Problem statement Most breast screening programmes worldwide have replaced Screen-Film mammography (SFM) with Full-field Digital Mammography (FFDM). While FFDM provides significant technical and practical advantages over SFM in the provision of population screening programs, whether this move has had beneficial effects on health outcomes remains unclear. An increase in screen-detection rates is only beneficial if the additional cancers detected would have otherwise presented at a later stage and caused morbidity and premature mortality. An indirect measure of this is an observed decrease in the interval cancer rate.

Methods/Results The objective of this study is to compare health outcomes before, during and after the transition from SFM to FFDM in women in NSW, Australia. To do this, we will use linked data to capture women’s journeys from screened/unscreened to cancer diagnosis/no cancer diagnosis to treatment/no treatment and to dead/alive. For the period 1988 to _, a cohort of women with records from Breastscreen NSW, and a second, but substantially overlapping, cohort of all women diagnosed with breast cancer will have their data linked with records in the NSW Central Cancer Registry, NSW Admitted Patient Data Collection, and the NSW Mortality Data. We will use interrupted time series regression to evaluate the effectiveness of the transition to digital mammography over time. We will use a segmented regression approach to analyse the effect of change in technology in mammography screening on each of the outcomes using an impact model.

We will present the results of our evaluation on whether the transition from SFM to FFDM was associated with changes in screen-detected cancer rates, interval rates, positive predictive values, recall rates and false positive recall rates. We will also evaluate changes in breast cancer tumour characteristics for screen-detected and interval cancers (stage at diagnosis, size, histological type, node status, grade), and differences in breast cancer treatment and breast cancer mortality rates. As well as evaluating overall changes, we will conduct analyses stratified by age, breast density, and initial and subsequent screening examinations where possible.

Conclusion This research will evaluate incremental benefits and harms of changes in breast cancer screening programs that can be translated to health policy recommendations as well as the information provided to women invited to screen. The innovative approach of estimating benefits and harms using rates of screen-detected cancers and interval cancers will allow for more timely and rigorous evaluation of changes in screening technologies and practice. Our methods will be able to be applied to future policy changes in breast cancer screening programs both locally and globally, and may be translated to other cancer screening programs beyond mammography.