dreams of our fathers II…

Posted on Sunday 13 May 2012

After the DSM-III and later DSM-IIIR were launched and in play, people began to look back at the origins and ask if the actual Manuals had lived up to their grand design. One obvious area of criticism was the derivation of the criteria. About the process that lead to the Feighner Criteria, they said that they came from the evaluation of actual cases using:

  1. Clinical Description
  2. Laboratory Studies
  3. Delimitation from Other Disorders
  4. Follow-up Study
  5. Family Study
… a claim I would have no reason to doubt, except for:
While no psychiatric syndrome has yet been fully validated by a complete series of steps, a great deal of work has been published indicating that substantial validation is possible. This communication is a summary of that work in the form of specific diagnostic criteria.

In addition, we in this department have carried out a study of interrater reliability and validation of reliability with an 18-month follow-up study of 314 psychiatric emergency room patients (to be published) as well as a seven-year follow-up study of 87 psychiatric inpatients (to be published), each of whom was interviewed personally and systematically. There were four different raters in the emergency room study. Agreement ranged from 86% to 95% about diagnosis with diagnostic criteria similar to those outlined in this report. There were two different raters in the inpatient study ; reliability between those raters was 92%. In the emergency room study and in the inpatient study, validity, as determined by correctly predicting diagnosis at follow-up by criteria such as those of this report, was 93% and 92%, respectively.

And exclusivity was also mostly a wish:
It will be apparent below that certain diagnoses are mutually exclusive (primary affective disorders and schizophrenia), while others may be made in the same patient (antisocial personality disorder with alcoholism or drug dependency ; hysteria or anxiety neurosis with secondary depression). More work will be necessary before the full significance of various diagnostic combinations becomes evident.
So the criteria came from their actual patients, some of whom they followed and checked for inter-rater reliability and diagnostic stability. But later critics noted that the DSM-IIIs didn’t follow even that lead, relying primarily on expert opinion, the very thing they’d complained about in the DSM-I and DSM-II. During the lead-up to the DSM-IV a decade later, Dr. Spitzer responded to some of those criticisms.
An outsider-insider’s views about revising the DSMs
by Spitzer RL.
Journal of Abnormal Psychology. 1991 100[3]:294-296.


…In discussing the development of DSM-III, Widiger et al. noted the increasing role of empirical validation in psychiatry and the five phases for validating a psychiatric diagnosis proposed by Robins and Guze. Their approach led Robins and Guze to recognize only 16 diagnoses that they believed had been validated by follow-up and family studies [Feighner et al., 1972]. Clearly, if the DSM-III Task Force had adopted this strategy, as Widiger et al. implied, it would not have recommended that DSM-III include over 200 categories — most of which were included on the basis of expert clinical judgment [face validity] alone. The Task Force recognized, correctly I believe, that limiting DSM-III to only those categories that had been fully validated by empirical studies would be at the least a serious obstacle to the widespread use of the manual by mental health professionals. The approach that was adopted by the DSM-III Task Force, from Robins and Guze’s recommendations, was the use of specified diagnostic criteria for virtually all of the disorders—the major innovation of DSM-III.

Expert Consensus Versus Empirical Basis: It is understandable that Widiger et al. (1991) emphasized the many ways in which DSM-IV can improve on the process involved in the development of DSM-III and DSM-III-R, such as by systematically reviewing the relevant literature, documenting the rationale for all changes, and conducting many focused field trials. The DSM-IV leadership is to be congratulated for the tremendous effort that is involved in these projects. However, I am troubled by the tendency [intended or not] to play down the major role that expert consensus will have in the final decision-making process for DSM-IV. My own prediction is that when final decisions are made about DSM-IV, they will still be based primarily on expert consensus, rather than on data, as was the case with the DSM-III and DSM-III-R…

Dr. Spitzer gave a straightforward answer. The critics were right. They had primarily relied on expert opinion. And in spite of the DSM-IV Task Force’s attempts to get on a more empirical footing with literature reviews and Field Trials, Dr. Spitzer predicted that the DSM-IV would end up doing the same thing – relying on expert consensus. The part they’d taken from Robins and Guze, or the Feighner Criteria was the use of specified diagnostic criteria. The list, it seems, was dwindling out of the gate:
  1. no a priori principles
  2. descriptive criteria
  3. follow-up
  4. family studies
  5. exclusivity
  6. reliability
  7. undiagnosed psychiatric disorder
Now we come to no a priori principles. We all know that the predominance of psychoanalytic thinking in psychiatry was the problem being dealt with in those days. But critics raised the question about the DSM-III thinking that replaced it. Was it really etiologically neutral? Spitzer responded in this paper a decade later:

