Classifying Psychiatric Illness

[Alemo]

Hi all,

Random musings of a soon-to-be graduating resident. I'm wondering what everyone's thoughts are on disease classification and nosology.

Obviously, there is great (patient and Pharma) demand for new therapies for "existing" illnesses as defined in DSM-5, and a lot of financial and academic work turns on these therapies being investigated and clinically used. But the cynical part of me thinks, "psilocybin and ketamine and Zuranolone and Trintellix augmented with Caplyta and exercise and XYZCBT therapies aren't really gonna do ****. The effect sizes aren't gonna change."

How can you treat a disease if you don't understand the pathophysiology? Looking back, it's insane to me that anyone in 1998 thought "genetic personalized psychiatry and the monoamine hypothesis of depression" was true. One hundred billion neurons and multiple subregions communicating with cortical areas with near constant stimulation of the organ (i.e. being awake and conscious) compared to something like the kidney which: A) sees your blood 2) and makes urine....it's really damn complicated!!! How do we know if treatment resistance in depression isn't "actually this person has another illness for which all the MDD treatments aren't effective?" We don't. But we'll continue to lump them in with every other MDD patient, even those that have mild two week symptoms that remit spontaneously.

The brain is by far and away the most complicated organ in the human body. Producing mental life is vastly more involved than myocytes contracting to move a limb. So why does our nosology suck? Why are we basing diagnostic constructs of something so complicated on yes or no questions? Shouldn't more effort be put into understanding diagnostic constructs, as opposed to a multi-site RDBCT of the newest patentable serotonin modulator that will never separate from the other twenty similar agents?

So where is the active research on nosology? Everything I read is just boilerplate language about RDOC, or descriptions of network dysfunction that is so complicated as to mean nothing. In a May 2020 review article in AJP titled "Toward Circuit Mechanisms of Pathophysiology in Depression," the authors write "a meta-analysis found that across a variety of tasks and experimental paradigms, depression is consistently associated with hypoactivity in the dorsolateral PFC, superior temporal cortex, insula, and cerebellum, and hyperactivity in the thalamus, caudate, visual cortex, and ventrolateral and anterior PFC." We've all read hundreds of such articles that seem akin to a cardiologist saying, "We don't really know what's happening in the heart but CHF involves dysfunction in the right and left atria, the right and left ventricle, and the aortic valve." Ok great, the mitral and triscupid valves weren't mentioned. Did we really learn anything here? If information is a reduction in uncertainty, how much uncertainty did we lose?

We don't diagnose myocardial infarction with Levine's sign and "it feels like an elephant is sitting on my chest." Those things might increase the prior likelihood of such, but we ultimately investigate the organ itself. The phrase "I'm depressed" has lost all specificity.

Obviously bench science is not necessarily translational work. I get that. People much smarter than I are trying to understand this stuff.

I guess I hope for the day we can investigate someone's mental functioning. Like, find someone that looks to be sufferings from what every agrees is "depression," investigate the functioning of their limbic system, find biosignatures, look for the same biosignatures in other people, call this "Depression Type A," and rebuild this DSM from the ground up.

Would love to hear the group's thoughts.


Additionally, I'm trying to think about how this interest of mine links to forensics because that's where I'm headed after graduation.

---

Members don't see this ad.

 

[clausewitz2]

The long version of my answer to your question is the short course on philosophy of psychiatry I teach to residents once a year. The short version is that you really need to be looking to history and philosophy of medicine generally and psychiatry specifically to find really meaningful engagement with this question.

Off the top of my head two books that are a good jumping off place are Sadler's Values and Psychiatric Diagnosis and Zachar's A Metaphysics of Psychopathology. Something more recent that is not strictly mental health focused but applies to how we conceptualize medical diagnoses and judge intervention effectiveness more generally is Stegenga's Medical Nihilism.

Rachel Cooper's work might also interest you, particularly Psychiatry and Philosophy of Science and Diagnosing the DSM.

