Mammography screening. Benefits, harms, and informed choice.

The rationale for breast cancer screening with mammography is deceptively simple: catch it early and reduce mortality from the disease and the need for mastectomies. But breast cancer is a complex problem, and complex problems rarely have simple solutions. Breast screening brings forward the time of diagnosis only slightly compared to the lifetime of a tumour, and screen-detected tumours have a size where metastases are possible. A key question is if screening can prevent metastases, and if the screen-detected tumours are small enough to allow breast conserving surgery rather than mastectomy. A mortality reduction can never justify a medical intervention in its own right, but must be weighed against the harms. Overdiagnosis is the most important harm of breast screening, but has gained wider recognition only in recent years. Screening leads to the detection and treatment of breast cancers that would otherwise never have been detected because they grow very slowly or not at all and would not have been detected in the woman's lifetime in the absence of screening. Screening therefore turns women into cancer patients unnecessarily, with life-long physical and psychological harms. The debate about the justification of breast screening is therefore not a simple question of whether screening reduces breast cancer mortality. This dissertation quantifies the primary benefits and harms of screening mammography. Denmark has an unscreened "control group" because only two geographical regions offered screening over a long time-period, which is unique in an international context. This was used to study breast cancer mortality, overdiagnosis, and the use of mastectomies. Also, a systematic review of overdiagnosis in five other countries allowed us to show that about half of the screen-detected breast cancers are overdiagnosed. An effect on breast cancer mortality is doubtful in today's setting, and overdiagnosis causes an increase in the use of mastectomies. These findings are discussed in the context of tumour biology and stage at diagnosis. The information provided to women in invitations and on the Internet exaggerates benefits, participation is directly recommended, and the harms are downplayed or left out, despite agreement that the objective is informed choice. This raises an ethical discussion concerning autonomy versus paternalism, and the difficulty in weighing benefits against harms. Finally, financial, political, and professional conflicts of interest are discussed, as well as health economics.