DSM5 -"HOW DOES PSYCHIATRY ESCAPE THE DSM5 'FLY BOTTLE' (CATCH ALL CATEGORIES) ?" - COURTESY OF DR RONALD PIES

The Ideal and the Real: How Does
Psychiatry
Escape The DSM-5
“Fly-bottle”?

Ronald Pies MD
SUNY Upstate Medical University; and Tufts University School of 
Medicine

        What is your aim in philosophy? To show the fly the way out of the fly-bottle.
                                            ...Ludwig Wittgenstein (Philosophical Investigations)

Introduction

    We all have our fantasies, and I confess that one of mine is a tad grandiose: I imagine being charged with creating a new system of psychiatric diagnosis, starting from scratch. My strong belief is that American psychiatry has become trapped, much like Wittgenstein’s famous fly, in a kind of conceptual fly-bottle, embodied in the DSM framework. Getting psychiatry out of that trap is the impetus behind my fantasy, and motivates part one (“The Ideal”) of this essay. On the other hand, I am realistic enough to know that we are likely to be saddled with some version of the present DSM framework for many years to come. With that in mind, I offer some specific comments on the DSM-5 framework, and on some of its prominent critics, in part two of this piece.

The Ideal
   
    So what’s my first move in creating a new diagnostic framework?  I would drop the “DSM” designation—where were all those “statistics” anyway?—and call my magnum opus the Manual of Brain-Mediated Disease, or MBMD.  I choose the term “brain-mediated disease” because I assume there is relatively little controversy in the claim that those conditions of greatest interest to psychiatrists are “mediated” by the brain. The concept of “mediation” gets me out of the fruitless Cartesian conundrum of “mind” versus “brain”, “mental” versus “physical” or somatic conditions, psychological vs. biological theories of disease, etc. My supposition is that—whatever the ultimate, ontological nature of what are now called “mental disorders”—the organ chiefly responsible for their manifestation is the brain, and not, say, the gallbladder.
    Technically speaking, the particular conditions listed in the MBMD would be considered “instantiations of disease”, rather than as “disease entities” or discrete “diseases.” I avoid the term “diseases” because this term carries with it the connotation of reified entities in the physical world, with the same ontological status as stones, trees, or sodium atoms.  I don’t believe disease entities sit in the same ontological category as these objects; rather, disease entities are essentially pragmatic constructs for making sense of human suffering.  In this regard, my view is closely related to that of Kendell and Jablensky, who wrote in 2003 that “…the mere fact that a diagnostic concept is listed in an official nomenclature and provided with a precise, complex definition tends to encourage this insidious reification.” [1]
    To be clear: my position does not deny that specific pathoanatomic lesions or pathophysiological dysfunctions may underlie many common instantiations of disease. But my position entails that such lesions or dysfunctions are not disease itself. The latter—derived from our subjective awareness of “dis-ease”—is a pre-scientific construct, available to men and women long before there were physicians, X-rays, or MRIs. (The English language once had the term “diseasy” to describe how people feel when experiencing dis-ease). The distinction I draw is one based on Virchow’s apparent distinction between Krankheiten (diseases) and die Krankheit (disease in general) [2]. I also draw on the important claim put forth by Kendell; namely, that “disease” is properly predicated of persons (“people”)—not of minds, brains, or bodies [3]. 
    My next move would be to separate clinical descriptions of disease —embodied in what I call “prototypes” [see below]--from research-oriented criteria aimed at fostering inter-rater reliability and uniform selection of research subjects. These research criteria would be relegated to one of the MBMD appendices, and would not be of primary interest to clinicians. The research-oriented criteria would be quite similar to those now used in the DSMs, and would consist of symptom checklists that “define” a particular condition, for research purposes. These criteria sets would indeed be “conservative”, as Dr. Frances would have it, to this extent: criteria would be altered from the previous DSM only if convincing scientific data supported the change—an issue to be determined by experts in research methodology, not just by experts in a particular area of study. By separating clinical from research-based descriptions, I try, in a rudimentary way, to deal with the important distinction Dr. James Phillips makes; i.e., “…the distinction between utility in practice and utility in research.” [4]. As Dr. Phillips rightly asks of the present DSM diagnostic criteria, “Who uses them?”
    The MBMD would be built upon six foundational principles, which I call the 6 “Ps”: privilege, prototypes, parsimony, pragmatism, pluralism, and phenomenology.  Roughly, these are defined as follows:    