It might also just be helpful looking at a different nosological tradition, like phenomenological approaches (real phenomenology, not 'list of symptoms' phenomenology). There is a Maudsley Reader in Phenomenological Psychiatry but any of Josef Parnass' work on anomalous self-experience in schizophrenia might interest you.

Welcome to a really fascinating area of thought/research, fellow asker of basic but obnoxiously difficult to answer questions.

 

[Alemo]

Thanks, I will check those out. That course sounds really interesting too…

 

[Stagg737]

Long post incoming Broke this up into multiple posts to make my thesis easier to read : P

Love this thread topic and could pontificate with you about this for days. I agree with Clause, that a great place to start would be with looking into the h/o psychiatry and psychiatric diagnoses more, as the historical context of our diagnoses is more relevant to where we are than most other fields of medicine (other than maybe Rheum, genetics, or ID). Some baseline thoughts/answers to some of your questions:

How do we know if treatment resistance in depression isn't "actually this person has another illness for which all the MDD treatments aren't effective?" We don't. But we'll continue to lump them in with every other MDD patient, even those that have mild two week symptoms that remit spontaneously.

Of course we don't, this is the problem of any field that uses clinical diagnoses without confirmatory testing. A question I like to pose to younger residents is "How many different disorders are we actually diagnosing as schizophrenia?" Does someone who presents as flat/avolitional with a Cotard-like delusion likely have the same neurochemical problems that someone who is severely agitated and trying to kill all the other patients on the unit because God's voice is telling them that they're the archangel Michael sent to cleanse the Earth? No. Are they both often diagnosed as schizophrenic? Yes. Imo, with the info we currently have, this isn't completely inappropriate as antipsychotics may be significantly helpful for both of them and the initial steps of the algorithm have decent utilitarian value. A lot of what I enjoy in psych is once you're past those first few steps and have to start asking "why didn't that work?" A lot of times the answer may be "I don't know", but that vast expanse that still needs more research and explanation is another major reason why I chose psych. There's still so much more to learn.

 

[Stagg737]

So why does our nosology suck? Why are we basing diagnostic constructs of something so complicated on yes or no questions? Shouldn't more effort be put into understanding diagnostic constructs, as opposed to a multi-site RDBCT of the newest patentable serotonin modulator that will never separate from the other twenty similar agents?

The nosology sucks because we still have such a poor understanding of the neurochemical and biochemical mechanisms of the pathology and continue to rely on a clinical diagnosis model. Being able to use simple tools like y/n screens and evals which will have a high sensitivity for which our medications will be helpful to a large percentage screened positive is the simplest implementation of treatment, especially when many patients are reliant on PCPs or non-psychiatrists/psychologist to diagnose and initially treat them. We all know there's a huge difference between getting a hand-off for a patient that's just "patient depressed, started lexapro" vs. a strong formulation which helps provide nuances to the specific aspects of a patient's mental status. I look at the Y/N constructs as a simple way to quickly drop someone into the bucket that would most likely benefit them.

It's up to the individual psychiatrist whether they want to pull each patient out of the bucket and really examine and treat them or if they just want to throw the right med into the corresponding bucket. For those wanting to really dive into the underlying psychopathology, I don't think we use the simple Y/N model as much. When I see a new psychotic patient on the unit, I'm not just saying "psychotic, start SGA". I want to know the etiology of the psychosis. Is it a "pure" psychotic disorder? Manic-psychotic? Psychotic depression? Is this severe neurosis presenting as an anxious psychosis or even OCD psychosis? Is there a component of BPD? Are there likely different physical pathologies for many of these patients? I think so, and figuring out that formulation in my mind helps me decide my treatment plans, even within the same class of medications at times. I don't go down the rabbit hole with every patient, but there have been plenty of times when it has really made a huge difference in the treatment plan and ultimately the patient outcome.

Yes, we absolutely should be trying to clarify and identify new diagnostic constructs. However, the RDBCTs will continue because SSRIs work for a lot of people and that's the easiest way to bring in the green.

