Psychiatry Remains a Science, Whether or Not You Like DSM5

By Ronald W. Pies, MD | February 25, 2010

Quick—which screening test or instrument has greater specificity for the target condition: the PSA (prostate specific antigen) test for prostate cancer, or the BSDS (Bipolar Spectrum Diagnostic Scale), for bipolar disorders?

Wait, one more question: Which physicians are more likely to agree with one another regarding a diagnosis: 2 radiologists inspecting a renal angiogram for arterial stenosis; or 2 psychiatrists using a structured interview to assess a patient for possible major depression?

Those of you who are used to “trick questions” on Board exams will probably not be surprised that the answers are, respectively, “The BSDS” (specificity 0.93 vs 0.33 for PSA); and “the 2 psychiatrists.” (kappa [inter-rater agreement]= 0.73 vs 0.43 for radiologists)^1-5 (Disclosure: The BSDS was devised by the present author, then refined and field-tested by Dr Nassir Ghaemi and colleagues; it is available free of charge on this website).

What does this very selective demonstration prove? Not much. Can we conclude that psychiatry is more “scientific” than urology or radiology? Hardly. The exercise was presented mainly to roil the waters surrounding those comfortable cynics who insist that psychiatry is not “scientific” or a “real science.” The numerous controversies surrounding the DSM5—covered at length in this journal—seem to have brought these critics of psychiatry out in force.

Whatever the failings of the DSM5, the notion that psychiatry is not a science is profoundly wrong. Proponents of that view persistently confuse science with “logical positivism”—-a largely discredited form of scientific fundamentalism⁶—and misunderstand the nature of the scientific enterprise. In fact, my gambit involving PSAs and angiograms was itself a bit misleading. To the extent we can identify the foundational principle of science, it has little to do with lab tests, and a great deal to do with the scientist’s mind-set and methodology.

To be sure, philosophers of science point out that there may be no single, valid definition of “science” or of the “scientific method.” As philosopher Samir Okasha puts it, “…science is a heterogeneous activity, encompassing a wide range of different disciplines and theories. It may be that they share some fixed set of features that define what it is to be a science, but it may not.”⁷

And so, to assert what science “is” or “is not” with great confidence—-or to declare categorically which medical specialties constitute “real science”—is to over-reach in one’s epistemology by a considerable stretch. This doesn’t mean that we are left utterly adrift, however, without even a notional definition of science. Recently, the British Science Council spent a full year developing a definition of “science.” Their work-product is succinct and yet radically insightful:
"Science is the pursuit of knowledge and understanding of the natural and social world following a systematic methodology based on evidence."⁸

Good heavens! No lab tests required for science? No MRIs? No demonstrations of cellular pathology? Why, if this barmy British Science Council has its way, fields as diverse as physics, meteorology, linguistics, and anthropology would qualify as sciences! And, yes, without question--so would psychiatry.

Let’s be clear: not all science is physical science. Although psychiatry is nowadays associated with “biological psychiatry”--with PET scans, MRIs, neurotransmitters and the like--the domain of psychiatry is broader, deeper, and more pluralistic. As my colleagues Nassir Ghaemi MD9 and Michael A. Schwartz MD¹⁰ have argued, psychiatry is fundamentally a science of meaning. Wiggins and Schwartz define “meanings” as “…mental processes and their intended objects.”¹⁰(p.49)

We acquire evidence of our patient’s mental processes through precisely the “systematic methodology” required by the Science Council’s definition: we take a personal and family history; we perform a mental status exam; we observe our patient’s facial expression, affect, mannerisms, and speech. And we ask countless questions of the patient, aimed at eliciting deeper levels of meaning within the felt experience of the patient’s world. In some instances, we supplement these “office based” methods with projective or neuropsychological testing. In selected cases, we ask the patient to complete screening questionnaires or (rarely) to participate in a structured clinical interview. And, consistent with our pluralistic model of “mind,” we order appropriate laboratory and somatic tests to rule out underlying medical or neurological disorders.

We then form hypotheses based on these methods,regarding the patient’s psychopathology, personality structure, and clinical diagnosis. We test these hypotheses against subsequent observations, and—if we detect inconsistencies—we revisit our initial formulation. What Okasha identifies as “some of the main features of scientific inquiry”⁷(p. 125)—induction, experimental testing, observation, theory construction—are all part of psychiatric methodology. In short, psychiatry is well within the orthodox definition of “science.”

