Background. Patients with schizophrenia have a winter–spring excess of births compared with the general population, the cause of which is unresolved. Fluctuations in the magnitude of the seasonal variation may provide clues to its aetiology.

Methods. All Finnish patients with schizophrenia born between 1950 and 1969 (N = 15892) were identified from two nationwide health-care registers. Their background demographic information was obtained from the Population Register Centre, which also provided monthly numbers of births in each municipality of Finland as multidimensional tables, with sex and year, month and place of birth as marginals. The incidence of schizophrenia was modelled using Poisson regression analysis, with sex, onset age, birth cohort, place of birth (urban/rural), trend and seasonal variation as explanatory variables. We also constructed a monthly time series and decomposed it into three components – seasonal, trend and remainder.

Results. Seasonal variation of births among patients born in the 1950s, especially between 1955 and 1959, was marked, but decreased among patients born in the 1960s. No interaction between place of birth or sex and seasonal variation was observed. The incidence was higher among the rural-born than the urban-born, but declined more slowly among the urban-born than the rural-born.

Conclusions. The intensity of the factor causing the seasonal variation of births in schizophrenia may be decreasing. Urban birth may be emerging as a risk factor for schizophrenia in Finland, as elsewhere.