My comments here are an elaboration of my response to this post:
There are many problems with the DSM system (and many of its counterparts), but the main problem, in my view, is that it operates with no coherent definition of what constitutes mental ‘disorder’. By extension, it likewise has no watertight notion of mental health. Instead, it determines disorder by committee. Even assuming for the best committee possible, this would remain an intellectual embarrassment. In any event, with the many vested interests at stake, the committees are far from optimal. Any diagnostic system which tries to evade this point – on the lack of a logical definition of disorder – cannot be taken seriously. Statistical tests of reliability and validity are very much of secondary importance here, since what is that value of undertaking statistical analysis on a nonsensical concept?
This problem is not unique to the DSM-5, either, though it is this latter document which has recently attracted criticism (in my opinion, thoroughly deserved). At this point, whilst the DSM is of bureaucratic importance in some parts of the world, it is now rather like a magician whose secret compartments have been revealed. It has lost its symbolic efficacy.
Now, any characterisation of a patient could function more or less like a diagnosis. The problem is not diagnosis per se, but the shoddy and untenable basis for it given by the DSM and ICD systems. This in turn leads to flippant, tick-a-box diagnosis for patients, with all the requisite consequences.
So, in short, I would abolish all diagnosis deriving from DSM-type recipe books, as well as those deriving from psychometrics.
A viable and much more rigorous alternative can be found in Lacanian psychoanalysis, in which everybody (not merely the ‘pathological’) fits into one of a small number of structural categories. (Of course, within each category, there is infinite room for variation). For instance, a key distinction is that between neurosis and psychosis, and the main determinant for diagnosis one or the other is the presence/absence of repression (to put this in Freudian rather than Lacanian terms, for now. There are many other determinants, but these are of somewhat lesser importance). Now, one will not necessarily observe repression in a questionnaire or an interview or two, and diagnosis can take months. (The absence of repression, on the other hand, in certain psychotics, for example, might be discernible rather quickly).
This system of diagnosis comes closer to the ‘formulation’ proposed by some in the UK in that it allows for richness and complexity without too much stigmatising and reduction. In fact, if one takes ‘formulation’ seriously, it is difficult to see how one can arrive at any detailed assessment of a person’s subjective interaction with their circumstances without a psychoanalysis, or something very much like it. This structural framework can, to an extent, be made to work with the categories of classical psychiatry (though not so much those of the DSM). For instance, melancholia/depression can be differentiated on the basis of whether the person experiencing it is psychotic or neurotic. Moreover, it is broadly consistent with the more robust areas of empirical sychology, such as attachment theory.
Further, merely listing and identifying symptoms is not really adequate for a thorough assessment. Anybody can diagnose a phobia – indeed, most phobics are perfectly capable of self-diagnosing – but at least part of thorough diagnosis should involve consideration of what the symptom is doing for the patient, namely, details of its history, phenomenology, sustaining factors, associations, support in certain fantasies, etc. Psychoanalysis – in particular, that inspired by Lacan – is a credible, and rigorously logical alternative to the shambles that is contemporary psychiatric diagnosis.