PIN151 Adaptation and Calibration of a Markov MODEL Framework in Hepatitis D

Abstract

Few models have been developed to assess the cost-effectiveness of treatment options for Hepatitis D virus (HDV) chronic infection. Due to a lack of specific data, models have required many assumptions, which have led to invalid results. The objective of this study was to validate an HDV model and characterize the related costs and outcomes associated in France. An existing Hepatitis C model was adapted using a calibration approach. Patient data for fibrosis (F0-F3) or compensated cirrhosis (CC) were entered into the model. Natural history of HDV was simulated through fibrosis progression, CC, decompensated cirrhosis (DC), hepatocellular carcinoma (HCC) and liver transplant (LT). Liver-related and general mortality were applied. Infection and treatment status (detectable or undetectable HDV RNA, with or without treatment and HBsAg loss) impacted disease progression by modulating transition probabilities (TP). TP were calibrated based on the observed rates of CC, DC and HCC in a population of sustained virologic response (SVR)/non-SVR and previously/never treated with Interferonα (IFNα) patients. TP related to LT were determined from the French Biomedicine Agency. Finally, the model was validated on alternative studies. Discrepancies led to new calibration of alternative TP, characterizing their uncertainty. French costs and utilities were applied to patients treated with Peg-IFNα with representative F0-F4 state distribution to generate cost-effectiveness results from a collective perspective with a 2.5% annual discount rate. Depending on the source for TP calibration, the model determined an average survival of 12.6 (11.6 - 13.1) years, corresponding to 9.7 (8.8 – 10.2) QALYs at a total cost of €42,318 (€38,067- €49,532). Adaptation and calibration of an HDV model based on disease specific studies led to valid estimates of HDV-related outcomes and costs. Calibration allowed to characterize the uncertainty around TP and model outcomes, thus accounting for the impact of geographic context and ethnicity on HDV outcomes.