In this paper, we have investigated the differences between the Cox and logistic regression models in cohort and case-cohort studies using simulations and SNP–CHD associations from EPIC-CVD. In the simulations, we examined a wide range of scenarios, including varying the amount of censoring (over a 20-year follow-up period), the RAF of the SNP, the strength of the association and the number of events (by changing the cumulative disease incidence). The results from the EPIC-CVD case-cohort show similar findings to the simulations. We believe that this is the first study to specifically compare the Cox and logistic regression models in case-cohort studies.
