Why DSM-III, IV, and 5 are Unscientific

October 14, 2013

If science is defined as some kind of systematic study of observed experience applied to hypotheses or theories, and then confirmation or refutation of those hypotheses or theories, followed by new hypotheses or theories that are further tested and refined by new observations – if this is the core of any scientific inquiry, I think that no objective observer can attribute the history of DSM-III, IV, and 5 to anything that approximates this process.

A Spirited Exchange about the Biology of Mental Disorders and DSM-5
 
James L. Knoll, IV, MD

The spirited exchange between Drs Ronald Pies and Nassir Ghaemi is, I contend, good for the field.  Rather than avoid the sensitive or uncertain issues, both psychiatrists turn to face them with their own thought provoking styles and piercing insight. Would that we had more such collegial exchanges in psychiatry. Drs Pies and Ghaemi model for us how to abandon our defensiveness in an objective, respectful and passionate discourse. 

As nobel laureate Erick Kandel points out, our field is highly complex, quite young and "our understanding of the biology of mental disorders has been slow in coming." But it is, in fact, slowly progressing. The question of how long it will take to gain a more precise understanding of the underpinnings of mental illness has no definite answer. Or as Dr Theodore Pearlman has stated - "is any psychiatrist's guess."

And so in this spirit of continued scholarly dialogue, I offer a well-reasoned, trenchant response to Dr Ghaemi by Dr Pearlman, which can be found at the end of Dr Ghaemi’s esssay.

 

Dr Ron Pies makes many insightful comments in the accompanying analysis. My view about why DSM revisions have been unscientific is based on concepts of science that are in agreement with much of what he describes. If science is defined as some kind of systematic study of observed experience applied to hypotheses or theories, and then confirmation or refutation of those hypotheses or theories, followed by new hypotheses or theories that are further tested and refined by new observations – if this is the core of any scientific inquiry, I think that no objective observer can attribute the history of DSM-III, IV, and 5 to anything that approximates this process.

Let’s review this history, so well documented now by historian Hannah Decker with archival, often unpublished evidence for DSM-III.1 In the 1970s, scientific studies that meet the above definitions were collected mainly by researchers centered at the Washington University of St. Louis, and about 2 dozen diagnoses were found to be definable based on such empirical evidence. These were published a few times, lastly in 1978, as the Research Diagnostic Criteria. Within 2 years, Robert Spitzer had taken those scientifically-based diagnoses as the basis for DSM-III, but, through an immense amount of political wheeling-and-dealing (documented in painful detail by Decker), he produces 292 diagnoses. Obviously, in 2 years, a huge amount of scientific research did not suddenly identify 270 new diagnoses. Fourteen years later, with DSM-IV, 365 diagnoses were produced, but the original 270 were little changed. Now, about 20 years later, we still have almost 400 diagnoses, with little change in the original 292 from the Groundhog Year of 1980. 

How were these other 200 to 300 diagnoses developed? Was it through a scientific process? As so well documented by Decker and historian Edward Shorter2 and others who observed the process, like Michael Alan Taylor,3 these other diagnoses were based almost entirely on the opinions and beliefs of leaders and interest groups in the psychiatric profession. Why do we have about 10 personality “disorders”? Because psychoanalysts believe in those ideas. Were those ideas tested with observational studies, and then revised based on confirmations and refutations of their content? Not before 1980, and hardly since. As an example, Taylor describes a DSM-IV conference on personality where a huge amount of scientific research was presented on personality traits, and then the DSM-IV leaders stated clearly that they would ignore that scientific evidence and would hardly change the DSM-III personality disorder definitions at all.3 Twenty years later, after literally thousands more studies with some of the best possible scientific evidence possible in experimental psychology, the DSM-5 task force had no choice but to admit the need to add personality dimensions to the nosology. It got all the way to the APA Board of Trustees, and within weeks of publication, was simply rejected tout court.

