The Jan. 23, 2001, issue of CMAJ made a real attempt to bring together several articles on breast cancer, a topic of considerable importance. However, I found the paper by Ruhee Chaudhry and colleagues to be seriously flawed.1
In this retrospective study, the women seen in community hospitals were markedly different from those seen in teaching hospitals. This could result in lead-time bias in favour of teaching hospital patients. There is indeed some evidence of this in the paper, as the tumours of women presenting to teaching hospitals tended to be smaller and less malignant tumours (ductal carcinoma in situ) than those of women presenting to community hospitals. Thus, they would have had better outcomes irrespective of location.
In addition, the authors failed to describe the manner in which breast cancer was detected. There is a better outcome for breast cancer detected through screening mammography than for breast cancer detected clinically.
Lastly, we don't know the proportions of women who had auxillary node dissections in each group. This procedure is used less often in community hospitals than in teaching hospitals, and thus there may be a greater potential for misclassification of the stage of disease in the community setting. Do the authors have any information on this important variable?
Reference
- 1.↵