Age-period-cohort modelling of the incidence of breast and cervical cancer In New South Wales

The incidence of breast and cervical cancer in New South Wales women, diagnosed between 1972 and 1991, is modelled using an age-period-cohort poisson regression model. The age-period-cohort models are fitted to the data without the age constraint, using 5-year age and cohort effects and annual period effects. These models show significant age, period and cohort effects for breast cancer and significant age and cohort effects for cervical cancer. Both models confirm expected aetiological patterns. A sharply increasing age effect is exhibited until age 45 with a change in slope thereafter, the effect at age group of 80-84 is approximately 11% higher than at age group of 60-64.