The Book of Woe: The DSM and the Unmaking of Psychiatry

by Gary Greenberg

  Spitzer’s solution was characteristically wily. He carved out an exception to the depression diagnosis and deftly inserted it at the back of the DSM-III. In a section called “Conditions Not Attributable to a Mental Disorder That Are a Focus of Treatment or Attention,” a clinician could learn that the recently bereaved may well have all the symptoms of depression, but because “a full depressive syndrome21 frequently is a normal reaction to the death of a loved one,” the patient’s condition was better understood as Uncomplicated Bereavement. Clinicians confronted with the absurdities of the descriptive approach were thus given cover to ignore it even as their prerogative to treat the recently bereaved was preserved (although insurers rarely if ever pay for treatment of those non-mental-disorder conditions).

  The bereavement exclusion22, as it came to be known, became part of the diagnostic criteria in the DSM-III-R, where it acquired a time limit: after two months of meeting five of the nine criteria, the bereaved became mentally ill. But certain thorny matters—notably, whether or not a clinician should distinguish among the death of a parent or child, the death of a celebrity, and the death of the family dog—remained up to the individual clinician to decide. Even so, it was a win-win, which may be why no one seemed to notice that the loophole amounted to saying that people who had all the symptoms of a disorder had the disorder—unless they didn’t. The criteria didn’t really add up to a mental illness until a doctor determined that a disorder was present and by this judgment transformed suffering into the symptoms of a disease.

  This is, in part, as it should be. Isn’t this what we go to doctors for—to learn, from someone who knows, the true nature of our suffering, to find out whether that nagging pain is the leading edge of something horrible or just a random twinge, to find out if our persistent malaise is grief or depression or maybe even a malfunctioning thyroid? Even before we ask for remedy, this is what we ask of our doctors: to lay bare the beginnings of our suffering, to elicit our present crisis, to tell us what is going to happen in the end. Without that story, we might not take their pills, and—since so much of our response to medication is the result of placebo effects, and placebo effects in turn depend on the patient’s belief in his or her doctor—the pills might not work.

  On the other hand, most of us won’t accept just any tale about our woes. We want our doctors’ stories about us to be based in fact, not opinion. We want them to make sense, which, if they start telling us that grieving the loss of a parent is an illness, they don’t. That’s why the bereavement exclusion was necessary: without it, the DSM loses its credibility, and the doctors who use it cannot perform their healing magic.

  • • •

  Spitzer did something else to juice the credibility of the DSM-III, something that no one else had done, at least not in a diagnostic manual: he tried to define disease23. This is harder, and a lot more audacious, than it might seem. Like life and obscenity, disease is one of those phenomena that you might recognize when you see it—but go ahead and try to define it.

  You have to admire Spitzer for making the attempt, for not simply cribbing Webster’s and then moving on to his list of diseases and the symptoms by which they would be known. But you also have to understand that he really had no choice. So long as psychiatry had no scientific knowledge about which ingredient was missing from the chemical soup roiling inside your head, so long, that is, as diagnosis was still a matter of a doctor deciding that you had a disease and then telling you which one it was, psychiatrists needed to be able to say with certainty how they made that decision, and why it wasn’t simply a matter of personal prerogative. They needed a definition that would serve as a gatekeeper to the kingdom of illness, that would reassure the public that the profession didn’t intend to claim sovereignty over all our troubles, that would keep homosexuality out and depression in—that would, as Spitzer put it in the introduction to DSM-III, “present concepts that have influenced the decision24 to include certain conditions and to exclude others.” Without that barrier, DSM would not be a medical text but a collection of old wives’ tales.

  Spitzer understood from the beginning that the commonsense definition of disease—“a progressive physical disorder with known pathophysiology”—simply couldn’t be stretched to cover mental illness. He finessed this problem by proposing that disease was only one of a number of medical disorders—conditions that had “negative consequences . . . an inferred or identified organismic dysfunction, and an implicit call to action.” Mental disorder, he argued25, was “a medical disorder whose manifestations are primarily signs or symptoms of a psychological (behavioral) nature.” This was a clever move on Spitzer’s part, acknowledging that mental illnesses were not diseases in the usual sense, even as he preserved their place in “real medicine.”

