Assignment

PAGE 2

The year 2013 marked the publication of DSM-5, the fifth edition of the Diagnostic and Statistical Manual of Mental Disorders (DSM), published by the American Psychiatric Association (APA). This was the first major revision in more than 30 years. Prior to 1980, diagnostic classification of mental disorders was an abstruse subject, of interest only to researchers and a few experts. But if mental disorders are medical diagnoses, they require a scientifically based classification. Moreover, since 1980, the DSM system has had a profound influence on all the mental health professions. The public, some of whom have been on the receiving end of a diagnostic process, also finds the subject fascinating, so revisions of psychiatry’s manual are front-page news. This book is a guide to the main features of the latest version of the manual. It will focus on three questions. First, what are the most important changes? Second, what are the implications of these changes for practice? Third, is DSM-5 better, worse, or equal to its predecessors? This book, as a critical guide for the intelligent clinician, will applaud the positive aspects of DSM-5 but underline its limitations. It will be supportive of some changes but be critical of others.

What DSM-5 Can and Cannot Do

The first two manuals published by APA, DSM-I (1952) and DSM-II (1968), did not have a great impact on psychiatry. They were used for statistical purposes, but they were not guides to clinical practice.

Page 4

In contrast, the third edition of the manual, DSM-III, published in 1980, was a major break with the past, as well as a best-selling book. The ideas behind this edition reflected a new paradigm for psychiatry, and the politics that made a radical revision possible are a fascinating story in their own right (Decker, 2013). DSM-III moved classification from clinical impressions to some degree of rigor. It increased reliability by taking an “atheoretical” position—that is, making diagnoses based on what clinicians can see and agree on as opposed to the abstract theories used in DSM-I and DSM-II. DSM-III, and its successors, found a place on the shelf of almost every psychiatrist, psychologist, and mental health professional. There were no major changes in the manual during the next 30 years. DSM-III-R, published in 1987, allowed a greater degree of overlap between diagnoses, and DSM-IV, published in 1994, added some important new diagnoses, including bipolar II disorder and attention-deficit hyperactivity disorder in adults. In 2000, a slightly edited version, DSM-IV-TR, appeared. The absence of major changes for so long could be seen as suggesting a need for a new system that could radically revise the diagnosis of mental disorders. This was the mandate given to the editors of DSM-5 by the APA. The work lasted 10 years, with a result that was initially claimed to be a “paradigm shift.” Is the DSM-5 system an improvement over previous editions? The answer has to be yes and no. One would like to believe so, but there are reasons for doubt. Some problems derive from the concept that psychopathology lies on a continuum with normality, making it difficult to separate mental disorders from normal variations and leading to a danger of overdiagnosis. Other issues derive from a strong attachment to the principle that mental disorders are brain disorders, even though knowledge is insufficient to develop a classification based on neuroscience. Although great progress has been made in research on the brain, the origins of mental illness remain a mystery. When one does not know enough, one should not invest in change for change’s sake. Sometimes it is better to keep a known system, however faulty, than make modifications with unpredictable

Page 5

consequences. Moreover, even the smallest changes to diagnostic criteria can have profound effects on research and practice. Finally, revisions with good intentions can still lack clinical utility. Revising DSM is an enormous job, and each edition has grown larger, more complicated, and thicker. Yet much of what is written in the manual may never be applied in practice.

