The design and analysis of case-control studies with biased sampling.

Area of Science:

• Epidemiology
• Biostatistics
• Clinical Research Design

Background:

• Case-control studies are fundamental in epidemiology for investigating disease causes.
• Traditional case-control designs may face challenges with subject recruitment and efficient variable ascertainment.
• The need for optimized sampling strategies that account for potential confounding factors is critical.

Purpose of the Study:

• To propose a new case-control study design integrating disease status and screening variables for subject selection.
• To enable efficient estimation of effects for variables influencing recruitment, including potential confounders.
• To offer flexible sampling strategies for various research scenarios, such as rare exposures.

Main Methods:

• The proposed design utilizes a joint selection process based on disease status and readily available screening variables.
• Independent Bernoulli sampling schemes are employed, with investigator-defined recruitment probabilities.
• The design supports frequency matching and sample enrichment for specific subject categories.

Main Results:

• Two valid logistic regression analysis approaches are presented for the proposed design.
• Efficient estimation of effects for screening variables that influence recruitment is demonstrated.
• Asymptotic and simulation-based comparisons of estimator properties are provided for large and small samples.

Conclusions:

• The proposed case-control design offers a flexible and statistically sound approach to subject selection.
• This methodology enhances the efficiency of estimating effects in the presence of recruitment bias.
• The design provides valuable tools for optimizing epidemiological studies, particularly for rare diseases or exposures.