Clearly the idea comes out of medicine and what is sometimes called “the medical model” but which I prefer to call the “diagnosis model” or the “disease model”. (Yes, I have a lot of ambivalence and dissonance about my first training and profession but sometimes I find a need to stand up for it”) Basically the diagnosis idea is that there are fairly unitary states, transient or enduring, that are dysfunctional for the person (or other animal: the term applies to veterinary work too!) and/or distressing for them. The model has been transferred into the psychological/MH/WB world rather problematically.
Details #
Even in the purely physical health field diagnosis has a broad spread from simple, persuasive and generally useful,through to much more problematical. At one end are things like a broken bone or a traumatic injury, genetic disorders and problems clearly of infectious origin. At the other end are much more difficult and complex issues: hypertension is probably a paradigmatic example of something that clearly threatens life expectancy but has no direct symptoms and is defined by medicine and epidemiology.
In the psychological realm there are many, including psychiatrists, who argue that the term should never be used for psychological issues and that there are no such things as psychological diagnoses. I disagree but would like to see the term used much, much less than it is and more as a discursive tool rather than an assertion or attribution. I believe that classifying some developmental disorders such as Down’s syndrome as diseases, using them as diagnoses, has some utility though it is still really just an aetiological statement and somewhat limited prognostic indicator. Huntington’s disease seems to me to one of very few conditions with huge psychological impacts (and physical) and a very clear genetic cause and some prognostic information, so much so that I don’t see any gain from denying it as a “disease” affecting the psyche, as a psychological diagnosis. However, regarding depression, anxiety and many other conditions in both the DSM (currently DSM-5) and the ICD (currently ICD-11) as “diseases” on a par with Huntington’s disease, TB or meningitis has dangers that perhaps outweigh the utility of this culture. That’s a topic of books and perhaps too huge a discussion for this glossary but I think it needs a bit more here.
Yesterday I was in a webinar in a series in which young psychotherapy researchers present their work. The theme was around the intersection of two diagnoses taken (via a structured interview) from the DSM: Major Depressive Disorder (MDD) and Borderline Personality Disorder (BPD). The presenter and the other members seemed to take the group design, based on those diagnoses, as not something needing comment. I felt a bit differently but as their total age was a bit less than mine and I am undoubtedly not a “young researcher” so I didn’t chip in. OK, I exaggerate the age issue but I think I was the oldest there by a long way and I suspect most weren’t born in 1984 when I started working in psychiatry and started learning and delivering “psychotherapy”. That’s a rather personal, narcissistic even, perspective but I wonder if it does matter: I think that the construction of MH diagnoses was very different back in 1984 from what it is now. Then my psychiatric training certainly stressed attention to diagnosis but held it as only part of the much more important “formulation” of the client’s situation which included genetics (a bit!), developmental history, family structure and relationships, education and employment histories, the “pre-morbid personality” and the crucial “history of the presenting complaint” and both “past medical history” and “past psychiatric history” as well as observational data (the “mental state examination”). It was an idiographic thing, unique to the individual even if it did, as it did for some I met, include a reality of clear Huntington’s disorder, a stroke, a post-partum bacteriaemia etc. In my psychotherapy training things varied but were largely reluctantt to use diagnoses though I think there were varying degrees of wariness about challenges in the presence of substance abuse/dependence, “psychotic states” and possible physical illnesses.
Since then the diagnostic model, mostly a DSM one, the exact version changing with the decades, seems to have come to dominate therapy research funding.
I think that’s a problem. Of course, no research designs are perfect, research designs simplify what we look at, based on simplifications in theory (“models”) and we can’t proceed without simplifying. However, if we relentlessly use more or less sophisticated vesions of the post-ANOVA group mean difference designs and if we make the independent/predictor variables creating the groups ones from diagnosis, we risk not acknowledging what may be being systematically missed.
To go back to the in so many ways excellent presentation, two things strike me:
- (post)ANOVA designs always have to cope with unreliability in the response/dependent variables but they are very, very vulnerable to loss of power if we have unreliability in the predictor/independent, the grouping, variables.
- in particular if we have heterogeneity within the groups, our model is starting to collapse and may be finding things that are true at the level of the grouping used, but of no real use, in terms of interventions, at the individual level.
The presenter was doing lovely work looking at physiological measurements, the ones presented were blood pressure variables. (Good to go back to bodies, the ego is a body ego anyone?) One nice comment was that clients like to have physical data, particularly perhaps MRI brain scans. But that got me remembering many clients I saw over 32 “clinical” years who would have picked up BPD diagnoses. Some were attracted to neurophysiological ideas, some were quite physiologically revolted by the idea of looking there. Some had very complex relationships with their bodies with long histories of often severe non-suicidally intended self-harm, others seemed reassured by neurphysiological ideas and “grounded” by thinking about their bodies in ways that seemed to reduce self-harm, some simply didn’t have the physical self-harm side of things but certainly had the psychological and interpersonal things in the diagnostic criteria for BPD. If we lump all these people together under “BPD” do we risk missing really potentially important and interesting issues about neurophysiology and therapies? Do we risk coming out of our statistical analyses with generalities that, as I suggested above, may be fine as group mean generalisations (within the limits of the sampling frame) but have near zero predictive value for individuals because crucial variables that maybe have strong links between psyche and neurophysiology were not measured?
Of course and we can never measure more than a few variables and never analyse statistical associations across many variables unless we have huge datasets. However, if we almost solely fund diagnostic group based research and forget human individuality, or not forget it but just push it into the “error” component in our models while perhaps having it centre stage in therapy work itself, are we not missing things?
All rather rhetorical but I hope it unpacks some of my concerns about diagnoses and how they too often dominate therapy research and routine therapy change data analyses. It also comes back to the personal/impersonal issue: perhaps we have been swung too far to the impersonal partly by the diagnosis model, and too far from the personal and idiographic?
For completeness, I should mention that DSM and ICD are not the only diagnostic frameworks, I found Foulds’s hierarchical model useful at times and there are psychoanalytically derived models and the SWAP (Schedler-Westen Assessment Procedure) as an interesting alternative system of categorisation. To be added some time I hope!
Try also #
Classification
DSM
ICD
SWAP (Schedler-Westen Assessment Procedure)
Chapters #
The issue runs through the entire book but, as here, we avoided diving right into it!
Online resources #
None currently.
Dates #
Started 10.ii.24 and got stuck(!), current version 21.viii.24.
