The rapid advancement of biomedical implant devices has raised new challenges to the failure analysis and reliability study of such new technologies. Based on the physics-of-failure study of stents, we develop a probabilistic reliability model that considers the stochastic processes presumed to describe two failure-generating processes, as well as the statistical interdependence between them due to exposure to the same external shock process. Different activity levels of patients may result in different levels of shocks and shock damages. Our new reliability models take into account the different effects from the shock process, and the statistical dependence between two failure processes for different patient groups. Based on the respective reliability models, customized maintenance strategies are then developed for different patient groups, with the aim to minimize the impact of unforeseen failures. A numerical example using data from literature is presented to illustrate the derived reliability and maintenance procedures. The results indicate the difference between the inspection timetables for patients with different activity levels. The reliability and maintenance models developed in this paper can facilitate continued advancement of implant devices, and provide fundamentally new perspectives on the application of reliability concepts to evolving medical devices.