Gems: An R package for simulating from disease progression models

Abstract

Mathematical models of disease progression predict disease outcomes and are useful epidemiological tools for planners and evaluators of health interventions. The R package gems is a tool that simulates disease progression in patients and predicts the e_ect of di_erent interventions on patient outcome. Disease progression is represented by a series of events (e.g., diagnosis, treatment and death), displayed in a directed acyclic graph. The vertices correspond to disease states and the directed edges represent events. The package gems allows simulations based on a generalized multistate model that can be described by a directed acyclic graph with continuous transition-speci_c hazard functions. The user can specify an arbitrary hazard function and its parameters. The model includes parameter uncertainty, does not need to be a Markov model, and may take the history of previous events into account. Applications are not limited to the medical _eld and extend to other areas where multistate simulation is of interest. We provide a technical explanation of the multistate models used by gems, explain the functions of gems and their arguments, and show a sample application.