Do the methods of the DSM5 conform to this paradigm? That is a more complicated question, since the DSM5 work groups do not obtain data in the direct, observational way clinicians do. However, if science is “the pursuit of knowledge and understanding of the natural and social world following a systematic methodology based on evidence", the DSM5 process is arguably working within the broad framework of science. Like all such endeavors, the DSM5 process is buffeted by external forces and pressures that may mar its objectivity and undermine its science. We shall have to wait and see. But whatever the merits or flaws of the DSM5, psychiatry as a profession remains a science—not a physical, but a human science, grounded in a pluralistic understanding of our patients’ “meanings.”

References
1. Ghaemi S, Miller CJ, Berv DA et al: Sensitivity and specificity of a new bipolar spectrum diagnostic scale. J Affect Disord. 2005 Feb;84(2-3):273-7.
2. Hoffman RM, Gilliland FD, Adams-Cameron M, et al: Prostate-specific antigen testing accuracy in community practice . BMC Fam Pract. 2002; 3: 19.
3. Schreij G, de Haan MW, Oei TK, et al. Interpretation of renal angiography by radiologists. J Hypertens 1999;17(12 Pt 1):1737–41.
4. Ruskin PE, Reed S, Kumar R et al. Reliability and acceptability of psychiatric diagnosis via telecommunication and audiovisual technology. Psychiatr Serv 1998;49:1086–8.
5. Pies R: How “Objective” Are Psychiatric Diagnoses? (Guess Again) PsychiatryMMC, 2007. Accessed at: http://www.psychiatrymmc.com/how-%E2%80%9Cobjective%E2%80%9D-are-psychiatric-diagnoses-guess-again/.
6. Hanfling O: Logical Positivism. In: Philosophy of science, logic, and mathematics in the twentieth century, by Stuart Shanker. Volume 9 of the Routledge History of Philosophy, New York, Routledge, 1996. pp. 193-213.
7. Okasha S: Philosophy of Science. Oxford University Press, 2002, pp. 16-17.
8. http://www.guardian.co.uk/science/blog/2009/mar/03/science-definition-council-francis-bacon
9. Ghaemi SN: The Concepts of Psychiatry, Baltimore, Johns Hopkins University Press, 2003, pp. 90-93.
10. Wiggins O.P., Schwartz MA: Is there a science of meaning? Integrative Psychiatry 1991;7:48-53.

by Richard A Engdahl | May 12, 2011 4:22 PM EDT

While I agree with what Dr. Pies has elegantly written , I wonder why it is necessary to give legs to a spurious set of ideas in the first place by even responding. Perceptions are reality to the perceiver and as Richard Feynman put it so well: "The first principle is that you must not fool yourself -- and you are the easiest person to fool.

by Ronald Pies | April 26, 2011 5:05 PM EDT

Polythetic Pluralism is Not Promiscuous Eclecticism!

I thank Dr. McLaren for his further comments, and I suspect that he and I would agree on more than one might guess from our exchanges here. We would agree, for example, that mental disorders are "real", and that they cause real pain and suffering. We would both agree that psychiatry does not have the properties of a "settled"or "mature" science, in the way that, say, thermodynamics is a mature science. Dr. McLaren describes psychiatry as a "protoscience", following his understanding of the influential philosopher of science, Thomas Kuhn. (In his book, Humanizing Psychiatrists (p. 103), Dr. McLaren describes psychiatry as a "failed science"). In contrast, I would characterize psychiatry as a "young" science which, nonetheless, utilizes many traditional scientific methods of empirical investigation and testing. (I believe that Dr. McLaren might be surprised at how fraught and controversial Kuhn's views remain, to this day, and would find the discussion of Kuhn in Prof. Samir Okasha's book, Philosophy of Science: A Very Short Introduction (2002), of great interest).

Dr. McLaren and I would no doubt agree on the desirability of a unified model of psychiatric illness (a term I prefer to "mental disorder"), but we differ on whether the existence of such a model is a prerequisite to psychiatry's status as a "science". I maintain that a unified model of mental illness is neither necessary nor sufficient for the pursuit of science-nor do I believe it is essential (though it may be helpful) for doing good clinical work with patients; that is, for relieving the pain and suffering all of us wish to relieve.