The original DSM-III personality disorders were almost completely based on psychoanalytic opinion, with hardly any scientific validity literature to support them, as documented well by Hannah in her archival research. In the intervening 30 years, a number of scientific validity studies (using the classic nosology validators of phenomnelogy, course, genetics, and biological markers) have invalidated most DSM-IV personality disorders; in other words, they have been falsified scientifically. In fact, this was the scientific conclusion of the personality disorder research summary provided for DSM-5 by the world’s most prominent personality researchers.4 And yet DSM personality disorders have remained little changed in DSM-5 by fiat. This is another proof of being unscientific: the DSM nosology refuses to accept the falsification of its cherished beliefs. 
 
This is the problem. It’s not complicated, and philosophically difficult. If you have opinion, and nothing else, it’s not science. If you refuse to change your opinions, it’s not science. Most of DSM has been based on opinion, and our profession has refused to change most of that opinion for 2 generations. How can anyone imagine that any profession would ever experience progress, much less scientific progress, if it refuses to change its opinions, themselves based on nothing but prior opinion?

We are much more ignorant than Hippocrates over 2 millennia ago. He knew that opinion breeds ignorance, while science is the father of knowledge. We mistake our opinions for science.

Dr Pies’ attempt to distinguish the terms “scientific” and “validity” is mistaken when applied to DSM, I think, partly because he uses the word “valid” in both English and scientific usage, which is not the same. “Validity” when applied to nosology means that a diagnosis is true; it exists; it is not wrong. Cancer is valid; drapetomania is not. This is different from the English usage, when Dr Pies says [in his footnote] that the steady-state theory of the universe was invalidated by the Big-Bang theory. In the latter example, physicists were testing hypotheses by evidence, and changing their theories based on those experimental results. That’s science.

In the DSM process, psychiatric leaders enforce their opinions, and then they refuse to change them at all based on any experimental research. That’s not science. There are some cases, much less common, where DSM changes have been based on scientific evidence. So my claim here is not absolute, but it does reflect the predominant approach in DSM changes for the 4th revision almost entirely, and for the 3rd and 5th revisions mostly.)  

Worse, the leader of DSM-IV is explicit that science is low in priority and should not be valued as the main method of revising diagnosis.5 Instead, he says, we should base revisions mainly on “pragmatism,” meaning what DSM leaders think is good for everyone else. (A direct quote: It seems clear to me that pragmatic concerns for patient welfare always trump ‘science,’ especially since the ‘science underpinning psychiatric diagnosis is so thin and subject to alternative interpretations.”)6

As usual, the disciples are far inferior to the master. Freud knew better when he wrote in Future of an Illusion: “No, our science is no illusion. But an illusion it would be to suppose that what science cannot give us we can get elsewhere.”7

Nosological validity means applying scientific methods of hypothesis and experiment, and – most importantly – changing theories and hypotheses based on observations and experiments. This is not how the original 292 DSM-III diagnoses came about, and it is certainly not how they’ve been handled for the last 3 decades. 

That’s why science and nosological validity (based on the 5 validators described by Dr Pies-see footnote) are synonymous in the field of psychiatric diagnosis. There’s no way to be scientific unless you use examine the evidence, which comes from those 5 areas.  That’s nosological validity; that’s science.

The issue is not whether scientific ideas are later “invalidated”; here Dr Pies is using the word in its English meaning, not in its technical meaning in the science of nosology. Validity in nosology means studies that use the 5 validators (or pathology); that is different from the use of the term in the general English meaning that something is tested and confirmed based on experiment. The KIND of experiment varies from science to science: in histology, validity is based on microscopes; in astrophysics, telescopes; in psychiatric diagnosis, the 5 validators. The use of the word “invalidated” is English, and different from the term “validity” in nosology.

If you divide the 2 dozen somewhat validated psychiatric diagnoses, by the 400 or so claimed by DSM-5, we can precisely state that DSM-5 is about 97% unscientific. We need to change the attitude of 2 generations in our profession: paying lip-service to science, and then ignoring it. 