    1. “Privilege” refers to strict limitations on what conditions are permitted into the diagnostic schema. If psychiatric nosology is conceived as a house with many rooms, only certain kinds of conditions would get through the front door; namely, only conditions that entail substantial and enduring suffering and incapacity. (To signal forward a bit, I believe my position is compatible with similar views advanced by Dr. Allen Frances). Furthermore, the “suffering” would need to be intrinsic to the condition, at least in part—not merely the result of societal disapproval, punishment, or “bad stuff” ensuing from a particular behavior [5].  In the Judaic tradition, the rabbis speak of tiruf hada’at—“mental anguish”. This is not a bad phenomenological starting point for the kinds of conditions we ought to let through our diagnostic “door” to disease. Accordingly, conditions like ego-syntonic “hypersexuality” or sociopathy—both presumably lacking intrinsic suffering—would not be admitted as instantiations of disease. This is not to say that such conditions should never be the focus of therapeutic intervention, or of social and legal regulation —but those thorny issues would take us far afield.
    2. “Prototypes” refers to the use of idealized models or archetypes of disease, rather than of “categorical” or “dimensional” methods of classification. These prototypes are similar to what Nassir Ghaemi has called “ideal types”, described as “…simplified version[s] of reality”[6]. The use of disease prototypes is one way to navigate around the Scylla of “nominalism” and the Charybdis of “realism” [7]. In a sense, prototypes are the diagnostic equivalent of “fuzzy logic”, and would make up the core of the new diagnostic descriptions. Each prototype would consist of a richly-detailed, generic, clinical case history, illustrating a particular clinical condition. The prototypes would be compatible with, but not identical to, the research-oriented criteria. Essentially, the research criteria would constitute a subset of features within the surrounding “fuzzy” construct of the prototype.  
    3. “Pragmatism” refers to the instrumental nature of the diagnostic schema; specifically, psychiatric diagnosis is seen fundamentally as a means to an ethical and humanitarian end—namely, the effective relief of certain kinds of human suffering and incapacity (“dis-ease”). Thus, my diagnostic categories would strive for “utility” in roughly the sense described by Kendell and Jablensky [1] . Specifically:  if the net amount of psychic misery in the world were reduced more by using prototype A than prototype B, prototype A gets priority, all other factors being equal. I do not use the term “pragmatic” in the debased sense that Nassir Ghaemi MD rightly castigates; i.e., as a hodge-podge of “…the beliefs of clinicians, the wishes of patients, our general ignorance about many scientific facts, the limitations of our treatments, [and] the needs of insurance reimbursement.” [8]
    4. “Parsimony” refers to the goal usually expressed in terms of Occam’s Razor; i.e., “entities should not be multiplied beyond what is necessary". Ideally, this principle would reduce the total number of psychiatric diagnoses, but without eliminating essential categories. However, I would not prejudice the scientific enterprise by aiming for a particular number of diagnostic categories, or even for the goal of increasing or decreasing the total number of such categories. 
    5.  “Pluralism” refers to the use of multiple types of evidence and levels of understanding, in answering Prof. Tim Thornton’s question; i.e., “What is it…for something to be a mental disorder?” [9] Or, in my terms, “What ought to count as an instantiation of brain-mediated (“psychiatric”) disease?” Pluralism allows for, but does not require, biologically-based criteria for specific instantiations of brain-mediated disease. Data on biological factors related to a particular condition would be appended to the basic prototype, as “Supporting Data.” Phenomenological data (see #6) would also “count” in identifying instantiations of brain-based disease.
    6. “Phenomenology” –i.e., the contents and structure of the patient’s felt experience—would be an important part of the prototypical descriptions in the MBMD, following the work of Husserl, Karl Jaspers, and various “existential” philosophers. Exemplary in this regard are Arieti’s classic descriptions of the inner world of patients with schizophrenia [10].

The Real: DSM-5 and its
Discontents: Areas of Agreement with Dr. Allen Frances

    I find many areas of agreement with Dr. Allen Frances’s positions, regarding the DSM-5.  For example, I agree with Dr. Frances that there is no “paradigm shift” involved in the ideas underlying the DSM-5. As several commentators in the AAPP Bulletin have already pointed out, Kuhn’s construct of the paradigm shift is in no way commensurate with the tactical tinkering proposed for the DSM-5, including the possible use of “dimensional” approaches to diagnosis.
     I also agree with Dr. Frances that there is a “threshold” problem with several of the diagnostic categories proposed for DSM-5; i.e., too many conditions that do not reach the level of “suffering and incapacity” seem to be getting through the door. Specifically, I agree with Dr. Frances that “… the difficulties people have in meeting society's expectations should not all be labeled as mental disorders…” [11], absent convincing evidence of intrinsic suffering and substantial incapacity; that is, the presence of dis-ease [5]. If society were suddenly to demand that we all function at the cognitive level of Steven Hawking, and began to label as “mentally disordered” those who couldn’t meet that expectation, something clearly would be amiss.  On the other hand, it is pragmatically necessary to specify certain (relatively) culture-neutral, rudimenatry abilities as defining the lower limits of functional capacity; e.g., the ability to get out of bed, feed oneself, maintain basic self-care, and perform certain essential cognitive tasks, such as remembering to turn off the stove. 