Are DSM-III and DSM-II1-R Atheoretical With Regard to Etiology? As is well known, the developers of DSM-III and DSM-IIIR claim that—with only a few exceptions, such as the organic mental disorders and adjustment disorder—the classification does not subscribe to any particular etiologic theories. For example, some investigators who have studied panic disorder believe that the disorder arises from learned avoidance responses to conditioned somatic symptoms of anxiety; other investigators believe the disorder results from a dysregulation of biological systems mediating separation anxiety. However, neither etiologic theory has any effect on the diagnostic criteria for the disorder, which are solely based on the descriptive features of the disorder. Therefore, I am puzzled by Millon’s (1991) statement that "despite assertions to the contrary, recent DSMs are a product of implicit causal or etiologic speculation"…
I accept that Dr. Spitzer believed what he said. However, he was not the only psychiatrist involved. There were others – lots of others. I was actually alive during most of this period. In the sixties, I was a medical student and later Internal Medicine resident in Memphis Tennessee. We all knew that the center of the biological psychiatry world was St. Louis. The reason we knew is that’s what we were told. In medical school, in the sparse behavioral science course, a lecturer drew a US map, and told us that the biological psychiatrists were in programs along the Mississippi River, putting a line around the center of the country that included St. Louis, Memphis, New Orleans. My friend Bill married a psychiatry resident who told us [every time she had too much wine]. Later, as a psychiatry resident in the 1970s, we heard the same thing. We had people in Atlanta trained in St. Louis, young psychiatrists on staff at the VAH mainly, who literally preached about the non-scientific-ness and non-medical-ness of psychoanalysis – talking only of biology and the neo-Kraepelinian creed, which is where I first heard it. They said all of these things, repetitively [with or without beverages]:
    1. Psychiatry is a branch of medicine.
    2. Psychiatry should utilize modern scientific methodologies and base its practice on scientific knowledge.
    3. Psychiatry treats people who are sick and who require treatment.
    4. There is a boundary between the normal and the sick.
    5. There are discrete mental illnesses.  They are not myths, and there are many of them.
    6. The focus of psychiatric physicians should be on the biological aspects of illness.
    7. There should be an explicit and intentional concern with diagnosis and classification.
    8. Diagnostic criteria should be codified, and a legitimate and valued area of research should be to validate them.
    9. Statistical techniques should be used to improve reliability and validity.
During the time when the DSM-IV Revision was coming to a close, the question of a priori principles in the DSM Revolution was still being debated:
On Values in Recent American Psychiatric Classification
by John Z. Sadler, Yosaf F. Huglus and George J. Agich
The Journal of Medicine and Philosophy 1994  19:261-277.

The DSM-IV, like its predecessors, will be a major influence on American psychiatry. As a consequence, continuing analysis of its assumptions is essential. Review of the manuals as well as conceptually-oriented literature on DSM-III, DSM-III-R, and DSM-IV reveals that the authors of these classifications have paid little attention to the explicit and implicit value commitments made by the classifications. The response to DSM criticisms and controversy has often been to incorporate more scientific diversity into the classification, instead of careful inquiry and assessment of the principal values that drive the nosologic process. Implications for psychiatric science and future DSM classifications are discussed.
It’s a long and deep article about a lot of things, but it gets around to the a priori principles point along the way:
One reason why value conflict is not seen as such by DSM-IV – and its predecessors, to be sure – is the apparent adherence of the Task Force to a particular view of classification and science. A large literature establishes that medical practice [and scientific practice as well], including classification, necessarily involves value commitments… For a psychiatric classification example, consider the above mentioned dispute over DSM-III/III-R’s descriptive terms. Psychodynamically oriented psychiatrists believe the nosology ignored theoretically important terms essential to their practice [such as "neurosis" and "defense mechanism"]. The syndromatic approach used by DSMIII/ III-R, however, met the descriptive needs of biological psychiatry much better… The implicit value choice made by the authors of DSM-III/III-R was that the biological descriptive approach was more important than the psychodynamic descriptive approach, presumably for a variety of reasons. We doubt that this preference for biological approaches that is implicit in the descriptive, syndromatic approach was consciously intended by the authors of the DSM-III and III-R. Instead, we suspect that this preference and its associated commitments were more the byproduct of a naive view of science and psychiatric nosology as value-free or value-neutral… Although the notion of value-free or value-neutral science has been discredited by a large number of authors, philosophers and scientists… the view nonetheless persists. The reasons for the rejection of value-neutral language are many, but can be summarily stated:
    Values, not cognitions, determine what we select as "important," "crucial," "central," "decisive," or "related." In other words, values lend structure to the field of attention, pre-defining background and foreground, and clustering disparate items into groups. Consequently, "descriptive" statements about psychopathology issue from presupposed value stances that conceal their own deeper sources, compatibilities, and incompatibilities…
It’s a complicated way of saying it, but it’s on point. Dr. Spitzer and his colleagues may not have wanted to make a choice between psychology and biology, but they for sure didn’t want to choose psyhology or psychoanalysis. So they chose the language of the biologists, and by doing so implicitly chose biology. These authors must’ve struck a nerve with Dr. Spitzer, because seven years later, he was still thinking about their article – some two decades after the publication of the DSM-III:
Values and Assumptions in the Development of DSM-III and DSM-III-R:
An Insider’s Perspective and a Belated Response to Sadler, Hulgus, and Agich’s “On Values in Recent American Psychiatric Classification”
by ROBERT L. SPITZER, M.D.
Journal of Nervous and Mental Disease. 2001 189:351–359.

