Observational study of Age, test THreshold and frequency on English NAtional Mammography screening outcomes (ATHENA-M)

Research Question

How does age of eligibility, screening interval and recall threshold affect the benefits and harms of breast cancer screening? Background

There debate about whether the benefits of breast cancer screening outweigh the harms. The UK Independent review estimated that Breast Cancer Screening in the UK has mortality benefit, but harms through overdiagnosis of disease that would never have become symptomatic within the woman’s lifetime, and through false positive recalls to assessment of women who do not have cancer.

International Variation in Breast Screening

Different countries give different versions of breast screening, due primarily to a lack of evidence about which gives most benefit and least harm to women screened. In the UK we recall 4% of women for further tests because their mammograms show suspicious signs, with other countries recalling as few as 2% (Denmark) or as many as 10% (USA). In the UK we offer screening every 3-years, which is the longest time between screens in the world.

In the US it is every one or two years and in most of Europe every 2-years. Within England there is a lot of variability between centres, because we don’t know what is best to offer women. The Research Plan

In this observational study we will analyse the records of women screened in England between 1990 and 2018, with follow up to whether they developed breast cancer (from the English cancer registry) and whether they died (from the Mortality and Births Information System). We will analyse how inviting women of different ages, with different frequencies, and recalling different proportions affects the benefits and harms of screening.

Analysis methods are designed to allow causal inference, using Hernan’s target trial approach. For screening interval we will use the clone, censor and weight approach, and for age extension we will split into hypothetical weekly trials. Short term outcomes include false positive recalls, cancers detected at screening, and symptomatic cancer detection after screening.

Long term outcomes will include mortality, breast cancer mortality, and overdiagnosis. We will also analyse the mechanism of action for any changes, how changes to number and type (grade, stage and size) of cancers detected at screening affect number and type of cancer detected symptomatically in the years after screening, numbers overdiagnosed and mortality.

We will explore how more or less detection of Ductal Carcinoma in situ affects benefits and harms of screening. Using Findings to Change NHS Practice

We will communicate findings to the UK National Screening Committee, feed into the process of redrafting the English quality assurance guidelines for breast cancer screening, and disseminate via national champions.