Dump the DSM>!

(Please see Counterpoint article by Michael First, MD, and Robert L. Spitzer, MD)

The American Psychiatric Association's DSM diagnostic system has outlived its usefulness by about two decades. It should be abandoned, not revised. Its primary achievement was to force American psychiatrists to recognize that not all patients presenting with florid psychoses had schizophrenia. More generally, it aimed to force the idea of operationally defined syndromes down the throat of a profession that was still, in the 1970s, dominated by the vague and archaic concepts of psychoanalysis at its American 1950s worst.

These goals have long since been accomplished, and, like the preceding analytic vagueness, it is time for the arbitrary, legalistic symptom checklists of the DSM to go. (The lag time would be about the same, 20 years or so.) Let me say at the outset that I do not wish to disparage all the hard and well-intentioned labors of the various work groups that developed the different sections of these books in their several editions or to deny the enormous amount of information summarized therein. But again, the aggregate is an awkward, ponderous, off-putting beast that discredits and diminishes psychiatry and the insight of those who practice it.

Consider the fact that your clinical practice is governed by a diagnostic system that:

• is a laughingstock for the other medical specialties;
• requires continual apologies to primary care doctors, medical students, residents, and the occasional lawyer or judge;
• most of our thoughtful colleagues privately rail against;;
• insists upon rigid categories that often serve only to confuse and misinform patients and their clinical workers (sometimes abetted by televised drug advertising);
• is so intellectually incoherent as to raise eyebrows among the well-educated, critical thinkers in our own psychotherapy clientele;
• persuades the world at large that psychiatry no longer has anything of interest to say about the human condition.

If it were within your power to do so, wouldn't you get rid of this system?

Let us make a tour of some of the diagnostic categories we all use and abuse. Schizoaffective disorder comes immediately to mind. Some argue that schizoaffective disorder should be a rare diagnosis. As unhappy as I am with the DSM-IV description, they reach an opposite conclusion to my own. Rightly criticizing the arbitrariness of the current criteria for this disorder (two weeks of hallucinations or delusions in the absence of prominent mood symptoms, but prominent mood symptoms for a "substantial portion" of the illness), one authority has even speculated that the confusion about schizoaffective disorder deters medical students from entering psychiatry!

I rather think that it is today's incarnation of the DSM, the DSM-IV, that deters medical students from entering psychiatry. Strict diagnostic criteria (which some want to repair by making them still stricter) reflect neither biological nor clinical reality. These realities simply will not yield to anyone's desire for precision. On the research front, recent studies suggest that there is considerable overlap in the genetic vulnerability for schizophrenia and for bipolar disorders. What is the point of false precision when the genes themselves are imprecise?

And when I make my weekly visit to the revolving-door world of Mental Health Clinicland, where brief med checks are the usual context for my best efforts, schizoaffective disorder is among the most common diagnoses I make. One reason, of course, is that reliable histories of precisely defined symptom clusters and periods of time are usually impossible to obtain, while many of my patients have both psychotic and mood disorder features most of the time, fitting neither of the DSM-IV's clear bipolar or schizophrenic pictures.

Another reason is that the clinical syndromes we treat, whatever their genetic underpinnings, are themselves changing. If the reader will permit a brief digression, the schizoaffective debate reminds me of debates over whether black lung disease was a real illness when I was in general practice in East Kentucky years ago. Some physicians, particularly those hired by mining companies' liability insurers, held that it was simply chronic obstructive pulmonary disease (COPD). Indeed, most of the miners who suffered from it were also smokers. But the fact was that black lung behaved differently than typical COPD. At least in my care, it seemed to have a restrictive, as well as an obstructive, nature and needed more and earlier steroid treatment. These men had spent their lives breathing coal dust.

Today, our young bipolar patients are spending their lives ingesting antidepressants, cocaine, methamphetamine, methylenedioxymethamphetamine (MDMA) and hallucinogens over prolonged periods. These chemicals change their brains, just as coal dust changes lungs. Inter-episode recovery, a hallmark of classic bipolar disorder, becomes a thing of the past. Delusions, hallucinations and mood-cycling become entrenched, and antipsychotic maintenance essential. I could try to be strict and pile up two or three Axis I diagnoses to describe this entity, thereby confusing everyone else involved in the patient's care. Instead, and with no apology, I call it schizoaffective disorder, which is more easily explained to nonpsychiatrists as a nonhomogeneous in-between category.

One can leaf through the DSM-IV and find countless howlers and paradoxes, as I am sure many readers have already done. More important are major problems like the relationship of narrowly defined posttraumatic stress disorder to the commonly found clinical entity of "chronic complex PTSD," which includes mood-cycling, dissociative and/or psychotic features, and predictable personality and boundary derangements. Again, we need to pile up several Axis I and II labels to strictly diagnose--and obfuscate--a familiar clinical presentation.

Aside from conversion disorder, the somatoform disorders are a mess (four pain symptoms, two gastrointestinal symptoms and so on). Primary care physicians never use these diagnoses, instead sticking with the clinical presentations they see, such as fibromyalgia syndrome. Anxiety disorders are artificially separated from the mood and psychotic disorders with which they are usually intertwined, yielding the frequent question, "If I have an anxiety disorder, why are you treating me with an antidepressant?" How many cases of pure generalized anxiety disorder have you seen? Of isolated social phobia? And so on.

