Overdiagnosis in breast cancer screening is an important issue.
•
There are many methodological pitfalls in studies of overdiagnosis.
•
Stringent research methods should be used.
•
Otherwise, studies of overdiagnosis may easily provide erroneous results.
Abstract
Overdiagnosis in breast cancer screening is an important issue. A recent study from Denmark concluded that one in three breast cancers diagnosed in screening areas in women aged 50–69 years were overdiagnosed. The purpose of this short communication was to disentangle the study's methodology in order to evaluate the soundness of this conclusion. We found that both the use of absolute differences as opposed to ratios; the sole focus on non-advanced tumours and the crude allocation of tumours and person-years by screening history for women aged 70–84 years, all contributed to the very high estimate of overdiagnosis. Screening affects cohorts of screened women. Danish registers allow very accurate mapping of the fate of every woman. We should be past the phase where studies of overdiagnosis are based on the fixed age groups from routine statistics.
