WHEN SHOULD END-STAGE LIVER DISEASE (ESLD) PATIENTS UPDATE THEIR MODEL FOR END-STAGE LIVER DISEASE (MELD) SCORE?

Tuesday, October 26, 2010
Sheraton Hall E/F (Sheraton Centre Toronto Hotel)
Z. Gozde Icten, MS1, Lisa Maillart, PhD2, Andrew Schaefer, PhD2, Atul Bhandari, PhD3 and Mark S. Roberts, MD, MPP4, (1)Industrial Engineering Department, Pittsburgh, PA, (2)University of Pittsburgh, Pittsburgh, PA, (3)SmartOps Coorporation, Pittsburgh, PA, (4)University of Pittsburgh School of Medicine, Pittsburgh, PA
  

Purpose: UNOS ranks ESLD patients by their MELD score and requires patients to update their score at a frequency that ranges from weekly to yearly depending on the patient's last reported MELD score. Therefore a tradeoff exists between reporting a higher (sicker) MELD score to UNOS to improve a patient's position on the waiting list, but at the same time shortening the time between required MELD score updates which may force the patient to report a healthier MELD score if their condition improves. We investigate the question of how an ESLD patient should decide, over time, whether to update her MELD score with UNOS, accept an organ offered for transplantation, or continue to wait (do nothing) to maximize her total expected life days.   

Method: We formulate an infinite-horizon discrete-time Markov decision process model with daily decision epochs. The state space is comprised of the patient's actual MELD score, the MELD score last reported to UNOS, the number of days remaining until the next required update, and the quality of the current liver offer.  The health transition probabilities, liver offer probabilities and post-transplant life days are estimated using UNOS and University of Pittsburgh Medical Center data.    

Results: We establish conditions under which the optimal policy exhibits control limit structure in each state element except patient health. We also show that under certain conditions waiting is always suboptimal when patient is sicker than the last recorded MELD score. Lastly, we draw comparisons between our results and those of a base case scenario in which the patient's MELD score is reported to UNOS every day. We find that patients who do not update daily should refuse lower quality organs compared to those who update their MELD score every day (Figure 1), and those who do not update daily can realize an increase in life expectancy of as much as 13.65%.   

Conclusion: Our results suggest that under the current required updating frequencies, patients should always update sooner than required if their MELD score increases; but, somewhat counterintuitively, there are situations in which it is optimal to update even if the MELD score decreases.

Figure 1. Patients who do not update daily, refuse lower quality organs compared to   those who update their MELD score every day.  

Candidate for the Lee B. Lusted Student Prize Competition