Implementation of a Microsimulation Model for the Clinical Pathway of Liver Transplantation in Italy Using Stata/MP©

OBJECTIVES: Liver transplantation (LT) remains the optimal therapeutic option for various liver complications and diseases. Due to the complexity of patients’ follow-up, deterministic cohort models are inadequate to reflect the decision problems related to LT, making individual-based models the most appropriate choice. This study aims to outline the implementation of a microsimulation model using the software Stata/MP© to replicate the clinical pathway of patients undergoing liver transplantation in Italy and estimate their associated lifetime costs.

METHODS: Following a pragmatic literature review to conceptualize the patient flow post-transplantation, a microsimulation model was developed in Stata/MP©. The model simulated 10,000 individuals who started their clinical pathway post-surgery and were followed-up until death. It consisted of 12 one-month cycles in the early post-transplant period and up to 44 one-year cycles subsequently, for a maximum time horizon of 45 years. In each cycle, patients could be subject to different events with a time-varying probability of occurrence and were assigned the associated costs. Each cycle was modelled separately using an iterative approach, by updating patients’ characteristics, health, and cost outcomes with respect to the previous cycles, thanks to the ability of microsimulations to track each patient’s individual history. The direct healthcare lifetime cost of an average patient undergoing LT was obtained by running 1,000 simulations of the model, to allow for sample standard errors convergence.

RESULTS: The model was suitable to replicate the clinical pathway of liver transplantation and successfully provided statistically stable estimates of the average cost for a patient undergoing LT in Italy. Its implementation in Stata/MP demonstrated flexibility and the average running time of each 10,000 patients’ simulation was less than 4 seconds.

CONCLUSIONS: In the context of LT, the implementation of a microsimulation model using the software Stata/MP© was a valuable alternative to more conventional cohort modelling approaches in Microsoft Excel.