Kendler Keeping It Real…..

 

Kenneth Kendler, MD needs no introduction to anyone even vaguely familiar with the psychiatric literature.  If you need to do your own research his accomplishments and scientific papers are widely available on the Internet. This post is to focus on his recent commentary in JAMA psychiatry (1) over the issue of psychiatric diseases and whether or not they are brain diseases.  He starts out with a 1867 quote from Griesinger stating that the brain is the only logical origin for symptoms of insanity. His analysis is at the level of “pathological and physiological” factors.

He briefly reviews two common arguments about whether psychiatric disorders are brain diseases.  The first Cartesian dualism that a mind emerges from the brain and is not the same as a brain. Since a brain is necessary for all mental phenomenon there is no specific answer to the question about whether the phenomenon observed with psychiatric disorders are diseases.  The second common argument is that grossly detectable brain diseases (lesions at autopsy and sophisticated imaging)  eventually became the purview of neurology.  To complement Kendler’s commentary, I would add that this has never been strictly true since both overt lesions and physiological brain dysfunction has always been studied by psychiatrists.  It has been a common antipsychiatry argument advanced by Szasz and others based on 19^th century concepts.  Ron Pies (2) has recently commented that it involved a misunderstanding of Virchow’s work on pathophysiology.   

An indirect way this problem has been handled is to suggest that it has to do with the vague definitions of disease (3).  Without a clear definition, anyone can use their own to declare that psychiatric disorders are not diseases. That has been a common tactic used to declare that not only that mental illnesses are not diseases – because of the lack of clear gross pathology they do not exist.  Dealing with the problem at this rhetorical level has not been very successful largely due to the lack of interest in rhetoric on the part of medical professionals and constant repetition by the rhetoricians.

More practical philosophical attempts at disease definition like loss of function models seem to not have much traction. Munson and Resnick (4) proposed one of these models and also suggested that the loss of function is related to programming errors in biological processes.    

Kendler suggests a clear path that has appeal to anyone who has studied pathophysiology and treated illnesses without clear lesions or with lesions that had to be the end product of some unknown pathophysiology.  That group of people would be anyone who has done an internship or residency in any medical field.  Anyone with that experience has seen a wide array of medical conditions that are polygenic in nature and have either an unknown or highly speculative pathophysiology.

The suggested path is genetics-> pathophysiology or more broadly “genetics -> brain -> schizophrenia.”  Rather than bemoaning all of the failed GWAS studies and Decade of the Brain, Kendler cites “the most robust empirical findings in all of psychiatry—that genetic risk factors impact causally and substantially on liability to all major psychiatric disorders.”  More specifically he cites a 2022 report that shows that gene expression (as mRNA levels) of risk variants for schizophrenia were noted in the brain and no other tissue.  That brings the brain expression in his causal link into clear focus. 

At that point he hedges and suggests that this may not be robust enough to suggest that a brain disease is occurring. For me it is plenty.  He goes on to suggest that there are 5 advantages of this approach including data driven rather than metaphysical, bypasses the 19^th century need for gross lesions, fits with pluralism or multiple possible etiologies, can potentially provide information about other diseases affecting the brain, and avoids a hard line of demarcation between normal and disease at the physiological level.  The last point has been elaborated in the more recent past as quantitative versus qualitative diseases and the associated variants. 

On the limitation side – a genetics only approach is the main consideration.  The antipsychiatrists that he has alluded to may be realizing that they need to finally modify their 19^th century rhetoric and I have seen the equally absurd claims that there are no genetic effects for psychiatric disorders.  The difference is that Kendler is an expert in the area – so only the most dedicated post modernists will claim that they did their own research and came to a different conclusion.  He does see the innovation of being able to detect tissue levels effects of genetic variants as a good starting point.  The goal is to elaborate the functional networks affected by these variants, describe mechanisms at the molecular level, and how those mechanisms are affected by variants (5).

This is really an inspiring commentary at a time when it is getting more fashionable to attack basic science research in psychiatry. I saw a comment just last week about how biological psychiatry was a drain on mental health research.  And there are frequent comments about how there should be more psychosocial research, even though there is no clear evidence that is necessary.

As a clinical psychiatrist and a physician first, my observations have been that most people go to medical school to gain knowledge about the human body and how to treat, prevent and where possible cure diseases. Speculative pathophysiology and mechanisms are all part of that starting in the first two years of basic science course and extending to clinical rounds at bedside during residency.  Philosophy and endless arguments about the nature of disease or psyche is not.  Psychiatry has lost its way many times due to an inability to recognize and respond to rhetoric. Kendler’s solution to the question of whether mental disorders are brain diseases is an elegant one and it is consistent with the way physicians are trained.  It also establishes a boundary that some questions in psychiatry are not answerable by philosophy.

