Monday, October 27, 2008

Psychiatric Diagnoses: Fact or Fiction?

Below is a guest post from Tim Desmond. I do not necessarily agree with all of the the contents of the post below, but I thought the topic was thought-provoking and controversial, so I have agreed to publish it. Feel free to add comments as you see fit; I will likely add my two cents in the next couple days...

I would like to contribute to the discussion on this blog by summarizing the work of Richard Bentall, psychologist and award-winning author, on psychiatric diagnosis and the DSM-IV. While we may be able to agree that long-term use of neuroleptics should not be the treatment of choice for schizophrenia, or that bipolar disorder is being over-diagnosed in children, I would invite us to question whether 'schizophrenia' or 'bipolar disorder' are valid diagnoses at all. Over the course of his career Richard Bentall has critiqued the medical model of modern psychiatric diagnosis and proposed instead a more personalized symptom-based approach.

The basic question is this: Do people suffer from a finite number of discrete psychiatric disorders/diseases or do people experience varying degrees of human suffering in their own idiosyncratic ways (which include spectrums of sadness, fear, dissociation, etc)? The modern mental health establishment clearly subscribes to the former as evidenced by the structure of the DSM and the theory of 'chemical imbalance.' This belief is so pervasive that even people who claim to disagree with the medical model of diagnosis often think within its terms. For example, the idea that one can be 'misdiagnosed' presupposes that a correct diagnosis could exist. Similarly, saying that schizophrenia is partially caused by psychological factors assumes that 'schizophrenia' is a valid way to group people.

The theory that psychological distress is caused by a finite number of psychiatric diseases can be attributed to Emil Kraepelin, who first published his Compendium of Psychiatry in 1883. Kraepelin believed that the psychiatric patients he treated suffered from diseases analogous to any treated by a practitioner of internal medicine. (Note the common comparison today between psychiatric diagnoses and diabetes forwarded by the pharmaceutical industry.) He said a specific disease process should generate identical symptoms, have identical pathological anatomy and identical etiology (or cause). According to Kraepelin, all that was needed was for these diseases to be discovered in order for diagnosis and treatment in psychiatry to catch up with the rest of medicine. Since it was far beyond the knowledge of his time (or ours for that matter) to find common pathological anatomies or etiologies, he chose to group symptoms. He believed that if he grouped symptoms 'correctly' the people grouped together would be sharing the same underlying disease. They would therefore have the same anatomical problems, the same etiology and respond more or less identically to treatment. He hoped that this kind of systematizing would lead to great advances in the efficacy of treatment.

While Kraepelin's categories have changed over time and grown from 3 to over 200, the basic idea persists to this day – that there are a finite number of discrete psychiatric disorders/diseases people can have and if we were to somehow group symptoms 'correctly' we would have isolated real disorders/diseases.

The problem with this idea is that it is not at all based in evidence. It began as Kraepelin's assumption and he was unable to provide any research in his lifetime to support it. However, it has been an idea so compelling to psychiatrists that they have tried in vain to support it for over one hundred years. In fact, there is a large body of research that directly contradicts this theory.

For example, you would expect that if one doctor diagnosed you with Strep Throat, you would be able to go to any other doctor and get the same diagnosis – and you'd be right. The reason for this is that Strep Throat is a real disease associated with an infection of streptococcal bacteria. You either have it or you don't and there are reliable ways to test if you do.

However, if you are experiencing severe psychological distress and one psychiatrist diagnoses you with "bipolar disorder" there is only a 50-60% chance that the next one you see would give you the same diagnosis. Why is this? Both psychiatrists would have been highly trained in diagnosis, and they would be using the same criteria to make their judgment. So if one says you have PTSD, another says bipolar and a third says brief psychotic disorder, which is the "correct diagnosis?" What do you really have?

