Two-stage estimation in copula models used in family studies

Area of Science:

• Biostatistics
• Survival Analysis
• Statistical Modeling

Background:

• Register-based family studies often involve large cohorts, necessitating efficient statistical methods.
• Copula models are crucial for analyzing multivariate failure time data with dependencies.

Purpose of the Study:

• To develop and analyze a two-stage estimation procedure for copula models in multivariate failure time data.
• To generalize existing estimation approaches for parametric and semi-parametric models.
• To derive methods for analyzing sampled cohorts from large family studies.

Main Methods:

• Two-stage estimation procedure for copula models.
• Derivation of asymptotic properties for parametric and semi-parametric estimators.
• Development of methods for analyzing sampled cohorts.
• Simulation studies to assess estimator efficiency.

Main Results:

• The proposed two-stage estimation procedure demonstrates high efficiency.
• Asymptotic properties of estimators are derived, generalizing previous work.
• A method for analyzing sampled cohorts from large studies is successfully developed.

Conclusions:

• The developed methods provide efficient estimation for copula models in multivariate failure time data.
• The techniques are suitable for large-scale register-based family studies and sampled cohorts.
• The methods were effectively applied to a twin mortality study.