Manufacturing depression (35 page)

Rank-and-file psychiatrists did eventually figure out that a big change was afoot, and much of the ensuing consternation focused on another word slated for elimination—
neurosis,
which, the new Committee on Nomenclature argued, should not appear in DSM-III because it
“assumed…an underlying process
of intrapsychic conflict resulting in symptom formation.” In other words, you needed a theory about mental illness, about what caused it and where it came from, to diagnose a neurosis, and theory—especially the Freudian theory in which neurosis played a central role—was exactly what had gotten the profession into its reliability troubles in the first place. It had to go.

Eradicating
neurosis
was not as easy as getting rid of
reaction
had been. Neurosis, and especially depressive neurosis, was the psychiatrist’s stock in trade, the general label for the everyday discontents that Meyer had long ago said were a proper indication for outpatient psychotherapy. The proposal to erase the word crystallized opposition to the remaking of DSM-III. Some doctors took a historical approach, pointing out that
neurosis had been first described
not by Sigmund Freud, but by Scottish physician William Cullen in 1769, and thus had earned a place in medicine. But others got right to the point.
“DSM-III gets rid of the castles of neurosis
and replaces it with a diagnostic Levittown,” one psychiatrist said; most psychiatrists knew where they would rather live. Another colleague dispensed with metaphors and appealed directly to doctors’ self-interest. Without neuroses of various kinds, he wrote,
“many patients
who are not in prolonged therapy will be said to have no disorder.” The whole world could not be insane, at least not reimbursably insane, if the net were cast so much more narrowly.

But that was exactly the committee’s intent—to prune the taxonomic tree of its less reliable branches, of which neurosis, weighed down with the Freudian idea of a dynamic inner world, was perhaps the most rotten. So when a psychiatrist lamented that proposed
changes would turn the DSM into
“a straitjacket
and a powerful weapon in the hands of people whose ideas are very clear…and the guns are pointed at us,” he wasn’t as ready for his own straitjacket as he sounded. Indeed, Donald Klein, a pharmacologist and prominent defender of DSM-III, only confirmed those suspicions when he proclaimed that opponents of DSM-III
“wish [neurosis] reinserted
because they wish a covert affirmation of their psychogenic hypotheses.” Taking such malcontents seriously could only spell disaster for
“scientists attempting to advance
our field via classification and reliable definition.”

Spitzer did take seriously the
“pro-neurosis forces”
—in the infelicitous term used by one of them—at least seriously enough to offer any number of compromises. He suggested allowing psychoanalytic-minded doctors to insert an
N
after the diagnostic code, indicating that the clinician thought the problem had something to do with conflict in the psyche. He floated the idea of adding
neurotic
as a descriptor after certain labels, but only in parentheses. He promised to allow the pro-neurotics
a large role in “Project Flower,”
which would produce a companion volume to DSM-III that would allow theoreticians to fill in the diagnostic picture beyond the criteria lists—and whose name, Spitzer said, was inspired by Mao’s aphorism, “Let a thousand flowers bloom.” He even invited psychoanalysts to add some of their Project Flower material to the DSM’s introduction.

In due course, however, the
N
modifier disappeared, the introduction idea was dropped as
“extremely embarrassing
and extremely divisive,” and Project Flower somehow failed to bloom. After their trip to the diagnosis wars, all that the pro-neurotics ended up with were lousy parentheses: anxiety disorder became
anxiety disorder
(or
anxiety neurosis
) and depressive neurosis became
dysthymic disorder
(or
neurotic depression
). And in April 1979, after five years of diplomatic nosology, after the Talbott Plan and the Offenkrantz Complaint and Washington Challenge and the Modified Talbott Plan, after the APA’s assembly elected to approve the DSM-III, the APA’s board of trustees once again voted on the existence of diseases.
This time, the stroke of their pen didn’t eliminate a single illness but rather a whole class of them, even as it created some fifty more that hadn’t previously existed. But these were new and improved diseases, the kind that could be reliably diagnosed without recourse to theoretical notions about how the mind works.

