The article introduces a new method called Extended Information Criterion (EIC) to compare different linear mixed effects models in clinical data analysis. This method helps researchers estimate variance components more accurately when dealing with complex data structures. The EIC approach, which is based on AIC and uses bootstrap techniques, allows for a more straightforward comparison of models with varying mean and covariance structures. The researchers conducted simulation studies and applied the EIC method to real clinical data sets, demonstrating its effectiveness in improving model comparisons and decision-making in data analysis.