The diagnostic dilemma: is revising the official manual of mental disorders a prescription for confusion?

The Diagnostic Dilemma

The God of the Old Testament advised, "Be fruitful and multiply." Developers of the "psychiatric bible" setting forth the guidelines for diagnosing mental disorders -- known secularly as the Diagnositc and Statistical Manual of Mental Disorders, or DSM -- have taken His counsel to heart. The 1980s have seen a rapid revision and expansion for all manner of mental ailments in the DSM, from schizophrenia to social phobia to sexual sadism.

There are, however, rumblings of discontent from researchers and clinicians faced with these "biblical" rewrites every half dozen years.

The DSM, published by the American Psychiatric Association in Washington, D.C., serves as the standard reference guide to psychiatric diagnoses in the United States. DSM-I, published in 1952 as a pamphlet, described nearly 60 mental disorders, and an expanded version, DSM-II, came out in 1968. These two editions contain fairly general descriptions of mental problems based on Sigmund Freud's concepts of psychoses (severe mental disorders such as schizophrenia) and neuroses (less severe forms of psychological conflict and anxiety).

The release of DSM-III in 1980 brought major changes. It largely dropped Freudian terminology and instead listed specific criteria for more than 150 disorders, based on statistical analyses of standardized interviews with thousands of psychiatric patients. Seven years later a revision, DSM-III-R, incorporated more than 250 disorders into the psychiatric fold. DSM-IV is slated to appear by the end of 1992.

The DSM has assumed increased importance in the 1980s because its diagnoses are often required by government and private insurers who pay for psychotherapy and other mental health services. It has also become the reference of choice for scientists studying mental disorders, although there is concern that the rapid appearance of new DSMs will muddy the research waters.

Some psychiatrists hail the impending arrival of the third diagnostic manual in 12 years, but other mental health workers, both in and out of psychiatry, question how fruitful it is for DSMs to multiply so rapidly. They say the flurry of new and changing diagnoses is based on inadequate data, confuses research on mental ailments and may be ignored by many clinicians who are comfortable with the diagnostic guides they originally trained with.

A number of reservations about the current and future diagnostic manuals are set out in the December 1988 ARCHIVES OF GENERAL PSYCHIATRY in an article titled "Why Are We Rushing to Publish DSM-IV?" Its author, Mark Zimmerman, a psychiatric resident at Chicago Medical School in North Chicago, has published numerous scientific studies in collaboration with psychiatrists at the University of Iowa School of Medicine in Iowa City.

"I think psychiatry will get itself into a lot of trouble by changing diagnostic criteria too quickly," Zimmerman says.

To begin with, he notes, the five-year span between DSM-III-R and DSM-IV guarantees that only limited data on the usefulness of current diagnoses will be available to guide revisers. In late 1987, Zimmerman points out, research began on the validity of DSM-III-R criteria, aimed at determining whether a diagnosis such as schizophrenia accurately predicts how a person will respond to specific treatments and functions several years down the road. Initial results will not be ready until 1990, about the same time as the first draft of DSM-IV appears.

Although early validity data no doubt will inform the final version, there will be no time to replicate results suggesting diagnostic changes are needed. Zimmerman asserts. And if past experience serves as a guide, he adds, members of DSM-IV committees in charge of reviewing diagnostic guidelines will tailor many changes to their own clinical judgments.

Zimmerman, a member of the DSM-III-R committee that revised the definition of melancholia, a form of severe depression, saya consensus based on individuals' clinical experiences -- not on empirical data -- governed the committee's thinking. For example, guilt was eliminated as one sign of melancholia because committee members agreed it was not specific to that disorder; no studies were cited, however, to justify the change.

"There's a real problem with people working on DSM-IV not having access to research on DSM-III-R," says psychiatrist Robert L. Spitzer of Columbia University in New York City. Spitzer and Columbia social worker Janet B.W. Williams were major forces in the development of DSM-III-R and its predecessor.

Spitzer nevertheless supports the timetable for DSM-IV so that diagnoses can be aligned with those in an upcoming revision of the International Code of Diseases (ICD), a listing of physical and mental disorders published by the World Health Organization.

The ICD, designed for use by physicians worldwide, is now in its ninth edition. A new edition is scheduled for publication around the time DSM-IV becomes available.

