E-poster Presentation 2014 World Cancer Congress

Overdiagnosis due to mammography screening programs: Evidence from South Australia (#992)

Kerri Beckmann 1 , Janet Hiller 1 2 , John Lynch 1 , Gelareh Farshid 1 3 , Nehmet Houssami 4 , David Roder 5
  1. University of Adelaide, Adelaide, SA, Australia
  2. Swinburne University of Technology, Melbourne, VIC, Australia
  3. BreastScreen SA, Adelaide, SA, Australia
  4. University of Sydney, Sydney, NSW, Australia
  5. University of South Australia, Adelaide, SA, Australia


Mammography screening is effective in reducing breast cancer (BC) mortality, however concerns have been raised that it may also lead to over-diagnosis, i.e. the detection of cancers that would never have emerged clinically in a woman’s lifetime had she not participated in screening. The extent of over-diagnosis due to mammography is contested, with estimates varying from 0 to 54%.


To estimate the extent of overdiagnosis due to mammography screening in South Australia using two different methodologies.


Method 1 used a case-control design to compare screening histories for women with and without BC. Odds ratios were determined across different time intervals after screening to allow for lead-time effects. Cumulative incidence (CI) was calculated by applying odds ratios to background reference rates, derived from projection of pre-screening incidence trends. Over-diagnosis estimates were obtained by comparing CI with and without screening for women aged 45-85yrs. 

Method 2 used a lead-time modelling approach that applied estimates of lead-time duration and screening sensitivity, and screening participation data to adjust the expected  background incidence, by iteratively adding the number of cancers expected to be brought forward by screening each year, then subtracting this number from the pool of cancers in future years.  Over-diagnosis was calculated by comparing lead-time-adjusted and observed CI for women aged 40-84yrs.


Estimates of over-diagnosis from the case-control study were 7.6% for invasive BC and 13.8% for all BC. These estimates are likely to be inflated due to higher breast cancer risk in screening participants compared with non-participants. Estimates from the lead-time modelling approach were 8.7% and 12.3% respectively, but these estimates assume no other influences on background rates.


Our findings suggest a modest level of over-diagnosis due to mammography screening, which is consistent with findings from screening trials and the recent UK review of mammography screening.