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

by Gary Greenberg

  Of the 1,308 NCS subjects with major depression, 157 reported that their symptoms had been set off by a bereavement and 710 said that a different kind of loss had triggered them. Wakefield then examined to what extent those two groups differed along two dimensions: disorder indicators, such as how many symptoms of depression they had or how long they lasted; and symptom group, which of the nine depression symptoms they reported. It turned out that there was virtually no difference between the two groups in either dimension. Divorce, death of a loved one, financial setback—each was as likely as the others to leave a person sad or apathetic or insomniac, as likely as the others to send a person to a therapist or a doctor for an antidepressant.

  Wakefield further divided his subjects into uncomplicated and complicated groups. Complicated subjects were those with the severe symptoms, such as suicidal thoughts or the inability to function, that would have disqualified them from the bereavement exclusion, and uncomplicated subjects . . . well, you figure it out. Not surprisingly, people in each group resembled one another more than they resembled people in the other group who had suffered the same kind of setback. Loss is loss, in other words. What differs is how we respond to it, not what we are responding to.

  Taken together, these two findings—that no matter what the precipitating event is, it has the same effects, and that the main difference is in the intensity of response—indicated that there was no scientific reason to treat the death of a loved one differently from other losses. The experiment was simple, elegant, and conclusive. The bereavement exclusion was arbitrary and incoherent, and if that was so, then the MDD criteria were not able to sort the well from the sick.

  In The Loss of Sadness, Wakefield and his coauthor, sociologist Allan Horwitz, made it clear that more was at stake than the bereavement exclusion. Even if we accept the inescapable fuzziness of psychiatric diagnoses, they argued, still a nosology must have some kind of integrity, lest the DSM “define every undesirable consequence7 of sadness as a disorder.”

  “Psychiatry has made immense strides8 in recent decades,” they wrote, but one thing it had not figured out was how to limit its tendency to “engulf all the problems that life poses.” The solution was not to abandon the depression diagnosis, even if the bereavement exclusion constituted a “compelling, clear and major violation of validity.” Instead, they wrote, the diagnosis “should be fixed.” And the best way to do that was to expand the exclusion beyond bereavement to “other major life stressors” so that clinicians could distinguish between disorder and distress, and diagnose and treat people accordingly. Forcing clinicians to consider the actual life experience of their patients would limit the damage that their diagnoses could do even as it maintained their ability to help patients.

  “There are few signs9 that changes in the MDD criteria are a high priority for the . . . DSM-V,” Wakefield and Horwitz concluded. But that was in 2007, before their book became a minor sensation, and just as a spate of other books and articles complained that depression was getting overdiagnosed and, not coincidentally, antidepressants alarmingly overprescribed. And as the DSM-5 revision geared up, it became clear that researchers concerned about psychiatry’s overstepping weren’t the only people disturbed by the bereavement exclusion.

  “It just doesn’t make any sense to me whatsoever10,” psychiatrist Sidney Zisook told me on the phone. “Maybe it did when I first got into the field, but not anymore.”

  But Zisook’s objection was different from Wakefield’s. It wasn’t the double standard—the fact that only the grieving were eligible for a reprieve from diagnosis—that bothered him. It was the assumption that bereavement-related depression (BRD), as he called the suffering covered by the exclusion, was somehow different from standard major depression (SMD). If the two were not different in nature, if, in other words, the only difference between BRD and SMD was in their cause, then there really was no difference between them, no reason to distinguish patients according to their recent experience.

  In 2007, working with Ken Kendler, Zisook did some data mining of his own11. He extracted information from about fifty studies conducted over three decades and comprising more than eight thousand depressed subjects, only some of whom met the exclusion criterion. They looked at the Robins and Guze validators—family history and other demographics, overall health (mental and otherwise), lab tests, response to treatment, and the course and outcome of the patients’ troubles. And while some of the evidence was inconclusive—it wasn’t clear whether BRD ran in families, for instance—most of the results showed that BRD was no different from SMD. Bereaved people who had the symptoms of depression were no different from other patients who had the symptoms. Wakefield had proved that loss was loss; now Zisook and Kendler had shown that depression was depression, regardless of its cause.

  Zisook and Kendler were careful with their conclusions, at least at first. Their analysis only “provides some support12, although indirect and limited,” they wrote, “that excluding recently bereaved individuals from the diagnosis of MDE [Major Depressive Episode] . . . may not be justified.” And pending further research, “the most propitious conclusion may be that, on average, these two syndromes [BRD and SMD] appear to be closely related.”

  Six months later, however, Zisook and Kendler, along with a third collaborator, published another paper on the same topic. Its data and findings were nearly identical to the earlier report’s. But there was something new in the second article. “Because work toward the DSM-V13 has begun,” they wrote, “it is timely to re-examine the DSM’s bereavement exclusion.”

  Reframed as a discussion of what should be done now that the revision was under way, the questions about the bereavement exclusion took on a new urgency and the authors demonstrated a new certainty. First, they pointed out, research had shown that the earlier depression is treated the better, and that antidepressant medication is particularly effective in the earliest stages of a depressive episode. Which means, they said, that the “validity of the bereavement exclusion14 . . . is not an academic issue.” Indeed, Zisook told me, we should also worry about what happens to people who don’t get diagnosed.

