An Exchange with Ruffalo & Pies: More on Diagnosis & Diagnostic Explanation

The following are comments from Mark L. Ruffalo, D.Psa., L.C.S.W and Ronald W. Pies, M.D. in response to my previous blogpost: "Can Symptoms Be Caused by Descriptive Syndromes? An Analysis".

As a reminder, my blogpost was written in the context of an article by Ruffalo and Pies ("What Is Meant by a Psychiatric Diagnosis?") that had been written in response to an article by Jonathan Shedler, Ph.D ("A Psychiatric Diagnosis Is Not a Disease"). Both these article were in Psychology Today.

I will post the comments from Ruffalo and Pies, followed by brief comments from me.

COMMENTS FROM MARK RUFFALO AND RONALD PIES

Dear Awais,

Thank you for the opportunity to respond to your thoughtful comments on our article in Psychology Today, and on the broader—and very complex—issues we are all raising.

Our position, roughly speaking, could be summed up in three basic principles:

1) Avoid definitional essentialism

2) Embrace clinical pragmatism; and

3) Respect ordinary language.

In turn, these principles lead us to three fundamental conclusions, which we flesh out in our article:

1) There is no “essential definition” (i.e., one specifying necessary and sufficient conditions) for terms like “disease”, though entities called by this name typically possess “family resemblances.”

(2) The most clinically relevant family resemblance among members of the class called “disease entities” is the presence of prolonged or substantial suffering (or distress) and incapacity (or impairment); and these issues constitute the central focus of clinical care and treatment.

(3) While the concept of “cause” and “causality” is complex—and, like “disease”, admits of no essential definition—it is quite consistent with ordinary language to say that at least some DSM diagnostic categories represent causes of a patient’s suffering and incapacity in the mental, psychological and behavioral realm. [https://www.iep.utm.edu/ord-lang/#H5 ]   

Any one of these principles or conclusions would merit a long paper, which would be far outside the bounds of a blog. Nevertheless, we can “unpack” our three principle conclusions and elaborate on some additional points that you raise in your own comments.

1) The concept of “disease”—its meaning, scope, definition, diagnosis and treatment—has been a source of controversy since the ancient Greek academies of Knidos and Kos competed with one another [https://link.springer.com/article/10.1007/s00381-010-1271-2 ]. We are not surprised the controversy continues to this day! Although some scholars proffer definitional distinctions among terms like “disease”, “disorder”, “illness”, “malady”, “morbus,” etc., these terms are actually used very loosely—if not promiscuously—in the medical literature, and in everyday medical practice. [https://www.psychiatrictimes.com/dsm-5/what-should-count-mental-disorder-dsm-v]. The commonly made binary distinction between “disease” (known pathophysiology) and “disorder” (unknown cause or pathophysiology) is of very limited clinical utility. Example: Kawasaki Disease is an inflammatory condition affecting children, whose cause/etiology remains unknown. [https://kidshealth.org/en/parents/kawasaki.html ]. But it is not called “Kawasaki disorder” on that basis. More importantly, many medical conditions do not lend themselves to the binary distinction between disease and disorder, given that knowledge of etiology and pathophysiology exists on a broad continuum.

2) Very few physicians contemplate their overcrowded waiting room and think to themselves, “Hmmm…I wonder which of these patients has a disease, and which has a syndrome, a disorder, a malady, or an illness?” The physician’s chief concern is with determining who is experiencing suffering and incapacity (in varying proportions); identifying a likely cause, whenever possible (it often isn’t!); and relieving the patient’s misery safely and effectively.

3) As you suggest in your comments, Awais, the term “cause” is used in various ways, depending on the context. And, as Ludwig Wittgenstein would remind us, the “meaning” of a word depends critically on how it is used, and for what purpose. [https://plato.stanford.edu/entries/wittgenstein/ ]  

We would distinguish at least three senses and contexts for the term “cause”:

a) etiopathological causation: this refers, ultimately, to the physical and physico-chemical mechanisms through which a disease process develops; e.g., the causal role of the tuberculosis bacillus in causing tuberculosis. In many ways, this is the “gold standard” to which medical science aspires, but which is often unrealized, particularly in psychiatric (and several neurological) disease entities.

