Our piece, “Mammogram Trials” , was a critique of the Canadian National Breast Screening Study (CNBSS). We stated shortly in to the piece “we will analyze the CNBSS for its signal, against the noise.” If our intent remained deceptive to some , we regret it.
The astute reader will have picked up that the plural, “trials,” was not accidental. Scientific inquiry seeks the objective truth. Such an inquiry can also be framed as a trial of the innocent. In this politicized corner of medicine, skepticism of mammography often evokes the latter sentiment. In hindsight, a better title would have been “Mammogram Trials and Tribulations.”
To recap, we believe that the CNBSS is not generalizable, it was underpowered to show a relative risk reduction (RRR) of 15–20%, all-cause mortality is not a practical metric to measure the benefits of screening, and the rate of overdiagnosis of 20% is probably accurate. It is unclear which part of our analysis of the CNBSS made Dr. Fowler “deeply distressed.” We are most at loss how our conclusion that CNBSS is “not compelling enough to stop screening” implies, as she has insinuated, “a single study with negative results could be used to influence health policy.”
We acknowledge the studies that Dr. Fowler presents. They show a range of RRR of breast cancer deaths. However, a dispassionate physician scientist must be careful so as not to cherry pick the most optimistic result and make the same error as those who wish to change policy based on a single negative study.
The disinterested scientist may also be intrigued why there is such a broad range of RRR. This is an epistemological issue. The reason for the broad bandwidth of estimates is because of the size of the treatment effect. How large is that treatment effect? One estimate is that if 10,000 women ages 40–49 are screened for 10 years, four breast cancer deaths are averted . Is this a large or a small treatment effect? This is best answered by a comparison to other interventions. The numbers needed to treat with statins for primary prevention is 96 over 5 years to reduce all-cause mortality , and for secondary prevention is 30 over 5 years . The numbers needed to treat for elective coronary artery bypass graft is 15 . The numbers needed to screen to save one death from lung cancer is 320 .
Dr. Friedewald disagrees with our conclusion that “CNBSS is compelling enough to never start mammogram but is not compelling enough to stop screening.” We suspect it is the first half of our conclusion that she disagrees with, which is that if CNBSS was the first randomized controlled trial to assess the benefits of screening, mammography would not have been approved. Note, this is hypothetical. One is, of course, entitled to one’s counterfactual history, but is it inaccurate pointing out that if the National Lung Screening Trial failed to show a mortality benefit, screening computed tomography for lung cancer would not have been approved?
Get Radiology Tree app to read full this article<
Get Radiology Tree app to read full this article<
Get Radiology Tree app to read full this article<
Get Radiology Tree app to read full this article<
Get Radiology Tree app to read full this article<
Get Radiology Tree app to read full this article<
Get Radiology Tree app to read full this article<
Get Radiology Tree app to read full this article<
Get Radiology Tree app to read full this article<
Get Radiology Tree app to read full this article<
Get Radiology Tree app to read full this article<
Get Radiology Tree app to read full this article<
References
1. Jha S., Ware J.B.: Mammogram trials. Acad. Radiol 2015; 22: pp. 973-975.
2. Fowler A.: Screening mammography: more than just one trial. Acad. Radiol 2015;
3. Price E.R.: Overemphasis on overdiagnosis. Acad. Radiol 2015;
4. Miller A.B., Wall C., Baines C.J., et. al.: Twenty five year follow-up for breast cancer incidence and mortality for of the Canadian National Breast Screening Study: a randomized screening trial. BMJ 2014; 348: pp. g366.
5. Breast Cancer Screening : USPSTF. Available at http://www.uspreventiveservicestaskforce.org/Page/Document/RecommendationStatementDraft/breast-cancer-screening1#tab3 Accessed October 5, 2015
6. Taylor F., Huffman M.D., Macedo A.F., et. al.: Statins for the primary prevention of cardiovascular disease. Cochrane Database Syst Rev 2013; CD004816
7. Scandinavian Simvastatin Survival Study Group: Randomised trial of cholesterol lowering in 4444 patients with coronary heart disease: the Scandinavian Simvastatin Survival Study (4S). Lancet 1994; 344: pp. 1383-1389.
8. The Veterans Administration Coronary Artery Bypass Surgery Cooperative Study Group: Eleven-year survival in the Veterans Administration randomized trial of coronary bypass surgery for stable angina. NEJM 1984; 311: pp. 1333-1339.
9. National Lung Screening Trial Research Team : Reduced lung-cancer mortality with low-dose computed tomography. NEJM 2011; 365: pp. 395-409.
10. Friedewald S.: Screening mammograms would have been initiated regardless of the CNBSS. Acad. Radiol 2015;
11. Moy L.: Acad. Radiol 2015;
12. Bleyer A., Welch H.G.: Effect of three decades of screening mammograms on breast-cancer incidence. NEJM 2012; 367: pp. 1998-2005.