  “I don’t think we’re doing patients a service by avoiding the label if they meet the criteria,” he said. “I think if we say that in addition to their bereavement they are also depressed, we might intervene more aggressively and sooner and save people a lot of pain, discomfort and, who knows, maybe even lives.”

  “Why should bereavement be singled out15 as the only stressful life event that excludes the diagnosis of MDE?” they asked. It was a rhetorical question, of course. To Wakefield and Horwitz, psychiatric diagnosis was a potential harm against which we need protection. To Kendler and Zisook, it was a benefit that we should not be denied.

  “Idiotic16,” said Jerry Wakefield, when I asked him what he thought of the proposal to remove the exclusion. “But clever in its own diabolical way—they used my evidence against me!”

  Wakefield was not on the mood disorders work group for the DSM-5. Neither was Michael First or Allan Horwitz. Ken Kendler, on the other hand, was, and Sid Zisook was a special consultant to it. Which may or may not explain why the work group decided, even in the acknowledged absence of direct evidence and despite the tentativeness of the conclusions—not to mention the obvious public relations problem—to eliminate the bereavement exclusion. At the very least, the proposal seems to confirm the notion that experts, given the charter to implement their pet theories as diagnostic changes, will tinker, and that their tinkering is likely to be in the direction of more rather than fewer diagnoses.

  Experts like Zisook are not unaware of the possibility of overdiagnosis. They’re just less worried about what they call false positives than about false negatives. Leaving an illness untreated is more dangerous, in their view, than making illness out of everyday life. And Zisook may be on the National Advisory Board of GlaxoSmithKline, for whom he once ran a study showing that GSK’s Wellbu
trin17 was effective for grieving people, but that doesn’t necessarily mean he’s shilling for the company by proposing the expansion of the diagnostic net, or even that he will hawk their products to more patients once the exclusion is lifted. “I don’t think everyone with major depression needs meds or formal therapy,” said Zisook. “I think sometimes watchful waiting is appropriate. Sometimes support and education is appropriate. Sometimes decreasing stress is appropriate.” A good clinician knows when to do which. Expertise is what guarantees that a doctor’s power is used beneficently.

  Zisook places great stock in expertise—and not just when it comes to treatment. “The DSM criteria don’t allow for judgment on symptoms,” he told me. “I think a really astute clinician will make decisions as to whether a behavior is a symptom of depression or not.” Removing the bereavement exclusion won’t stop a good doctor from withholding a diagnosis from a grieving patient even if she meets the criteria.

  Consider, Zisook said, a patient who is recently widowed and not only sad and withdrawn but also guilt-ridden over all he should have done to prevent the loss of his wife. “That guilt would count toward depression,” Zisook acknowledged, but “when I have this patient in front of me, I don’t count that as a symptom of depression.” Expertise can be relied upon to determine whether guilt is a symptom or just guilt, and whether the criteria add up to a disorder. Astute clinicians, like the crackerjack doctors who staff Zisook’s elite university medical center, don’t need bereavement exclusions to limit their diagnostic powers. They don’t need to slave away at checklists like common clerks.

  But Zisook evidently does not want to extend this same latitude to the rank-and-file clinicians, the ones who spend their days seeing patients with a range of psychiatric disorders—not to mention the general practitioners who do most of the prescribing of antidepressants—and who aren’t supported by research grants and generous faculty salaries, who don’t have an hour or so to spend talking with their patients or administering carefully calibrated diagnostic tests that can be scored by assistants or worrying over the fine points of what it means to say that the miserable person in front of them actually has a mental disorder. Run-of-the-mill doctors, it seems, can’t be trusted with a bereavement exclusion, any more than the layman can be trusted with all that Wizard of Oz stuff. They must remain loyal to the fundamental tenet of the DSM: that, as Zisook put it, the “criteria are the same regardless of context. If someone meets them, I should give them the label of major depression because that’s what they have”—although evidently only so long as the “I” in question is not Zisook or any other clinician too astute to take the DSM too seriously.

  • • •

  “The DSM-IV position is not logically defensible18,” Ken Kendler wrote in a statement the APA posted on the DSM-5 website shortly after Frances’s Times op-ed appeared and a public furor broke out.

  Kendler wasn’t talking about the incoherence of saying that depression is depression unless an elite clinician determines that it is not. He was talking about the way the exclusion contradicted the basic idea of the DSM—that mental disorders can (and, for now anyway, must) be known by their symptoms alone. “The grief exclusion criterion needs to be [either] eliminated or extended so that no depression that arises in the setting of adversity would be diagnosable.” And, since “the majority of individuals” who get the diagnosis develop it “in the setting of psychosocial adversity,” he added, this curtailing of the diagnosis would “represent a major shift . . . in the nature of our concept of depression.”

