Affordable Access

Publisher Website

An analysis of detection bias and proposed corrections in the study of estrogens and endometrial cancer

Authors
Journal
Journal of Chronic Diseases
0021-9681
Publisher
Elsevier
Publication Date
Volume
34
Identifiers
DOI: 10.1016/0021-9681(81)90002-3
Disciplines
  • Mathematics
  • Medicine

Abstract

Abstract We present an algebraic analysis of detection bias and the Horwitz-Feinstein correction in the study of estrogens and endometrial cancer. This analysis sets forth precisely how the validity of conventional study results depends on the cancer detection rates, and how the validity of the Horwitz-Feinstein ‘alternative’ approach depends on both the detection rates and the rates of the medical intervention employed as the matching factor. Numerical examples demonstrate that the observed results of both conventional and alternative studies are mutually compatible with a broad range of true values of the association under study, and that there is no a priori basis for assuming that either set of results is closer to the true value. Nevertheless, estimates of exposure-specific disease detection rates can be used to evaluate the extent of bias present in conventional results.

There are no comments yet on this publication. Be the first to share your thoughts.