Abstract
The article by Autier et al. 1 shows a most elementary lack of understanding of screening and is simply wrong.
The authors’ assumption that, “during post-intervention periods, because screening (or absence of screening) activities are similar in the screening and in the control group, cancer detection rates in the two groups are also similar” is nonsense since a large number of cancers in the ASP will have been screen detected in the intervention period, which otherwise would have been detected in the post-intervention period. Thus, in the post-intervention period, cancer detection rates in the ASP will be lower than in the control group, and breast cancer mortality also will be lower. This is what they found, and is what they should expect if screening reduced breast cancer mortality.
This fundamental point also invalidates their argument against the closure screen of the control group; a closure screen of the control group is conservative, albeit less biased than the authors’ preferred method. 2
When arguing against the closure screen, the authors’ adjusted estimates are wrong, since they subtract the deaths from cancers detected at screening of the control group, but not those detected contemporaneously in the study group. In the Two-County Trial, the reduction in mortality from breast cancers prior to the closure screen has been in the public domain since 1985, 3 and was 31%, similar to the reduction in mortality observed when the control group includes deaths from cases diagnosed in the closure screen and their counterparts in the study group. 4 Why resort to speculation when the empirical data are already published?
Their assumption that the 10% of breast cancer deaths in the control group in the Two-County Trial from cancers detected in the closure screen can be applied to other trials also is naïve, as is their argument that the smaller numbers of advanced cancers in the study group contemporaneously with the control group’s closure screen somehow invalidates the design and analysis. The PSP screen is a prevalent screen, whereas at that time, the ASP is in incident screen mode. The appropriate comparison is with the prevalent screen of the ASP, which was published in 1992 and shows similar results to the PSP closure screen. 4
The criticism of the Swedish Two-County Trial on the grounds of imbalances in missing values is inaccurate, not only are we unable to verify these figures from the trial data, we cannot find them in the paper that Autier et al cite as the source. 4
The issue of potential bias in cause of death has been examined time and again and shown to be a red herring.5–7 Indeed, the Swedish overview has published the excess mortality analysis which does not require classification of cause of death and found essentially the same mortality reduction as in the cause-specific analysis. 8
The paper does not contribute to the debate on the value of mammographic screening, but confuses the discussion due to fatal errors that negate their conclusions.