have been diagnosed by doctors using DSM.

As editor of the British Journal of Psychiatry, Tyrer says he repeatedly reminds American authors “that the UK is one of the last DSM-free zones in the world, because we haven’t been influenced by the heavy selling of the DSM.”

He says he has two jokes about what “DSM” stands for—“though they’re not only jokes, they are also serious [points].” One is “Diagnosis as a Source for Money, because it brings lots of income to the APA, because of all they copies they sell, whereas [the ICD] from WHO is open access.”

The second is “Diagnosis for Simple Minds, because the big attraction of DSM is it allows a tickbox mentally, where you can say these are ‘operational criteria’ and if you’ve got three or four out of seven then it fact you’ve got the disease—which is a funny way of diagnosing people in medicine.”

Nevertheless, “people use DSM as their core currency in diagnosis—not just in research, where of course you can say it may be appropriate because ICD is less research oriented, although we’re trying to change that now—but in ordinary clinical practice. If someone says, ‘This person has a DSM diagnosis of major depression,’ that carries much more weight than saying, ‘This person has a ICD diagnosis of mild depressive illness.’”

This was shown by a high profile British trial in 2005, in which a man accused of killing his parents and embarking on a spending spree with their credit cards avoided a murder conviction by admitting manslaughter with diminished responsibility on the ground that he had narcissistic personality disorder.24 The disorder appears only in DSM: “Narcissistic has never been in the ICD classification and I am confident it never will be,” says Tyrer. “It is a diagnosis of vanity for both patients and their doctors.”

DSM’s defence

One man who will be attending the King’s DSM-5 conference next month is David Kupfer, chair of the DSM-5 task force, who clearly believes the UK merits a visit during what for him must surely be a frenetic post-publication period—and who gives not the slightest impression of being a man in a diabolically hot seat.

“DSM is widely used around the world for teaching psychiatric diagnosis, training clinicians, and specifying criteria for people entered into research studies,” he says. One of the key messages he is keen to get across is that DSM-5 remains “the best science available [and] the best manual for helping clinicians care for patients”—and, unsurprisingly, he does not accept the notion that DSM should step aside in favour of ICD.

“DSM and ICD can be thought of as companion publications,” he says. “They are cross-linked so that a clinician using DSM can use the ICD diagnostic coding system required by most of the world’s health systems. It’s important to note that ICD does not include descriptive diagnostic criteria, only a listing of disorders. DSM-5 is the best possible clinical guidebook for the diagnosis of mental disorders.”

Likewise, despite the vehemence of the National Institute of Mental Health’s attack on DSM, he insists that “DSM-5 and the National Institute’s RDoC represent complementary, not competing, frameworks. Once the RDoC effort starts to take shape, any information or findings stemming from its research agenda will be integrated into future DSM editions to further strengthen patient care—just as all advances up until now have done.”

He also sidesteps a question about the difficulty of having his predecessor as chair of DSM-IV as one of the fiercest critics of DSM-5.

“While criticism is bound to occur,” he says, “we think it is important to stay focused on the fact that DSM-5 was developed over the course of more than a decade with the input of more than 1500 of the best experts in the field and through extensive reviews of the scientific literature, original research, and nationwide field trials.”

Those field trials, of course, have caused their own share of controversy. For Farley and others, all they showed was that “the science [behind DSM-5] isn’t good enough.”

Between November 2010 and November 2012 the DSM-5 criteria for 23 disorders were tested by hundreds of psychiatrists working with 3500 patients across the country. Six diagnoses, according to the American Psychiatric Association, “had relatively low reliability scores, which the report characterised as questionable but acceptable”; among them were two of the DSM’s most common diagnoses: major depressive disorder and generalised anxiety disorder.

“At least some of the reason why the scores were not at the level of previous field trials,” commented Kupfer at the time, was that “depression and anxiety represent symptom clusters that can fluctuate during several weeks, and the rigorous trial design of the field trials—employing a test-retest design requiring two separate clinicians to evaluate the same patient on different days—may have contributed to a lower reliability score than was found in previous field trials.”25

Legislative threat

All the debates over what should and should not go in DSM-5 are now academic. But the future of the manual may face a longer term threat in the form of the mandatory shift in October 2014 from ICD-9 to ICD-10 code sets for all healthcare providers covered by the US Health Insurance Portability and Accountability Act (HIPAA), which embraces the Medicare national insurance programme and Medicaid, the means tested health programme.

The mandate, says the American Psychological Association, “represents a fundamental shift for many psychologists and other mental health professionals, who are far more attuned to the American Psychiatric Association’s Diagnostic and Statistical Manual of Mental Disorders (DSM).”

As with ICD-9, the US has modified ICD-10 for use in its own medical system, but the difference with ICD-10-CM is that it will pull many psychiatrists and psychologists away from the orbit of DSM for the first time.

“Currently, many psychologists utilise the DSM-IV-TR when diagnosing patients, and the corresponding DSM-IV-TR codes to submit a health insurance claim,” noted the psychiatric association in a practice update to members in February. That was fine under ICD-9 because its diagnostic codes aligned with those of DSM, but the ICD-10 codes will be different and “anyone who bills DSM codes instead of ICD-10-CM codes presumably will experience claim denials.”26

Kapur expects “a well-rounded discussion” at the King’s conference next month.

“We are not interested in just having a big controversy—we were sure there would be enough of that going around. We want to make sure we hear the different perspectives from which these issues are important, because, after all, DSM is not just for psychiatrists or psychiatric researchers, it has implications for patients, GPs, carers, funders, and so on.”

It is, he says, easy to criticise, but “I think many people are really ahistorical, because they can find all the faults with DSM but are totally naive about the total mess that the diagnosis of psychiatric disorders used to be about 40 years ago . . . What you called schizophrenia in New York wasn’t schizophrenia in London, and we have made a huge transition since then.”


Cite this as: BMJ 2013;346:f3256


  • Competing interests: I have read and understood the BMJ Group policy on declaration of interests and have no relevant interests to declare.

  • Provenance and peer review: Commissioned; not externally peer reviewed.


View Abstract