Bentall argues the problem is that Kraepelin's main assumption – that there are a finite number of discrete psychiatric disorders – is just not true. You don't have any of those disorders because they are not real. Instead he argues that any psychiatric patient is experiencing a high level of emotional distress that is expressing itself through a range of symptoms and these symptoms can be better understood as extreme expressions of normal human responses to distress.

Bentall advocates for abandoning psychiatric diagnoses altogether. He claims that psychiatry's stubborn attempt to treat mental distress as a medical problem is what has led to its inability to improve treatment outcomes over time. Citing a large body of research, Bentall shows that symptoms from depressed mood to hallucinations can be accounted for psychologically and that doing so is not only more in line with science but more humanizing to patients. Therefore he favors what he calls a 'complaint-oriented' approach in which each patient would be assessed according to his or her unique symptomology. The focus becomes the symptoms themselves and we avoid trying to groups them into arbitrary non-existent disorders. Symptoms can be understood and treated, while disorders cannot because they are not real.

To learn more about Bentall's work, read his 'Madness Explained' which won the British Psychological Society's Book Award.

Author Note: Tim Desmond offers phone counseling and training for therapists through his website at www.coherencecounseling.com

14 comments:

Radagast said...

Well, I've said before how I think we may understand the concept in layman's terms, but I'll restate: "mental illness is behaviour that one doesn't like the look of." It may also be behaviour that causes a person to do harm to him or herself/others, but that's beside the point, when we're trying to apply a definition that we can understand.

People are required to operate in a world where the vast majority of the information that they need to operate effectively is denied to them (ie, they're told to do something, and then not given instruction as to how to do it, and then berated when they get it wrong). That's likely to fuck anybody up. It did me (and I'll fucking crucify my ex-employer for it, however long it takes me).

Matt

Anonymous said...

Perhaps Bentall's never met someone with schizophrenia. Anyway, his theory is quite obviously rubbish since many diseases treated by psychistrists which were once as unexplained as manic depression are now known to be biological in origin - insanity/dementia caused by syphilis, for example - and many diseases have visible physical signs, like the brain abnormalities in schizophrenia, though AFAIK the causal chain is not known.

It's easy to pick on the badly-defined mental disorders that we do not yet fully understand or even know if we have correctly categorised, such as bipolar disorder, but we forget that many diseases treated as psychiatric in the past are now treated as neurological - general paresis in syphilis, or epilepsy - so there's no reason to suppose that we won't eventually find strong neurological bases for diseases like manic depression and schizophrenia, at least in the definite cases.

Bernard Carroll said...

This post does little to advance the understanding of psychiatric disorders. It is a rehash of the familiar contrast between idiographic and nomothetic approaches to assessment and diagnosis. The former emphasizes each patient’s unique experience, symptoms, and circumstances. The latter focuses on formal diagnostic categories, because these, rather than symptoms per se, dictate treatment. In the practice of wise clinicians, both approaches are maintained simultaneously – each patient’s experience of bipolar disorder or of schizophrenia or of obsessive compulsive disorder or of anorexia nervosa or of PTSD is indeed unique.

Unwittingly perhaps, Tim Desmond’s focus on symptoms plays into the hands of those who collude with Pharma in pushing atypical antipsychotic drugs for treatment of nonpsychotic unipolar depression: If we don’t have diagnoses then we don’t have disorder-specific treatments like antidepressants and antipsychotics, and anything goes. Here are 2 links to a fuller discussion of that problem.

http://hcrenewal.blogspot.com/2008/01/variations-on-theme-of-sleaze.html

http://hcrenewal.blogspot.com/2008/06/medscapes-cme-ethics-part-ii.html

Bernard Carroll

CL Psych said...

I'm not very familiar with Bentall's work so it's hard for me to comment on its merits. Yes, the syphilis-mental illness connection is a good example of a biologically based mental condition. However, we are lacking in reliable biological diagnostic markers for the vast majority of diagnosed individuals. Look at how serotonin was hyped as the big neurotransmitter in depression and how that has been debunked.

