Abstract:
|
The adverse event is an important topic under safety analysis in clinical trials. It is usually summarized by descriptive statistics with the subgroups defined at baseline. When taking the time to the adverse event into consideration, the cox proportional hazards model incorporating the covariates is frequently used. In reality, however, adverse events could repeatedly occur on a subject during the whole treatment duration; meanwhile, using baseline values of either subgroups or variables of interest cannot accurately profile the changed situation after the start of the treatment. Therefore, the authors utilize the survival models for the recurrent events and time-varying variables in order to accommodating the recurrence of adverse events and reflecting the temporal change in the subgroups and variables. The presentation aims to compare the performance of the well-developed recurrent events survival models fitted with the proposed dummy variables which test for the drug-drug interaction effect base on simulation scenarios.
|