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

by Gary Greenberg

  But when it came to revising that language, they were now the outsiders. Spitzer and Frances had been placed on the enemies’ list. First had been pushed aside. Pincus had never been involved. Wakefield (who, as a nonpsychiatrist, was never as inside as the rest) had had his evidence used against him. They were still shouting at one another, still teasing, enjoying conversation as if it were a contact sport. But their shared love of argument, their common language, and the DSM they had fashioned—the one that, in true Talmudic fashion (and it was hard not to notice that these men were all Jews by birth), they were both proud and skeptical of—was now the language of opposition, of resentment, and, increasingly, of powerlessness and cynicism. The event they’d all come together to commemorate signaled not just the end of a career, but the passing of an era.

  There was one word in the discussion with deeper meanings that an outsider might have grasped, if only by the seething anger with which it was uttered: “Darrel.”

  Darrel Regier is not a shouter or a teaser (or, as it happens, a Jew). He is a mild-mannered man, a Midwest-educated bean counter whose time in New York had begun and ended with his medical internship at Montefiore Medical Center in the Bronx, who has no apparent use for irony, and whose ability to insulate his sentences from objection with thick layers of bureaucratese was probably more useful in his ascent to the rank of vice admiral in the U.S. Public Health Service than shouting and teasing ever would have been. After all, it’s pretty tough to argue with a sentence like this one, taken from an article he wrote.

  Moving forward2, I believe the more critical issue for work group members is how to avoid mind-body dualism in which mental disorders are moved in neurological classifications as more precise pathophysiological, neuroimaging, genetic, nutritional, infectious, traumatic or other etiological characteristics are discovered.

  Not that Regier lacks the ability to fight. “Darrel knows how to throw an elbow,” one psychiatrist told me. Indeed, just a month or so before the gathering for Spitzer, Regier’s pugilistic skills had been on full display at a conference on psychiatric nosology held in Copenhagen and sponsored by the Danish National Research Foundation’s Center for Subjectivity Research—a gathering originally scheduled for April, but delayed when a volcanic eruption in Iceland closed European airports.

  Regier’s talk was in many ways a career retrospective. He told his colleagues that he had entered the profession in the early 1970s, in the bad old days when the Midtown Manhattan Study and other pre-DSM-III efforts were turning up inordinate rates of mental illness in the population, and when “the clinical judgment of a psychiatrist3 was the gold standard of psychiatric diagnosis.” The DSM-III had made matters a little better, but not much, he said—the 50 percent lifetime prevalence rates reached in some analyses of his ECA project were still too high and the National Comorbidity Survey4, the one that had shown that 27 percent of diagnosable people had more than one DSM-III disorder, were a blot on the nosological record. Diagnosticians may have gotten more scientific, but their methods still yielded results that flew in the face of common sense.

  Regier recounted the two epidemiological studies from the 1950s that had yielded much lower (and, Regier thought, realistic) prevalence rates. These projects shared one important characteristic: they relied on doctors to identify the cases in a community. Ask doctors who under their care is sick, it seemed, and you get a much lower number than if you ask randomly selected people about their symptoms. “We had much greater confidence5 in the clinical assessments of physicians with some longitudinal knowledge of all persons requesting care,” Regier said.

  So perhaps that old gold standard, clinical judgment, had been too hastily discarded in the name of reliability. Intimate knowledge of a patient’s history and circumstances, coupled with sound medical knowledge, seemed to provide a check on diagnostic exuberance, identifying within the group that met the criteria those individuals who were truly ill. But how to incorporate that into a reliable nosology? Or to put it another way, how could epidemiologists cut the vagaries of physician judgment out of the loop while still maintaining doctors as the gatekeepers to diagnosis and treatment?

  Regier’s solution was to try to establish when a person has crossed the threshold from healthy to sick. In their introductions, Spitzer and Frances had instructed users of the DSM that if the symptoms weren’t clinically significant, there was no diagnosis to make; no matter how many symptoms you had, you weren’t mentally ill unless your suffering was clinically significant. But that term had never really been defined; establishing the threshold had been left up to the astute clinician. To find a definition that could be consistently and reliably applied to individual cases was to answer this question, or so Regier proposed. It would allow doctors to specify when the criteria signified the presence of illness. And since those cases were sure to be only a subset of all the people who met the criteria, the prevalence estimates would come down from the stratosphere.