References
1. Decker HS. The Making of DSM-III®: A Diagnostic Manual's Conquest of American Psychiatry. New York: Oxford University Press; 2013.
2. Shorter E. Before Prozac: The Troubled History of Mood Disorders in Psychiatry. New York: Oxford University Press; 2009.
3. Taylor MA. Hippocrates Cried: The Decline of American Psychiatry. New York: Oxford University Press; 2013.
 4. Widiger TA, Simonsen E, Krueger R, et al. Personality disorder research agenda for the DSM-V. J Pers Disord. 2005;19:315-338.
5. Frances A. DSM in Philosophyland: Curioser and Curioser. 2010; http://alien.dowling.edu/~cperring/aapp/bulletin.htm. Accessed November 8, 2010.
6. Frances A. DSM-5 and practical consequences. Psychology Today. December 23, 2010. http://www.psychologytoday.com/blog/dsm5-in-distress/201012/dsm-5-and-practical-consequences. Accessed November 9, 2013.
7. Freud S. Future of an Illusion. Vol 22. London: Hogarth Press; 1927.

In Defense of DSM

As an elder psychiatrist practicing before and after DSM-III became the standard diagnostic handbook for psychiatrists in the 1980s, I take issue with Dr Ghaemi’s position that DSM lacks scientific credibility. Dr Ghaemi wants psychiatry to confine its knowledge to diagnosing and treating only some dozen of the 265 DSM listed disorders-the conditions which he believes are the only true diagnosable psychiatric diseases. The rest Dr Ghaemi asserts, are essentially psychogenic entities, formulated through self-deception and agreed upon falsehoods by members of the DSM task forces in order to seemingly promote reliability but lacking true validity.

I believe that far from deviating from scientific truth, the task forces of DSM were spurred towards seeking a truthful classification of psychiatric illnesses for sound research and treatment purposes. Various criteria influenced this authenticity. Firstly, the DSM represented an atheoretical classification of mental disorders and avoided non-creditable speculation as to causation of psychiatric illness. Nowhere in the manual does the term disease associated with a functional psychiatric disorder appear; which would erroneously imply proven structural brain pathology. Furthermore, in accordance with the concept of true illness, each diagnosis required a level of malfunction causing clinically significant distress or impairment in social, occupational or other important areas of functioning.

Before the advent of DSM-III, clinical psychiatry was bogged down by unproductive emphasis and debate on what appeared biological and genetic, versus psychological reactions to adverse environmental or personality factors. Hence as Dr Ghaemi has inferred, Sir Martin Roth, past doyen of British psychiatry and knighted by Queen Elizabeth for his services to British psychiatry, asserted that endogenous depression [phased in DSM-III into the nomenclature of Major depression] was exclusively a biological disease. Reactive depression termed neurotic depression by Roth was considered exclusively psychogenic in origin and to be managed with psychotherapy alone.2

The problem with the Roth concept that endogenous depression is a structural brain disease is that the argument fails to mandate any proof of brain pathology. Instead there is speculation about the existence of innate brain molecular dysfunction which science thus far has failed to elucidate. Despite the unquestionable efficacy of antidepressants and ECT with endogenous depression, its pathology remains largely unknown. Symptom relief with SSRI and SNRI drugs or ECT does not directly imply an innate deficiency of brain amines, much as aspirin relieving fever in cases of strep throat does not imply deficiency of innate anti-fever substances in the body. How long psychiatry will take to discover the structural brain pathology of severe depressive illness, like bacteriology concerning the etiology of strep throat discovered centuries gone by, is any psychiatrist’s guess. In the meantime, while we wait for discovery of brain molecular pathology underlying endogenous depression, the DSM classification of depressive disorders on a continuum of diminishing severity namely from major to dysthymia, and both treatable with integrated biological and psychological treatment, remains psychiatry’s closest alignment to creditable scientific methodology and towards the ongoing search for the whole truth.

Theodore Pearlman, MD
Irvine, California
 

References
1. Diagnostic and Statistical Manual of Mental Disorders, Fourth Edition; American Psychiatric Association, Washington D.C. 1994.
2. Slater E. Roth M. Clinical Psychiatry, Bailliere Tindall and Cassell: London: 3rd Edition, p. 10-13 1969.