NBSS: opportunity to compromise the process
CMAJ 1997;157:247-8
Re: "Randomization in the Canadian National Breast Screening Study: a review for evidence of subversion" (CMAJ 1997;156:193-9 [full text / résumé])
In response to: D.B. Kopans
Dr. Kopans persists in raising concerns, most of which have previously been shown to be unwarranted.13
The recent review of randomization in the NBSS was initiated after Kopans made a charge to the National Cancer Institute of Canada of scientific misconduct by one of us. This serious charge was based on hearsay from a radiographer previously employed at an NBSS centre; the radiographer had begun her employment after randomization had ceased, as Bailar and MacMahon discuss. In spite of assurances of confidentiality, the employee refused to respond to Drs. Bailar and McMahon's request for confirmation of her claim. In the face of unconfirmed hearsay evidence, Bailar and MacMahon chose not to accede to Kopans' demand that they interview NBSS centre coordinators.
Randomization in the NBSS was not "open."3 Individualized randomization was achieved by a process in general use before distributed computing and electronic mail were available. Instead of telephone operators consulting prearranged lists, we had specially trained administrative staff handle our randomization process. Only they had access to the lists. The screen-examiners did not conduct the process, nor did they have access to the lists.
The NBSS is the only screening study in the world that can completely document balanced randomization in the 2 allocation arms.4 Three other screening studies have used cluster randomization, which often yields imbalanced distribution of variables between arms. Such imbalance has been reported in the Edinburgh trial.5
Two external evaluations of randomization in the NBSS have failed to find evidence of falsification.6 No other screening study has been subjected to equivalent scrutiny, although questions should have been raised not only by the Edinburgh trial but also by the recently published Gothenburg trial, in which screening did not detect a higher rate of breast cancer than in the control group.7
It is not a "revelation" or an "imbalance," as Kopans claims, that women in a usual-care group, in whom breast cancer is mainly detected on clinical grounds, are treated at different institutions than those receiving screening mammography. What may have been a revelation to Kopans was that women with breast cancer in the usual-care group fared no worse than those who had been screened with mammography, although they had lesser degrees of axillary dissection and less extensive histologic examination of resected tissue.
Kopans refers to "mistakes" in the data we submitted for the NIH consensus conference.8 At the conference, we reported 82 deaths due to breast cancer in the mammography arm and 72 in the usual-care group, not 82 and 67, as Kopans states. What Kopans fails to acknowledge is that at the conference other investigators presented revised figures that superseded the data in their abstract submitted months before. The purpose of all presentations at the conference was to give the most recent data.
The NBSS has revealed clearly what other studies have only hinted at: namely, mammography's failure to demonstrate a prompt and substantial reduction in the mortality rate among younger women who volunteer to be screened.9 Mammography is an inadequate technology; tumours for which the prognosis is good are detected early, but those for which the prognosis is poor are not detected early enough to benefit the women affected.10 Radiologists such as Kopans, who rely on good survival from screen-detected case series to establish that a benefit exists,11 are unhappy because women 40 to 49 years of age with mammographically detected breast cancer in the NBSS achieved a 90% 10-year survival rate, and yet these good survival data do not translate into a reduced rate of death due to breast cancer. Kopans' zeal may be excessive.12
Anthony B. Miller, MB
Cornelia J. Baines, MD, MSc
Teresa To, PhD
National Breast Screening Study
Department of Preventive Medicine and Biostatistics
University of Toronto
Toronto, Ont.
References
- Miller AB, Baines CJ, Sickles EA. Canadian National Breast Screening Study. Am J Radiol 1990;155:1133-4.
- Miller AB. The breast cancer screening controversy continues [letter]. Ann Intern Med 1993;118:748-9.
- Baines CJ. The Canadian National Breast Screening Study: a perspective on criticisms. Ann Intern Med 1994;120:326-34.
- Miller AB, Baines CJ, To T, Wall C. Canadian National Breast Screening Study: 1. Breast cancer detection and death rates among women age 40 to 49 years [published erratum appears in CMAJ 1993;148:718]. CMAJ 1992;147:1459-76.
- Alexander FA, Roberts MM, Lotz W, Hepburn W. Randomization by cluster and the problem of social class bias. J Epidemiol Community Health 1989;43:2-36.
- Cohen MA, Kaufert PA, MacWilliam L, Tate RB. Using an alternative data source to examine randomization in the Canadian National Breast Screening Study. J Clin Epidemiol 196;49:1039-44.
- Bjurstam N, Björnheld L, Duffy SW. The Gothenburg Breast Screening Trial: results from 11 years' follow up. Monogr Natl Cancer Inst. In press.
- Miller AB, To T, Baines CJ, Wall C. The Canadian National Breast Screening Study -- update on breast cancer mortality. Monogr Natl Cancer Inst. In press.
- Kerlikowske K. Efficacy of screening mammography among women aged 40 to 49 years and 50 to 69 years. Monogr Natl Cancer Inst. In press.
- Miller AB. Screening for cancer: Is it time for a paradigm shift? Ann R Coll Physicians Surg Can 1994;27:353-5.
- Stacey-Clear A, McCarthy KA, Hall DA, Pile-Spellman E, White G, Hulka C, et al. Breast cancer survival among women under age 50: Is mammography detrimental? Lancet 1992;340:991-4.
- Taubes G. How one radiologist turns up the heat. Science 1997;2275:1057.
 |
Send a letter to the editor responding to this letter
Envoyez une lettre à la rédaction au sujet de cette lettre |
Letters