The Validity of Psychiatric Diagnosis

DSM-III aimed to make diagnosis more reliable, but reliability is not validity. During the next 33 years, constant use of the manuals gave clinicians the impression that their categories were valid. That was not true. The DSM system lacks the data to define mental disorders in the way that physicians conceptualize medical illnesses. Diagnoses in medicine can also be vague, but psychiatry is far behind other specialties in grounding categories in measurements that are independent of clinical observation. Almost all DSM-5 diagnoses are based entirely on signs and symptoms. Although some disorders have support for their validity, and although observation can be made more precise through statistical evaluation and expert consensus, most other areas of medicine use blood tests, imaging, or genetic markers to confirm impressions drawn from signs and symptoms. Psychiatry is nowhere near that level of knowledge. No biological markers or tests exist for any diagnosis in psychiatry. For this reason, any claim that DSM-5 is more scientific than its predecessors is unjustified. In 1980, I was a strong supporter of the paradigm shift introduced by DSM-III. It was progressive to make diagnosis dependent on observation rather than on theory. But this provisional stance became frozen in time, and progress during the succeeding decades has been slow. Radical changes in classification would require much more knowledge about the causes of mental disorders. And that is just what we do not have.

Page 6

Psychiatry and Neuroscience

Psychiatry has bet on neuroscience as the best way to understand mental disorders, to solve problems in diagnosis, and to plan treatment interventions. Only time will tell how this wager will pan out. Some psychiatrists claim that the field is on the verge of a great breakthrough. If one were to believe the hype, a biological explanation—and a biological cure—for mental illness lies just around the corner. (Or as one wag put it, every few years we are told that answers are just a few years away.) Although progress in brain research has been rapid and impressive, its application to psychiatry has thus far been very limited. Brain scans are impressive (even if one keeps in mind that the colors are artificial), but all they tell you is that activity is different at different sites. The precise meaning of these changes is unclear, and none are specific to any diagnosis. We do not know enough about the brain, or about the mind, to develop a truly scientific classification, and it could be 50–100 years before we can even get close. It is understandable that psychiatry, so long the Cinderella of medicine and desperate for respectability, wanted to plant its flag on the terrain of neuroscience. But the promise of the 1990s (“the decade of the brain”) for research on mental disorders has not been fulfilled. Neuroscience has shed much light on how the brain functions, but we do not understand the etiology or the pathogenesis of severe mental disorders. We know that most are heritable, but we have no idea about which (or how many) genes are involved. Although some disorders are associated with abnormalities on brain imaging, the findings are neither specific nor explanatory. Although psychopathology can be associated with changes in neurotransmitters, the theory that chemical imbalances cause mental disorders is too simple or plain wrong. Ultimately, it may be impossible to fully explain mental disorders as brain disorders. The neuroscience model attempts to reduce every twisted thought to a twisted molecule, but it devalues studying the mind on a mental level. Considering that it will take many decades to unravel these mysteries, the current situation is nothing to be ashamed of. The DSM-5

Page 7

task force, as well as the leaders of the National Institute of Mental Health, believe that psychiatry should give up its traditional mission, which was both scientific and humanistic, and redefine itself as the clinical application of neuroscience. To paraphrase a famous line from the Vietnam War, they want to destroy psychiatry in order to save it. It is of course true that mental phenomena reflect the activity of the human brain. But the brain is the most complex structure known in the universe. There are more synapses in the brain than stars in the galaxy. This is a project for a century, not a decade, and its results may never provide a full explanation of mental illness.

Unsolved Problems in Psychiatric Diagnosis

Lack of Knowledge About Mental Disorders: DSM-5 is not “the bible of psychiatry” but, rather, a practical manual for everyday work. Psychiatric diagnosis is primarily a way of communicating about patients. This function is essential but pragmatic—categories of illness can be useful as heuristics without necessarily being “true.” The DSM system is a rough-and-ready classification that brings a degree of order to chaos. But it describes categories of disorder that are poorly understood and that will be replaced with time. Moreover, current diagnoses are syndromes, not true diseases. They are symptomatic variants of broader processes defined by arbitrary cutoff points. Thus, although classifications serve a necessary function, psychiatrists can only guess how “to carve nature at its joints.” That phrase (attributed to Aristotle) describes an impossible task. We do not know if it is possible to find joints to be carved. Even in medicine, diagnoses are not always cleanly defined or related to a specific etiology. In contrast, mental disorders greatly overlap with each other—and with normality. The Need for Biological Markers: In the absence of a more fundamental understanding of disease processes, DSM-5, like its predecessors, had no choice but to continue basing diagnostic criteria on signs and symptoms. But observation needs to be augmented by biological markers, as has been done in other medical specialties.