 

[Stagg737]

So where is the active research on nosology? Everything I read is just boilerplate language about RDOC, or descriptions of network dysfunction that is so complicated as to mean nothing. In a May 2020 review article in AJP titled "Toward Circuit Mechanisms of Pathophysiology in Depression," the authors write "a meta-analysis found that across a variety of tasks and experimental paradigms, depression is consistently associated with hypoactivity in the dorsolateral PFC, superior temporal cortex, insula, and cerebellum, and hyperactivity in the thalamus, caudate, visual cortex, and ventrolateral and anterior PFC." We've all read hundreds of such articles that seem akin to a cardiologist saying, "We don't really know what's happening in the heart but CHF involves dysfunction in the right and left atria, the right and left ventricle, and the aortic valve." Ok great, the mitral and triscupid valves weren't mentioned. Did we really learn anything here? If information is a reduction in uncertainty, how much uncertainty did we lose?

There is some research out there trying to do this. HITOP model is one which is attempting to do this, however it continues to utilize clinical observation and appears to be mostly trying to rename and re-organize the diagnostic buckets. I have little faith that this will provide significant breakthroughs that you're referencing though. It looks to be more of a shift in focus on how we respond and treat various mental conditions as opposed to gaining an actual understanding of the physical workings of the mind and how neurobiochemical states result in various presentations.

My current research is on the alternative model of PDs and we are currently looking into the specific characteristics of certain disorders and attempting to identify clinical correlates between specific traits, measured level of personality functioning, and clinical screens and outcomes with various treatment forms. So less about the overarching diagnoses, but looking into specific traits (several of which can be associated with or attributed to certain brain structures and neuronal pathways) and pinning down their clinical relevance and treatment options.

There also has been a lot of research into understanding more specific aspects of psychopathology. There's constant and ongoing research with TMS and identifying specific targets (with some groups researching personalized targeting beyond just mapping). There's been a lot of studies attempting to identify specific genes and gene mutations which may pre-dispose patients or cause psychopathology. The problem is, most of this research doesn't pan out or the results are less than earth-shattering. There has been breakthroughs though. In the late 1960s limbic encephalitis was identified and around 2000 it was determined that there are autoimmune etiologies for it and that it is not just 2/2 cancer. A couple years later around 2005-07 anti-NMDA receptor encephalitis was identified which was pretty ground-breaking imo as we now have a known diagnosis which can be confirmed with testing and imaging for why some "schizophrenic"/"psychotic" patients get worse when dopamine antagonists are started.

I don't think that there is going to be a major revolution that shifts the whole field like germ-theory was for infectious disease for a few reasons. I think it's unrealistic to consistently objectify thoughts and the human thought process and emotional experience. We can look at fMRIs and hormonal levels, but short of a multi-million dollar work-up, I don't see this happening anytime soon and likely not in my lifetime. I also don't think that the plethora of diagnoses we consider "psychiatric" are really within the same area of medicine. The first thing we're taught to do is rule out "medical" problems, but if we believe in the neurobiological model (aka, every thought, idea, action, etc can be determined by chemical interactions) then everything is "medical". Psychiatric symptoms may involve almost any system as we classify them; neuro, endocrine/hormonal, rheumatology (inflammation), ID, genetics, GI, etc. If we could identify the true physiological mechanisms causing various symptoms or 'disorders', will anything really be considered psychiatry anymore, or would be just be absorbed by neuro into some subspecialty dealing less with gross structures and pathways and more of a biochemical/molecular approach?

I also like the question of reduction in uncertainty, as I think this is the most practical way to systematically move forward. Even small chunks of uncertainty removed, like anti-NMDA receptor encephalitis, gives us a better understanding of other mechanisms involved in pathology as well as what symptoms aren't accounted for by the identified pathology.

