An introduction to instrumental variables--part 2: Mendelian randomisation

Area of Science:

• Epidemiology
• Statistical Genetics
• Causal Inference

Background:

• Randomised controlled trials (RCTs) are gold standard but often impractical or unethical for exposure studies.
• Instrumental variables (IV) analysis is increasingly used to infer causal effects from observational data.
• IVs act as proxies for treatment allocation, mimicking RCTs under specific assumptions.

Purpose of the Study:

• To outline the Mendelian randomisation (MR) approach, a specific type of IV analysis.
• To discuss the advantages and disadvantages of employing MR in epidemiological research.
• To provide a statistical primer on causal inference methods beyond RCTs.

Main Methods:

• Utilising genetic variants as instrumental variables to estimate the causal effect of an exposure on an outcome.
• Ensuring the chosen genetic variant (instrument) is associated with the exposure, not directly with the outcome, and independent of confounders.
• Applying MR principles to observational data to approximate the conditions of an RCT.

Main Results:

• MR can provide robust estimates of causal effects when traditional methods are limited.
• The validity of MR relies heavily on the strength and assumptions of the chosen genetic instrument.
• Understanding the assumptions and limitations is crucial for accurate interpretation of MR findings.

Conclusions:

• Mendelian randomisation is a valuable tool for causal inference in epidemiology, especially for non-modifiable exposures.
• Careful selection and validation of genetic instruments are paramount for reliable results.
• MR analysis offers a complementary approach to RCTs for investigating health-related exposures.