The increasingly rigorous attention paid to the diagnosis of schizophrenia has not yet been extended to the development of reliable and valid criteria for describing the course of the illness. Over the past two decades the concept of 'relapse' has become the pre-eminent measure of the success of community treatment programmes of schizophrenia. However, to date there has been little effort to examine this measure critically, or to develop other measures to reflect the wide range of clinical outcome in schizophrenia. This issue is not new. Kraepelin has been faulted for his lack of clarity in discussing the association between the diagnosis of dementia praecox and a poor clinical outcome (WHO, 1975). Stressing that 'the general course of dementia praecox is very variable', he outlined several typical courses of the illness, all of which proceeded to terminal dementia (Kraepelin, 1919). He recognized the lack of specifically defined outcome criteria, citing reports showing recovery rates varying from 0 to 25%.Although Kraepelin remained sceptical that full recovery without any residual defect was possible, Meyer (1922) contended that functional recovery was possible and proposed that, in many cases, schizophrenia could be regarded as episodic, with relapses representing fresh attacks of the disease. Kraepelin acknowledged that 'more than half exhibit marked improvement', but noted that these recoveries 'give way sooner or later to a relapse'. In support of his thesis of incomplete remission, he observed that subsequent relapses almost always consisted of symptoms similar to those presented in the original episode. This consistency of symptom patterns was also reported by Bleuler (1911), who supported the notion that complete restoration of premorbid mental health was a rarity. Bleuler cautioned against attributing much significance to published reports regarding 'cure' rates, not only because of the subjective nature of such determinations, but also because of' the varying conditions of admission and relapse for each institution', which often play a greater role in defining discharge status than objective changes in psychopathology.Concern about the natural course of schizophrenia led these pioneers to focus discussion more upon the definition of remission, than its counterpart, relapse. Although concern regarding the course of the illness has continued (Vaillant, 1964;Huber et al. 1981), little attention has been paid to this matter in studies investigating the effectiveness of clinical interventions. With the advent of newer pharmacological and psychosocial interventions, emphasis has shifted to factors surrounding clinical exacerbation after the patient has been discharged from the hospital in a state of partial or complete clinical remission. The effectiveness of community aftercare has been measured in terms of failure to sustain these 'remissions' and, more often than not, by readmission of the patient to hospital. The latter is primarily a social intervention which, although usually associated with symptom e...