If we take a more psychologically non-deterministic approach, then I don't think we can ever truly have a comprehensive understanding of the suspected pathophysiology of any mental state as it would imply a non-physical component of some "diseased states". In which case we are back to figuring out what the physical role of the disease is and what portion is physical and potentially objectively measurable vs aspects attributable to transcendent thought processes.

 

[Stagg737]

Additionally, I'm trying to think about how this interest of mine links to forensics because that's where I'm headed after graduation.

Look up Kent Kiehl's work on fMRIs and psychopathy if you're not familiar with him. "The Psychopath Whisperer" is his commercial book and is an easy audiobook if you commute to get you started. He was a PhD fellow under Robert Hare at the University of BC and is currently a psychology professor at U of NM. He did a lot of work on brain imaging and EEGs in incarcerated psychopaths and came out with some of the earliest imaging data that showed that psychopaths do seem to have significant different brain activity than most "normal" individuals. He was also involved with an EEG project that found a unique "weird P wave" in individuals who were diagnosed as psychopaths. Obviously there's still a lot to do, but I've found his research to be some of the most interesting, especially in terms of neurobiological exploration of forensics populations.

 

[hallowmann]

Hi all,

Random musings of a soon-to-be graduating resident. I'm wondering what everyone's thoughts are on disease classification and nosology.

Obviously, there is great (patient and Pharma) demand for new therapies for "existing" illnesses as defined in DSM-5, and a lot of financial and academic work turns on these therapies being investigated and clinically used. But the cynical part of me thinks, "psilocybin and ketamine and Zuranolone and Trintellix augmented with Caplyta and exercise and XYZCBT therapies aren't really gonna do ****. The effect sizes aren't gonna change."

How can you treat a disease if you don't understand the pathophysiology? Looking back, it's insane to me that anyone in 1998 thought "genetic personalized psychiatry and the monoamine hypothesis of depression" was true. One hundred billion neurons and multiple subregions communicating with cortical areas with near constant stimulation of the organ (i.e. being awake and conscious) compared to something like the kidney which: A) sees your blood 2) and makes urine....it's really damn complicated!!! How do we know if treatment resistance in depression isn't "actually this person has another illness for which all the MDD treatments aren't effective?" We don't. But we'll continue to lump them in with every other MDD patient, even those that have mild two week symptoms that remit spontaneously.

The brain is by far and away the most complicated organ in the human body. Producing mental life is vastly more involved than myocytes contracting to move a limb. So why does our nosology suck? Why are we basing diagnostic constructs of something so complicated on yes or no questions? Shouldn't more effort be put into understanding diagnostic constructs, as opposed to a multi-site RDBCT of the newest patentable serotonin modulator that will never separate from the other twenty similar agents?

So where is the active research on nosology? Everything I read is just boilerplate language about RDOC, or descriptions of network dysfunction that is so complicated as to mean nothing. In a May 2020 review article in AJP titled "Toward Circuit Mechanisms of Pathophysiology in Depression," the authors write "a meta-analysis found that across a variety of tasks and experimental paradigms, depression is consistently associated with hypoactivity in the dorsolateral PFC, superior temporal cortex, insula, and cerebellum, and hyperactivity in the thalamus, caudate, visual cortex, and ventrolateral and anterior PFC." We've all read hundreds of such articles that seem akin to a cardiologist saying, "We don't really know what's happening in the heart but CHF involves dysfunction in the right and left atria, the right and left ventricle, and the aortic valve." Ok great, the mitral and triscupid valves weren't mentioned. Did we really learn anything here? If information is a reduction in uncertainty, how much uncertainty did we lose?

We don't diagnose myocardial infarction with Levine's sign and "it feels like an elephant is sitting on my chest." Those things might increase the prior likelihood of such, but we ultimately investigate the organ itself. The phrase "I'm depressed" has lost all specificity.

Obviously bench science is not necessarily translational work. I get that. People much smarter than I are trying to understand this stuff.