Page 8

In the absence of independent measures of this kind, we cannot be sure that any category in the manual is valid. We should not therefore think of current psychiatric diagnoses as “real” in the same way as medical diseases. Also, listing them in a manual does not make them real. For example, broad categories such as “major depression” in no way resemble diseases. Even the most “classical” concepts in psychiatry, such as the separation of schizophrenia from bipolar disorder, have not fully stood up to scrutiny. In summary, psychiatrists must make diagnoses, but they do not need to reify them. They are best advised to stay humble and to avoid hubris. Boundaries Between Mental Disorder and Normality: This is one of the most nagging problems in psychiatric diagnosis. Every edition of DSM has expanded this frontier, taking on increasingly more problems of living as diagnosable disorders. Psychiatric classification has become seriously overinclusive, and the manual grows ever larger with each edition. DSM-5 also errs on the side of expanding boundaries—mainly out of fear of “missing something” or not including problems that psychiatrists treat in practice. The result is that people with normal variations in emotion, behavior, and thought can receive a psychiatric diagnosis, leading to stigma and inappropriate and/or unnecessary treatment. Diagnostic Validity and Research: Because we have to live with a diagnostic system that is provisional—and that will almost certainly prove invalid in the long run—much of the research on mental disorders has to be taken with a grain of salt. For example, although a massive amount of data has been collected on the epidemiology of mental illness, almost all its findings are dependent on the current diagnostic system. Similarly, studies of treatment methods in psychiatry that target specific disorders are sorely limited by the problematic validity of categories. Most treatments, from antidepressants to cognitive behavioral therapy, have broad effects that are not specific to any diagnosis. Comorbidity: One of the most troubling problems with the DSM system is that it yields multiple diagnoses in the same patient. That is not the way medicine usually works. It is possible for patients to suffer from more than one disease. But in psychiatry, if you follow

Page 9

the rules, the same symptoms can be used to support two or three diagnoses. Thus, “comorbidity” is little but an artifact of an inexact system in which criteria overlap. The sicker a patient, the more mental disorders will be identified. DSM-5 considered severity ratings and diagnostic spectra to address this problem, but these procedures could not resolve underlying questions about boundaries. Algorithmic Diagnosis:  Another source of uncertainty is that diagnosis in psychiatry does not depend on “pathognomonic” signs and symptoms that define specific diseases. The algorithmic approach of the DSM system has been rightly popular: It uses “polythetic” criteria—making a list and then requiring a given number to be present. These quantitative thresholds are superior to asking clinicians to determine whether the patient’s condition resembles a prototype. But if a typical DSM diagnosis requires, for example, five out of nine criteria, nobody knows whether four or six criteria would have been more or less valid. Few categories have absolute requirements for any criterion, and no system of weighting takes into account the most characteristic features. The DSM system has been jocularly called a “Chinese menu” approach to diagnosis. But most clinicians need to consult the menu, and they would be hard put to remember all criteria for any category. Dimensionalization: The editors of DSM-5 thought that the solution to the comorbidity problem is to view disorders as dimensions—spectra of pathology that can be scored in terms of severity. All previous editions have classified mental disorders as specific categories, much like general medicine. One of the main ideas behind DSM-III was the revival of a model based on the work of the German psychiatrist Emil Kraepelin (1856–1926). Categories are consistent with the view that psychiatry concerns itself with mental illness, not with unhappiness or life itself. They also imply that psychopathology falls into a set of categories or natural kinds, much like tuberculosis or most forms of cancer. DSM-5 sought to overthrow this “neo-Kraepelinian” approach and replace it with a model in which normality and illness lie on a continuum. The rationale is that research suggests the underlying biology of mental disorders is more dimensional than categorical. But measuring the