Finally, what is still lacking?  I think that ultimately, we want medicine and psychiatry to be part of a comprehensive view of human biology. We need more comprehensive theories about human biology and how things really work at the physiological and molecular level.  That knowledge is currently spotty across all specialties. Biology theory rather than biological psychiatry is really the goal here and we can use more input from theoretical biologists of all specialties.

George Dawson, MD, DFAPA

References:

1:  Kendler KS. Are Psychiatric Disorders Brain Diseases?-A New Look at an Old Question. JAMA Psychiatry. 2024 Feb 28. doi: 10.1001/jamapsychiatry.2024.0036. Epub ahead of print.

2:  Pies R.  Did Szasz Misunderstand Virchow’s Concept of disease? Psychiatric Times. Feb 21, 2024.  https://www.psychiatrictimes.com/view/did-szasz-misunderstand-virchow-s-concept-of-disease

3:  Pies RW, Dawson G.  Epistemic Humility in Psychiatry: Why We Need More Montaigne and Less Savonarola.  Psychiatric Times.  Oct 19, 2023.  https://www.psychiatrictimes.com/view/epistemic-humility-in-psychiatry

4:  Albert DA, Munson R, Resnik MD.  Reasoning in Medicine: An Introduction to Clinical Inference.  Baltimore, Maryland: The Johns Hopkins University Press, 1988: 150-180.

5:  van Dongen J, Slagboom PE, Draisma HH, Martin NG, Boomsma DI. The continuing value of twin studies in the omics era. Nat Rev Genet. 2012 Sep;13(9):640-53. https://doi:10.1038/nrg3243

.

Scientific realism versus instrumentalism

 

“A philosopher who is not trained in a scientific discipline and who fails to keep his scientific interests alive will inevitably bungle and stumble and mistake uncritical rough drafts for definitive knowledge.  Unless an idea is submitted to the coldly dispassionate test of scientific inquiry, it is rapidly consumed in the fire of emotions and passions, or else it withers into a dry and narrow fanaticism” 

Karl Jaspers

Way to Wisdom, p. 159

 

I encountered 2 philosophical ideas today that I did not have any disagreement with and decided I would take that as a sign and discuss them here on my blog. Readers may have noticed that I am very skeptical about what philosophy adds to psychiatry and have posted numerous examples over the years. Today the above quote was posted on Twitter and it is an idea that I completely agree with. One of the problems that I have with philosophy in psychiatry is that is generally written as a rhetorical attack on the field. That is easy to do when you control the premise and that premise is generally false. There are numerous examples but the most obvious one is defining what you consider a disease to be and then concluding that no mental illnesses are diseases. More specific examples are available such as using similar definitions for addiction and then concluding that addiction is not a disease. Many of the people posting these arguments are not scientists or clinicians but in some cases are clinicians who have yet to have ever seen the patient they are talking about or who have never seen a patient – period.  To a large extent I think this is what Jaspers is focused on in this quote. It comes from a chapter in the reference book that is labeled as “Appendix 1: Philosophy and Science” and states that it was reprinted by permission from the Partisan Review.

The second opinion piece was from Kenneth Kendler, MD a leading expert in psychiatry, genetics and psychiatric research. He also has written extensively on the evolution of psychiatric thought over the decades and how philosophy applies to psychiatry. I have probably read at least a hundred papers written by Dr. Kendler in the past 30 years – and that is a small number of the papers he has written. I have also read his book Why Does Psychiatry Need Philosophy wherein he and his co-authors are focused on issues of phenomenology, nosology, and the degree of explanation as the subtitle suggests.