If the reliability of diagnoses is an issue, we have to consider the circumstances under which diagnoses are typically made. If you take the time to thoroughly perform a diagnostic interview, reliability of many psych conditions is pretty high. So I'm necessarily buying that the poor reliability of psychiatric diagnoses means they are invalid, because assessments are often done very quickly, so they're going to miss many things.

The argument that diagnostic categories don't tell you anything very useful for treatment planning might be true in many cases. So a person is diagnosed as depressed -- who cares? That leaves the door open to dozens of treatments, but which one is best tailored to the individual patient? There's a problem... We have little clue which of the treatments that have demonstrated some semblance of efficacy will work for any given patient. Might that relate to a diagnostic system that puts a very wide variety of people into a diagnostic group that are all then labeled as "depressed"?

Mark p.s.2 said...

LOL at anonymous. I'm a diagnosed schizophrenic. Mental illness has so many aspects , with everyone having a different definition of each aspect. Psychosis or mania is it perminate or temporary? Medications for life? Is it a choice to be mentally ill?
Where are the lab tests? Tests that would turn it from a mental to physical illness.
Is it justice to imprison law abiding adults for crimes they have not commited? When does thinking and believing stupid turn into mental illness? When a psychiatrist gets involved.
I think the defining of schizophrenia (negative and positive symptoms) makes the diagnoser fit the definition of schizophrenia with its overly complicated explination of a behaviour.
Not to forget it is the only illness where the "patient" can be forcefully treated and defined as a "patient" by one person who has no lab tests to confirm or deny the illness.

Daniel Carlat said...

Fascinating discussion, I can't resist joining in.

While I agree that we do not know whether psychiatric diagnoses correspond to specific anatomical or chemical defects, this hardly renders them invalid. The realm of thought and emotion is fantastically complicated, and what psychiatrists and psychologists have done is to map out this territory incognito . DSM is simply a map of the terrain. There are, indeed, symptoms that tend to cluster together, like the “SIGECAPS” of depression, the somatic correlates of a panic disorder, the persecutory delusions of schizophrenia. There’s nothing wrong with giving names to these clusters, so that we have some way of communicating, both to our patients and to other clinicians about patients.

Bentall is correct that every patient is unique, but this doesn’t mean that there aren’t important similarities in presentations. And by the way, this is true in the rest of medicine as well. No two heart attacks are exactly the same, they all vary in important ways, and they all need to be treated differently. But that doesn’t mean that we should banish the label “heart attack.” Without such agreed-upon shorthand descriptions, our treatment system and research enterprise would fall apart. Nobody could communicate. Determining the appropriate treatment, whether medication of psychotherapy, would be rendered laborious and time-consuming without this method of quick communication.

No, I believe the DSM diagnoses are useful, as long as we see them for what they are: agreements. They aren’t biological entities. The problem is that some clinicians become fixated on the diagnostic criteria, thereby ignoring all the messy psychosocial data hiding underneath them.

CL Psych said...

I'm not very familiar with Bentall's work so it's hard for me to comment on its merits. Yes, the syphilis-mental illness connection is a good example of a biologically based mental condition. However, we are lacking in reliable biological diagnostic markers for the vast majority of diagnosed individuals. Look at how serotonin was hyped as the big neurotransmitter in depression and how that has been debunked.

If the reliability of diagnoses is an issue, we have to consider the circumstances under which diagnoses are typically made. If you take the time to thoroughly perform a diagnostic interview, reliability of many psych conditions is pretty high. So I'm not necessarily buying that the poor reliability of psychiatric diagnoses means they are invalid, because assessments are often done very quickly, so they're going to miss many things.