Page 10

severity of depression is not like taking blood pressure. The definition of dimensions is based on observation rather than biological markers, and it can only be provisional. Dimensional diagnosis also runs the risk of being overinclusive. Even normal people have some symptoms of disorder but do not deserve a formal diagnosis. Because differences in degree can become differences in kind, categories are necessary. Expert Consensus: DSM-5 is not a scientific document but, rather, a product of consensus by committees of experts. Sometimes the outcome depends on who was put on these committees. Where experts disagree, there is a way to “fix” results in advance—by ensuring that membership reflects a preexisting point of view. There are many scientific disputes affecting diagnosis, but most reflect a lack of basic knowledge. As the American physician Alvan Feinstein once remarked, the consensus of experts is the source of most medical errors. In summary, DSM-5 was a noble attempt at a revision in line with current research, and it can be considered as a draft for future editions based on more data. What has to be kept in mind is that the new manual only begins to develop a better framework for research and practice. Psychiatry has to put off scientifically based definitions of mental disorders to a future time when it knows more.

The Constituencies of DSM-5

A diagnostic manual serves many purposes and has many potential constituencies. Let us consider each of them. Clinical Practice: The most important consumers of a diagnostic manual are mental health clinicians: psychiatrists, psychologists, social workers, and family physicians. Practitioners use the manual on a daily basis but are strapped for time. Researchers do not mind if procedures for reaching a diagnosis are complicated, but clinicians do. If the manual is not user-friendly, it will never be used as intended—or not used at all.

Page 11

Clinicians do not have the time or the inclination to open a book and count criteria. (They are even less likely to score symptoms on rating scales.) That is why previous editions of DSM were never applied in a systematic way. Many (if not most) users of DSM-5 may prefer an electronic version that can be easily searched. Even so, given that previous printed editions have not been used as intended, it is unlikely that the text of DSM-5 will be regularly consulted, even when available on mobile devices. Clinicians need to reach conclusions rapidly. They rarely follow algorithmic procedures, and they prefer to make diagnoses intuitively. Most have a prototype in their mind as to what any disorder should look like. The more closely a patient fits this model, the more likely it is that a diagnosis will be made. In this light, DSM-5 will not make as much difference to real-world practice as one might imagine. Because the details of DSM-III and DSM-IV were complicated, clinicians were happy to leave systematic diagnosis to researchers. Previous editions of DSM were poor guides to therapy. To be fair, the system was never intended to guide treatment. (This principle was explicitly stated in DSM-III, but clinicians did not seem to believe it.) As mental health practice becomes increasingly evidence-based, it could eventually develop specific treatments for diagnoses based on research. Doing so is not possible now. Only a few well-established links are known between any diagnostic category and specific therapeutic options. Research: This is the area in which the DSM-5 manual will be followed most systematically. Diagnosis is an essential tool for clinical investigation, and researchers need to establish that their studies examine populations that correspond to accepted rules of classification. However, changing criteria for mental disorders leads to discontinuities, making it more difficult to compare older studies to newer ones. The question is whether changes in DSM-5 can move research forward or only create unnecessary confusion. Another issue is whether the more complex procedures described in DSM-5 should be reserved for researchers if they are the people most likely to apply these procedures.

Page 12

The Pharmaceutical Industry: This constituency has a strong interest in DSM-5. Companies are interested in maximizing their profits; one way to do so is to get physicians to prescribe more drugs for more people. Any change that encourages wider diagnosis of mental disorders as a whole is in their interest. Specifically, the way DSM defines schizophrenia, bipolar disorder, and major depression could have an impact on industry profits. Some of the most problematic trends in modern psychiatry have resulted from attempts to make patients fit into categories that justify the prescription of drugs. But overinclusiveness tends to make the pharmaceutical industry happy. The Legal System: Lawyers and judges are also interested in DSM-5. Psychiatric diagnoses have found their way into the court system, affecting everything from criminal responsibility to custody rights and insurance payments. Although the science behind diagnosis does not justify any of these practices, they are widespread. The General Public: Finally, DSM-5 has and will continue to influence the way everyone views mental illness. Patients (and nonpatients) have access to published criteria and sometimes diagnose themselves (or their relatives). This amplifies the danger that too many people will receive psychiatric diagnoses. The way consumers and families view mental problems will also have an influence on practice, most likely in the direction of more aggressive treatment. The one certain thing about DSM-5 was that it would be another best-seller. What is less certain is whether it would lead to improved mental health care.