In the opinion piece (3), he takes a brief look at the philosophy of science and how that applies to the DSM.  He describes the main philosophical divisions as scientific realism and instrumentalism as they apply to diagnoses.  With scientific realism whatever the diagnostic construct - it is accepted at face value.  They are thought to really exist that way. The best examples I can think of from the DSM are Dissociative Identity Disorder and Intermittent Explosive Disorder.  Both are highly problematic and yet – there they are in the DSM. Further - the people who believe they exist seem undeterred in their use and they seem to be just carried along with subsequent revisions of the DSM.  Instrumentalism on the other hand sees diagnostic constructs as a work in progress.  They are not accepted as ultimate diagnoses but are subject to the scientific process of validation by empirical evidence. A good example in that case would be schizophrenia subtypes.  Previous editions of the DSM had 5 subtypes of schizophrenia including paranoid, catatonic, disorganized, residual, and undifferentiated.  Psychiatrists treating acute schizophrenia noticed that the subtypes were not consistent over time and the same person could be diagnosed with different subtypes. On that basis, the DSM-5 revision dropped the subtype classification but unfortunately implemented a schizophrenia spectrum of disorders.  I think that applying a physical concept to heterogeneous group of biologically determined disorders is probably a step in the wrong direction and that an instrumentalist approach will eliminate spectrums in the future along with Dissociative Identity Disorder and Intermittent Explosive Disorder

Kendler goes on the discuss 5 arguments in favor of an instrumentalist approach.  Before I raise those points, why would everyone not be in favor of this approach? Certainly, math and science majors would. Even though you can specialize in physical and biological science before medical school and students at that level don’t get much explicit instruction in the history of science – it happens nonetheless. Most high school students in the US are exposed to Darwin, Lamarck, and the DNA double helix as sophomores in high school. Almost all of the main concepts in physics and chemistry include some discussion of innovation and how earlier theories were rejected. That approach is more notable in medical school where some of the timelines and necessary technology become clearer. All of that information greatly favors an instrumentalism over scientific realism.  Although psychiatry is a relatively new discipline, it is clear that diagnostic systems have been greatly modified over the past 100 years from the Unitary Psychosis model o the 19^th century.

His first argument - pessimistic induction highlights the history of changes in diagnoses in the past and suggests that we should expect the same pattern in the future. A counterargument is that significant refinement has occurred and we can expect fewer errors than in the past but at any rate these observations need to be made.  The second argument is that given the nature of descriptive diagnoses rather than direct test observation determining validity will always require an instrumentalist approach.  The third argument is that the uncertainty about two competing diagnoses can be empirical or conceptual. Kendler favors the latter and that means changing the construct. The fourth argument is that scientific advances in psychiatry cannot be predicted or anticipated and are potentially transformative.  An empirical approach is required to test the future approaches against the current approaches and make the indicated modifications.  The fifth argument reduces a reverence bias toward existing diagnoses, much like clinicians use the provisional diagnosis term to reflect diagnostic uncertainty. 

He touched on the problem of how polygenic heterogeneous disorders can lead to the philosophical problem of multiple realizability.  That is - multiple genotypes leading to the same phenotype and the implications that has for moving to a diagnostic system that includes etiology. He points out that psychiatric disorders have higher degrees of polygenicity.  He briefly alludes to the potential problem of scientists with epistemic privilege studying nosology and phenomenology – but concludes on a more positive note about current research methods not available to previous generations.

Al things considered this did not seem like a powerful argument for philosophy in psychiatry. The current arguments in favor of instrumentalism seem like the general process of science.  The exceptions mentioned for psychiatry did not seem that specific relative to other specialties diagnosing complex polygenic disorders. I really wonder who the psychiatrists are who accept DSM diagnoses at face value?  I don’t think that I have ever met one.

That leads to the question of whether there are philosophical approaches that might be useful to psychiatry.  In my research for this post – I did find a fairly interesting one called constructive empiricism by von Fraassen (7,8).  Simply defined constructive empiricism states:  “Science aims to give us theories which are empirically adequate; and acceptance of a theory involves as belief only that it is empirically adequate.”  This is a departure from scientific realism and the premise that science is giving us the truth and belief in the theory means believing it is true. There is debate regarding the empirical adequacy of a theory with the critics using circularity arguments and the defenders pointing out that it is determined by scientists for specific goals. Philosophical debates tend to be endless and there are seldom any clear answers.  To me constructive empiricism seems to be an accurate description of what happens in psychiatry both at the biological and phenomenological levels. It certainly applies to the diagnoses of schizophrenia spectrum as opposed to subtypes and Dissociative Identity Disorder used in the original paper and many other papers written by Kendler on the evolution of various diagnoses over time. 

The critics of psychiatry are another story. There are people in the world right now who attack the entire diagnostic system of psychiatry. They either don’t have alternatives or the suggested alternatives have not been widely validated or adopted. In some cases, the theoretical basis for their proposed system is highly questionable. These same critics always seen to caricature the diagnostic process as one that is based on neither scientific realism or instrumentalism. The best example I can think of is any paper written that characterizes the DSM as the Bible of psychiatry.  That speaks not only to a general level of ignorance about how it is regarded in the field but also the philosophical bias suggesting the approach to the DSM has been static and not based in reality - even though major disorders have been present for centuries.  The critics also have an associated lack of knowledge about the biological constraints – even as they are briefly outlined in this opinion piece.

