Modern classifications of schizophrenia and their limitations

ICD-10 and DSM-IV provide the criteria for diagnosing schizophrenia and other psychotic disorders. At best, they can only be considered an 'arbitrary but well informed consensus on the definition of schizophrenia aimed at reliable communication' (Andreasen & Carpenter 1993). These reflect our current understanding of the concept of schizophrenia.

These diagnostic concepts must be considered provisional constructs intended to fulfil the need for international communication and research. Therefore, a need for constant revision based on epidemiological, pathophysiological, aetiological validation and evaluation of emerging neurosciences and genetic data cannot be denied. This must be an ongoing process.

These classification systems currently use descriptive diagnostic criteria based on the intensity and duration of systems. They are operationalized to a variable extent (more in DSM-IV than in ICD-10) with explicit exclusion and inclusion criteria.

The conditions recognized by applying the criteria laid down in DSM-III, and its successors DSM-IIIR and DSM-IV (American Psychiatric Association 1993), and also in ICD-10 (WHO 1993), are not described as diseases but as disorders.

The rules for schizophrenia laid down in the Diagnostic Criteria for Research in subchapter F20 of ICD-10 are far from describing a disease concept. They list most of the symptoms described by Kraepelin but do not include a long-term course or a particular outcome, or refer to a pathology or a cause. The distinction from bipolar disorder if both are present is limited to a clinical judgement as to which type of symptom occurs first.

Schizophrenia in ICD-10 is not a disease but a disorder. The introduction explains that this terminology is adopted: so as to avoid even greater problems inherent in the use of terms such as 'disease' and 'illness'. 'Disorder' is not an exact term, but it is used here to imply the existence of a clinically recognizable set of symptoms or behaviour that in most cases is associated with distress and with interference with functions.

An article in the Schizophrenia Bulletin illustrated the position clearly (Flaum & Andreason 1991). The authors listed DSM-IIIR and ICD-10 criteria, and three further versions then under consideration for DSM-IV. It is unlikely that a disease concept will change its nature by choosing two of one kind of item and three of another, rather than three of the first kind and two of the second. In fact, in the version eventually adopted (DSM-IV), the chief distinctions from ICD-10 (the requirement for a 6-month course and deterioration in social functioning) remain. Nevertheless, the coding system for DSM-IV is still mapped to that of ICD-9. However, both sets of criteria should be applied, using standardized instruments such as SCAN/PSE10 and the Composite International Diagnostic Interview (CIDI) (Robins et al. 1988) in research and public health projects, in order to foster international comparisons and comparisons with locally favoured alternatives.