Matched case-control data analysis with selection bias

Area of Science:

• Epidemiology
• Biostatistics

Background:

• Case-control studies are valuable for hypothesis testing but are susceptible to selection bias.
• Selection bias in case-control studies can invalidate results and is difficult to assess.
• Existing methods for evaluating and correcting selection bias are limited.

Purpose of the Study:

• To propose novel bias-corrected estimators for case-control studies.
• To address the challenge of selection bias in epidemiological research.
• To provide a statistically sound method for improving the accuracy of case-control inferences.

Main Methods:

• Development of bias-corrected estimators using a joint estimating equation approach.
• Application of the method to data from the Northern Manhattan Stroke Study (NOMASS).
• Utilizing telephone survey data to inform selection probabilities and potential bias.

Main Results:

• The proposed joint estimating equation approach yields bias-corrected estimates.
• Standard statistical software can be used to obtain the bias-corrected estimate and its standard error.
• The methodology provides a practical tool for researchers dealing with selection bias.

Conclusions:

• The proposed bias-corrected estimators enhance the validity of case-control studies.
• This approach offers a feasible solution for mitigating selection bias in epidemiological research.
• The method facilitates more reliable inferences from case-control study designs.