A log-linear model for binary pedigree data.

Area of Science:

• Epidemiology
• Statistical Genetics
• Biostatistics

Background:

• Familial aggregation of diseases is a key area in epidemiology.
• Log-linear models provide a framework for analyzing binary data in familial studies.
• Understanding disease concordance is crucial for genetic and public health research.

Purpose of the Study:

• To develop and apply a pedigree model for binary data to examine familial aggregation of disease status.
• To assess the impact of ascertainment and population prevalence on disease concordance estimation.
• To explore the incorporation of ancillary risk factors into familial disease models.

Main Methods:

• Developed a pedigree model motivated by log-linear modeling for binary disease data.
• Utilized odds ratios to express disease concordance between relatives.
• Applied the model to breast cancer data, considering ascertainment through an affected proband.
• Incorporated population prevalence estimates and ascertainment correction for accurate concordance estimation.

Main Results:

• Disease concordance estimation is critically dependent on accurate population prevalence and ascertainment correction.
• The developed model demonstrated the multiplicative nature of relative risks in pedigrees.
• The model's flexibility allows for the inclusion of ancillary risk factors, such as genetic markers.

Conclusions:

• The proposed pedigree model provides a robust method for assessing familial aggregation of diseases.
• Accurate estimation of disease concordance necessitates careful consideration of ascertainment bias and population disease probabilities.
• Future studies should collect comprehensive risk factor data, including genetic information, for all individuals in a pedigree.