Socio-demographic and behavioural characteristics are associated with delayed diagnosis and disease progression in HCV-infected persons. However, many analyses focused on single variables rather than groups defined by several variables. We used latent class analysis to study all 4488 persons enrolled in the Swiss Hepatitis C Cohort Study. Groups were identified using predefined variables at enrolment. The number of groups was selected using the Bayesian information criterion. Mortality, loss to follow-up, cirrhosis, treatment status and response to antivirals were analysed using Laplace and logistic regressions. We identified five groups and named them according to their characteristics: persons who inject drugs, male drinkers, Swiss employees, foreign employees and retirees. Two groups did not conform to common assumptions about persons with chronic hepatitis C and were already in an advanced stage of the disease at enrolment: 'male drinkers' and 'retirees' had a high proportion of cirrhosis at enrolment (15% and 16% vs <10.3%), and the shortest time to death (adjusted median time 8.7 years and 8.8 years vs >9.0). 'Male drinkers' also had high substance use, but they were well educated and were likely to be employed. This analysis may help identifying high-risk groups which may benefit from targeted interventions.