Mixed effect regression analysis for a cluster-based two-stage outcome-auxiliary-dependent sampling design with a...

Area of Science:

• Biostatistics
• Epidemiology
• Biomedical Research Methodology

Background:

• Two-stage designs offer cost-effectiveness in biomedical studies, especially when exposure variables are costly or difficult to measure.
• Recent advancements allow outcome-dependent sampling in one or both stages, enhancing efficiency and reducing study costs.
• Outcome-dependent sampling (ODS) and outcome-auxiliary-dependent sampling (OADS) are key strategies for efficient data collection.

Purpose of the Study:

• To develop a semiparametric mixed effect regression model for two-stage designs utilizing an outcome-auxiliary-dependent sample (OADS) scheme.
• To incorporate and account for cluster or center-effects within the study subjects.
• To enhance parameter estimation efficiency and reduce overall study costs.

Main Methods:

• Development of a semiparametric mixed effect regression model.
• Utilizing an estimated likelihood function for regression parameter estimation.
• Employing an outcome-auxiliary-dependent sample (OADS) scheme for second-stage data collection.

Main Results:

• The proposed two-stage OADS design with center-effects demonstrates greater study efficiency gains.
• The method effectively accounts for cluster or center-effects in the analysis.
• Simulation studies confirm the efficiency improvements compared to alternative sampling schemes.

Conclusions:

• The developed semiparametric mixed effect model provides an efficient approach for two-stage outcome-dependent sampling designs.
• The inclusion of center-effects and OADS further optimizes study efficiency and cost-effectiveness.
• The method is applicable to real-world biomedical data, as illustrated by the Collaborative Perinatal Project dataset.