Mammograms do not significantly reduce deaths from breast cancer and should be abandoned. So concluded a 1999 Danish review of research on the safety and accuracy of mammogram screening. [1]

The review's methodology was widely criticised, so it was rerun using a method designed by the Cochrane Collaboration, an international group whose evaluations of medical treatments are highly respected. The re-run came to the same conclusions.

The re-run by Ole Olsen & Peter Gotzsche from the Nordic Cochrane centre in Copenhagen (Denmark), routine breast cancer screening of women over 50 did not reduce deaths. It might actually do harm, they suggested, because they often led to surgery to remove tumours, some of which are so slow growing that they would never have developed into cancer in the women’s lifetimes. They found that women who had been screened had suffered a 20% above average level of mastectomy and a 30% above average level of combined mastectomy and tumour operations.

This time it was the Cochrane Collaboration themselves who objected. They refused to publish it unless changes are made. In response, The Lancet took the unusual step of publishing the original study (20.10.01) accompanied by a blistering comment from editor Dr. Richard Horton condemning Cochrane for interference that "eroded academic freedom".