Problem 1: Questionable deaths were more probably be called breast cancer deaths in the control group.
The primary outcome of any breast cancer screening study is breast cancer death. While it is obvious that some women with breast cancer die from breast cancer, for others the reason for death is questionable. In the HIP study, 71 deaths were judged questionable in the breast exam/mammography group, versus 73 in the control group.
That is about what one would expect: deaths were about equally likely to be considered questionable in the two groups. Acommittee then reviewed the deaths and decided that 13 questionable deaths in the mammography group were due to breast cancer, compared with 35 in the control group. That's not what one would expect: an extra 22 breast cancer deaths were assigned to the control group.
For perspective, the total number of breast cancer deaths was 147 in the mammography group and 192 in the control group. In other words, the difference observed in the entire study was 45 breast cancer deaths. If the swing of 22 deaths in the death review process represented bias, then almost half of the observed effect of breast exam/ mammography may not really exist.
Problem 2: Women with a prior history of breast cancer were less likely to be excluded from the control group.
Ideally, women who have already had breast cancer would never be enrolled in a study of breast cancer screening (since screening is for those who have never had the cancer being tested for).
But the HIP didn't evaluate women as they entered the study; instead it simply randomized women from the health plan's patient list. The result was that numerous women with a prior history of breast cancer were enrolled. Investigators later tried to exclude these women. Although one would expect that the mammography and the control group would contain about the same number of these women, somewhere between 400 and 500 (once again, the numbers are not consistent from publication to publication) more women were excluded for a prior history of breast cancer from the mammography group than from the control group.
How did that happen? There was systematic data collection on women receiving intervention. Thus, women with prior breast cancer were quickly identified and excluded from the group. But the control group did not receive such close attention (remember, these women did not even know they were in a study), and consequently women with prior breast cancer were less likely to be identified and excluded.
The thing is that women with a prior history of breast cancer are usually to die from breast cancer. Although the investigators tried to remove these women when they reviewed deaths, that is hard to do reliably. Hence, the Cochrane Collaboration concluded that because the difference observed in the HIP study 45 more breast cancer deaths in the control group than in the intervention group could be explained merely by the imbalance of women with prior breast cancer (whereas a balanced study might well have shown that mammography combined with clinical breast exam had no effect), the study should be excluded from their systematic review.
The reviewers raised similar concerns with three of the four Swedish trials, concerns that were subsequently contested by the investigators concerned. 14 Edinburgh was thrown out for the reasons alluded to earlier. Of the eight studies reviewed, therefore, only three even achieved the rating "medium quality": Malmö and the two Canadian trials. The issues raised by the Cochrane Collaboration highlight how difficult it is to run a good randomized trial of screening. Based on their analysis, the authors argued that "screening for breast cancer with mammography is unjustified."15 The general response of the medical and public health community to the Cochrane report was to dismiss the concerns and reaffirm previously held positions.
The National Cancer Institute, U.S. Preventive Services Task Force, and American Cancer Society reiterated their view that there was sufficient evidence of the effectiveness of regular mammography. The World Health Organization issued a similar statement. Only an obscure group of experts from leading American medical institutions and government agencies had a different view. The PDQ Screening and Prevention Editorial Board, charged by the National Cancer Institute to produce data summaries for doctors and patients, concluded that they could not be sure about the true effect of mammography.
Our website is not responsible for the information contained by this article. Webworldarticles.com is a free articles resource thus practically any visitor can submit an article. However if you notice any copyrighted material, please contact us and we will remove the article(s) in discussion right away.
This article was sent to us by:
Fiona Landman at
08152010
1. Understanding the Mesothelioma Cause
All articles in this directory are property of their respective authors. Additionally, read our Privacy Policy
© 2010 WebWorldarticles.com - All Rights Reserved. Partners: Gunblade Saga