These critics whether they are antipsychiatrists or not seem to believe that what is in the DSM is accepted as the truth based on blind belief by psychiatrists and there is no evidence that is true now or at any point in time.  The observable changing diagnostic criteria over time and teaching future generations about all of the constraints is the best way to address empirical adequacy.

         

George Dawson, MD, DFAPA

 

References:

1:  Jaspers K.  Way To Wisdom. Yale University Press, New Haven, 1951: p. 151.

2:  Kendler KS.  Why does psychiatry need philosophy? In: Kendler KS, Parnas J, eds.  Philosophical Issues In Psychiatry: Explanation, Phenomenology, Nosology. Baltimore, MD; The Johns Hopkins University Press, 2008: 1-16.

3:  Kendler KS. Potential Lessons for DSM From Contemporary Philosophy of Science. JAMA Psychiatry. 2022 Feb 1;79(2):99-100. doi: 10.1001/jamapsychiatry.2021.3559. PMID: 34878514.

4:  Chakravartty, Anjan, "Scientific Realism", The Stanford Encyclopedia of Philosophy (Summer 2017 Edition), Edward N. Zalta (ed.), https://plato.stanford.edu/archives/sum2017/entries/scientific-realism

5:  Eronen MI. Psychopathology and Truth: A Defense of Realism. J Med Philos. 2019 Jul 29;44(4):507-520. doi: 10.1093/jmp/jhz009. PMID: 31356663.

6:  Kendler KS. Toward a limited realism for psychiatric nosology based on the coherence theory of truth. Psychol Med. 2015 Apr;45(6):1115-8. doi: 10.1017/S0033291714002177. Epub 2014 Sep 2. PMID: 25181016.

7:  Monton, Bradley and Chad Mohler, Constructive Empiricism. The Stanford Encyclopedia of Philosophy (Summer 2021 Edition), Edward N. Zalta (ed.), https://plato.stanford.edu/archives/sum2021/entries/constructive-empiricism.

8:  von Fraassen, Bas. Constructive Empiricism Now. Philosophical Studies, 2001; 106: 151–170https://www.princeton.edu/~fraassen/abstract/docs-publd/CE_Now.pdf

Concept Credit:  Dr. Ahmed Samei Huda a colleague from the UK came up with the concept that the critics of psychiatry are functioning at the level of scientific realism when psychiatrists are not.  That occurred during a Twitter discussion. 

Supplementary 1:  In the philosophy world there are much more detailed and varying definitions of scientific realism (4) than what Kendler discusses in the opinion piece.  The most accessible article I could find on the subject is by Eronen (5) that is more or less a refutation of Kendler and Zachar’s position on scientific realism. I say more or less because the author takes various positions to illustrate that scientific realism is necessary or at the minimum his Kendler and Zachar’s position on scientific realism may be closer to his that not. What I like about the Eronen paper is that he uses very clear examples with clear diagnoses like anorexia nervosa to make his point. My longstanding arguments about the validity of major psychiatric diagnoses is that they have always been there and more than anything that has driven the need for psychiatry and psychiatric care.

Graphics Credit:

Karl Jaspers downloaded from WikiMedia Commons on 2/27/2022 per the following:

Universitätsbibliothek Heidelberg, CC BY-SA 4.0 <https://creativecommons.org/licenses/by-sa/4.0>, via Wikimedia Commons  at the following URL: https://commons.wikimedia.org/wiki/File:Karl_Jaspers_(HeidICON_33479).jpg

Indexing Versus Diagnosis - A Non-trivial Difference?

There was an interesting article written by Kenneth S. Kendler, MD in this month's American Journal of Psychiatry.  It addresses a phrase that I have seen in typed evaluations that causes me to cringe: "The patient does/does not meet criteria for major depression."  I was asked why that phrase bothered me so much and I basically pointed out that I don't care whether a person "meets criteria" for a specific DSM criteria - the overall assessment was more important to me.  Kendler describes the difference in terms of phenomenology - are there other aspects of depression that merit further description than what is in the DSM?  That seems true to me as well as, the time domain of symptoms development and how some of the critical symptoms may have developed.  On a developmental basis - sleep, anxiety, and depression all may have different time frame and given the length of time that the DSM approach has been around - there has been very little discussion of some of the key convergences and divergences.  Is primary insomnia beginning in middle school associated with depression in the twenties - the same problem as no insomnia in childhood and depression in the twenties.