  But it wasn’t clever enough to sneak past the members of the American Psychological Association, who immediately recognized the proposal as a way to maintain physician dominion over mental suffering, and they sent a letter protesting it.

  “These guys have some chutzpah26,” Spitzer groused to the APA’s president as he prepared a letter in response. But he didn’t dispute the psychologists’ conclusion. Indeed, he may have gotten a fight he’d been spoiling for all along. He suggested that the exchange of letters be “made public to our membership, as it would be another way of demonstrating our conviction that psychiatry is a specialty within medicine. It would also make it clear to our profession that DSM-III helps psychiatry move closer to the rest of medicine.” If they saw their generals aggressively moving to consolidate their power, Spitzer thought, the morale of the rank and file might improve.

  The definition Spitzer finally settled on wasn’t quite so chauvinistic as the original. But it had plenty of its own chutzpah.

  In DSM-III each of the mental disorders27 is conceptualized as a clinically significant behavioral or psychological syndrome or pattern that occurs in an individual and that is typically associated with either a painful symptom (distress) or impairment in one or more important areas of functioning (disability). In addition, there is an inference that there is a behavioral, psychological or biological dysfunction, and that the disturbance is not only in the relationship between the individual and society. (When the disturbance is limited to a conflict between the individual and society, this may represent social deviance, which may or may not be commendable, but is not by itself a mental disorder.)

  Mental disorder occurs, Spitzer seems to be saying, when something has gone wrong in the mental apparatus, and the result is distress or disability. But what is the tip-off that something has gone wrong? The presence of distress or disability, of course. And how can a doctor determine whether it is clinically significant? The definition doesn’t specify, but one obvious method is to say that the mere fact that a patient’s suffering is significant enough to make him or her show up at the clinic suffices. And once the patient has arrived, who gets to decide whether the disturbance is limited to a conflict between individuals and society, whether, say, an impoverished person’s distress is caused by poverty or internal dysfunction? Doctors, of course, who might be no more reliable at judging whether a person needs psychiatric treatment than barbers are at judging whether a person needs a haircut.

  This definition, an obvious response to the homosexuality debacle, was an attempt by the general to fight the last war, and it doesn’t really make diagnosis any less circular. Indeed, it just places the already circular definitions of the individual mental disorders inside the larger circle of clinical significance. And in case you weren’t dizzy enough, the DSM-III-R added one more loop: “The syndrome or pattern28 must not be merely an expectable and culturally sanctioned response to a particular event.” For its part, DSM-IV devoted seven of its 886 pages29 to a list of “culture-bound syndromes,” including ataque de nervios, an “idiom of distress” common in Latin American groups that an unsuspecting doctor might mistake for an anxiety disorder; pibloktoq, an “abrupt dissociative ep
isode” found among Eskimos, in which a person might rend his or her garments, break furniture, or eat feces, but does not suffer from mania or Dissociative Identity Disorder; and koro, an “intense anxiety that the penis will recede into the body and possibly cause death” that sometimes afflicts Malaysians but which should not be mistaken for a psychosis or Depersonalization Disorder.

  The impetus here is obvious. The gay activism that led to the deletion of homosexuality was on the leading edge of the identity politics that took hold in the 1980s and 1990s. The DSM had to keep up with the times, to reassure the public that psychiatrists were not out to pathologize mere difference or to declare certain identities inherently sick. But how is a physician to know what responses we should expect of ourselves when confronted by unprecedented events like 9/11 or the financial meltdown of 2008? Why is a response that is neither expectable nor culturally sanctioned, but clearly justified, such as, say, occupying Wall Street, any more disordered than blithely taking home multimillion-dollar bonuses for running a company into the ground? And why do doctors get to decide which, if any, of those behaviors is symptomatic of Antisocial Personality Disorder?

