Background. Although studies investigating changes in diagnosis between psychotic episodes have differed in design, some consistent findings have emerged. This study seeks to clarify and extend these findings by describing and comparing clinical and operationally defined diagnostic stability in a group of subjects with multiple episodes of functional psychotic illness.

Methods. The OPCRIT programme was applied to case notes of 204 subjects with multiple admissions for psychotic illness. Clinical and operationally defined diagnoses were compared and the spread and stability of diagnoses determined.

Results. An increase in the frequency of diagnosis of schizophrenia from initial to subsequent episodes was demonstrated. High levels of stability were found for schizophrenia (58 to 98%), moderate levels for affective disorders (24 to 83%), low levels for other non-organic psychotic conditions (27 to 54%) and atypical psychosis (27 to 53%), and very low levels for schizoaffective disorder (5 to 39%) and other conditions (0 to 4%).

Conclusions. The stability levels for schizophrenia and affective disorders are adequate, but the low levels for a range of other psychotic conditions raise questions regarding their predictive validity.