The DSM-III was a huge hit. Purged of theory, of any pretense to saving the world, and of any claim to know how the mind worked or what caused mental illnesses, the book was invaluable to psychiatrists’ attempt to secure their place in “real medicine.” Thanks to the descriptive approach, there would no longer be any question about who was schizophrenic and who was manic-depressive, or, for that matter, who had major depressive disorder (MDD), as it was now called, and who was merely unhappy. Nine out of ten doctors using the criteria agreed on diagnoses, a spectacular improvement over the old days of theory-laden nosology.

The DSM criteria for MDD
were straightforward: take one from column A (“dysphoric mood or loss of interest or pleasure in all or almost all usual activities and pastimes…sad, blue, hopeless, low, down in the dumps, irritable”), four from column B (“poor appetite or significant weight loss…or increased appetite or significant weight gain…insomnia or hypersomnia…psychomotor agitation or retardation…decrease in sexual drive…fatigue…feelings of worthlessness, self-reproach, or excessive or inappropriate guilt…diminished ability to think or concentrate…recurrent thoughts of death, suicidal ideation, wishes to be dead or suicide attempt”), and rule out the symptoms in column C (“mood-incongruent delusion…[or] bizarre behavior,” which are indications of other disorders), and you’ve got your diagnosis. A similar process could lead to
dysthymic disorder
or
adjustment disorder
with depressed mood.
“Clerks rather than experts
can make this kind of classification,” one psychiatrist grumbled. But of course that was exactly the point.

These criteria weren’t original to the DSM. In fact, Spitzer and his committee had lifted them, sometimes word for word, from
the Feighner criteria
, invented by a group of researchers at Washington University in St. Louis who, in 1972, had developed descriptive diagnostic standards for depression (and fourteen other psychiatric disorders). The Washington team was the first to achieve those excellent reliability numbers, and other researchers were soon scrambling to hitch their wagons to the Feighner star, using the criteria to back their own studies or as a model for their own tests. By 1989, the paper introducing the criteria had become
the single most commonly cited article
in the psychiatric literature.

What the Feighner criteria didn’t address was the old Kraepelinian problem, the one about the symptoms constituting the diseases and the diseases comprising the symptoms. All the reliability in the world does not add up to validity. That’s an especially glaring omission when you consider that the paper came out at the height of the battle over homosexuality, a condition whose presence doctors could reliably agree upon even without fancy criteria. Not to mention that in the particular case of MDD, the Feighner criteria bore a strong resemblance to the items on the Hamilton Depression Rating Scale, whose own author had long cautioned that his test was not valid for making diagnoses.

The authors tried to gloss over the issue by conflating reliability and validity.
“This communication will present
a diagnostic classification validated primarily by follow-up and family studies,” they wrote, as if those follow-up studies could do more than show that people could be reliably grouped by their symptoms—something Kraepelin had already shown, but that didn’t prove anything about whether or not those common qualities constituted a disease.

The team also wrote that their
“criteria for establishing diagnostic validity
in psychiatric illness have been described elsewhere,” as if others had settled the question. But thirty-five years later, scholars
Allan Horwitz and Jerome Wakefield examined
those other studies and found that their authors never claimed that they had built
a bridge from symptom to disease. One researcher warned that he couldn’t guarantee that his “depressed” patients’ symptoms weren’t the result of some other illness, another concluded his paper by noting that defining “clinical entities by symptom pictures” remained a “serious problem in psychiatry,” and the final source pointed out that in the absence of some theory about what causes depression, it was impossible to sort out the unhappy from the sick. Taken together, Horwitz and Wakefield concluded, “these sources neither justify nor even address the validity of the specific definition” offered by the Feighner criteria. Indeed, the only reason to believe that descriptive psychiatry had solved the validity problem was wishing it was so—which the industry had plenty of incentive to do.

