Application of Markov Models to Cost-Effectiveness Analysis in the Selection of Patients for Liver Transplantation

Background: Liver transplantation is the most effective curative treatment for patients with hepatocellular carcinoma. Due to the scarcity of cadaveric donor livers, several selection criteria have been established; however, these criteria are highly restrictive. In this study, we compare alternative selection tools with the standard selection criterion, the Milan Criteria. We conducted a cost-effectiveness analysis from the perspective of the U.S. healthcare system to determine which criterion provides the greatest benefit to the health system. Methods: An innovative non-homogeneous Markov model was developed to simulate the health trajectories of patients with hepatocellular carcinoma who underwent liver transplantation over five years. The model incorporated time-dependent transition probabilities, enabling the simulation to capture the evolving risks of recurrence and mortality. Transition probabilities, costs, and QALYs were obtained from published studies, while recurrence probabilities were estimated using the Kaplan–Meier method based on a cohort of 149 patients. We evaluated mean recurrence-free survival, life years gained, quality of life, and the incremental cost-effectiveness ratio (ICER) relative to the Milan Criteria. Results: HepatoPredict yielded the most significant benefits but incurred higher total costs than the other criteria. The ICERs of HepatoPredict Class I and Class II relative to the MC were $14,689.58/QALY and $39,542.98/QALY, respectively. Both values were below the cost-effectiveness threshold (U.S. GDP per capita: $81,632.25/QALY), indicating that HepatoPredict is cost-effective in the U.S. healthcare system. Conclusions: HepatoPredict stands out as the most cost-effective criterion and optimises organ allocation, an especially important consideration given the scarcity of donor livers. This represents a substantial advantage for healthcare institutions.