Dr. McLaren wants to situate my claims on the slippery slope of promiscuous "eclecticism", at the bottom of which lie such horrors as blood-letting and lobotomy. I would encourage Dr. McLaren to take a deep breath, and to rest assured that I am arguing nothing of the kind. I believe that his misapprehension of my views rests on a confusion of two superficially related but quite distinct concepts: promiscuous eclecticism, on the one hand, and what I would call "polythetic pluralism", on the other.

The distinction may be grasped by returning to Dr. McLaren's "smorgasbord" metaphor-or is it a "model"? The promiscuously eclectic eater grabs whatever foods he likes, without any attempt to distinguish the whole-grain bread from the cream-filled twinkies. The polythetic pluralist selects according to defined principles of health, nutrition, calories, and, yes, even flavor. A "polythetic" class is based on Wittgenstein's model of a "long rope twisted together out of many shorter fibers." No one fiber defines the rope, but many overlapping fibers do so. If readers want to gain a better understanding of polythetic pluralism in psychiatry, a good place to begin is with the psychiatrist and philosopher, Karl Jaspers (1883-1969). As my Tufts colleague, Dr. Nassir Ghaemi, explains:

"[Jaspers'] approach to spiritual and existential notions in his thinking built on, rather than negated, an appreciation for science. In psychiatry, his biologically-oriented views to many conditions are most clear, and the linkage he then makes with an existential appreciation of other psychiatric states shows us that Jaspers was neither a biological reductionist nor a phenomenological/hermeneutic radical. He was a pluralist, a thinker who held that different methods were needed in different settings, but he was not simply eclectic (another misconstrual), allowing for any and all methods in whatever circumstances."
--Existenz, Vol. 3, No. 2, Fall, 2008; link to
http://www.bu.edu/paideia/existenz/volumes/Vol.3-2Ghaemi.html

Dr. McLaren is clearly familiar with Jaspers, and writes of him with some approval in chapter 5 of his book, Humanizing Psychiatrists. Yet Dr. McLaren does not seem to find much value in the kind of pluralism Jaspers embraces, which Ghaemi characterizes as "biological existentialism."

By the way, Dr. Ghaemi-like Dr. McLaren-is a prominent critic of the "biopsychosocial model" (BPSM)-but not because it is pluralistic in the "Jaspersian" mode. Rather, Ghaemi criticizes the BPSM because it tends to be employed in a way that ignores the evidentiary basis for effective treatment of a particular condition. That is, some practitioners use the biopsychosocial model by throwing "a little of this and a little of that" at the patient's problem-a little medication, a little psychotherapy, etc.-without carefully evaluating what the specific condition calls for, based on carefully controlled studies.

This point speaks to Dr. McLaren's mischaracterization of my own view of psychiatric treatment, which is not founded on "good intentions" or on promiscuous "picking and choosing"; but rather, on a polythetic model that draws upon science, the humanities, and the best available evidence. We do not need a single, unified model of bipolar disorder to know that lithium is an effective treatment for this condition. We do not need a single, unified model of unipolar major depression to know that cognitive-behavioral therapy is often effective for it, though less so when there are pronounced melancholic and psychotic features. At the same time, we can understand some important aspects of our bipolar and unipolar depressed patients by understanding their psychological and spiritual "world view"-in the vocabulary of phenomenology, the structure and contents of the patient's experience [see Pies R: The Anatomy of Sorrow, PEHM
http://www.ncbi.nlm.nih.gov/pmc/articles/PMC2442112/ ].

Jaspers rightly regarded an exploration of phenomenology-the patient's way of "being in the world"-- as an empirical activity. It was part of his polythetic pluralism, and it is what helps make psychiatry not just humane, but also a humane science. -Ronald Pies MD

P.S. If Dr. McLaren wishes to have the last word in this stimulating debate, he may be my guest. I believe I have said enough!

