BACKGROUND: Data on cervical cancer screening programs that have covered a whole nation over a prolonged time are scarce. The effectiveness of a 40-year established nationwide cervical screening program has been evaluated to define optimal age limits and screening intervals. METHODS: Trends in incidence and mortality by calendar time, age, histology, stage and attendance during 1964-2002 and the predictive power of calendar year, age, stage and histology on the cause-specific mortality rate were analyzed. RESULTS: The rate of squamous cell carcinoma decreased significantly, but the rate of adenocarcinoma increased. The age-specific incidence and cause-specific mortality decreased significantly for all age groups except those women aged 20-29 years. An increased age-specific incidence rate, confined to stage I, was observed in the age group 20-39 years after 1980 and a positive correlation was observed between early attendance and the rate of microinvasive squamous (stage IA) cell carcinoma and adenocarcinoma in this age group. The cumulative incidence of invasive disease started to increase two years after the last negative smear. Stage was the strongest risk factor, followed by age and calendar time, and to a lesser degree histology. CONCLUSIONS: The results confirm the effectiveness of the screening program and support the recommendation that screening should commence below age 25 with a maximum of 3-year initial screening intervals. The interval can then be extended after age 40 and stopped after age 65.