Nonparametric estimation of cause-specific cross hazard ratio with bivariate competing risks data

We propose an alternative representation of the cause-specificcross hazard ratio for bivariate competing risks data. The representationleads to a simple plug-in estimator, unlike an existing ad hocprocedure. The large sample properties of the resulting inferencesare established. Simulations and a real data example demonstratethat the proposed methodology may substantially reduce the computationalburden of the existing procedure, while maintaining similarefficiency properties.