While many would agree that the current classifications of psychiatric disorders are not ideal and would be happy to see them improved, it is unclear how best to achieve this. This lack of clarity reflects a deeper problem with our understanding of the pathogenesis of most mental disorders. In the absence of that understanding, the empirical foundation upon which to erect a revised new set of descriptions of those disorders is either not available or at best noticeably weak 1. A common consensus therefore is that, at least for the time being, a conservative approach to the revision process makes the most sense 2.
This humility is not without basis. Many are still quite aware of the damage that the free-wheeling approach to diagnosis of mental disorders did to psychiatry as a scientific branch of medicine. Indeed, the derision with which the field is often held by both fellow medical colleagues and, sometimes, by the lay public is partly to do with what is perceived as the amorphous nature of our diagnostic entities. It is important that we should not lose sight of that recent past as we make efforts to build on what we currently have.
It is precisely in response to that past that the designers of DSM-III pitched their tent strongly in the achievable domain of improving reliability 3. The fashioning of explicit inclusion and exclusion criteria, and the specification of the number of symptoms as well as the duration of their occurrence, in making psychiatric diagnosis was entirely to achieve the goal of reliability. Most would agree that, largely, that goal has been achieved with our current classificatory systems. But the achievement of this common language has not been without its occasional discontents. Thus, even today, some researchers still indulge in the idiosyncratic definition of what constitutes some mental disorders. While it is probably less fashionable now to talk, for example, of “masked depression”, it is still the case that, just by slicing what is decidedly a dimensional construct at an arbitrary point, some would claim to be describing a unique disease entity. So, if you cut the depression dimension at any arbitrary point and claim you are describing, for example, a unique African depressive syndrome, some, including prestigious academic journals, could be impressed. It is therefore clear that, even with the current rigid, albeit also arbitrarily defined symptom clusters upon which our consensus has been built, there are still instances when clear assault is launched at that common language, simply by going fishing, nosologically speaking.
This, for me, is the important backdrop against which to consider the proposal for the use of prototypes for making diagnosis of mental and behavioural disorders. Given our current state of knowledge and the limitations of our classificatory systems today, the question is worth asking: what do we seek or should we expect to achieve in a revision exercise? What gains must we protect even as we struggle to improve on what is obviously an imperfect system? Everyone would agree that we should seek to achieve improved validity for the disorders in our classificatory systems. However, most would also accept that the goal of validity remains a distant aspiration 4. Indeed, by the nature of the phenomena we deal with in psychiatry, many would probably acknowledge that there will always be a limit to which the road to the validity paradise could take us. However, while we strive to reach the validity nirvana, we have to keep an eye on the possibility of losing reliability and utility. It is precisely because our achievement in having reliably defined entities has also improved the clinical utility of our current classificatory systems, that there is now an admission that, at least for the ICD system, the pursuit of improved utility is the beginning of the revision wisdom 5. So, would the use of prototypes help the attainment of improved reliability and utility?
While Drew Westen presents a compelling reason for us not to dismiss the value of prototypes, embedded in his description is the source of my concern. May it lead us to the danger of losing some of our current gains whilst we are in pursuit of yet another forlorn diagnostic destination? For example, inferences from the same clinical presentation, in regard to clinician interpretations of that observation, can be protean and subject to a lot of subjectivity. Also, we may need several prototypes to capture the range of presentations that a disorder may have. For example, without explicit criteria, patients with schizophrenia with diverse admixtures of negative, positive and disorganization symptoms may not be unambiguously captured with a few prototype descriptions. Answering the question “Does this patient have personality disorder, schizophrenia, or delusional disorder?” may be problematic without some specific symptoms and the duration of their occurrence.
In many respects, the editorial by Mario Maj, presaging this forum 6, has identified some of the possible problems associated with the use of prototypes. I only wish to add that it may also embolden those who are keen on nosological exoticism to create “new” entities, hiding under the claim of taking cultural relativity into account, and develop versions of the template prototypes that suit their fancy. That way, we would have lost our imperfect but useful common language and created a Tower of nosological Babel.