Methodological issues in case-control studies: validity and power of various design/analysis strategies

Area of Science:

• Epidemiology
• Biostatistics

Background:

• Case-control studies are crucial for investigating chronic diseases.
• Estimating relative risk requires careful consideration of study design and analysis strategies.

Purpose of the Study:

• To evaluate the statistical power and efficiency of various case-control study designs and analysis methods.
• To determine conditions for consistent relative risk estimation in general models.
• To assess the practical implications of design and analysis choices.

Main Methods:

• Computer simulations were employed to model different design and analysis combinations.
• Mathematical derivations were used to establish conditions for consistent relative risk estimators.
• Efficiency was measured by statistical power.

Main Results:

• The study found that suboptimal designs or analyses often result in small efficiency losses in real-world scenarios.
• Theoretical statistical considerations may be less critical than practical aspects in strategy selection.
• The conditions under which different analytical approaches yield consistent relative risk estimators were identified.

Conclusions:

• Matching in case-control studies for chronic diseases is generally not justified.
• Practical considerations should guide the selection of study designs and analysis strategies over purely theoretical statistical ones.
• The efficiency of various strategies is robust enough to accommodate minor deviations from optimal choices.