…Let us broadly divide etiological perspectives into two major … groupings: according to the biological perspective, the causes of mental disorders will ultimately be shown to be disturbances in biological functioning that are relatively independent of life experience; according to the psychological perspective, the major causes of mental disorders will ultimately be shown to be disturbances in life experiences. The author challenges anyone to show how grouping disorders together on the basis of their shared descriptive features … inherently suggests favoring either perspective.

…I recall a psychoanalyst and chair of a DSM-III oversight committee who commenting on a draft of DSM-III said, “There is so much more that we know.” By this, he meant that DSM-III did not include all of the knowledge that his fellow clinicians had painstakingly learned about human behavior and motivation from the intensive study of patients in long-term psychotherapy. In a sense, the real controversy about DSM-III was a controversy about who were the leaders in the profession and whether progress in our field was most likely to come from empirical research studies or from clinical wisdom collected by intensive long-term psychotherapy. It is hard to see how the controversy would have been conducted at a higher level if the DSM-III committee had made any clearer their value commitments.

Sadler et al. are correct when they assert that basic values, assumptions, and commitments determine how developers of a classification system of mental disorders approach their difficult task. In this paper, we have presented those values, assumptions, and commitments, which were, for the most part, widely known and were contained in the ongoing DSM-III and DSM-IIII-R literature. It is not true that DSM-III and DSM-III-R gave greater emphasis to reliability than to validity, and it is not true that the DSM atheoretical approach with regard to etiology is implicitly biased toward a particular etiological perspective [organic or behavioral].

He stood by his claim of neutrality with the counter that the alternative couldn’t be proven, but added, "In a sense, the real controversy about DSM-III was a controversy about who were the leaders in the profession and whether progress in our field was most likely to come from empirical research studies or from clinical wisdom collected by intensive long-term psychotherapy," which was, of course, the real central question in his mind. So I’ll accept what Dr. Spitzer thinks about his own compromise. But as for Psychiatry as a whole, I buy the implicit choice argument. I was alive then too, and my immediate thought when I got around to reading the DSM-III was, "This is that St. Louis thing." I trust that thought. So from my perspective, revisiting the dream, we now have:
  1. no a priori principles
  2. descriptive criteria
  3. follow-up
  4. family studies
  5. exclusivity
  6. reliability
  7. undiagnosed psychiatric disorder
There were other criticisms, one of which is almost too big to even talk about – Validity. Were the disorders of the DSM-III and its followers valid? I’ll punt that one down the road for the moment…
  1.  
    Bernard Carroll
    May 13, 2012 | 4:30 PM
     

    Also lost in the ovations for DSM-III was the uncomfortable fact of diagnostic instability or inconsistency. DSM-III and its successors never confronted this issue, but the data are dismal. The St. Louis group itself documented that fewer than half of outpatients given diagnoses of primary unipolar depression received the same diagnosis on follow-up some years later. Here is their summary:
    Journal of Affective Disorders 1992 Oct;26(2):111-6.
    Unipolar depression: diagnostic inconsistency and its implications.
    Clayton PJ, Guze SB, Cloninger CR, Martin RL.
    Department of Psychiatry, University of Minnesota, Minneapolis 55455.
    AbstractMajor depressive disorder using Feighner et al. (Arch. Gen. Psychiatry 26, 57-63, 1972) and DSM-III or DSM-III-R criteria has proven to be a heterogeneous diagnosis. It apparently includes a wide variety of clinical conditions. This report, based upon the results of a multi-year blind follow-up of 500 randomly selected psychiatric outpatients focuses on certain problems associated with the diagnosis of primary unipolar affective disorders. At index, 141 patients received diagnoses of primary unipolar depression. At follow-up, only 62 (44%) of these received the same diagnosis, with an additional 14 (10%) receiving a diagnosis of undiagnosed: questionable primary unipolar depression, and 5 (4%) a diagnosis of bipolar disorder. Thus, about 43% received other diagnoses at follow-up: 35 (25%) diagnoses of secondary depression and 25 (18%) other diagnoses without indication of an affective component. Bipolar patients’ stability was significantly better for those who were manic at intake.
    PMID: 1447428 [PubMed – indexed for MEDLINE]

    The switches to bipolar disorder over time are understandable and acceptable, but what about all those other diagnostic changes? This degree of longitudinal instability makes a mockery of cross-sectional reliability studies and their associated Kappa results (such as these are). As for giving research studies a firm foundation… forget about it!

    As for ‘major depressive disorder,’ here we have an acknowledgment of the heterogeneity yet DSM-IV barreled ahead without any basic re-framing and, to the stupefaction of many, DSM-5 is making no basic changes, either. It looks like obtuse and self defeating perseveration to me.

Sorry, the comment form is closed at this time.