Adopting the Feighner criteria, the DSM-III committee also adopted this wishful thinking. But reality soon intruded. A psychiatrist had discovered that
many people who had recently been bereaved
met all the Feighner criteria. For all their atheoretical purity, those standards, now incorporated into the DSM, couldn’t even distinguish between the diseased and the merely bereaved. And this psychiatrist, Paula Clayton, wasn’t just some pro-neurotic dead-ender. She was on the faculty of Washington University and a member of the DSM-III Task Force on Affective Disorders.

So when
the DSM-III committee were reminded
that, according to Clayton, grief was indistinguishable from depression, when, in other words, the validity problem emerged from the avalanche of reliability statistics under which it had been buried, neither she nor the committee should have been terribly surprised. Neither could they simply ignore it, even if they wanted to. A diagnostic manual that turned a person in the throes of grief into a mental patient was a scientific nightmare and a potential public relations disaster. It threatened Spitzer’s strategy of rescuing psychiatry through a return to Kraepelin, to make all the professional blood spilled in the name of reliability a vain sacrifice.

The committee’s response was to solve the public relations problem, if not the scientific one, by establishing a loophole in the
definition of MDD—the bereavement exclusion.
“A full depressive syndrome,”
the DSM-III eventually said, “is a normal reaction to the death of a loved one,” so a recently bereaved person does not have major depression, even if he is depressed. Instead, he is suffering from uncomplicated bereavement, which doctors could still treat if they liked, but, because it was listed in the section of the DSM-III devoted to “conditions not attributable to a mental disorder,” they were unlikely to get reimbursed for. There was good news on this front, however. The exemption was time limited. After two months, uncomplicated bereavement could become major depressive disorder.

It’s not clear why the exclusion expires after two months. There is some statistical evidence that
grief begins to wane
, on average, about ten weeks or so after a loss, but while this says something about the usual course of mourning, it hardly proves that an unusual course is a disease. Even more important, it’s not clear why bereavement is the only exempt condition, why, for instance, misfortunes like betrayal by a lover or severe financial loss or political upheaval or serious illness—or for that matter a noncatastrophe, the slow accretion of life’s difficulties or a loss of faith in one’s government or simple existential despair kindled by an awareness of mortality—do not also spare people from the rolls of the diseased. If the whole point of the DSM-III was to eliminate considerations of the nature and causes of a condition in favor of pure description, if indeed depression was no more or less than its symptoms, then why was it suddenly, and only in this one case, acceptable to talk about nature and causes?

The scientific answer is that there is no reason. The bereavement exclusion is like the epicycles that Ptolemaic astronomers added to their models of planetary motion—little loops within the orbit of planets that allegedly explained why they showed up in places where Ptolemaic astronomy, with its insistence that heavenly bodies moved in perfect circles, said they shouldn’t be. Epicycles worked on paper, sort of, but they did a much better job at keeping
astronomers respectable and their models intact than at describing the actual movements of heavenly bodies; they have come to be known as the epitome of bad science.

Doctors couldn’t ignore Clayton’s findings any more than ancient astronomers could ignore the actual orbits of the planets, but if they had responded by trying to figure out which other setbacks kindled responses indistinguishable from depression, if they generated a list of exemptions, then their diagnoses would have gotten awfully unwieldy and unscientific sounding. Not only that, but people might well have started marching and raising hell about why their particular cause for grief should (or should not) be on the no-diagnosis list. If, on the other hand, the committee had left the winnowing of diseased sadness from healthy sorrow to the judgment of doctors, then
the profession would have been back to the bad old days
when the circumstances of an individual’s life mattered, when one doctor’s “normal reaction” could be another doctor’s disease. The bereavement exclusion—a single, time-limited exception based in common sense, as if it was just an unfortunate coincidence that bereavement mimicked depression—steered a middle course between these hazards.

Even as a one-time deal, however, the bereavement exclusion doesn’t really work to keep theory, opinion, and judgment out of the clinical picture. It doesn’t, for instance, specify the degree of relationship a patient must have to the deceased in order to qualify. Should a person really be allowed as many days to get over the death of, say, Michael Jackson as that of his own mother? Wouldn’t the diagnosis still require a clinician to make a judgment about just how important those people ought to be to the patient?