by Niall McLaren | April 26, 2011 8:15 AM EDT

I would like to reply to each point in Dr Pies' response but that would take a long time. Suffice it to say that the outline he has given of a scientific approach is most assuredly not that of Thomas Kuhn. There is no doubt about Kuhn's position: the presence of competing models of explanation in any field, where each of them explains a small part of the observations but none of them explains all, and each of them floats freely in conceptual space, unrelated to the others, is the hallmark of protoscience. I do not know any serious philosopher of science who would disagree with that.
In brief, the approach he favours, in which a clinician can pick and choose from whatever approach he wishes, merely licenses whimsy. I addressed just this point in my 1996 paper, "The myth of eclecticism." The essence of the scientific method (as distinct from the model of a science) is that the practitioner is bound to a larger set of rules than personal choice. The rules tell us what we must accept as (probably) true, we are not permitted to treat the phenomena as a smorgasbord, picking and choosing the bits that suit us and disregarding the rest. So Dr Pies feels grief may contribute to depression? I know plenty of psychiatrists who would totally disagree with him. And that's just a simple case: what about anxiety contributing to depression? What about the relationship of personality factors to mental disorder? Psychiatry doesn't even have a model of personality, let alone a model of personality disorder, yet we expect courts and governments to take us seriously. Be assured that, to our detriment, increasingly, they do not.
Dr Pies ends his comments with the soothing balm of how hard we try to do the right thing by our patients. I have no doubt that Walter Freeman was trying to do the right thing, as were the physicians who bled Washington to death. Good intentions are no defence against the charge of intellectual ineptitude.
Each of his points is interesting but they have been more than adequately answered in the literature. Psychiatry is not a science. It is at best a protoscience. I think our patients deserve better.

by Ronald Pies | April 23, 2011 9:30 PM EDT

I would like to provide some very preliminary responses to Dr. McLaren's claim that
"Psychiatry is most definitely NOT a science. It fails the first requirement of any field claiming to be scientific, an articulated model of the subject matter. Accordingly, to qualify as science, psychiatry would need a model of mental disorder, which it doesn't have."

First, though, a few caveats. Dr. McLaren has written several very ambitious books-of which I have read only excerpts and reviews-on what he calls the "biocognitive model"of the mind; and, by extension, of so-called mental disorders. Dr. McLaren defines the biocognitive model (BCM) as essentially one in which mental disorder is seen as the outcome of disturbed psychological constructs in the setting of a physically normal brain. He characterizes the BCM as "...far and away the most complex and far-reaching model in the history of psychiatry" (p. 205
Humanizing Psychiatry: The Biocognitive Model).

I am not in a position to comment on the BCM as such, and the interested reader may find useful discussion on Wikipedia [http://en.wikipedia.org/wiki/Niall_McLaren#Part_I:_Psychiatry_in_Crisis:_Intellectual_Failure_in_the_Science_of_Mental_Disorder] and by going to Google Books. Certainly, whether one agrees with Dr. McLaren's model or not, one must credit him with thinking long and hard about issues in the philosophy of mind and the nature of psychiatric illness.

But as to his claim about what is or is not "scientific" or "science"-here I must respectfully disagree with Dr. McLaren. Let me preface my comments by stating that, when Dr. McLaren refers to "the first requirement" of science, I assume he is speaking conceptually and not chronologically. That is, I take him to mean that "an articulated model of the subject matter" is the most central or critical element of a science, and not the "first requirement" in the chronological development of a given science. Few historians of science would credit the latter view, since "models" in science typically develop quite late in the history of a particular discipline, and do not define the initial activities of the nascent or "young" science. For example,
the "planetary model" of the atom, developed by Rutherford, came only after many painstaking empirical observations of how various elements combine, carried out by Dalton, JL Proust, and others.

But even on a conceptual level, I believe it is misleading to assert that a discipline is not operating in a "scientific" manner unless and until it advances a univocal and well-articulated "model" of its subject matter. The seminal investigations of Lavoisier, Proust, and others could hardly be criticized as "unscientific", simply because they preceded a fully-articulated view of atomic structure. And few would argue that Lavoisier and Proust were not "scientists", working in the early science of chemistry, merely because an articulated atomic model of matter did not yet exist. In my view, contrary to Dr. McLaren's assertion, a "model" of one's subject matter is neither necessary nor sufficient to define either the scientific method, or, indeed, a scientific discipline. That said, Dr. McLaren and I would doubtless agree that a model (or perhaps several complementary models) is an important component of a "mature" science, such as mechanical engineering or microbiology.

In short, it is important to distinguish the features of a mature science from what I would call "the scientific attitude" and "the scientific method." Few would suggest that psychiatry is a "mature science"; but it is wrong, in my view, to claim that psychiatry is no science at all, or that it cannot claim to be "scientific." Furthermore, as one philosopher of science recently wrote me, "If you want to have a strict demarcation rule, requiring a model is not a bad proposal, but…the notion of a model is nearly as vague, shifting and vexed as the notion of a science."

