|
Activity Number:
|
459
- Rethinking Intercurrent Events and Estimators Within the ICH E9(R1) Estimand Framework
|
|
Type:
|
Topic Contributed
|
|
Date/Time:
|
Wednesday, July 31, 2019 : 8:30 AM to 10:20 AM
|
|
Sponsor:
|
Biopharmaceutical Section
|
|
Abstract #304800
|
|
|
Title:
|
Some Thoughts on Recurrent Event Estimands and Estimators
|
|
Author(s):
|
Dong Xi* and Jiawei Wei and Tobias Muetze
|
|
Companies:
|
Novartis and Novartis and Novartis
|
|
Keywords:
|
Estimand;
Estimator;
Recurrent event
|
|
Abstract:
|
Recurrent event data are multivariate failure time data where the individuals experience repeated occurrences of the same type of events. In clinical trials where recurrent events are clinically meaningful and where treatments are expected to impact the first as well as subsequent events, Negative Binomial (NB) model and Anderson-Gill model with robust variance estimator (LWYY) are the most commonly used methods to evaluate the treatment effect. The recent ICH (2017) guideline emphasizes that trial objectives and statistical approaches (estimators) should be aligned by clearly defining the estimand of interest. Under this framework, we discuss estimands for various intercurrent events including non-terminal and terminal events. We then investigate the parameters targeted by NB and LWYY models to better understand whether they align with the estimand, both numerically and by simulation.
|
Authors who are presenting talks have a * after their name.
