Activity Number:
|
400
- Survival Analysis and Clinical Trials
|
Type:
|
Contributed
|
Date/Time:
|
Tuesday, August 1, 2017 : 2:00 PM to 3:50 PM
|
Sponsor:
|
Biometrics Section
|
Abstract #323882
|
View Presentation
|
Title:
|
Dealing with Competing Risks in Clinical Trials: What Are the Proper Questions to Ask and How Best to Answer Them?
|
Author(s):
|
James Troendle* and Eric Leifer and Lauren Kunz
|
Companies:
|
National Institutes of Health and NIH and NIH
|
Keywords:
|
Censoring ;
Clinical Benefit ;
Cumulative Incidence ;
Hazard
|
Abstract:
|
We investigate different approaches to evaluating efficacy in a clinical time to event trial when there is a competing event in addition to a primary event. We first consider two approaches based on separate causes of events. The first is to use cumulative incidence to estimate the probability of an event of a specific type. It is shown that this approach is not relevant to determining benefit of treatment. The second approach is to estimate the cause specific hazard ratios. It is shown that this approach has a potential drawback: estimates of the cause specific hazards are biased and this bias can be arbitrarily large. We then investigate the use of composite events. It is noted that a composite event approach is not affected by informative censoring from the competing event. However, if treatment affects the competing event differently than it affects the primary event, the resulting hazard ratio can be severely diluted by forming a composite. Thus, deciding between a cause specific analysis of the primary event and an analysis of a composite endpoint that includes the competing event is difficult and hinges on weighing the impact of the drawbacks of each approach.
|
Authors who are presenting talks have a * after their name.