Is mammography screening an effective public health intervention? Evidence from a natural experiment

Population-based breast screening programmes aim to improve clinical outcomes, alleviate health inequalities, and reduce healthcare costs. However, while screening can bring about immediate changes in mode of presentation and stage at diagnosis of breast cancer cases, the benefits and harms of these programmes can only be observed at a population level, and only over a long enough timeframe for the cascade of events triggered by screening to culminate in disease-specific mortality reductions. In this paper we exploit a natural experiment resulting from the phased geographic rollout of a national mammography screening programme to examine the impact of screening on breast cancer outcomes from both a patient cohort and a population perspective. Using data on 33,722 breast cancer cases over the period 1994-2011, we employ a difference-in-differences research design using ten-year follow-up data for cases diagnosed before and after the introduction of the programme in screened and unscreened regions. We conclude that although the programme produced the intended intermediate effects on breast cancer presentation and incidence, these failed to translate into significant decreases in overall population-level mortality, though screening may have helped to reduce socioeconomic disparities in late stage breast cancer incidence.