Consequentialism in the DSM-5

    As Dr. Alan Stone has noted (personal communication 12/24/09), medical ethics partakes of both deontological (duty-based) and consequentialist (outcome-based) elements.  Dr. Frances espouses a largely “consequentialist” ethos, in his urging that psychiatrists consider the possible adverse effects that may flow from changes in the diagnostic schema.  Thus, in a recent blog, he writes: “Much has been written about the "validators" of psychiatric diagnosis and how they should influence DSM...To my mind, by far the most important validator is how will any decision help or harm patient care, given the forseeable circumstances under which it will be used.” (italics added) [12]
     Indeed, on a purely ethical level, I agree with Dr. Frances’s position. By way of analogy: in the Talmud, the principle of pikuach nefesh (“danger to life”) overrides all religious laws except those involving murder, idolatry, and prohibited sexual unions. [13]. In psychiatric nosology, I would argue that demonstrable “danger to life” (e.g., “harm to patients”) should also override virtually all other concerns, if there is strong empirical evidence of such a danger.  For example,  if we had  well-founded, empirical evidence—say, from actual clinical experience in Europe—that criteria set A  for diagnosing ADHD inevitably leads to substantial harm to patients, whereas criteria set B does not, the latter ought to be favored, all other things being equal—even if criteria set A were directly linked with specific biomarkers, endophenotypes, etc.
    But it is a different matter when—despite persuasive scientific evidence to the contrary—we preemptively manipulate our diagnostic categories, in order to head off some anticipated form of substandard medical care.  In my view, this amounts to well-intentioned but misguided nosological gerrymandering. Thus, the fear that criteria set A will harm patients, based solely on hypothetical scenarios of slipshod medical practice, should not be sufficient to overcome criteria set A’s scientific superiority to criteria set B.   
    For this reason, I am very uneasy with Dr. Frances’s speculation that certain proposed changes in the DSM-IV criteria will lead to excessive prescribing of stimulants [11] or antidepressants [14].  In this regard, both Dr. Sidney Zisook and I differ with Dr. Frances on the issue of the “bereavement exclusion” for major depressive disorder, which now appears likely to be eliminated in DSM-5 [14, 15]. Zisook and I believe that concerns regarding overzealous prescribing practices are best addressed thorough careful drug labeling information; and especially, through intensive continuing medical education aimed at improving prescribing practices.  Indeed,  I am not aware of any medical specialty that routinely determines if condition X is, or is not, a “disease”;  or whether symptom X ought or ought not to be part of the criteria for a disease, based on anticipated problems with physicians’ prescribing habits or the marketing strategies of pharmaceutical companies.

What is Harm Avoidance?

