Professor Stephen Duffy: Overdiagnosis in Breast Screening
In addition to evaluating the effect of breast screening programmes on mortality, there is a need to estimate the extent of potential harms, such as overdiagnosis. This is not a simple issue, as even if there is no overdiagnosis at all, the introduction of a screening programme will inevitably be accompanied by an apparent increase in incidence due to lead time. In this presentation, we examine data from randomised trials and service screening programmes to estimate rates of overdiagnosis. We find these rates to be relatively low, ranging from 1% to 7%. Implications for screening and treatment are discussed.