(b) clinical causation: this refers to the clinician’s identification of a “good fit” between the patient’s presenting signs and symptoms, and a recognized clinical entity; i.e., the patient’s signs and symptoms “map onto” a particular syndrome, disease, disorder, etc. and not to some alternative condition (a critical point you make in your discussion). This condition may or may not be understood etiopathologically.^^ For example, the patient presents with severe, unilateral head pain accompanied by nausea and extreme sensitivity to light. The patient asks the physician, “What’s causing my problem, Doc?” The physician—having ruled out, e.g., a brain tumor—replies, “I believe you are having migraine headaches.” Interestingly, the word “migraine” is probably a corruption of the Latin, hemicranium, meaning, “half the skull”. The etiopathology of migraine—though clearer now than 100 years ago—is still not well or completely understood; i.e., “The exact cause of migraines is unknown…” [https://www.nhs.uk/conditions/migraine/causes/]

So, a “wiseass” (or very clever) patient might retort, “So all you’re telling me, Doc, is that my one-sided headaches are caused by pain on one side of my head!” Of course, there is much more to the neurologist’s understanding of migraine—e.g., what triggers it, its course and prognosis; genetic factors that make it more likely, how it responds to treatment, etc. These factors represent some of the “external validators” that help define disease categories.** The same type of validation applies to most of the major DSM diagnoses; e.g., schizophrenia, bipolar disorder, major depression, and yes—even generalized anxiety disorder [see, e.g., https://anxietyinstitute.com/what-we-treat/anxiety-disorders/generalized-anxiety/]

c) causation in ordinary language. Awais, you explicitly take this up in your discussion; i.e., you rightly note that, “In ordinary language, we sometimes use “cause”…as [meaning] something along the lines of “this is how to best make sense of it.” So: when we ask, “What was the cause of Smith’s panic attack?” it is perfectly understandable to reply, in ordinary language, “It turns out Smith has Panic Disorder.” This, of course, doesn’t mean that other, perhaps subsidiary or contributing causes can’t be posited or discerned; e.g., “Smith was under a lot of pressure at work,” or “Smith had just been evicted from his apartment.” We may even go so far as to posit unconscious causes that would lend themselves to psychoanalysis! But none of these additional causes impugns-- or renders in any sense tautological—our “ordinary language” claim that Smith’s problem is due to, caused by, or a consequence of his having Panic Disorder. That is, to use your formulation: his problem is “best made sense of” by this diagnosis.

Conclusion

Meanings (b) and (c) coalesce in this statement from the Anxiety Institute, which also happens to summarize our concept of “disease”:  

“Generalized Anxiety Disorder causes its sufferers great distress and trouble functioning in several different areas, such as school, work, at home with family, or in social contexts with friends.” [italics added] https://anxietyinstitute.com/what-we-treat/anxiety-disorders/generalized-anxiety/

Much more could be said, Awais, but we do not want to overstay our welcome on your blog site, or anesthetize our readers! We hope we have provided sufficient context here to render our article at least a bit more convincing. Thank you again for this opportunity.

Footnotes:

^^While etiopathologic diagnosis is usually the goal in medical diagnosis, there remain many conditions called “idiopathic” in general medicine; i.e., “We don’t know what causes this.” A good example is idiopathic facial paralysis [IFP] sometimes called Bell Palsy, for which there are many possible explanations, but no clear etiology. Nevertheless, physicians routinely cite IFP as “the most common cause of unilateral facial paralysis.” [emphasis added; see: https://emedicine.medscape.com/article/1146903-overview#a1]