  Regier thought the answer was hiding in plain sight—in data collected, but lying dormant, in the ECA and NCS studies. Researchers in both had asked subjects about how much their symptoms had interfered with their lives, and especially whether or not they had been so bad that they had gone to see a doctor about them. The answers had not been used to determine whether or not the people were mentally disordered, but in the late 1990s, Regier’s team began to dig them out6. They fashioned an algorithm, applied it to the data, and presto! Prevalence rates came in at 18 percent—still higher than where Regier thought they should be, but still as much as two-thirds lower than the original findings.

  Regier was sure he’d threaded the needle, but others disagreed, sometimes loudly. “To put it mildly7,” he told the Copenhagen crowd, “howls of protest arose from Spitzer, and later from Jerome Wakefield and others.” The first skirmish was at a meeting of the American Psychopathological Association in 2002, and two years later, when Regier published the results, a fight broke out in the pages of Archives of General Psychiatry. Wakefield and Spitzer pointed out8 that of the nine questions used to generate Regier’s data, seven were about whether people had sought treatment; only two were about how debilitating the symptoms were. But plenty of sick people never saw a doctor, and plenty of people who saw doctors weren’t sick. Regier had confused treatment seeking with clinical significance, they said, which meant he had reduced prevalence by gerrymandering the patient population rather than actually specifying who was sick and who simply met the criteria.

  “What is striking about this debate9,” Regier continued, “is that it became even more personal” when Ronald Kessler, who headed the NCS study, warned in a 2003 Archives article that it would be a mistake to use Regier’s methods to remove mild disorders from the DSM-5—an agenda Kessler thought he had sniffed out in Regier’s research. Kessler wrote that disorders falling short of the threshold Regier was proposing could and often did lead to more severe disorders. A person who confessed to anxiety or depression symptoms on a survey, but wasn’t distressed enough to go to a doctor today, might well end up in the emergency room next month with a panic attack or after a suicide attempt. Here again, the problem was a kind of gerrymandering. The way to solve high prevalence rates, Kessler concluded, was not “to define the problem out of existence10” by suggesting that mild illness wasn’t illness.

  It is a little hard to discern what exactly was personal about Kessler’s or Wakefield and Spitzer’s papers. They are workaday journal articles, all charts and graphs and eye-glazing details leading to careful conclusions. But there’s no question that the articles got Regier’s goat, at least to judge from his response, which he recalled for the audience in Copenhagen.

  In response to this scientific critique11, we thought it would be helpful to highlight the central issue in DSM-5 revisions very clearly and chose the James Carville approach by responding with a letter to the editor in Archives entitled, “For DSM-5, It’s the ‘Disorder Threshold,’ Stupid!”

 
; Now that’s personal.

  “Bob did put up quite a fuss at that conference [in 2000],” Jerry Wakefield told me. “But Darrel never forgave him.” And ten years later Regier was still settling old scores, or at least revealing scores he’d been keeping for a long time. “I wasn’t aware that he had interpreted12 my remarks that way,” Ronald Kessler said when I relayed Regier’s comments to him. “I honestly don’t know what he means about it being personal.” But, he added, it wasn’t entirely surprising. “Darrel reacts very strongly to people who disagree with him.”

  And by the time Regier got to Copenhagen, Allen Frances was on the top of that list.

  Certainly some of the loudest concerns13 in the blogs and print media are that we will unleash a wave of false epidemics—with the past editor of the DSM-IV claiming credit for creating false epidemics of pediatric bipolar disorder, ADHD, and autism . . . He also proclaims little confidence that the current editors can fix this problem or do better in the next edition.

  Regier went on to denounce Frances in a way that seemed more inspired by Joseph Biederman than James Carville. Just look at the history of “progress14 in the calibration and validation of scientific constructs in astronomy, biology, and psychiatry,” he said. “Defenders of the current construct do not yield easily to suggested changes in paradigms.”

