A model for deceased-donor transplant queue waiting times (Q2263381)

The paper presents the analysis and application of a queueing model for liver transplant patients of each ABO blood type, reflecting the sickest patient first aspect and allowing abandonments. It shows how queueing theory can produce a model generating the key performance measures: the likelihood of successful transplantation and that of abandonment or death while waiting. The queueing model includes a single server providing service to two classes of transplant requests served on the FCFT basis. The organ availability is a limiting factor, the service time is the interval from when the wait-listed patient reaches the head of the queue until an organ becomes available. Depending on the transplant urgency, two priority classes (HP and LP) are considered, the two classes having different preemptive priorities depending on the health state. The arrival processes to the two classes are of Poisson type with given rates. The patients from both queues can leave the list or die, those from the lower priority queue can join the higher priority one. The following performance measures are determined: the steady-state joint queue length probabilities, the cumulative distribution function and moments of waiting time in the case of successful completion for both priority queues, the same in the case of a promoted LP patient and the probabilities that a patient reneges prior to receiving an organ. The queueing model is a state-dependent quasi birth-and-death process, the computations are done by using the matrix-analytic technique. The model is checked using a liver transplantation wait-list data of a regional health centre of Canada.