An Application of the Cure Model to a Cardiovascular Clinical Trial

Intermediate events play an important role in determining the risk of a medical condition over time and should thus be accounted for in survival analysis. Myocardial infarction (MI) is one such condition whose hazard also depends upon the possible occurrence of an intermediate event—acute coronary syndrome (ACS). Accounting for the role that a possible ACS event plays in altering the hazard of MI becomes complicated when there is a cured fraction in the population. Data from the Clinical Outcomes Utilizing Revascularization and Aggressive Drug Evaluation (COURAGE) trial presents the scenario where the existence of a cured fraction is highly likely. In this article, we model the risk of developing an MI, while properly accounting for the effect/impact of a probable intermediate ACS event on that risk in the presence of a cured fraction. We adapt a maximum likelihood estimation approach to estimate the regression coefficients of this multi-part cure model. Simulation demonstrates satisfactory performance of the proposed estimator. We also utilize this dataset to explore the use of a proportionality constraint to help reduce the dimensionality of this multi-part model. The analysis yields novel findings that can be useful in guiding clinical practice.