  These debates were not going to end with a pinch on the cheek and a kiss.

  • • •

  But if Frances was playing the pope to Regier’s Galileo, did that mean that the paradigm shift was back on? It is possible the DSM-5 leaders couldn’t make up their minds, that the APA, like a squirrel in the road, knew enough to know it was in trouble, but not enough to know what to do about it.

  There was one way in which a paradigm shift was in the making. It was not the sweeping change to a brain-based nosology for which they so devoutly wished, but it was nonetheless far more than a small technical matter. And it was the subject, beneath the vitriol, of Regier’s talk in Copenhagen.

  Regier’s argument with Wakefield and Kessler was, as his Carvillian title pointed out, about the disorder threshold, about how a doctor determines that a person has become a patient. And here was one place that Regier agreed with Frances. Not that he would use barnyard epithets, or epithets of any kind, but he did not think the solution was to fabricate a definition of mental disorder. “It may be of interest15 to this audience,” he said, after reading aloud the DSM-IV definition, “that almost none of the DSM-V Task Force or Work Group Meetings struggled with these definitional issues as they evaluated the research literature to determine the evidentiary basis for revisions.” The committees weren’t wasting time on philosophy. “Our plans for DSM-V are to provide a range of cross-cutting measures that will identify continuous measures of emotional, cognitive, addictive, and other domains,” he said. “Most of our efforts will focus on the individual diagnostic measurements that permit better assessments of the thresholds between normal and pathological states.”

  Regier was confirming his intent to forge ahead with the new paradigm. (Don’t feel bad if you missed it. If Galileo could have buried his meaning so deeply, his heresy might have been lost on the bishops.) “Since the broad definition16 [of mental disorder is] almost impossible to test,” he said, “most of our efforts will focus on the individual diagnostic criteria and dimensional measurements that permit better assessment of the thresholds between normal and pathological states.” The DSM-5 was going to abandon the attempt to define mental illness or to derive clinical significance by proxy measures like treatment seeking. Instead, it would work to overcome “one of the clearest limitations of our current diagnostic criteria . . . the lack of quantitative measures.” It would provide the means to “set diagnostic thresholds for disorders.” Indeed, it would require clinicians to do so by administering tests that would assess the severity of the disorder, as well as the presence of symptoms—anxiety, sadness, obsessiveness, and so on—that might not fit the diagnosis but were nonetheless relevant to treatment. Numbers, not words, were going to guard the gates to the land of mental illness.

  • • •

  Not every psychiatrist believes it’s necessary to keep such a tight watch on those gates. Roger Peele, APA trustee and DSM-5 task force member, for example, acknowledged that the “DSM-IV has a label for everyone you might want to treat17.” But, he added, “ENTs [ear, nose, and throat doctors] don’t worry about the fact that everyone is going to have a cold.” Peele wished the “APA would get over their hang-up” and stop trying so hard to “avoid implying that everyone has had or will have a mental illness.”

  Peele is eighty years old. He’s tall and handsome and has a sure, long stride and a regal bearing. He’s wearing an ascot tucked into an open-collar shirt that matches his blue eyes. But maybe it’s his low-cut, old-school sneakers, or his humble office with its government-issue furniture and less-than-commanding views of the Rockville, Maryland, strip malls, or the fact that he works for a county government, but somehow his suggestion that the APA’s nosological restraint is a neurotic symptom doesn’t seem designed to enhance his power or give the APA permission to unleash its id in unrestrained pursuit of financial satisfaction. Rather, it seems in service of the truth, at least as he sees it: that we all suffer psychological setbacks, that we could all at some point benefit from some time spent with a headshrinker, that the DSM-IV’s universal appeal, and its implicit provision for universal treatment (at least for the insured), are good for all of us, doctor and patient alike.

  On the other hand, Peele ran for APA president in 2000 and 2009 and lost. The organization might not be ready for this kind of candor or to give up once and for all on the quest for a way to distinguish the sick from the merely suffering.

