Related Experiment Videos
Human teratogens, prenatal mortality, and selection bias.
M J Khoury1, W D Flanders, L M James
1Division of Birth Defects and Developmental Disabilities, Centers for Disease Control, Atlanta, GA 30333.
American Journal of Epidemiology
|August 1, 1989
Summary
Selection bias from prenatal mortality can distort studies on birth defects. This bias may overestimate or underestimate the true risk of teratogens (harmful agents) in case-control studies.
Area of Science:
- Epidemiology
- Reproductive Health
- Genetics
Background:
- Etiologic studies of human teratogens typically rely on case-control designs.
- These studies are often conducted on live births or spontaneous abortion data.
- Prenatal mortality associated with teratogens and defects can introduce selection bias.
Purpose of the Study:
- To analyze how prenatal mortality influences selection bias in teratogen studies.
- To derive relationships between true and observed odds ratios (OR) considering prenatal mortality.
- To assess the impact of bias on etiologic inferences for birth defects.
Main Methods:
- Mathematical modeling to relate true OR to observed ORs.
- Analysis of selection bias as a function of prenatal mortality rates.
- Examination of bias under different interaction patterns between teratogens and defects.
Main Results:
- Selection bias can lead to both overestimation and underestimation of teratogen-defect associations.
- Increasing prenatal mortality effects amplify bias, increasing observed OR in spontaneous abortions but decreasing it at birth.
- Weak associations (OR 0.3-3) may be artifacts of bias; weak teratogens (OR < 3) might be missed in live birth studies.
Conclusions:
- Prenatal mortality-induced selection bias is a critical factor in interpreting teratogen studies.
- The observed odds ratio can deviate significantly from the true odds ratio.
- Careful consideration of selection bias is essential for accurate etiologic inferences in birth defect research.