Objectives: The aim of this study was to estimate the cost-effectiveness of rituximab in patients not responding adequately to the first tumor necrosis factor (TNF) inhibitor using a model constructed to predict resource consumption and health outcomes in a population-based registry of biological treatments in Southern Sweden (SSATG).

Methods: The model was developed as a discrete event simulation model, using SSATG data for the years 1999–2007. The data set included 1,903 patients with complete data on treatments (up to three treatment lines), functional capacity (HAQ), disease activity (DAS28), and utility (EQ-5D). Resource consumption was based on a regular population-based survey of patients in Southern Sweden. Rituximab was incorporated as second line treatment, using effectiveness data for the active group (N = 311) in a clinical trial comparing rituximab to placebo (REFLEX). It is thus compared to the mix of second line biologics used in SSATG. The analysis starts after failure of the first TNF inhibitor. Results are reported as costs (€2008) per quality-adjusted life-year (QALY; both discounted 3 percent), for the societal perspective in Sweden.

Results: Total costs in the rituximab strategy are estimated at €401,100 compared with €403,000 in the TNF-inhibitor arm. Total QALYs are 5.98 and 5.78, respectively. The findings were found to be robust in extensive sensitivity analysis.

Conclusions: In our model, a strategy where rituximab is used as second line treatment after failure of the first TNF inhibitor provides a small saving (essentially due to the lower price of rituximab) and a QALY gain (due to better effect than the mix of second line TNF inhibitors).