To study the issue Kendler looks at 19 textbooks and 18 symptoms domains described in each of those textbooks.  He has specific criteria for textbook selection that resulted in 19 texts from 5 countries published between 1899 to 1956.  He included full criteria published by Wendell Muncie in 1939 and Aubrey Lewis in 1934 as being particularly instructive.  He looks at the number of authors describing a particular symptoms and additional descriptors of the symptoms they found in depressive states.  Just looking at neurovegetative states, the results are interesting.  The atypical depressive symptoms of hypersomnia and increased appetite and weight gain were essentially absent. Fourteen authors described sleep problems as initial insomnia or non restorative sleep.  Three authors described early morning awakening.  Poor appetite was listed by 10 authors and weight loss by 9 authors.  Anhedonia was listed by seven authors.  Kendler provides a detailed analysis of the remaining symptoms and how many of the authors consider these symptoms to be representative of depression.

One of the most instructive aspects of the paper was a direct comparison with DSM5 criteria across 18 symptoms, whether they are covered in the DSM and to what extent.  The symptoms not covered included volition/motivation, speech, other physical symptoms, anxiety, and depersonalization/derealization.  Experienced clinicians commonly encounter depressed people with all of these symptoms in everyday practice.  As an example, some of the most severely depressed people that I have treated were somatically focused to one degree or another on a continuum to the point of somatic delusions.  Adhering to DSM5 criteria would leave out the most important feature of their illness.  A more complete description of these symptoms allows for a better demarcation of the line between depression and psychotic depression - a critical line for developing a treatment plan.  Another critical aspect is the relationship between anxiety and depression.  The DSM tends to sacrifice a broader phenomenological approach for narrow, easily determined diagnostic markers.  Instead of describing the anxiety associated with depression, depressed patients often end up with additional diagnoses of anxiety disorders and in the DSM field trials it appears that anxiety and depression morph into on another and the criteria appear to have low diagnostic reliability in that context.

The broader concept from Kendler's philosophical perspective is whether meeting criteria is that same thing as having the disorder.  From a phenomenological perspective it is certainly possible to produce detailed analysis of patients that do not resemble one another in many regards.  Considering the fact that depression is always mapped onto unique conscious states that should not be too surprising.  Kendler's idea is that the DSM5 criteria do a "reasonable but incomplete job of assessing the prominent symptoms and signs of depression in the western post Kraepelinian tradition  ".  The idea of indexing cases of depression from what is not depression is relevant here.  I think that he should have been a little more specific in his criticism.  I don't think they do a reasonable job when they are not part of a psychiatric assessment by a psychiatrist who has enough time to do a good job.  Taking the DSM5 criteria and converting them into a checklist and applying that to the masses comes to mind.  This is probably the most absurd conclusion to the indexing concept, surpassed only by telling those who are screened that "you meet criteria for depression and that meets our health plan criteria to start citalopram".

 Kendler points out the consequences of reifying diagnostic criteria to the point that they become distinct disorders in the absence of any quantitative markers.  Andreassen made similar arguments about DSM technology and the death of an interest in psychopathology in a previous paper (2).  Both authors seem to miss the mark in terms of what is really missed here.  The diagnostic nosology has shifted from a relative simple mind based paradigm to one that purports to pick up extreme conditions at the fringe of human behavior.  The accuracy of those diagnoses is less as the described disorders get more common.  Human consciousness appears to be the critical variable here and there remain very few commentators on this issue.  Psychiatric disorders become very complex once the psychiatrist goes far beyond symptom lists and even personality disorders and what is commonly considered personality traits to recognizing that the person in front of you is a truly unique conscious state associated with a unique neurobiological state.

              

George Dawson, MD DFAPA



References:

1: Kendler KS. The Phenomenology of Major Depression and the Representativeness and Nature of DSM Criteria. Am J Psychiatry. 2016 Aug 1;173(8):771-80. doi: 10.1176/appi.ajp.2016.15121509. Epub 2016 May 3. PubMed PMID: 27138588.

2: Andreasen NC. DSM and the death of phenomenology in America: an example of unintended consequences. Schizophr Bull. 2007 Jan;33(1):108-12. Epub 2006 Dec 7. PubMed PMID: 17158191; PubMed Central PMCID: PMC2632284. (full text)


Attribution:

Quote at top is from reference 1 by Dr. Kendler.