The DSM contains many more mental disorders than ICD, and cultural differences inevitably lead to variations between the U.S. and international manuals. But according to psychiatrist Allen Frances of Cornell University Medical College in New York City, who is directing work on DSM-IV, developers of the two guidebooks are discussing ways to coordinate their diagnoses.

"We're mindful of the need to be conservative and respect past diagnoses," Frances says. "Our major goal is to increase the documentation supporting DSM-IV diagnoses and explain the reasons for any changes to the people who use the manual."

Frances does not share Zimmerman's worry that changes in the new manual will stand on shaky scientific ground. Diagnoses in each broad category, such as childhood disorders and anxiety disorders, will be evaluated by committees of from 50 to 100 people, according to Frances. DSM-IV committee members will conduct an exhaustive review of all available studies on the usefulness of every diagnosis, he maintains.

Yet researchers have reason for concern about upcoming changes in DSM-IV, says psychiatrist Wayne S. Fenton of the Chestnut Lodge Research Institute in Rockville, Md. Seemingly minor alterations may substantially reduce or expand the number of patients assigned a particular diagnosis and frustrate attemtps to build on past studies.

For example, Fenton and his colleagues find that, compared to DSM-III, DSM-III-R reduces the number of patients diagnosed as schizophrenic by about 10 percent. Of 182 Chestnut Lodge patients defined as schizophrenic by DSM-III, only 164 fit the diagnosis of schizophrenia in the revised manual, they report in the November 1988 AMERICAN JOURNAL OF PSYCHIATRY. The rest are reclassified under the category "atypical psychosis."

In DSM-III, schizophrenia refers to a number of related disorders marked by illogical thinking, inappropriate emotions, severely disorganized behavior, hallucinations and delusions. The revised schizophrenia criteria exclude individuals with delusions that are not flagrantly bizarre, such as delusions of grandeur, which occur in the absence of hallucinations, incoherence or inappropriate emotions. While this is not a far-reaching change, it accounts for most of the reduction in schizophrenic diagnoses, Fenton says.

The Chestnut Lodge patients -- who were released from the hospital between 1950 and 1975 -- were contacted an average of 15 years after discharge. There was no difference in overall functioning between those meeting DSM-III or DSM-III-R criteria for schizophrenia. This suggests the revised definition is no more "valid" than the original, Fenton contends.

"More thought has to be given to the nature and frequency of diagnostic changes," he says. "We need diagnoses that build on previous diagnostic systems so we can compare current and past psychiatric research."

In a similar study reported in the same journal, Yale University psychiatrist Fred R. Volkmar and his co-workers find that the DSM-III-R definition of childhood autism is considerably broader than the DSM-III version. The innovations have some merit for clinicians dealing with this complex disorder, they say, but autism researchers may want to use DSM-III to minimize the possibility of inadvertently including children with other severe disorders in their studies.

"Changes in diagnostic concepts [of autism] may complicate the interpretation of future and past research studies," the scientists conclude.

Another area of concern involves DSM's reliability, the degree to which independent clinicians agree on the diagnoses assigned to particular patients.

In the September 1988 HARVARD MEDICAL SCHOOL MENTAL HEALTH LETTER, two social workers maintain the reliability of DSM-III diagnoses was poorly documented. Furthermore, they contend, there was no attempt to improve the situation with DSM-III-R.

"Independent researchers have often found lower reliability scores than those reported in the [original DSM-III] field trials," write Herb Kutchins of California State University, Sacramento, and Stuart A. Kirk of the State University of New York at Albany.

The field trials, sponsored by the National Institute of Mental Health, involved the use of DSM-III diagnoses by volunteer clinicians with 12,667 psychiatric patients. Each member of a smaller group of about 900 patients was seen by two clinicians in an effort to estimate the reliability of various diagnoses.

These diagnostic duos often consisted of close colleagues who discussed clinical impressions with each other, a violation of experimental rules, Kutchins and Kirk contend. Even so, reliability was low for diagnoses in areas such as childhood disorders and personality disorders. The latter group of diagnoses is composed of personality traits, including paranoia and hostility, that create severe, long-standing problems in relations with others.

Questions remain about the reliability of specific personality disorders as well as some other diagnoses, respond Spitzer and Williams in the October 1988 HARVARD MEDICAL SCHOOL MENTAL HEALTH LETTER, but "the reliability of DSM-III is considerably improved over that of its predecessors."

