A quantitative study of the bias in estimating the treatment effect caused by omitting a balanced covariate in survival models | Synapse