I guess I hope for the day we can investigate someone's mental functioning. Like, find someone that looks to be sufferings from what every agrees is "depression," investigate the functioning of their limbic system, find biosignatures, look for the same biosignatures in other people, call this "Depression Type A," and rebuild this DSM from the ground up.

Would love to hear the group's thoughts.


Additionally, I'm trying to think about how this interest of mine links to forensics because that's where I'm headed after graduation.

Interesting topic, the answer/solution of which I don't feel that much closer to in my fifth year in this field than I did in my first.

A few quick thoughts:

I think psychiatry keeps me humble with regards to how little we truly know in medicine, not just psychiatry and psychiatric illness. Literally in every field we struggle to understand even some basic aspects of the field and associated pathology. There are a lot of things we don't know that we just "accept" as the most plausible or most useful information necessary to help people. What the hell are antibody negative (presumably) autoimmune disorders like those of the biliary tree, what's antibody-negative NMSOD, or even what truly is Alzheimer's? How many conditions are just phenotypically similar presentations of completely different conditions? There is so little we know that its actually amazing we even ever feel confident in anything we do in medicine. But that confidence has a purpose, and that's to treat patients as best we can as individuals with slightly more knowledge about their health conditions.

In Psychiatry, we have a collection of diagnoses that all of us recognize as both simplistic clinical diagnoses and insufficient to properly characterize many (most?) individuals with psychiatric illness. That said, its hard to treat people if we struggle to even name or categorize them in some (clearly imperfect) way. Its hard to study treatments when we can't even attempt to sort them and define what "better" means. I recognize the necessity of what we do, and I guess I've accepted that I will likely die long before an amazing breakthrough is psychiatric understanding occurs, but I'll try to stay humble and keep trying to identify ways in which patient's differ and how that can inform their treatments.

Its also unfortunately a bookkeeping thing - basically insurance companies and laws that define allocation of resources explicitly require me to define some type of diagnosis, NOS or otherwise, so I oblige our evil overlords.

Also, your statement about "kidney sees blood and makes urine" is obviously meant to be simplistic, but I think its worth it to acknowledge that there is a ton of complexity to the way our kidney's function, and there are still a number of conditions associated with them that we really don't understand. The only way we've been able to understand it more has been through both clinical practice and significant advances in imaging. I know you were a bit tongue in cheek here, but my Nephrology colleagues would drop their salt shakers in bewilderment at such a simplification.

And just in case my pessimism regarding psychiatric advancement in the near future bothers you, keep in mind there's still a lot of advancement and diagnostic clarification happening even within our imperfect system now. I mean just the fact that as you said in 1998 the monoamine hypothesis of depression made sense to many, and now its clearly ridiculous in its level of simplicity is (hopefully) a testament to continued improvement in our understanding of psychiatric pathophysiology.

We all know there's a huge difference between getting a hand-off for a patient that's just "patient depressed, started lexapro" vs. a strong formulation which helps provide nuances to the specific aspects of a patient's mental status. I look at the Y/N constructs as a simple way to quickly drop someone into the bucket that would most likely benefit them.

It's up to the individual psychiatrist whether they want to pull each patient out of the bucket and really examine and treat them or if they just want to throw the right med into the corresponding bucket. For those wanting to really dive into the underlying psychopathology, I don't think we use the simple Y/N model as much. When I see a new psychotic patient on the unit, I'm not just saying "psychotic, start SGA". I want to know the etiology of the psychosis. Is it a "pure" psychotic disorder? Manic-psychotic? Psychotic depression? Is this severe neurosis presenting as an anxious psychosis or even OCD psychosis? Is there a component of BPD? Are there likely different physical pathologies for many of these patients? I think so, and figuring out that formulation in my mind helps me decide my treatment plans, even within the same class of medications at times. I don't go down the rabbit hole with every patient, but there have been plenty of times when it has really made a huge difference in the treatment plan and ultimately the patient outcome.

