Confidence intervals for the risk ratio under cluster sampling based on the beta-binomial model

Area of Science:

• Epidemiology
• Biostatistics
• Statistical modeling

Background:

• The risk ratio (RR) is a key metric in cohort studies for evaluating disease risk factors.
• Cluster sampling introduces intraclass correlation, potentially invalidating standard RR interval estimators.
• Accurate estimation is crucial for understanding disease etiology and intervention effectiveness.

Purpose of the Study:

• To develop and evaluate new interval estimators for the risk ratio (RR) that account for intraclass correlation in cluster sampling.
• To compare the performance of these novel estimators against existing methods using Monte Carlo simulations.
• To provide a practical method for analyzing clustered data in epidemiological research.

Main Methods:

• Development of four asymptotic interval estimators for the RR based on the beta-binomial model.
• Application of Monte Carlo simulation to assess the finite-sample performance of the proposed estimators.
• Evaluation across various scenarios to determine robustness and accuracy.

Main Results:

• The beta-binomial model effectively accounts for intraclass correlation in clustered data.
• The interval estimator utilizing a logarithmic transformation demonstrated superior performance across most simulated situations.
• The developed estimators provide a more appropriate approach for RR estimation with cluster sampling.

Conclusions:

• The logarithmic transformation-based interval estimator is recommended for calculating risk ratios in cohort studies employing cluster sampling.
• Accurate estimation of risk ratios is vital for reliable epidemiological inference, especially when dealing with correlated data.
• The study provides valuable tools for researchers analyzing clustered health and behavioral data.