Similarly, the issue of harm avoidance—often expressed by the maxim, “First, do no harm”—also needs careful analysis. As Dr. Ghaemi has noted in his discussion of the Hippocratic philosophy, the complete quote attributed to Hippocrates is, “As to diseases, make a habit of two things —to help, or at least to do no harm." [16]  (italics added). Note that the first Hippocratic principle entails actively helping the patient. Ghaemi goes on to observe that, for Hippocrates, “ethics grows out of science”, which for Hippocrates meant knowledge of disease. “The ethical principle, standing by itself, is not at all what Hippocrates taught.” [16] 
    Thus, Hippocratic “harm-avoidance” begins with good science: i.e., deciding whether or not the patient is actually suffering from disease; if so, determining what kind of disease; and then offering the appropriate treatment. Hippocratic harm-avoidance is not a solicitation to “gaming the system” by pre-arranging our disease criteria so as to avoid certain feared socioeconomic or behavioral outcomes.
    To be sure, Dr. Frances rightly calls attention to the risk of overmedication that might very well accompany “over-diagnosis” [11, 14]. But this is merely one element of an overall, clinical risk-assessment. The inherent morbidity and mortality of a disorder must also be weighed in the balance. For example:  Dr. Frances maintains that reducing the number of days required for the diagnosis of hypomania (from the current 4 to 2 days) will result in over-diagnosis of bipolar disorder; over-prescription of atypical antipsychotic drugs; and adverse medical outcomes, such as metabolic syndrome[17]. This putative pathway to “harm” might or might not come to pass. But assuming, as Dr. Ghaemi argues, that the 2-day hypomania criterion is scientifically well-founded, its use could lead to robust societal “goods” that may well overcome the putative harm envisioned by Dr. Frances. 
    For example, the suicide rate in bipolar disorder is roughly 15-20 times that of the general population [18]. It is quite possible that by reducing the hypomania criteria from 4 to 2 days, we would vastly increase bipolar patients’ access to lithium, which appears to reduce suicide risk substantially in bipolar populations [19].   To be sure, this is a pharmaco-epidemiological prediction that depends in part on the behavior—and proper education—of physicians.  We would need several careful, medical-epidemiological studies, analyzing, say, the risk of antipsychotic-induced metabolic syndrome compared with the risk of missed bipolar disorder and subsequent suicide. My point is that merely positing iatrogenic harm to patients should not, by itself, overcome the scientific merits of refining our criteria for hypomania.  Only well-documented, demonstrable harm to the public interest should “trump” scientific data showing that a change in criteria is justified. As Drs. Waterman and Curley persuasively argue, we must consider “…the negative consequences of leaving largely unchanged a taxonomy we know to be inadequate at best and simply wrong at worst.” [20]
    As a very crude mathematical representation of these complex calculations, we can write:
J=    S1B
    S2R
    Where J is the justification for changing a specific criterion; S1 is the scientific evidence supporting the change; B is the known benefits of diagnosing and optimally treating the condition (e.g., reduced suicide rates, disability, etc.); S2 is the scientific evidence arguing against the change, and R is the known risk of over-diagnosis and excessive or inappropriate treatment.  Obviously, this theoretical calculation would be exceedingly difficult in practice. But it is equally evident that retaining a dubious DSM-IV diagnostic criterion requires much more justification than a mere prediction of iatrogenic harm to patients.

What is a “False Positive” in
Psychiatry?

    I also find Dr. Frances’s use of the term “false positive” [14] in the psychiatric context deeply problematic. Indeed, some psychiatrists have appropriated this term from the fields of pathology or infectious disease, without thinking through its epistemological meaning in psychiatry.  For example, when an infectious disease expert says, “False positive FTA-ABS tests [for syphilis] can occur in Lyme borreliosis,”[21]  the expert is invoking a very different epistemological “deep structure” than that invoked, implicitly, by Dr. Frances. Thus, when Dr. Frances expresses concern that eliminating the bereavement exclusion will increase the rate of “false positive” diagnoses of MDD [14], it is unclear to what veridical standard he is appealing. What is our laboratory “test” for a true positive in MDD? Until we have a widely agreed-upon criteria set for MDD, for which a specific biomarker or endophenotype has been identified, the term “false positive” is, at best, a wishful metaphor; and at worst, a kind of Rylean “category mistake.” The term misappropriates, from the epistemic structure of the physical sciences, a term that has little if any “physical” meaning in psychiatry—at least, in the sense that an infectious disease expert would use the term “false positive.” One hopes, of course, that this situation will change as neurobiological knowledge advances [22]. 
    That said, the term “false positive” can have a coherent conceptual meaning in psychiatry, when, for example, it describes a problem with the application of specific diagnostic criteria.  For example, the statement, “Failure to recognize cocaine intoxication can lead to a false positive diagnosis of DSM-IV-defined mania” is at least conceptually sound, since we can point to cocaine as a confounding factor in applying our present criteria for the diagnosis of mania. 
    Finally, it is fallacious to assume, following a change in condition X’s diagnostic criteria, that an increase in the total number of cases of X necessarily indicates an increase in “false positives” for X.  To assert this is merely to beg the question of what constitutes a false positive; indeed, absent a veridical biomarker for condition X, the increase in cases may simply reflect a legitimate increase in the identification of the illness. Thus, if elimination of the “bereavement exclusion” leads to an increase in the total number of diagnosed cases of major depressive disorder, this does not necessarily point to an increase in “false positives”[15].

