OBJECTIVES: The objective of this study was to do a cost-effectiveness analysis of infliximab for the treatment of rheumatoid arthritis (RA) in Swedish clinical practice, based on data from the STURE (Stockholm TNF-alpha follow-up registry). METHODS: Real world data from the STURE registry was implemented in a Markov cohort model, in which health states of disease severity were classified according to HAQ (five categories) and DAS28 (active or non-active disease). The transition probabilities between HAQ and DAS states in the treatment arm as well as discontinuation rates were based on data from the registry on patients taking infliximab as their first-line biological treatment. The data covered 637 patients with a mean follow up of 5.1 years since initiating infliximab therapy, giving a high degree of external validity to the results. After discontinuation, patients were assumed to revert to their baseline HAQ score and resume the natural disease progression rate. The transition probabilities in the comparator arm (no biologic treatment) were based on published literature on natural HAQ progression. The analysis had a societal perspective, i.e. all relevant costs were included irrespective of who incurred them, and efficacy was measured in quality-adjusted life-years (QALYs). In base case, patients were assumed to be 54 years at start of treatment and were followed for 20 years. RESULTS: Infliximab was associated with an incremental gain in QALYs of 1.06 and an incremental cost of SEK201,286 compared to natural progression, producing an incremental cost-effectiveness ratio (ICER) of SEK190 431. Sensitivity analyses of input parameters and model assumptions produced ICERs in the range from SEK150 000 to SEK400 000. CONCLUSIONS: The results indicated that it is cost-effective to treat patients with infliximab compared to no biologic treatment in patients with RA in Sweden.