Recurrent Event Data Approaches in Cardiovascular Outcome Trials: A Case Study

• Activity Number: 117
• Type: Topic Contributed
• Date/Time: Monday, August 5, 2013 : 8:30 AM to 10:20 AM
• Sponsor: Biopharmaceutical Section
• Author(s): Mouna Akacha*+
• Companies: Novartis
• Keywords: Recurrent events ; time to event analysis ; dependent censoring ; cardiovascular outcome trials

Abstract

Cardiovascular outcome trials usually assess composite primary endpoints with a mortality and a morbidity component (disease-related hospitalizations). Traditionally, these studies have been analyzed using a time-to-first-event analysis approach. A recurrent event approach, in which all events are included, i.e., death as well as recurrent, disease-related hospitalizations, more accurately captures the patient's clinical course and better reflects the true burden of the illness on the patient and the healthcare system. From a statistical point of view, such an analysis additionally provides a more efficient alternative to the classical approach. The price to pay for these advantages, are the numerous statistical challenges which arise with a recurrent event analysis. With a non-negligible rate of death and expected positive correlation between death and disease-related hospitalizations, a core question in the analysis is the handling of death which causes dependent censoring. Various analysis methods will be applied to a cardiovascular outcome study. The merits and drawbacks will be discussed from a statistical and clinical point of view.