This report has described three types of bias introduced into analyses of studies of pregnancy. In each case the bias came from selection of patients for final analysis. The application of the examples of bias described is not restricted to studies of pregnancy outcome. Any comparative study continued over an interval of time is open to similar opportunities for selection leading to similar problems of bias. It is hoped that this report will emphasize the need to control selection during the study or to use adjusted rates during analysis.
ASJC Scopus subject areas
- Obstetrics and Gynecology