We investigated prostate cancer incidence in East Anglia from 1971 to 2000. Using age-period-cohort modelling, the number of cases expected in 1991-2000, based on pre-PSA trends, 1971-1990, was compared with that observed. Based on pre-1991 trends, 9203 new cases were expected in 1991-2000, but 9788 cases were observed, an excess of 6%.