  Darrel Regier certainly isn’t. “That’s the problem I went into the field to address,” he told me at the outset of our interview. To a numbers guy like Regier, concern over the findings of the Midtown Manhattan Study—the infamous 85 percent—is no mere hang-up. The result is a statistical impossibility. We can’t all be abnormal any more than everyone who lives in Lake Wobegon can be above average. And if most of those people never seek treatment, then how sick could they be? Moreover, as Regier likes to point out when people accuse the APA of disease mongering, there is no way that the current mental health treatment system, with its 45,000 psychiatrists and half a million psychologists, social workers, and counselors could possibly keep up with all that demand. Bad as it is that the numbers are likely inaccurate, it would be worse if they were not, because then we would truly be awash in mental illness, our treatment resources swamped.

  Regier, a lifelong public health doctor, has to think about planning and policies and other unglamorous concerns that cannot be met without knowing how many people are sick and what diseases they have. That’s reason enough to devote a career to finding the elusive threshold, or so it seemed when I asked him to tell me exactly what the problem was with overdiagnosis. I wondered how an elite psychiatrist viewed the question, exactly what kind of harm he thought his profession could cause by labeling us. I wanted to know if he thought that a doctor who tells a widower that his grief is an illness is potentially not only labeling, stigmatizing, and medicating him, but also shaping the patient’s understanding of his loss, of himself, of the meaning of his life.

  But Regier didn’t seem to comprehend my question. He seemed to think I wanted him to explain what false positive meant. He might not be the philosophical sort or, after a lifetime of cultivating bureaucratic habits, he might have lost sight of philosophy. He might have only wanted to not get so far away from his talking points with me. Or maybe he thought the answer was self-evident: sky-high prevalence rates are an affront to reason, to everything that scientific medicine, with its aspiration to carve nature at its joints, stands for.

  Jay Scully was not so reticent.

  We went through the usual reasons that psychiatrists give when they worry about
the overdiagnosis problem: overtreatment, stigmatization, the credibility of the profession. But I wanted more from him. I wanted him to tell me why, from a psychiatrist’s point of view, it was a bad thing to, as he had just put it, “overpathologize the human experience.”

  “Because it limits the other potential ways to fix things. It limits the whole spectrum of the human experience,” he said. “I mean really bad things happen that are not necessarily psychopathological. You’d have to be crazy to kill your family. Well, do you? Maybe not, but how do we know? Where are the sharp dividing lines? Maybe we never really know.”

  It was the first time in our half hour together that Scully seemed genuinely interested, reflective, even spontaneous. “So,” I asked him, “the false positive problem is the problem of where psychiatry’s limits are?”

  “Yes,” he said. “You know, we always need to be humble.”

  As I sat with the CEO of American psychiatry, with his $600,000-per-year salary18 and his large office with its bird’s-eye view of the Potomac River and the Capitol, it was good to hear that word. It was good to know he thought that humility was important and to have heard him say, “Our treatment and ability to relieve suffering need work.” But he got quickly back on message, and the PR flack was looking at her watch and telling us our time was up, so I couldn’t ask him exactly what psychiatrists need to be humble about—whether it’s only the gap between suffering and their ability to relieve it, or if he was aware of that other, deeper chasm, the one between opportunity and knowledge, the one that gives psychiatrists the power to say what that dividing line is, and to cash in on saying it, even though they themselves don’t really know. Still, I was glad to hear this much.

  Because as Montesquieu once said, “The spirit of moderation19 is what we call virtue in an aristocracy. It supplies the place of the spirit of equality in a popular state.” Perhaps medicine is inescapably an aristocracy, and self-imposed moderation the only check on power, the only way to prevent beneficence from becoming oppression. Surely medicine can’t be a democracy; we can’t vote on whether or not we are sick or elect which illness we have. If we can’t depend on definitions of mental disorder or on statistical measures of clinical significance to prevent the DSM from cataloging the entire spectrum of human experience, from turning all mental suffering into illness, then perhaps we are dependent on the sophistication, the restraint, the discretion—in short, the virtue—of our psychiatrists.