As I noted in my original article, it is probably fallacious to claim that there is a single, "essential" definition of science or the scientific method, as Ludwig Wittgenstein taught us in his later work on language. With regard to a "family resemblance" (to use Wittgenstein's term), it is fair to say that the "family" of activities we call "scientific" are generally characterized by these properties: (1) the careful observation of natural phenomena (including those we would now call "mental" or "experiential"); (2) attempts to discover the causes of these phenomena; (3) a careful analysis of the putative causative agents; (4) development of a hypothesis to explain the phenomena using the putative causal principles; and (5) testing and verification of the hypothesis, and in some cases, attempting to disprove the hypothesis via experimentation. These features-first developed by Robert Grosseteste (1168-1253), have remained central to the scientific method for eight centuries, despite many controversies in the philosophy of science.
I believe that psychiatry easily fulfills these basic elements of the scientific method, notwithstanding the absence of a single, univocal "model" of either the mind, or of so-called mental disorders (what I would prefer to call, "brain-mediated disease"). Psychiatry is still a "young" science, to be sure, but it is indeed a science. It is also a human encounter of a more subjective nature, and-in its clinical application-also an "art".

In our everyday work as psychiatric clinicians-as distinct from the work of theoreticians-we actually bring to bear a number of partial "models" as ways of guiding our work. We view psychiatric illness as essentially "brain-based" or "brain-mediated", in the straightforward sense that without a brain, there would be no psychiatric illness (and indeed, no "person", so far as we are able to determine). This does not mean that our explanations are always based in neurophysiology or neurochemistry-but explanations should not be confused with ontological claims about the nature of "mind" or "matter", or about the nature of psychiatric illness.

That is to say: we can hypothesize that the brain is both necessary and sufficient for the existence of psychiatric illness, without asserting that we can (or should) "explain" psychiatric illness in purely neurological or neurochemical terms. By analogy: we can argue that "ice" is nothing over and above a configuration of hydrogen and oxygen molecules, without claiming that this constitutes an "explanation" of ice in meaningful, human terms. The latter would involve experiential terms like, "cold", "hard", "brittle", "wet to the touch when heated", etc. Thus, in psychiatry, while we may make ontological claims about the brain and its chemicals-and even argue that there are causal links between neurochemical abnormalities and psychiatric illness-we do not and should not "explain" so-called mental illnesses in purely neurochemical terms.

We do not, for example, say to a patient, "Mrs. Jones, you are depressed today because neurocircuits in your cortico-limbic regions are deficient in serotonergic function." Rather, our "explanations" are usually in terms that involve mental and psychosocial constructs and referents, such as "You are depressed because you miss your deceased daughter and have no close friends." There is nothing "unscientific" about either of these "models"-the neurochemical or the psychosocial. Each is based on observations of different levels of phenomena. Each may be subject to empirical testing. For example, if Mrs. Jones tells us, "I feel better since I made friends with my next door neighbor," that may count as evidence in support of our (partial) psychosocial model. If Mrs. Jones feels better after taking a SSRI, that may count as evidence in favor of our (partial) neurochemical model. Mrs. Jones, of course, couldn't care a whit about what "model" we bring to bear on her depression-she simply wants to feel better. And, indeed, in my view, our ultimate goal as physicians is to relieve suffering and incapacity in a humane way-not to provide a univocal and "articulated model" of how and why our methods produce such a desired outcome. Such a model-devoutly to wished, of course-is merely a means to the humane end we seek as psychiatric physicians: decreasing the amount of misery in the world, and enhancing the emotional life of our patients. --Ronald PiesMD

by Ronald Pies | April 14, 2011 1:12 PM EDT

I appreciate Dr. McLaren's spirited rejoinder, with which I respectfully--but emphatically--disagree. In a few days, I will try to provide a substantive response as to why "an articulated model of the subject matter"is not the essential or fundamental element of a "science", and how--in our everyday activity--we in psychiatry work with a multi-modal "model" of how human beings come to develop neurobehavioral (brain-mediated) disease. --Best regards, Ron Pies MD

P.S. None of what I have argued asserts that psychiatry is "purely" or "exclusively" a science. As Dr. McLaren and I would surely agree, psychiatric care is also a human encounter of a subjective and mysterious sort...and an "art"!

Article Comment Pages: 1 2 Next

Psychiatry Remains a Science, Whether or Not You Like DSM5

Join the Conversation