Ad-hoc method for causal estimation with binary outcomes, where 2nd stage regression is substituted with a logistic regression.
Genetic variants must satisfy the IV assumptions. Genes each exert a small effect on the exposure in order to approximately identify the same causal effect. Non-collapsibility of odds ratio means that population average rather than individual-level causal effects are estimated.
References
- Burgess S, Labrecque JA. Mendelian randomization with a binary exposure variable: interpretation and presentation of causal estimates. European journal of epidemiology 2018;33:947-952.
- Bowden J, Vansteelandt S. Mendelian randomization analysis of case-control data using structural mean models. Statistics in Medicine 2011;30:678-694.
- Palmer TM, Thompson JR, Tobin MD, Sheehan NA, Burton PR. Adjusting for bias and unmeasured confounding in Mendelian randomization studies with binary responses. International Journal of Epidemiology 2008;37:1161-1168.