The personality disorders section, categorical as it is, has been very effective in stifling nascent psychodynamic thinking among our trainees. Many of these "disorders" are extreme forms of various dimensions of normal personality. Their very extremeness is often, indeed, state-dependent with regard to other psychiatric illness and general stress level. The truly dynamic aspects of some of the more interesting ones, such as narcissistic and borderline personalities, have been bled out of the DSM categories. For example, a "deflated" aspect of narcissism, which presents with outward "low self-esteem" around a grandiose core, exists in variable equilibrium with the outwardly grandiose aspect described in DSM-IV, wherein it is nowhere to be found (see my essay "The Endless Walk of the Fool" in The Thaw [The Analytic Press, 2002]). And although observing how these character-disordered patients respond in treatment is the sine qua non of psychodynamic diagnosis, fundamental descriptions of such responses, such as the discussions of borderline personality organization by Otto Kernberg, M.D., are omitted.

In fact, psychiatrists who specialize in any one of the major subject areas in the DSM-IV seem almost universally frustrated by "their" section of the book. This is equally true of analysts specializing in personality disorders, clinical researchers in major mental illness, traumatologists and neurobiological investigators. Overlapping dimensions, or spectra, of pathology much more accurately reflect clinical reality, whether we are talking about the narcissistic/borderline personality spectrum or the bipolar, schizophrenia, obsessive-compulsive or autistic spectra. When they can, biologically oriented researchers come up with their own criteria (like the negative/positive symptom clusters in schizophrenia), while psychodynamic and cognitive-behavioral writers put forward alternative ways to look at personality function in therapy settings. The DSM-IV's relationship to all this is as a Berlitz phrase book is to the Tower of Babel.

And then there are those horrible Axes. I suppose they were designed with good intentions. They made us think always of the medical picture (Axis III) and originally tried to save a place for psychodynamics (Axis II). Today, we all know to think of the medical situation, and the Axis I=biology/Axis II=psychology distinction is blurred: patients with schizotypy have the eye-tracking abnormalities of schizophrenia, people with chronic PTSD often have borderline psychodynamics.

The Global Assessment of Function (Axis V) is good for generating the following quip: "Subtract 20 if you want to get the treatment plan approved!" And Axis IV, "psychosocial problems," shows no appreciation of the varying symbolic import of life events. These subjective judgments are not part of a diagnosis. They belong in a good narrative note with specific examples and a little originality in the use of adjectives!

Granted, there are disclaimers in the current manual's introduction about not taking categories literally and not regarding DSM diagnoses as a complete understanding of the patient. But these muted caveats do not help to reverse the fact that the DSM, by its very existence in its present form, implicitly encourages the entire mental health care system to do those very things. It is a tacit endorsement of false precision and superficial literalism in psychiatric assessment.

Two questions arise from this situation, and the first is, Why? Why maintain and elaborate a diagnostic system that no one is happy with? A skeptic need look no further than the catalogues full of DSM-IV treatment guides and companions, DSM-IV-keyed textbook editions, DSM-IV software and the like that fill every psychiatrist's mailbox. The DSM-IV is a big moneymaker for the APA. Who dares practice--indeed who can practice--without the reigning bible close at hand? I maintain that the APA is holding back the development of the profession it represents by maintaining its income and its institutional hegemony over American mental health care with the DSM system.

A final question is: What do I suggest instead? My regular readers well know that I am better at tearing down than at building, but I will try.

Researchers, of course, will always need research diagnostic criteria and are good at coming up with them. These same researchers ought to also pay some attention to operationalizing and studying the real, dirty categories that real clinicians use. This is already happening to some extent, as the pitfalls of excluding most of the patients we treat from research studies have become apparent.

The IDC-9 or ICD-10 would provide a perfectly good alternative for billing and coding purposes, and perhaps the APA could let the World Health Organization take back the job of developing future code bibles, rather than duplicating the task. While we will always need an administrative diagnostic system, the APA's resources are better spent finding ways to get people excited about--and interested in--the rich scope of the field it represents, instead of endlessly rehashing an arid and intimidating set of menus.

As clinical psychiatrists communicating among ourselves and to other specialties and concerned parties, we need not take diagnostic categories literally. We can save the major valid diagnostic syndromes like paranoid schizophrenia, or panic disorder with agoraphobia, but append other features freely and, most important, change our basic diagnostic stance to a dimensional rather than a categorical one. Arbitrary checklists and time cutoffs ("more than two weeks," "less than six months") can be dispensed with in favor of our best global diagnostic impressions. The focus of psychiatric treatment should be a single diagnosis--a single person--in most cases, with no tiresome Axes involved. Impressions of personality contributions, "stressors" (how I hate the word--its generic tone invites us to leave the patient's story out!), relevant medical illness and so on can go back into a narrative note to be discussed in a nuanced way. Attempts to quantify functioning can be confined to research and otherwise left to lawyers, government agencies, insurance companies and the psychiatrists they employ.

Will this ever happen? As my French-Canadian grandmother used to say, "Don't hold your breath!"

(Interested readers are advised to go to the Web site by Paul McHugh, M.D., [www.hopkinsmedicine.org/jhhpsychiatry/perspec1.htm] and read his systematic and cogent 1992 discussion of these same issues. From his lofty position as chair of psychiatry at Johns Hopkins University, he has long advocated for change in our diagnostic system.)