Yes, we absolutely should be trying to clarify and identify new diagnostic constructs. However, the RDBCTs will continue because SSRIs work for a lot of people and that's the easiest way to bring in the green.

I love this. These are exactly the thoughts that run through my head when making a formulation especially during intakes/diagnostics, and its one of my favorite parts of the field. This is what is often missed when I have too many patients to do good work, or when I get an intake written by a someone with less training that basically just lists the diagnoses attached to the encounter.

 

[Alemo]

No. Are they both often diagnosed as schizophrenic? Yes. Imo, with the info we currently have, this isn't completely inappropriate as antipsychotics may be significantly helpful for both of them and the initial steps of the algorithm have decent utilitarian value.

It's encouraging that these medications do typically offer some benefit, despite the obvious differences in presentations and crude diagnostic categories that subsume unique people. It makes you look back in horror about how it must have felt to treat these patients before the advent of neuroleptics. Descriptions like those provided by Kraepelin and Karl Leonhard are so invaluable as it sounds like patients were essentially a "captive audience" who could be observed (unfortunately likely "untreated" aside from basic psychotherapeutic environments).

I think so, and figuring out that formulation in my mind helps me decide my treatment plans, even within the same class of medications at times. I don't go down the rabbit hole with every patient, but there have been plenty of times when it has really made a huge difference in the treatment plan and ultimately the patient outcome.

Yes, we absolutely should be trying to clarify and identify new diagnostic constructs. However, the RDBCTs will continue because SSRIs work for a lot of people and that's the easiest way to bring in the green.

Good points. We have seen that perverse incentives (money, need for ease of use in research, "public health" a la putting Prozac in the water supply [also known as PCPs prescribing everyone everywhere SSRIs]) drive RCTs/me-too drug development/simplistic diagnostic schemes. My fear is that at some point, our understanding of mental illness will not only stagnate because of this, but actively damage progress ("it's not progress if you're going the wrong way").

It could be argued that there have been no real significant breakthroughs in understanding "depression" in the last forty years. At what point does using the ARTIFICIAL self-referential construct as first developed in DSM-III start to become synonymous in everyone's mind with REALITY? This is a huge problem. I completely agree with your point about necessity of formulations, because that is the only way to describe phenomena as they exist, not as (in the case of MDD) part of a 5/9 item checklist.

People like to say, "Sure we know the DSM is imprecise, these are just guides, blah blah blah." But the more we use it, the more reified it becomes. Maybe @clausewitz2 can chime in about how linguistic development leads to stereotyped thinking patterns. My point is we can't have entire generations of psychiatrists thinking this way. It's gets harder and harder for the field to say, "You know what, maybe we've been thinking about all this the wrong way." You know who doesn't say the DSM-5 is just a guide? The researchers publish in high-impact journals. They use operationalized guidelines (for obvious reasons) and contribute to the idea that "depression is what we say it is."

I think it's unrealistic to consistently objectify thoughts and the human thought process and emotional experience. We can look at fMRIs and hormonal levels, but short of a multi-million dollar work-up, I don't see this happening anytime soon and likely not in my lifetime. I also don't think that the plethora of diagnoses we consider "psychiatric" are really within the same area of medicine. The first thing we're taught to do is rule out "medical" problems, but if we believe in the neurobiological model (aka, every thought, idea, action, etc can be determined by chemical interactions) then everything is "medical". Psychiatric symptoms may involve almost any system as we classify them; neuro, endocrine/hormonal, rheumatology (inflammation), ID, genetics, GI, etc. If we could identify the true physiological mechanisms causing various symptoms or 'disorders', will anything really be considered psychiatry anymore, or would be just be absorbed by neuro into some subspecialty dealing less with gross structures and pathways and more of a biochemical/molecular approach?

Great points again. Re: other classification schemes, I would like to read more about HiTOP and the PDM (personality disorders manual). "Conceptual and historical evolution of psychiatric nosology" by Aftab and Ryznar (2020) mentioned these as well.

