Psychiatric diagnosis and its discontents

The classification and diagnosis of psychiatric disorder continues to be a contested area for empirical (doubts about the boundaries of different conditions) practical (doubts about the clinical utility of diagnostic... [ view full abstract ]

The classification and diagnosis of psychiatric disorder continues to be a contested area for empirical (doubts about the boundaries of different conditions) practical (doubts about the clinical utility of diagnostic constructs) and moral reasons (worries that diagnoses are dehumanising). In recognition that 'schizophrenia' is perhaps the most contested diagnosis of all, there have been intensive debates about the inclusion of the term in the names of ISPS and the journal Schizophrenia Bulletin (in the former case, the name was actually changed to remove mention of the 'S' word).

In this paper we will review recent empirical and conceptual research on diagnosis, focusing on our own work on psychosis. The evidence that psychotic experiences exist on continuua with normal functioning continues to build and has been supported by research using advanced statistical techniques such as taxometrics. Meanwhile, investigations of the structure of psychiatric symptoms has pointed to two possible solutions: a single 'psychosis' syndrome or (apparently inconsistent with this) five separate symptom factors (positive, negative, cognitive disorganisation, depression and mania). Our own recent studies with patient and population samples have supported a bifactor model that manages to combine these two perspectives.

Overshadowing these developments are two important conceptual issues. First, does the existence of empirical syndromes (e.g. the positive syndrome) imply that the relevant symptoms are the product of a common underlying disease process? We argue that the answer to this is 'No' and that syndromes can emerge from networks of relationships in which symptoms influence each other. Second, must we hope for a single one-size-fits-all method of classification? Again, we argue that the answer is 'No' and that researchers, clinicians and patients should accept a promiscuous realist position in which different (or no) classifications are fit for different purposes.