OBJECTIVES: Constructing a Markov model can be challenging as available data can be limited. In hepatitis C, more complex models that includes F0, F1, F2, F3 and F4 fibrosis states are required by HTA agencies. However transition probabilities (TP) can only be available for non-cirrhotic to cirrhotic states (simple model). Estimating separate TP for Fn to Fn+1 (Fn,n+1) can be challenging as Markov models are non-linear. The objective of this study was to estimate Fn,n+1 TP from a single non-cirrhotic (F0F3) to cirrhotic (F4) TP.  METHODS: Results in markov models are driven by sum of cycles spent in each state. Thus, the method was built to produce a similar number of cycles in the non-cirrhotic for both the simple and complex model. This was done by calculating: (1) the expected number of cycles in F0F3 before transitioning to F4 (E) in the simple model; (2) the minimum number of cycles from Fn to F4 adjusted for the fibrosis stage distribution; (3) the (I-T)^-1 matrix equal to the average number of cycles in each Fn state based on (1) and (2); (4) the T matrix corresponding to the transition probability for Fn to Fn+1. RESULTS: Based on a F0F3 to F4 TP of 0.04 and a fibrosis stage distribution of 23% F0 & F1, 27% F2 & F2, we obtained Fn to Fn+1 TP = 0.097. For both the simple and complex models E was equal to 25. However, the sum of cycles in the non-cirrhotic states after only 50 cycles were 20.12 and 21.92 for the simple and complex models respectively. Taking into account a 2% discounting the sums were 13.83 and 8.59.  CONCLUSIONS: Markov models are sensitive to their structure, even when properly fitting the TP. For HCV, changing from a simple to a complex model is not trivial.