Ten Highlights of DSM-5

Although many changes in the new manual are relatively minor, I highlight areas in which major revisions have been made, particularly those that have aroused controversy. All will be discussed in detail in this book.

Page 13

1. DSM-5 is reorganized into a new set of chapters that either reflect common clinical features or fall within a spectrum. Because most mental disorders begin early in development, there is no separate section for disorders first apparent in childhood.

2. The multiaxial system introduced in DSM-III has been eliminated. There is now no such thing as five axes, and personality disorders are considered in the same way as other categories.

3. Functioning, previously coded on Axis V, is now rated by scores on an instrument developed by the World Health Organization.

4. The criteria for attention-deficit hyperactivity disorder in adults have been expanded somewhat (by changing the requirements for a childhood onset).

5. Personality disorders are defined in the same way as in DSM-IV, but an alternative system based on trait dimensions can be found in a separate section of the manual.

6. Substance use disorders now describe cases using the term addiction, no longer distinguishing between dependence and abuse.

7. Highly moody and aggressive children can be given a new diagnosis—disruptive mood regulation disorder with dysphoria.

8. Autism spectrum disorders consider classical autism and Asperger’s syndrome as being on the same spectrum. 9. Dementias are classified as neurocognitive disorders, rated by severity on a spectrum. 10. Somatic symptom disorders replace somatoform disorders and are classified differently.

Some of the most controversial changes that were proposed earlier in the DSM-5 project were either dropped or greatly diluted. Thus, attenuated psychosis syndrome (which might have led clinicians to treat people who have mild symptoms but never develop schizophrenia) was moved to a section of conditions requiring further study. A proposal to reduce the length required for a hypomanic episode from 4 days to 1 or 2 days was not adopted. The potential

Page 14

impact of dropping the grief exclusion in diagnosing depression was diluted—by warning clinicians to avoid making a diagnosis when the course of mourning appears relatively normal. The range of autism spectrum disorder has been kept limited. A radically different system for classifying personality disorders was not accepted. Although it is assumed that most mental disorders lie on a spectrum with normality and have subclinical forms, dimensionalization was not consistently applied. Formal severity ratings for diagnoses, too complex for clinical use, were considered to require further study.

The Structure of This Book

The first part of this book is devoted to broader issues. Chapter 2 reviews the history of psychiatric diagnosis, Chapter 3 how diagnostic manuals are prepared, Chapter 4 how diagnoses are validated, Chapter 5 how mental disorder can be separated from normality, and Chapter 6 how dimensional assessment could be used. Chapter 7 examines clinical utility. The second part examines the major diagnostic groups in DSM-5. Chapters 8–14 present separate discussions of the most frequently used diagnoses: psychoses, bipolar disorder, depressive disorders, anxiety disorders, obsessive–compulsive disorder, neurodevelopmental disorders, impulse control and conduct disorders, eating disorders, sexual disorders, and personality disorders. Chapter 15 takes a briefer look at neurocognitive disorders, somatic symptom disorders, dissociative disorders, sleep–wake disorders, elimination disorders, and adjustment disorders. In Part III, Chapter 16 examines responses to DSM-5 from the scientific and clinical communities, and it comments on an alternative system proposed by the National Institute of Mental Health. Chapter 17 discusses how clinicians can use the manual in practice. Chapter 18 examines the future of psychiatric diagnosis, suggesting guidelines for the practical use of DSM-5 in clinical work and underlining issues that need to be resolved for the next edition—DSM-6.