This Article  • References  • Full text PDF  • Send to a friend  • Save in My Folder  • Save to citation manager  • Permissions  Citing Articles  • Contact me when this article is cited  Related Content  • Similar articles in this journal  Social Bookmarking   What's this?

Noel S. Weiss

Arch Intern Med. 1996;156(10):1113.

Abstract
By using records of the US Department of Veterans Affairs, Müller and Sonnenberg¹ were able to undertake a very large case-control study of the efficacy of endoscopic screening for colorectal cancer. However, the disadvantage of using data from this source is the inability to characterize individual endoscopic examinations regarding their intent, ie, screening vs response to a symptom or sign. The inclusion of nonscreening examinations in the analysis would serve to minimize the estimated efficacy of screening, since patients with fatal colorectal cancer would be far more likely to have had an examination of this type than controls without colorectal cancer. Recognizing this potential bias, Müller and Sonnenberg restricted their comparison to examinations performed before those that led to the diagnosis of colorectal cancer. However, this approach has its own drawbacks, for it will effectively ignore the most relevant portion of the screening history of those fatal cases in which

CiteULike    Connotea    Del.icio.us    Digg    Reddit    Technorati    Twitter     What's this?