What do you mean by the plethora of diagnoses we consider psychiatric are really within the same area of medicine? I am of the belief that everything is medical. All mental processes and experiences are (we have to assume) embodied within the neurologic architecture. I think we could "objectify" (did you mean quantify?) thoughts and experience. At the very least, we could come up with reliable valid correlates, such as activity patterns in limbic substructures (hand wavey but I am speculating).

Why do I think that? Well the phenomena or expressed phenotypes repeat themselves. Obviously behavioral expression is absolutely unique, but not so much so that patterns (or exceptions) cannot be recognized (as we practice psychiatry now). There is massive variability everywhere in biology and anatomical structures that other specialties deal with all the time, but speaking biologically, patterns exist. Like I said in my original post, we know how depression and subjective experiences are described. We have no shortage of that material. What we need are biological correlates to those experiences. And here's the kicker I think: you can't find correlates among those biologically distinct but experientially similar patients that have been wrongly lumped together.

We need a high, high, high degree of specificity in describing these conditions. We need to include in studies only those participants that present nearly exactly the same way, nearly prototypical. Sure, the studies will be difficult and underpowered. But that is the only way I see to make progress.

Re: the fear that "everything will become neurology one day," I think that is unfounded. Obviously the distinction between all medical specialties is a construct, and many overlap. I think neurology honestly has enough to worry about with gross pathology (strokes, MS, neuro-onc) to take on more "brain maladies." That being said, I think we need to step up and learn to safely treat those psychotic symptoms caused by things such as auto-immune encephalitis ourselves, not pawn the task off to "real doctors" that use "real drugs" like "steroids" and "IVIG."

Finally, the idea of "what is a mental illness" is honestly beyond the scope of this reply, but I'll say that we can easily (wrongly? correctly?) make an analogy to any other function/metric of the body's processes that fall on a spectrum. I think it's pretty clear that mental illness is highly unlikely to be categorical, as everyone experiences symptoms of mental illness on some level all the time (salience, mood, anxiety, attraction, irritability, repetitive thoughts). Most things in medicine could be construed as dimensional; even the presence or absence of a tumor requires relative levels of mutations/oncogenetic factors we all possess, but perhaps not in the organization to develop a tumor.

I'll admit you do have to be able to quantify something about the illness, and "feeling sad for two weeks" is a pretty pathetic quantification (again I know it's the best we have). But think about blood pressure. The guidelines exist to say, "Hey, at this point the way we describe vascular tone in the body seems to be approaching the level at which we've seen other people have other problems. The number is related to the problems. So we should intervene to make the number lower."

I don't see any reason why we couldn't have metrics like that for say, salience machinery in the brain that would allow us to interpret a test and say to patients, "Your salience function is higher than average. Have you been noticing things more frequently and intensely? You may be at risk for psychosis."

 

[Alemo]

Look up Kent Kiehl's work on fMRIs and psychopathy if you're not familiar with him. "The Psychopath Whisperer" is his commercial book and is an easy audiobook if you commute to get you started. He was a PhD fellow under Robert Hare at the University of BC and is currently a psychology professor at U of NM. He did a lot of work on brain imaging and EEGs in incarcerated psychopaths and came out with some of the earliest imaging data that showed that psychopaths do seem to have significant different brain activity than most "normal" individuals. He was also involved with an EEG project that found a unique "weird P wave" in individuals who were diagnosed as psychopaths. Obviously there's still a lot to do, but I've found his research to be some of the most interesting, especially in terms of neurobiological exploration of forensics populations.

I'll certainly do this. Psychopathy is a such a strange psychophenotype that has dramatic consequences for everyone involved with that person.

 

[Alemo]

Literally in every field we struggle to understand even some basic aspects of the field and associated pathology. There are a lot of things we don't know that we just "accept" as the most plausible or most useful information necessary to help people.

