Figure 2

Possible state transitions in the surveillance and re-treatment submodel. An individual-based Markov submodel simulates the follow-up testing which may lead to detection of asymptomatic recurrence, and the treatment which may occur after recurrence diagnosis. The arrows represent possible state transitions which could occur at each three-month time step. Each patient begins the model in the “No known recurrence” state wherein they undergo surveillance testing according to a sepcified schedule. If a true recurrence is discovered, the patient will proceed to one of the two states: “Recurrence curatively treated” or “Recurrence palliatively treated”, depending on whether the current model cycle occurs before or after U_i (U_i is assigned as a continuous value based on the disease progression submodel). Patients in the “Recurrence curatively treated” state continue to undergo surveillance testing after treatment. During each cycle, individuals may move to the “Dead due to other causes” state from any other living state based on a background transition probability of mortality from other causes. Additional file 1: Figure S1 provides a more detailed depiction of the contingencies which drive state transitions in the surveillance and re-treatment submodel.