Familial relative risk estimates for use in epidemiologic analyses

Area of Science:

• Epidemiology
• Biostatistics
• Genetics

Background:

• Traditional family history assessments for disease risk often overlook crucial factors like family structure and age at disease onset.
• The widely used Family History Score, while accounting for some factors, does not estimate familial relative risk but rather statistical significance.

Purpose of the Study:

• To develop and evaluate an empirical Bayes method for estimating familial relative risk.
• To compare this new method against existing family history summaries, including the Family History Score.

Main Methods:

• Utilized a two-level hierarchical model within an empirical Bayes framework.
• The first level models familial relative risk, and the second level uses a Poisson count for affected relatives based on this risk.
• Applied the methodology to a population-based case-control study of breast cancer.

Main Results:

• The new empirical Bayes estimates of familial relative risk showed a stronger association with case-control status than traditional methods.
• The proposed method more effectively identified effect modification of environmental risk factors by familial relative risk.
• Demonstrated superior performance compared to the Family History Score and other common family history summaries.

Conclusions:

• The empirical Bayes approach offers a more robust and informative estimate of familial relative risk.
• This method enhances the understanding of genetic contributions to disease and their interaction with environmental factors.
• The findings suggest this new methodology can improve risk assessment and identify gene-environment interactions in epidemiological studies.