Influence of mammographic screening on breast cancer incidence trends in South Australia

Abstract

PURPOSE: To examine breast cancer (BC) incidence trends in relation to mammographic screening and risk factor prevalence in South Australia (SA). MATERIALS AND METHODS: Trends in annual BC incidence rates were calculated using direct standardisation and compared with projected incidence derived from Poisson regression analysis of pre-screening rates. Annual percentage change and change time points were estimated using Joinpoint software. Biennial mammography screening participation rates were calculated using data from BreastScreen SA. Trends in overweight/obesity, alcohol use and hormone replacement therapy (HRT) use were examined using 1991-2009 Health Omnibus Survey data. Trends in total fertility were examined using data from the Australian Bureau of Statistics. RESULTS: BC incidence increased around the time BreastScreen commenced and then stabilised in the mid-1990s. However rates have remained higher than projected, even though the proportion and age distribution of first time screening attendees stabilised around 1998. A decrease in BC incidence was observed among women aged 50-59yrs from the late-1990's but not among older women. Obesity and alcohol use have increased steadily in all age groups, while HRT use declined sharply from the late-1990s. CONCLUSIONS: BC incidence has remained higher than projected since mammography screening began. The sustained elevation is likely to be due to lead time effects, though over-diagnosis cannot be excluded. Declining HRT use has also impacted incidence trends. Implications: Studies using individual level data, which can account for changes in risk factor prevalence and lead time effects, are required to evaluate 'over-diagnosis' due to screening.