**Many psychiatric diagnoses also embody predictive validity. Contrary to some claims, this does not entail tautological confirmation, such as when someone with schizophrenia—for which hallucinations are one of the diagnostic criteria—is found to have hallucinations two or three years after diagnosis. A diagnosis of schizophrenia also predicts, for example, a higher than expected likelihood of a dementia diagnosis, up to ten years later, even though dementia is not part of the diagnostic criteria for schizophrenia. See, e.g., https://ps.psychiatryonline.org/doi/10.1176/appi.ps.201900325

COMMENTS FROM AWAIS AFTAB

I would like to thank Mark and Ron for taking the time to respond to my blogpost and to further elaborate their point of view. I will keep these comments brief, largely because I don't have major disagreements with what they have outlined above. I do have some clarifications and something to add about why this debate is important. 

Mark and Ron talk about external validators and predictive validity of psychiatric diagnoses. I don't disagree with that. Major psychiatric diagnoses are more than mere labels. They contain explanatory content; they are not vacuous. Psychiatric diagnosis gives us important information about clinical features, clinical course, and response to treatments, even though this information may be general and not specific to the individual in front of us. 

They write: "at least some DSM diagnostic categories represent causes of a patient’s suffering and incapacity in the mental, psychological and behavioral realm". I don't have any disagreement with that, but I don't think seeing the diagnosis as a cause of patient's suffering and incapacity was ever the point of contention, either on my end or Dr Shedler's.

The issue that I was specifically looking at was whether symptoms that constitute the description of a descriptive syndrome can be said to be caused by that syndrome. My conclusion was that we can say so in instances where the diagnosis can also meaningfully refer to some underlying causal structure. This doesn't necessarily mean that we have to know the exact etiology of the syndrome; I think the more specific, the more distinct, the more fleshed out a construct is, the more meaningfully it refers to some underlying causal structure, even if that causal structure is unknown.   

Mark and Ron describe three senses and contexts for the term "cause": etiopathological, clinical, and ordinary language. What they refer to as "clinical causation" might also be described as a diagnostic explanation. In the context of my above discussion, I would say that the diagnosis can offer an explanation to the extent that the diagnostic construct is fleshed out.

Ken Kendler states that diagnostic criteria index the disorder, they do not constitute the disorder, i.e. a disorder is more than just the criteria.

Imagine that the information that constitutes the disorder construct is X

Imagine that the information that constitutes the diagnostic criteria is Y

Imagine that the information that constitutes the description of a single sign/symptom is Z

X is always greater than Y, and Y is always greater than Z, however, for a lot of diagnoses, X is not that much greater than Y, and Y is not that much greater than Z. For me, this is the situation for diagnoses such as obesity, essential hypertension, and generalized anxiety disorder (GAD). Essential hypertension is not that much greater than its diagnostic criteria, and the diagnostic criteria are not that much greater than the sign of high blood pressure. 

For some diagnoses, X is substantially greater than Y, and Y is substantially greater than Z. For me, this is the situation for diagnoses such as migraine, diabetes mellitus, schizophrenia, and bipolar I disorder, Migraine is a lot more than just a headache, schizophrenia is a lot more than just hallucinations. Is GAD a lot more than anxiety in the same way? I personally don't think so. I think it is comparatively more meaningful to say that headache is caused by migraine, or hallucinations by schizophrenia, because these constructs are more fleshed out, and thereby they offer more of an explanation.

Readers may reasonably ask: what's the point of this discussion? The reason I think it is important is because the clinical and ordinary language sense of causation can easily give us the illusion of scientific explanation. It seems to us as if we have explained something, the language we use gives us that impression, when in fact we have done little of the sort in the scientific sense.

When we say that Mr. Smith's anxiety is caused by GAD, yes, we have diagnosed Mr. Smith, and yes, by virtue of that diagnosis we have non-trivial information at our disposal regarding Mr. Smith's anxiety, yet we have not explained Mr. Smith's anxiety in any scientifically meaningful sense, any more than we have explained high blood pressure by diagnosing it as essential hypertension. That is why our language of causation is important: if we are not careful, our diagnoses can easily obfuscate more than they clarify, and end up as a way of hiding our ignorance.