Use of causal inference methods in case-control studies: a methodology review

Area of Science:

• Epidemiology
• Biostatistics
• Observational Studies

Background:

• Causal inference methods are widely adopted in cohort studies but less so in case-control studies.
• This review focuses on causal inference techniques specifically within the context of case-control study designs.
• Understanding these methods is crucial for advancing epidemiological research.

Purpose of the Study:

• To provide a comprehensive review of causal inference methods applied in case-control studies.
• To examine the practical applications of these methods in existing literature.
• To identify trends and gaps in the use of causal inference in case-control research.

Main Methods:

• A systematic literature search was conducted using the MEDLINE database.
• Original peer-reviewed case-control studies published between March 2014 and March 2024 were included.
• Methods such as intercept-adjustment, propensity scores, and doubly-robust estimators were investigated.

Main Results:

• Out of 418 identified studies, 23 met the inclusion criteria.
• Most studies utilized case-control matching and focused on incident cases.
• The covariate-conditional odds ratio was the primary parameter estimated, with 65% of studies adjusting for sampling bias.

Conclusions:

• The application of causal inference in case-control studies is still in its nascent stages.
• Further development and implementation of these methods, including addressing time-varying confounding, are recommended.
• Enhanced causal inference techniques can significantly improve the reliability of case-control study findings.