In RCTs in patients with serious cardiovascular (CV) diseases, it is common to observe recurrent CV events as well as substantial mortality. Often the risk of death is dependent on patient’s recurrent event history and death is a competing risk for recurrent events. In such setting, CV events and mortality can be defined as a composite endpoint with two components further analyzed to demonstrate consistent treatment effects in components. For example, in the ATTR-ACT trial which supported FDA approval of tafamidis for a serious CV disease, the primary endpoint was a composite endpoint and the 2 components were analyzed separately: Poisson regression for CV hospitalizations; Cox regression for mortality. The drawback of separate analyses is that dependency between death and recurrent events isn’t accounted for. Especially, unbalanced informative censoring occurs when there are differential mortality rates between treatment arms. Joint frailty models (JFM) are promising because they model the dependency between recurrent events and death. We evaluate JFM and traditional methods in the unique context of recurrent events in the presence of death under various settings via simulation.