Article
Weighted Composite Time-To-Event Endpoints with Recurrent Events: Comparing Three Analytical Approaches
Search Medline for
Authors
-
Ann-Kathrin Ozga
- Institut für Medizinische Biometrie und Epidemiologie, Universitätsklinikum Hamburg-Eppendorf, Hamburg, Germany
-
Heiko Becher
- Institut für Medizinische Biometrie und Epidemiologie, Universitätsklinikum Hamburg-Eppendorf, Hamburg, Germany
-
Geraldine Rauch
- Institute of Biometry and Clinical Epidemiology, Charité – Universitätsmedizin Berlin, corporate member of Freie Universität Berlin, Humboldt-Universität zu Berlin, Hamburg, Germany
| Published: | February 26, 2021 |
|---|
Outline
Text
Background: In many clinical studies the interest lies in the comparison of a treatment to a control regarding a time-to-event endpoint like time to myocardial infarction or time to death. Commonly, only one of those endpoints is considered in an analysis. A composite endpoint is an alternative approach where endpoints can be considered jointly. Usually the time to the first occurring event for an individual is thereby analyzed. However, an individual may experience more than one non-fatal event. By including all observed events in an analysis the effect estimates are based on more complete information. Thus, analytical methods for recurrent events are of interest where several event types, often of different clinical relevance, are considered. In such a case, weighting the event types regarding their clinical relevance was proposed. Such weight-based methods include the Wei-Lachin multivariate procedure [1], the weighted hazards approach [2], [3], and Bakal's weighted composite endpoint [4]. There exists no systematic comparison of these methods.
Methods: We therefore provide a simulation-based comparison of methods for weighted composite endpoints combining events of different clinical relevance; one recurrent non-fatal and one fatal. The methods are also applied to data from the GENESIS (‘Genetic, Socio-economic and Inflammatory Determinants of Ischemic Stroke and their Interdependence’) study conducted in Ludwigshafen, Germany.
Results: For all approaches a closed formula test statistic is provided based on similar assumptions but different modeling ideas. For the Wei-Lachin approach and the weighted hazards method a weighted effect measure is described but not for Bakal's approach. Confidence intervals can only be easily gained for the Wei-Lachin effect measure. Differences in the empirical power can be seen as well as a smaller mean squared error for the weighted all-cause hazard ratio compared to the effect of the Wei-Lachin approach. The Bakal approach lacks understandable interpretation.
These results are supported by the application.
Conclusion: This general comparison and simulation study helps to understand the features of methods proposed for the analysis of composite endpoints combining (recurrent) events of different clinical relevance.
The authors declare that they have no competing interests.
The authors declare that a positive ethics committee vote has been obtained.
References
- 1.
- Lachin JM, Bebu I. Application of the Wei-Lachin multivariate one-directional test to multiple event-time outcomes. Clinical Trials. 2015;12(6):627-633.
- 2.
- Rauch G, Kunzmann K, Kieser M, Wegscheider K, Koenig J, Eulenburg C. A weighted combined effect measure for the analysis of a composite time-to-first-event endpoint with components of different clinical relevance. Statistics in Medicine. 2018;37(5):749-767.
- 3.
- Ozga A, Rauch G. Introducing a new estimator and test for the weighted all-cause hazard ratio. BMC Medical Research Methodology. 2019;19(118):1-16.
- 4.
- Bakal J, Westerhout C, Armstrong P. Impact of weighted composite compared to traditional composite endpoints for the design of randomized controlled trials. Statistical Methods in Medical Research. 2015;24(6):980-988.