The argument that diagnostic categories don't tell you anything very useful for treatment planning might be true in many cases. So a person is diagnosed as depressed -- who cares? That leaves the door open to dozens of treatments, but which one is best tailored to the individual patient? There's a problem... We have little clue which of the treatments that have demonstrated some semblance of efficacy will work for any given patient. Might that relate to a diagnostic system that puts a very wide variety of people into a diagnostic group that are all then labeled as "depressed"?

CL Psych said...

My comment above adds a word in bold that I missed earlier due to poor attention. Oops.

Anonymous said...

I am happy to see this post has generated such passionate responses. I apologize that I was unable to do justice to Bentall’s work in such a short post. My hope is that this discussion will provoke people into reading Bentall for themselves, because this is not enough space here to provide an adequate discussion of all necessary evidence.

I would like to respond to three assertions made so far: first, that conditions such as bipolar disorder (called manic depression by anonymous) have a biological basis; second, that the DSM’s disorders are the most logical way group symptoms; and third, that the problems with reliability in the DSM are confined to data collected in non-ideal conditions.

While space does not permit going into a detailed discussion of the lack of evidence for biologically-based psychiatric disorders, I would like to recommend reading Wyatt, W. J. & Midkiff, D. (2006). Biological psychiatry: A practice in search of a science, Behavior and Social Issues, 15, 132-151. This paper provides an excellent overview of the available evidence and shows how biologically based theories have been advanced by pharmaceutical companies.

More interesting to me is the question of how to group symptoms. Bentall reviews 70 years of research using factor analysis to discover naturally occurring symptom clusters in psychiatric patients. The naturally occurring symptoms clusters are scientifically robust and consistent over time, while the DSM has not been. Just in the realm of psychosis, hallucinations and delusions make one cluster, flat affect and social withdrawal make another, and all forms of cognitive disorganization make a third. The low correlation between these symptom clusters (A person with a well-formed and internally consistent persecution preoccupation would rate high in hallucination/delusion and low in cognitive disorganization while another patient might exhibit no hallucinations or delusions but speak in a highly disorganized way.) seems to undermine the idea of a shared disease causing them. Bentall recommends using a convention that just gives the patient’s ratings on major symptom clusters.

Finally, Kirk, S. & Kutchins, H. (1994). The myth of the reliability of DSM, Journal of Mind and Behavior, 15, 71-86 shows that using Robert Spitzer’s own field trials and his own definitions of acceptable reliability (a kappa value of 0.7), the DSM-III and IV are unacceptable.

I am happy to correspond with anyone interested in these matters, and I hope that many of you will take the time to read ‘Madness Explained.’ The arguments I have outlined here clearly have many holes that I believe Bentall is able to fill in with evidence over the course of 500 pages.

Anonymous said...

I would echo Dr Carlat's above comments, and add this, starting with regarding the use of the analogy regarding Strept throats used in the posting.

Bacteria have different strains, and so the need for different antibiotics to treat them. Any responsible physician would insist on culturing the infection before starting an antibiotic, so if the strain was resistant, another medication could be substituted. And note, even when treating an infection, one does not just rely on an antibiotic alone; antipyretics like tylenol or motrin are advised, as are rest and fluids. And, exploring what could be the cause to the intitial exposure to such an infection should be pursued to protect others
from further infections, or reinfection in the said patient.

The point? All illnesses fall into the treatment paradigm of biopsychosocial interventions, as causes are multifactorial. Hypertension is not just a biological disorder, so just throwing antihypertensives is not just oversimplification, but minimizing. Heart disease is not just a biological disorder, so just sending the patient to a cardiologist is not the simple treatment. Multiple Sclerosis is not just a biological illness, so focusing on immunosuppressants is not the primary treatment.