Great point. Psychiatry only seems more relatively poorly understood because the diseases are only experienced (and narrated by patients), instead of experienced AND observed. Rheumatologic problems might be poorly understood, but you can't argue with histopathology of inflammation or EMG suggesting neuropathy.
And yes, I'm aware we observe patient's behavior, but it takes a significant amount of time and effort (a la Kraepelin) to say [without any narrative evidence from the patient] "I've been watching this person and they aren't acting quite right and I'm positive it's not due to any environmental causes/is related to a spectrum of normal human responses to life and therefore I'm diagnosing them with a psychiatric problem."

I apologize about the kidney slight and thanks for being charitable. I was trying to describe the relative lack of biological understanding in psychiatry, whereas we know pretty damn well what the kidneys are for, how they evolved, what the cytoarchitecture is for, and how blood is filtered to create urine.

Good points about formulation too. In a sense we are filling in gaps, or the negative space, from what the patient said or didn't say. It's incredibly (and I mean that emphasis) sad to me when I see ****ty work (from mid-levels or MDs/DOs or overworked residents such as myself) that amount to "I asked the patient if they felt depressed. Me and the patient don't really understand what that means and have wildly different conceptions that don't match up, but they said yes. I diagnosed them with depression."

 

[Osminog]

Psychiatry only seems more relatively poorly understood because the diseases are only experienced (and narrated by patients), instead of experienced AND observed.

Really? Isn’t observation of patient behavior a basic, crucial component of the psychiatric examination? There are behaviors related to psychiatric illness that are largely separate from patients’ narratives. Is a flat affect not “observed” in the same way that a facial droop is? Is pressured speech not “observed” in the same way that dysarthria is? Psychiatric illnesses generally consist of both observable behaviors and subjective symptoms, not unlike many illnesses found in other specialties.

In my view, the most pronounced fundamental difference between psychiatry and other specializes (with exceptions, e.g., pain medicine) is that psychiatry is at a drastically more primitive phase of biomedical understanding. Science doesn’t understand the brain well enough to know precisely how it produces personality, identify, beliefs, fears, worries, etc., so psychiatrists have to almost purely depends on what other specialties would describe as “clinical diagnoses,” diagnoses based purely on what is said and directly observed in the clinical setting. And as a consequence of this poor biomedical understanding, psychiatrists often have to prescribe drugs that, quite frankly, suck; it’s hard to discover effective drugs for psychiatric illness when you’re not entirely sure what the ideal target should be. But again, none of this means that psychiatric illnesses are all “only experienced” and lack clinically observable qualities.

 

[Alemo]

Re: observation, sure I agree that we observe. The next sentence I wrote after what you quoted explains my thoughts. But to elaborate more, I’m not so sure that observing behavior is necessary and sufficient for diagnosis.

If someone appears euthymic in clinic and tells you a story about their depression, would we not diagnose them as such? It could be argued that behaviors might be “observed” in their story (I can’t sleep, I’m in bed all day, I’m eating less, I’ve written a suicide note), but these are most often narrated (not clinically observed, again, unless you are Kraepelin watching the disease manifest).

If someone’s affect appears constricted, there are several reasons it could appear like that at any given time (boredom, fatigue, shock, shyness, sadness, annoyance—all of which could be related to several different diagnostic constructs: MDD, personality disorders, autism, adjustment problems, psychotic spectrum, etc). It’s completely different than a facial droop, that suggests a palsy/paralysis which must involve pathology of cranial nerves or brainstem). No such “pathology” is readily observable in psychiatry—it’s only with a reported narrative that we link said affect construction with a “disorder.”

I guess acute mania does involve clinical behaviors necessary for diagnosis (it’s hard to tell a history that suggests recent/active mania without demonstrating associated behaviors). But the behaviors aren’t sufficient, because you could just be a high energy talkative person (hyperthymic temperament). So even in that case, I don’t know if the behaviors are sufficient (as would be in the case of diagnosing Parkinson’s disease ).

I agree wholeheartedly with your thoughts on differences between psychiatry and other specialties.