  Like the bereavement exclusion, these definitions don’t really serve to limit psychiatry’s prerogative to decide what is sick and what is healthy. Instead, they daub whitewash over the fractures in its conceptual infrastructure. And the result is an edifice that holds up only if you don’t place any weight on it.

  In this respect, all these loopholes are not unlike epicycles, the little curlicues Ptolemaic astronomers built into the orbits of planets to account for why the heavenly bodies were not where they should have been if they moved, as Ptolemy said they must, in perfect circles. Epicycles, not unlike the codicils and caveats in the DSM, are a way to stave off the challenge of the enemy, intended more to preserve the authority of a profession and the dominion of its paradigms than to get to the truth. Unlike the DSM, epicycles have already gone down in history as the epitome of bad science.

  Chapter 8

  If psychiatry’s attempt to close the gap between opportunity and knowledge with a definition of mental disorder will always yield bullshit, psychiatry’s aspirations to scientific respectability are still not doomed—at least not according to Allen Frances. But you have to be willing to accept one premise: that, as he puts it, “psychiatric classification is necessarily1 a sloppy business.”

  But even if the definition of mental disorder is bullshit, Frances thinks, the mental disorders themselves are not. The categories may be arbitrary, their existence impossible to prove, and the lines between them as artificial as the lines between countries, but the fact is that an identifiable group of people do, for instance, have “recurrent and persistent thoughts2” that are “intrusive, inappropriate and cause marked distress,” that are “not simply excessive worries about real-life problems,” that can only be suppressed by some “other thought or action,” and that are recognized as “a product of [the patient’s] own mind.” To say, as the DSM-IV does, that those people have Obsessive-Compulsive Disorder is not, in Frances’s view, to make any grand claims about how (or even whether) mental illness exists in nature. It is only to glean from research what unites this population of sufferers and then to capture it in language that helps clinicians communicate with patients and colleagues and that provides researchers with categories for their work in developing treatments. A DSM diagnosis may be a construct, in other words, but it is not only a construct.

  Neither does using a label require some guiding definition about whose troubles are illness and whose mere suffering. Implementing labels requires only a faithful observation of people who come to doctors’ attention, a careful sorting of patients, a scrupulous attention to detail in fashioning the criteria, and then a highly skilled, careful clinician, one to whom the sloppiness of classification is a reason to exercise caution. And if all those conditions are in place, the criteria will indeed detect populations who can then be served by doctors alerted to the contours, if not the exact nature, of their patients’ troubles. The diagnostician, to use one of Frances’s favorite metaphors, is not so much a pathologist looking for a virus in a blood sample as he is a baseball umpire trained to call balls and strikes—even if an agreement to abide by an ultimately arbitrary tradition is the reason that pitches have those names and the strike zone has the exact boundaries that it does.

  That’s why Frances was so galled by the ambition that he saw in his successors, and why they seemed so reckless to him: they had failed to account for the fragility of a system that hinges on rules inscribed in language rather than on lab tests encoded in numbers. They were heedless of the possibility that once doctors started speaking the revised language, all kinds of hell could break loose. And even worse, they seemed to have lost track of the people who would be consigned to that hell: the patients.

  “A diagnosis is a call to action3 with huge and unpredictable results,” he said. “No decision can be right on narrow scientific grounds if it winds up hurting people.”

  The Bipolar II epidemic was a case in point. There was no question in Frances’s mind that making a new diagnosis was the correct decision on narrow scientific grounds. Research clearly showed that people who became manic after starting antidepressants tended to have a history of hypomanic episodes and that those spells tended to last for less than the week required for a diagnosis of mania. But he failed to consider how many people would get hurt if the diagnostic threshold was lowered, how easy it would be for a harried doctor to render the diagnosis and write a prescription with the pen supplied by the drug company rep who had just taken her to lunch. Real-life psychiatric diagnosis could not take place in a bell jar filled with experts and their pet theories. There would always be unintended consequences. That’s what Frances said he was trying to say to the DSM-5 leaders. “I just wanted them to learn4 from my mistakes.”