Frances adds that DSM-IV developers will study closely the reliability of a number of diagnoses, such as generalized anxiety disorder, dysthymic disorder (mild depression) and some of the personality disorders.

But problems with diagnostic reliability often are not readily apparent, Zimmerman argues. For instance, he and his colleagues found significant differences in the interpretation of criteria for DSM-III melancholia by research groups at three universities. Members of the same team were trained to agree with one another, Zimmerman notes, but the research centers -- located in Iowa, Pennsylvania and Texas -- were far from unanimous as to which patients were melancholic.

If DSM-III and its progeny provide a common language for studying mental disorders, he says, "the this language is spoken with many different accents. I fear that if the revision process does not slow down soon, the accents will eventually differentiate into separate languages."

A separate language for mental disorders is indeed in the works. Officials at the American Psychological Association, unhappy with DSM changes and the procedures behind them, last year began work on their own diagnostic manual.

"We're not satisfied with either the reliability or validity of any of the diagnostic categories in DSM-III-R right now," says psychologist Lenore Walker, a private practitioner in Denver who was elected by the psychology association's board of trustees as one of the two liaisons to the committees working on DSM-IV.

DSM-III was largely accepted by the psychological community, Walker says, but questions about its scientific credentials emerged as planned changes for DSM-III-R became known. The American Psychological Association's initial resistance was to three disorders the revised edition now describes as "proposed diagnostic categories needing further study."

One diagnosis, late luteal phase dysphoric disorder (popularly known as premenstrual syndrome), attracted objections from a number of psychologists and psychiatrists, many of them women. Feminist psychotherapists were aghast and formed a lobbying group they dubbed the Coalition Against Misdiagnosis. In their view, a gynecological problem was being branded a mental illness.

The other two diagnoses were also opposed by the psychological association and feminist clinicians. Self-defeating personality disorder includes women who repeatedly enter into relationships with men who abuse them as well as men who persist in taking menial jobs far below their vocational training. Sadistic personality disorder often refers to rapists or men who are violent and abusive toward others. In other words, Walker says, abused women are labeled as mentally ill and their abusers are given a mental diagnosis with which to defend themselves from criminal prosecution.

Controversy over the three diagnoses led American Psychological Association officials to review the other DSM-III-R categories, resulting in "a sense of general alarm" concerning their usefulness, Walker notes. The alternative diagnostic manual, now in the drafting process, will drop many categories in the psychiatric version, she says, but there will probably not be dramatic revisions in the remaining diagnoses.

The alternative manual would not be necessary, according to Walker, if the psychiatric association were willing to work with the psychologists' association on an equal basis in designing the new diagnostic guidebook.

While there is no official American Psychological Association involvement in DSM-IV, Frances says a number of psychologists are participating in work on the new manual.

But more disturbing than the psychologists' alternative manual to the America Psychiatric Association would be a widespread rejection of DSM-IV by front-line psychiatrists treating patients across the country.

"I'm concerned about a possible backlash from clinicians over a new DSM every six to seven years," Spitzer says. "But I think most of them see the changes in DSM-III-R as careful and responsible."

Zimmerman is not so sure. "I suspect clinicians will use the criteria with which they trained and they will not readily embrace new criteria," he asserts.

The only survey of how psychiatrists actually use a diagnostic manual appeared in the February 1986 AMERICAN JOURNAL OF PSYCHIATRY. V. Chowdary Jampala of Chicago Medical School and his colleagues mailed questionnaires concerning DSM-III to 1,000 psychiatrists. The majority of the 557 respondents approved of the manual in general, although 35 percent said they would stop using it if it were not required of them. Nearly half expressed doubt in the validity of DSM-III diagnostic criteria.

Zimmerman and Jampala are now conducting a similar questionnaire survey of 1,000 randomly selected psychiatrists to determine their attitudes toward DSM-III and its revised version, whether they prefer one manual over the other and if they think new guidebooks are coming out too fast.

"The issue is whether the diagnostic glass represented by DSM-III and DSM-III-R is half empty or half full," wrote Spitzer and Williams in the HARVARD MEDICAL SCHOOL MENTAL HEALTH LETTER last year. "We think it is best described as half full."

On the other hand, warn Zimmerman and his colleagues in the July 1988 ARCHIVES OF GENERAL PSYCHIATRY: "The Bible may tell us so, but the [diagnostic] criteria don't. They are better than what we had, but they are still a long way from perfect."