  It’s not clear where Kendler got the idea that the only alternative to eliminating the exclusion was to make it impossible to diagnose depression in the presence of a stressor. It couldn’t have been from Wakefield, who never said that depression kindled by adversity should go undiagnosed. To the contrary, he was trying to lay the groundwork for a return to a traditional, pre-DSM-III idea of depression: “that sadness in response to loss19 is natural and normal, and that the indication of disorder lies in the sadness being without sufficient cause in given environmental contexts or being disproportional to actual loss.” Depression, according to this abandoned notion, was sick only to the extent that it occurred without regard to a person’s life circumstances.

  Wakefield was proposing, in other words, that if you want to know if a person is really sick rather than simply reacting to a loss, you have to look at the relationship between the stressor and the person’s reaction. You have to try to understand what the loss means to the patient. And if it is out of proportion, then you have to figure that you are in the presence of a genuine illness—something gone wrong in the body (in this case, the brain)—rather than normal sorrow. Indeed, Wakefield was only suggesting that the prerogative Zisook reserved for the astute (but would deny to everyone else by removing the exclusion) was actually an obligation for all clinicians: to pay attention not only to diagnostic checklists but also to the patient in front of them.

  Right now, if you are not a psychiatrist or any other kind of mental health worker, or maybe even if you are, you are probably thinking, “Duh!” You’re probably wondering how it is possible that something as obvious as this even needs to be said, let alone argued about. But that’s only because you have decided to make your living in some other fashion, one that does not require you to satisfy an insurance company or your colleagues in other specialties or yourself—if you take seriously your involvement in the medical-industrial complex—that you are treating “real illnesses.” You haven’t signed on for a profession whose badly tarnished reputation was once restored by applying the thinnest veneer of scientific rhetoric and polishing it vigorously, and which now has to protect itself from the sullying effects of a proposal like Wakefield’s, even if it means taking the absurd position that there is no essential difference between a depression kindled by the loss of a child or parent and a depression brought on by the cancellation of a favorite television show or the death of a pet hamster—or between any of those and a depression that seems to come on out of nowhere.

  Good for you, I say. Because if even a man of integrity like Sid Zisook, who is as affable and humane and charming a guy as you might like to meet, and surely the kind of psychiatrist—empathic, articulate, knowledgeable—whom, if you needed one, you’d want to see, if such a man finds himself, in the space of two minutes, totally contradicting himself, talking out of both sides of his mouth, and unwittingly becoming the best evidence of the incoherence of his own position merely by acknowledging that he pays attention to his patients, when smart and compassionate doctors spill buckets of ink over the question of whether or not they should consider what we are going through before they tell us what we are suffering from and what we should do about it, and when the proposal to do so moves Bob Spitzer to say to me, “If we did that, then the whole system falls apart,” then you know that you are knee-deep in a setting of psychosocial adversity.

  And, indeed, all of us in the mental health industry know this, from the APA trustee who shook his head and told me in a mournful voice, “We’re just so stuck,” to Allen Frances to Steve Hyman to lunchbucket therapists like me who hold our noses when we put those codes onto the bill. We may not be sophisticated enough to grasp the nobility of the lie that sustains the DSM or to be trusted as its keepers. We may not be astute enough to distinguish between depression and grief. But we all know that the DSM is at its best a clumsy and imperfect field guide to our foibles and at its worst a compendium of expert opinions masquerading as scientific truths, a book whose credibility surpasses its integrity, whose usefulness is primarily commercial, and whose most ardent defenders are reduced to arguing that it should be taken less seriously even as all of us—clinicians, researchers, and copyright holders alike—cash in on the fact that it is not.

  Chapter 12

  On December 17, 2010, instead of the usual Friday grand rounds, the department of psychiatry at Columbia University honored Bob Spitzer on the occasion of his re
tirement. Colleagues from past and present paid tribute with recollections and praise. A historian put Spitzer’s forty-nine-year career into historical perspective. Allen Frances spoke from the same lectern from which he had railed against DSM-5 eleven months earlier. This time he really did hold back. He scarcely mentioned the revision. He didn’t tease Spitzer. He just talked about his contribution, about how he was one of the most important psychiatrists who had ever lived, placing him in the company of Freud and Kraepelin. He couldn’t resist reminding the audience about the limitations of the descriptive method, but still, he said, it is “the best we can do1, and Bob has been and will remain our master describer.”

  Like many such celebrations, the proceedings were more than a little funereal—a tone deepened by Spitzer’s halting, unsteady ascent to the stage, his quavering voice, his obvious frailty. After warning the audience that his Parkinson’s disease had worsened his tendency to burst into tears, he did exactly that. He didn’t get everything said that he intended to say. He had a little trouble remembering exactly how many children he had. But between gusts of weeping he told stories and recounted arguments in a way that conveyed both the charisma and the toughness that had allowed him to become the man who saved psychiatry.

  That night Spitzer joined some of his colleagues and collaborators and their wives for dinner at a downtown restaurant. Harold Pincus, Michael First, Jerry Wakefield, and Allen Frances were there. The DSM was far from the main topic of the loud and lively conversation. When the subject did come up, it was in a dense shorthand, the code that signals deep agreement, words such as “field trials” and “polythetic criteria” and “prevalence rates” that carried a world of meaning impenetrable to outsiders. Theirs was a shared, private language, much as they thought the DSM’s should be.