  • • •

  Even if definitions of mental disorder weren’t bullshit, they wouldn’t solve the validity problem, at least not by themselves. To declare a boundary between illness and health is not to guarantee that any particular category of illness is real. In fact, definitions could worsen the problem. “One of the reasons5 that diagnostic classification has fallen into disrepute,” Eli Robins and Samuel Guze wrote in 1970, “is that diagnostic schemes have been largely based upon a priori principles rather than upon systematic studies.”

  Robins and Guze were leaders of the team at Washington University that developed the descriptive approach that Spitzer adopted for DSM-III. They recognized that the history of science, and especially of medicine, was littered with examples of prejudice-blinded researchers following their often unacknowledged traditions and principles down dead-end alleys. These beliefs aren’t always as dunderheaded as the ones that shaped Samuel Cartwright’s understanding of a slave’s thirst for freedom or Freud’s notions about same-sex attraction. In the nineteenth century, for instance, doctors believed that illnesses should be classified by their signs and symptoms—a conviction that had prevailed since Hippocrates had given birth to Western medicine, and which was not unreasonable, given that doctors had little else to go on. So there was really no reason to doubt that patients with genital sores were suffering from a disease different from what patients with a skin rash had, and patients with general paresis, a form of dementia, had yet another illness. There wasn’t even a reason to think that this scheme was based on any a priori principle, that it was anything other than a faithful account of how nature itself sorted diseases.

  That all changed when some doctors, notably Louis Pasteur and Robert Koch, began to insist that there was more to disease than met the unaided eye. Beneath the appearances, the pustules and the fevers and the complaints, was a microbial world populated by the real sources of illness. And if the detectable presence of viruses and bacteria was not convincing enough, the successes of pasteurization and anthrax inoculations soon had doctors abandoning those first principles
and peering into microscopes to find the germs that caused diseases. Among the first organisms they spotted was a spirochete, a spiral-shaped germ they named Treponema pallidum, which was present in patients with sores, rashes, and dementia alike. They concluded that T. pallidum6 was the natural formation that united those scattered particulars, which they now recognized as different stage of syphilis. By century’s end, doctors were asking questions about bacteria and viruses in addition to signs and symptoms, and seeking cures in drugs that targeted those microbes rather than remedies tailored to those outward appearances. Unfettered by archaic beliefs, they were free to find the truth about what ailed us.

  But a century after the advent of the germ theory, as Robins and Guze knew too well, psychiatrists had yet to discover a “schizococcus” bacterium or a “depressenza” virus or anything else that would reduce the profession’s dependence on a priori principles, and the disasters of the late 1960s and early 1970s were the result. So the two men proposed a solution that they thought could keep descriptive psychiatry safe from belief, at least until those bugs could be found: a five-step process toward validity7 that, so they said, required no assumptions, that purely through the accretion of evidence would converge to confirm (or disconfirm) that an alleged disease really existed.

  Start with clinical description, Robins and Guze said, with a careful account of how patients present themselves, and establish the criteria that link similar patients. Add laboratory studies—including psychological tests—that will confirm (or not) that those people belong together. Develop exclusion criteria so that a patient who is, say, depressed but also has hallucinations and delusions gets grouped with the schizophrenics rather than the depressives. Do follow-up studies to make sure that the people you’ve grouped together have similar outcomes, as you would expect if they were suffering from the same disease. And study the patients’ families to see if their members share symptoms, which would indicate that there is some genetic link among the patients. By working all of these angles, they argued, doctors would eventually accumulate enough evidence to say which mental disorders were valid and which were only figments of an enthusiastic doctor’s imagination.