Estimating the utility of matching in case-control studies

Area of Science:

• Epidemiology
• Biostatistics

Background:

• Matching is a common technique in case-control studies to enhance validity and efficiency.
• Assessing the utility of matching in completed studies is essential for methodological evaluation.

Purpose of the Study:

• To examine methods for estimating the utility of matching in completed case-control studies.
• To differentiate between frequency-matching and pair-matching utility estimation.

Main Methods:

• Developing methods to estimate the utility of frequency-matching in completed studies using broad strata.
• Demonstrating the technique with a literature example.
• Showing the impossibility of estimating pair-matching utility in completed studies.

Main Results:

• Frequency-matching utility can be estimated if population stratum proportions are known.
• Pair-matching utility cannot be estimated from a completed study.
• Unmatched analysis on completed studies is inappropriate for estimating matching utility.

Conclusions:

• Estimating frequency-matching utility is feasible in completed studies with appropriate data.
• Pair-matching utility assessment requires different approaches than post-hoc analysis.
• Accurate reporting of stratum proportions is recommended for frequency-matched studies.