As I have said even before I began commenting at blog sites like this, everyone plays a role in the growing failures of medicine, psychiatry as just one example. And I find it both curious and pathetic how patients get so outraged when I include them in the group of causes to the downward spiral, as I do include doctors as well. Face it, patients want and expect quick fixes, be it diagnosis, treatments, and, dare it be used, CURES! And, WHEN it fails after this foolish pursuit, someone has to be blamed. And, who better to target than those who care, and, loath until proven otherwise, to take on fights as it is against our training? Psychiatrists and other mental health professionals. In my opinion, that is where training has failed for so many years, as we have to be the "mensch". Well, as a doctor, and as a jewish person, I am tired of the hostile, misguided rhetoric of the antipsychiatry crowd, who seems to be growing like a cancer in this medium of the internet these days. It resembles nazism-type rhetoric, and I think physicians need to step back and start mounting a challenge to anti-mental health positions that just rail for the profession to be eradicated.

Think what I offer above is so off-beat and inappropriate? Look at history and relearn how this type of attack had such horrendous outcomes; forget the Nazi analogy and look at other examples of discriminating and championing for isolation and eradication against things like religions, races, political ideologies, and other groups that dare to be different and seek inclusion that has occurred in this and other cultures since our species learned how to herd, for the readers who sense my concern has some applicability.

If I am wrong, and this is just a pursuit for rectifying wrongs and applying healthy and productive corrections, then I will be happy, because truth and reason will be shown to prevail. In my opinion, mental health is shunned because people can't handle reality. It is subjective and gray, and the usual suspects who cry out for retribution and retaliation cringe at the sight and sound of these concepts of individuality and uniqueness. Doesn't fit into their soundbites, eh? Politics of hate only thrive by divisiveness. That is why I love George Carlin; while he used sarcasm and humor, his message was, we suck when we assemble. That is not so funny when it is said so plainly, true?

So, now we can attack and diminish the use of terminology as provided by the DSM, or, more responsibly, realize it allows us to have a beginning to look for some common ground, and the responsible provider can tailor the treatment needs once the general problem can be identified. As I was taught in my medical training, don't complain unless you are going to bring some ideas for solutions or alternatives if the problem does not fit. I.E, don't bitch unless you have a path to lay for possible improvement

Just watch how the continuing market woes will spill into provider offices for quick healing needs. I'm seeing it!

It will be sad, and irresponsible, if the caretakers of DSM V are stupid enough to give this a label. My overall point: responsible, credible providers need to reject all the elements that demean and simplify and return to what psychiatry should be: helping people regain function from psychological dysfunction, by using a multifactorial intervention process.

Can this happen? only if we care!

Just an opinion. Hopefully, a thread that will provoke dialogue, not mob chants!

Anonymous said...

I'm currently reading another book which Bentall co-authored, with Read and Mosher, called Models of Madness. It's excellent. Amongst its many interesting contents is the revelation (to me) that Kraepelin was originally funded, to the tune of $325,000 by Rockefeller, the oil magnate, who had just discovered that there was money to be made from turning petroleum by-products into medicine.

Anonymous said...

Count me as in agreement with Desmond. I'd like to add a comment by C G Jung, made more than 50 years ago and as valid now as when he made it:

It is generally assumed in medical circles that the examination of a patient should lead to the diagnosis of his illness, so far as this is possible at all, and that with the establishment of the diagnosis an important decision has been arrived at as regards prognosis and therapy. Psychotherapy forms a startling exception to this rule: the diagnosis is a highly irrelevant affair since, apart from affixing a more or less lucky label to a neurotic condition, nothing is gained by it, least of all as regards prognosis and therapy. (Jung, CW 16, p.86)

Bernard Carroll said...

Tim Desmond has responded with further comments on three issues.

First he repeats his challenge to the notion that major psychiatric disorders like bipolar disorder have a biological basis. He does not support his challenge with evidence. Rather, he smears the notion of a biological basis using guilt by association: “…biologically based theories have been advanced by pharmaceutical companies.” We need to be very clear that the key theories arose out of clinical research in the 1950s and 1960s. They were not ginned up by pharmaceutical companies. In the instance he chose, bipolar disorder, there are many lines of evidence for a biological basis – the natural history, the genetic data, the occurrence of secondary cases associated with right-sided brain lesions, the precipitation of manic episodes by antidepressant drugs, and above all the preventive action of mood stabilizers like lithium. We do not have a complete understanding of the biology of bipolar disorder, but there is no doubt what ballpark we are in – it is not primarily a psychological disorder. It is a biological disorder with physical, psychological and behavioral manifestations.

