Impact of Measurement Errors on Gene-Environment Interactions
Author Information
Author(s): Darren C. Greenwood, Mark S. Gilthorpe, Janet E. Cade
Primary Institution: University of Leeds
Hypothesis
How do measurement errors in covariates affect the estimation of gene-environment interactions?
Conclusion
Measurement error in confounders does not bias interaction coefficients but can bias estimated effects of exposure.
Supporting Evidence
- Measurement error in confounders can lead to biased estimates of exposure effects.
- The interaction regression coefficient remains unchanged under certain conditions.
- Statistical power to detect interactions is reduced by measurement error in exposures.
Takeaway
If we make mistakes when measuring things that affect health, it can change how we understand the relationship between genes and the environment.
Methodology
Simulated data was used to investigate the effects of measurement error in covariates on gene-environment interactions.
Potential Biases
Potential bias in estimating exposure effects due to measurement errors.
Limitations
The results are based on simulations and may not apply to all real-world scenarios.
Participant Demographics
Participants were from the UK Women's Cohort Study, aged 35-69.
Digital Object Identifier (DOI)
Want to read the original?
Access the complete publication on the publisher's website