Bmj Usa: Education And Debate

Commentary: Breast self examination

BMJ 2003; 327 doi: (Published 19 November 2003) Cite this as: BMJ 2003;327:E198
  1. Robert A Smith, director of cancer screening (Robert.Smith{at}
  1. American Cancer Society.

    Do we really know what we think we know?

    From BMJ USA 2003;Mar:168

    Austoker salutes the publication of the final results of the Shanghai Trial of Breast Self Examination (BSE),1 and she echoes the conclusion reached by Harris and Kinsinger2 that “routinely teaching and doing breast self examination is dead.” This regrettable choice of words seems to signal a sense of triumph: Now that results from a randomized trial show no benefit from BSE instruction, gold standard science has prevailed over intuition, observational data, and the “breast self examination lobby.”

    However, while celebrating the decisiveness of these findings, Austoker and others overlook the challenge of earlier detection of palpable tumors. The fundamental aim in the control of breast cancer is the application of therapy while tumors are still small.3 4 The distinction between non-palpable and palpable breast cancers is a good measure of this challenge, but it is equally true that it is preferable to treat a smaller rather than a larger palpable tumor. This is important because current imaging technologies lack perfect sensitivity, access to breast imaging is limited among specific age-groups, and adherence to screening mammography is highly variable. Patients and health care professionals contribute further to delays in “diagnosis of palpable tumors.5 Thus, a sizable proportion of breast cancers still are detected after symptoms have become evident. Among all incident breast cancers diagnosed in the world, most are diagnosed only after becoming symptomatic.

    The Shanghai trial offers important insights, but we should be cautious about concluding that this study answers the question of BSE efficacy once and for all. First, as the authors remind us, this was a trial of BSE instruction, not of BSE per se. What was unclear in the first publication,6 and clearer in the second,1 was the limited potential for BSE to improve on the average stage at diagnosis. More than half of the incident cases in the control group involved lesions classified as T1 or smaller, which suggests that Shanghai women may have already had a heightened sense of awareness about breast cancer. Moreover, BSE instruction resulted in only marginally improved technique in the instruction group compared with the control group.

    Other unexplained observations may have influenced the results. Although the investigators observed no difference in breast cancer mortality between the two groups, the instruction group had 10% fewer deaths from all causes, a difference that is statistically significant. The deficit in all-cause deaths was large (n=594), exceeding the number of breast cancer deaths in both arms of the study, and is not explained by the investigators. In the absence of any demonstrated benefit, the observation that the ratio of total biopsy specimens to breast carcinomas was much higher in the instruction group compared with the control group (4.2 versus 2.7) represents a net cost to women and the health care system, but it must be put in context. First, the excess of 1.5 involved specimens; fewer women per diagnosed cancer were subjected to biopsy. Second, most of this excess occurred early in the study; in the final years the two groups were quite comparable. Perhaps adherence waned in the instruction group, but other explanations are plausible.

    A fair conclusion about the results of the Shanghai study is that a program of BSE instruction did not significantly reduce breast cancer mortality in a population that was already highly responsive to breast symptoms. These results are likely applicable to similar populations elsewhere, including those of most countries that have had ongoing awareness campaigns promoting early detection.

    We should not conclude, however, that BSE is ineffective. BSE tends to be performed infrequently7 8 and with low proficiency.9 It would be expected to have less potential to downstage tumors in populations where women already have a heightened awareness of symptoms and quickly seek evaluation. There is a persuasive literature indicating an advantage for women who regularly perform BSE proficiently.10 Even among these women, it is likely that the value of BSE diminishes over time in the presence of regular, high-quality mammography. As noted by the authors of the Shanghai trial, more research is needed to assess the value of BSE as a strategy to identify early-stage palpable tumors in a setting where women commonly present with very advanced disease: “It should not be inferred from the results of this study that there would be no reduction in the risk of dying from breast cancer if women practiced BSE competently and frequently.”1 The proper interpretation of the Shanghai trial is not that BSE is ineffective, but that we need to think about BSE and self detection differently.

    In a perfect world, few breast cancers would progress undetected by mammography. But this world is not perfect: Mammography will not benefit some women, and others cannot benefit because they lack access to breast imaging. For these women, much remains to be learned about the factors that contribute to heightened awareness and to delays in diagnosis. Austoker and others2 11 have asserted that awareness is still important, even if BSE is not—as if the path to (and meaning of) awareness were self evident. It clearly isn't. We may need to expend on promoting awareness the resources that we have spent on BSE.

    Ultimately, BSE may be shown to have value in some settings, and in others it may offer no measurable advantage over promoting awareness about symptoms of breast cancer and the importance of promptly seeking medical care. But the jury is still out.


    • Competing interests None declared.


    1. 1.
    2. 2.
    3. 3.
    4. 4.
    5. 5.
    6. 6.
    7. 7.
    8. 8.
    9. 9.
    10. 10.
    11. 11.
    View Abstract

    Log in

    Log in through your institution


    * For online subscription