Concordance correlation coefficients estimated by modified variance components and generalized estimating equations...

Area of Science:

• Biostatistics
• Longitudinal Data Analysis

Background:

• Existing variance component estimators for longitudinal overdispersed Poisson data are biased.
• Generalized estimating equations are not typically applied to longitudinal overdispersed Poisson data for agreement estimation.

Purpose of the Study:

• To propose a modified variance component approach for unbiased concordance correlation coefficient estimation in longitudinal overdispersed Poisson data.
• To develop generalized estimating equation indices for intra-, inter-, and total agreement in longitudinal count data.

Main Methods:

• Developing unbiased estimators for concordance correlation coefficients using a modified variance component approach.
• Extending generalized estimating equations to handle correlation structures in longitudinal count data.
• Conducting simulation studies to compare the proposed methods.

Main Results:

• The modified variance component approach yields unbiased estimators with low mean square errors and accurate coverage rates.
• The generalized estimating equation approach demonstrates flexibility in modeling correlation structures for satisfactory agreement estimation.

Conclusions:

• The modified variance component approach is highly effective for estimating agreement in longitudinal overdispersed Poisson data.
• Generalized estimating equations offer a flexible alternative for agreement estimation in longitudinal count data.