OBJECTIVES: The fibrosis progression of chronic hepatitis C may crucially influence the cost-effectiveness of HCV antiviral agents because advanced fibrosis leads to more frequent hepatocellular carcinomas and progression to liver cirrhosis. The aim of this study was to evaluate the impact of the difference of natural history models on the cost-effectiveness of antiviral agents for chronic hepatitis C genotype 1 patients. METHODS: We developed and evaluated the appropriateness of two types of Markov natural history models: chronic hepatitis model with fibrosis stages (F model) and a model without fibrosis stages (CH model). We used the values by the algorithm of multivariable meta-regression (F-mv model) from a published meta-analysis for the progression rates of the fibrosis. We conducted external validations of the models using two published cohort studies that included Japanese HCV patients, comparing the survival rates and cumulative incidences of LC. To evaluate the impact of the difference in the models on the cost-effectiveness of antiviral treatment for HCV patients, we compared the incremental cost-effectiveness ratios (ICERs) of treatment by the two hypothetical antiviral HCV agents: one with high sustained virological response (SVR) and high drug price, and the other with low SVR and low drug price. RESULTS: The F-mv model with appropriate calibration of the parameters yielded the closer predicted values compared to the actual values in both cohorts. The impact upon the absolute difference in ICER became greater as the difference in SVR became smaller, and the cost difference increased. CONCLUSIONS: Our results suggest that the difference in the natural history models has a significant impact on the QALYs of the natural history and the ICERs of the cost-effectiveness of antiviral treatments for HCV patients.