Multivariate analysis vs Proportional hazard model
Hello, when we are exploring certain predictors for bad outcome (mortality), we usually run univariate, followed by multivariate models (including all univariate analysis with p < 0.20) with backward elimination, where we report outcomes of this final model with specific OR including CI and p values.Now recently , when I explored proportional hazards models, as we do have for all patients time to ...