|
Abstract:
|
In many multi-center, randomized, controlled clinical trials in the device industry, the primary endpoint may be a time to event for several important events such as myocardial infarction, revascularization, death, or cardiovascular death. There may be a need to analyze these data for intervals that begin at a time other than Day 0 as well as multiple occurrences of an event within the interval being considered. Correctly addressing these statistical issues is critical for proper interpretation of treatment effect in this setting. This investigation is to consider different model building strategies which will look at assessing a fitted Proportional Hazards model with possible selection of model covariates that would be appropriately used for this analysis.
|
