Background. Several studies have replicated the finding of increased incidence of schizophrenia and related psychoses in first and second generation migrants from the Caribbean. The finding has remained consistent in studies employing different methods, but concern has been expressed about indirect methods of calculating the population at risk. This study aims to overcome these shortcomings.

Method. A further prospective study was undertaken in Nottingham assembling an inception cohort of psychotic patients (N=168) presenting from a defined catchment area. The 1991 census, which includes codings for self-ascribed ethnic origin, was used to calculate the denominator, employing correction factors for potential under-enumeration. Case-ascertainment was based upon all service contacts and subjects had in-depth assessments including the SCAN. Collateral history was obtained from informants.

Results. Subjects born in the Caribbean, or who had one or both parents born in the Caribbean, had a greatly elevated risk (incidence ratios above 7) for all psychotic disorders and for ICD-10 (DCR)-defined F20 Schizophrenia.

Conclusions. The size of the increase and the methodological safeguards employed support the validity of this now highly replicated finding. A personal or family history of migration from the Caribbean is a major risk factor for psychosis; the consistency of this finding justifies a systematic evaluation of potential aetiological factors. Any hypothesis derived from the evidence so far must explain: increased incidence in first and second generation migrants; increased risk for all psychoses (including affective psychoses); and an effect specifically associated with a migration history from the Caribbean to Northern Europe.