The review of randomization in the Canadian National Breast Screening Study. Is the debate over?

The randomization procedure in the Canadian National Breast Screening Study (NBSS) is assessed in this issue (see pages 193 to 199) by Drs. John C. Bailar III and Brian MacMahon. They conclude that although there was ample opportunity for the randomization process to be subverted, no evidence of subversion was found. This is unlikely to allay all concerns about randomization, because there are still puzzling differences between the arms of the NBSS in a number of baseline variables. For example, the existence of prior health claims for breast cancer for women who entered the NBSS in Manitoba has raised the possibility that subversion occurred. Although the question may never be resolved, one lesion is clear: randomization in clinical trials should be managed in a manner that makes subversion impossible. As for the clinical implications of the NBSS for women in their 40s, physicians may now look to the results of randomized trials that have been published more recently. A meta-analysis of these results suggests that screening mammography reduces deaths from breast cancer among women in their 40s, but continued follow-up over the next few years will be needed to settle the debate.