Desmond next argues for dimensional factors rather than diagnostic categories. He overstates how “robust and consistent over time” are empirically derived symptom factors. These are crucially dependent on the choice of items for the analyses. He also overlooks that symptom factors do not generate diagnostic symptom patterns – they mostly serve just to quantify symptom profiles that have already been recognized. No clinical disorder has ever been identified through factor analysis. Desmond’s comments on the low correlations between symptom clusters are puzzling. The essence of factor analysis is to identify symptom clusters that are independent (orthogonal, in the technical terminology). Kraepelin understood that 100 years ago when he described the variety of complicated states in bipolar disorder – a patient in a manic stupor looks nothing like a patient in a dysphoric mixed manic-depressive state, yet both undoubtedly suffer from bipolar disorder. And a patient in a florid manic episode looks little like the same person who has cycled down to a depressed phase – but she is the same person and the diagnosis endures through both phases of illness. The variability of clinical appearance results from varying mixtures of several independent dimensions of the disorder.

The third issue taken up by Desmond is diagnostic reliability. There is no doubt that the reliability of many DSM-IV diagnoses is disappointing. In part this is because DSM-IV does not clearly recognize psychiatric disorders as processes: Cross-sectional diagnoses will always be less reliable than longitudinal diagnoses. Disorders evolve over time, and the early diagnostic uncertainty shrinks. In this sense, the initial cross-sectional reliability of psychiatric disorders is not worse than that of, say, multiple sclerosis. Moreover, the accuracy, reliability, and validity of diagnoses always increase as one moves from the hurried primary care setting to the specialty setting to the tertiary care setting. That is true throughout medicine. Reliability is not a single number – it must be understood within the Bayesian context of the assessment. And it is amusing to observe some groups carping about reliability when their own track record is nothing to brag about.

Bernard Carroll

Anonymous said...

I'm in agreement with the post. I'm in agreement with lots of 'anti-diagnostics' but am, like the rest, cognisant of a mental health phenomena.

Several things do not appear to have been considered:
1. As a question - do psychological abnormal events become "mental illness" when functioning is compromised or when a diagnosis is made based on simply on existence of symptoms?

2. The biological model will always have a 'theory' to rely on. We know chemicals are responsible for how the brain functions so it is inevitable that altering this will alter function. What we don't know is cause or effect. Do neuroleptics actually bring about change in the process or merely symptomatically relieve it? Psychopharmacology may not be ( and for me, is not) the answer to mental disorder. But none of other options hold any better prognosis either.

3. Why is there no study or theory evident into the psychological causes of mental disorder? Surely, there could be some pre-morbid studies to identify the psychological stressors that occur prior to symptoms. My hypothesis would be to identify those life stressor events that biological theorists might consider to be a 'trigger' for a predisposed condition as being the causality of mental disorder. Even if found to be psychological tho - the type of disorder that manifests and why is something that is as debatable as the biological/psychological aeitiologiy debate.

My gravest concern, and in line with 'therapyfirst's assertions that all illness is biopsychosocial, is in how we persist in aiming to bring out one singular modality as being of greatest importance.

Personally, I use the symptomatic relief model - aka Recovery model - the person is broken, we aim to fix that best we can.
Many physical conditions cannot be 'cured' but can be modified or contained to permit a least disrupted existence and this is best achieved by using the eclectic array of options.
I am not aware of any psychosis that has been 'cured' by psychiatry - tho I am aware of many that have 'gone away' of their own accord. Just like grief can, or heartache, or happiness or feeling suicidal.