Abstract
Clinical endpoints such as stroke in cardiovascular trials or disease progression in oncology trials are often assessed with uncertainty. The conventional approach is to classify each potential endpoint as true or false by an endpoint adjudication committee via a voting procedure, and only include the first confirmed endpoint with a majority of votes for each patient in Cox regression analysis. To retrieve this uncertainty information, Snapinn (1998) proposed a weighted Cox regression model and showed substantial gain in power over the conventional approach. In this research note, we try to complement this work by (1) demonstrating the impact of adjudication on the conventional approach; and (2) providing a theoretical explanation for why the weighted method works better.