G.Saka,"In Which OPOs Should an End-Stage-Liver-Disease Patient List?"
  • FENS
  • G.Saka,"In Which OPOs Should an End-Stage-Liver-Disease Patient List?"

You are here

Faculty of Engineering and Natural Sciences
FENS SEMINARS


In Which OPOs Should an End-Stage-Liver-Disease Patient List?

Görkem Saka
Department of Industrial Engineering, University of Pittsburgh


Abstract: There are nearly 17,000 end-stage-liver-disease (ESLD) patients in the US transplant waiting list, and transplantation is the only possible treatment for ESLD patients. Since livers are a scarce resource, the allocation of livers is highly competitive among these patients. Therefore, patients tend to join multiple waiting lists at different organ procurement organizations (OPOs) in order to increase their chances of receiving a liver. This practice is referred to as “multiple listing”. Multiple listing is a very controversial topic because of its potential effects on equity in patients’ access to transplantation. There has been controversial debate about banning multiple listing in 1988 and again in 1994 in US. However neither dispute came to a final vote.

Patients cannot be listed in every OPO due to various costs and restrictions, so they choose subsets of OPOs to be listed in. The subset of OPOs that a patient chooses to be listed in is an important decision to make since it determines the liver arrival probability matrix to the patient.

We explore the potential sets of OPOs a patient can be listed in using the branch-and-bound method. Imposing problem specific branching and bounding strategies, we implement an efficient algorithm. At each node of the branch-and-bound tree, the liver arrival matrix will be distinct and therefore the optimization problem to be solved will be different. We formulate the accept/reject decision faced by the patient at each node as a continuous-time, infinite-horizon Markov decision process (MDP). We construct the uniform equivalent discrete-time MDP at each node and solve this MDP using the value iteration algorithm. We discuss the structural properties and solve this problem computationally using clinical data.

Bio: Görkem Saka is a PhD student at the Industrial Engineering Department at University of Pittsburgh. She obtained her M.S. degree from the Industrial Engineering Department at University of Pittsburgh in 2005 and her B.S. degree from the Industrial Engineering Department at Bilkent University, Turkey in 2003. Her research and teaching interests are in the areas of medical decision making, stochastic programming, completely and partially observed Markov decision processes, game theory and simulation. Her publication “Use of dynamic microsimulation to predict disease progression in patients with pneumonia-related sepsis” in “Critical Care” journal was selected as the seventh most accessed paper in the journal in June 2007. She is a member of INFORMS and IIE.


August 15, 2007, 13:40, FENS G032