In the natural history of hepatitis C virus (HCV) infection, a person can clear the virus (transient) or progress to chronic infection, which can remain asymptomatic for decades before the development of liver disease. The diagnostic tests cannot distinguish between transient and chronic infections, and there may be different ways to define the chronic infection endpoint based on serial detection tests over time (e.g. detectable HCV over 6 months). Therefore, HCV vaccine trial planning can benefit from assessment of the endpoints that are based on the periodically observed HCV statuses at study visits. In this work, a multi-state model for transient and chronic infections is postulated based on the known HCV natural history, and the type I error and the power of log-rank tests for the vaccine efficacy against chronic infections are evaluated in the framework of composite hypotheses.