Conclusion

    I have tried to look at the problems of DSM-5 from the standpoints of the “ideal” and the “real”. Ideally, in my view, the field of psychiatry would scrap the basic “categorical” structure of the DSMs, and start from scratch. A new diagnostic system, in my view, would be guided by the six basic principles enumerated earlier:  privilege, prototypes, parsimony, pragmatism, pluralism, and phenomenology.  The foundational principle of this idealized system is that only those conditions that entail substantial, intrinsic suffering and incapacity “count” as instantiations of disease. Disease descriptions would consist of “ideal types” (prototypes)—not symptom check-lists specifying necessary and sufficient conditions or “essential definitions.”
    In reality, we are probably stuck with tinkering around the edges of our present, seriously flawed diagnostic system. Even so, our tinkering should at least be guided by the best available science. Absent convincing empirical evidence that a change in criteria will harm those we treat, our diagnostic criteria should follow the principle, “Go where the best science leads you.” That direction just might lead psychiatry out of the diagnostic fly-bottle!

References

    1. Kendell RE, Jablensky A: Distinguishing Between the Validity and Utility of Psychiatric Diagnoses. Am J Psychiatry 2003; 160:4-12
    2. Pies R: On myths and countermyths: more on Szaszian fallacies. Arch Gen Psychiatry 1979; 36:139-144.
    3. Kendell RE :The myth of mental illness. In: Szasz Under Fire. Edited by JA Schaler. Chicago, Open Court, 2004, pp. 29-48
    4. Phillips J: The muddle that is DSM-V. AAPP Bulletin, 2010; 17:10-12.
    5. Pies R: What should count as a mental disorder in DSM-V? Psychiatric Times. April 14, 2009. Accessed at: www.searchmedica.com.
    6. Ghaemi SN: The Concepts of Psychiatry. Baltimore, Johns Hopkins, 2003, p. 179.
    7. Coppleston FC: Medieval Philosophy. New York, Harper Torchbooks, 1970.
    8. Ghaemi SN: The secret of DSM-IV: The danger for DSM-5 .How DSM-IV attitudes could harm DSM 5. Psychology Today blog, published on April 28, 2010 Accessed at:
http://www.psychologytoday.com/blog/mood-swings/201004/the-secret-dsm-iv-the-danger-dsm-5
    9. Thornton T: On wishing for a paradigm shift. AAPP Bulletin, 2010; 17: 17-19
    10. Arieti S: Interpretation of Schizophrenia. New York, Basic Books, 1974
    11. Frances A: DSM5 Plans To Loosen Criteria For Adult ADD. Psychiatric Times, April 20, 2010. Accessed at: http://www.psychiatrictimes.com/depression/content/article/10168/1556462
    12. Frances A: Should Practical Consequences Influence DSM5 Decisions?  Yes, of course. Psychology Today blog, April 27, 2010, Accessed at: http://www.psychologytoday.com/blog/dsm5-in-distress/201004/should-practical-consequences-influence-dsm5-decisions.
    13. Kottek S: The practice of medicine in the Bible and Talmud. In: Pioneers in Jewish Medical Ethics. Edited by F. Rosner. Northvale NJ, Jason Aronson, 1997, pp. 7-2
    14. Frances A:  How To Avoid Medicalizing Normal Grief In DSM5.  Psychiatric Times March 16, 2010. Accessed at:www.searchmedica.com.
    15. Pies R, Zisook S: Disentangling Grief and Depression: Rejoinder to Dr Frances. Psychiatric Times, March 16, 2010.  Accessed at: www.searchmedica.com
    16. Ghaemi S: First Do No Harm and DSM - Part I: An empty slogan? What does First Do No Harm mean?  Psychology Today blog, May 6, 2010. Accessed at: http://www.psychologytoday.com/blog/mood-swings/201005/first-do-no-harm-and-dsm-part-i-empty-slogan
    17. Frances A: Bipolar II Revisited-Always Take The Experts With A Grain Of Salt. Psychology Today blog, April 26, 2010. Accessed at: http://www.psychologytoday.com/blog/dsm5-in-distress/201004/bipolar-ii-revisited-always-take-the-experts-grain-salt
    18.  Pompili M, Rihmer Z, Innamorati M et al: Assessment and treatment of suicide risk in bipolar disorders. Expert Review of Neurotherapeutics  2009;  9: 109-136
    19. Tondo L, Baldessarini RJ.Long-term lithium treatment in the prevention of  suicidal behavior in bipolar disorder patients. Epidemiol Psichiatr  Soc. 2009;8:179-83.
    20. Waterman GS, Curley DP: Doing no harm: the case against conservatism. AAPP Bulletin, 2010; 17:19-20
    21. Smith JL, Crumpton BC, Hummer J.  The Bascom Palmer Eye Institute Lyme/syphilis survey. J Clin Neuroophthalmol. 1990 Dec;10(4):255-60.
    22. Maletic V, Raison CL Neurobiology of depression, fibromyalgia and neuropathic pain. Front Biosci. 2009; 14:5291-338.