BACKGROUND: Several treatment options are now available for rheumatoid arthritis patients that have inadequate response to methotrexate alone. These agents are different in terms of their efficacy, safety, cost, and ease of administration. This makes it essential to perform a cost-effectiveness analysis taking into account the important clinical and cost differences. This model focuses on two combinations with proven efficacy, in adequate well controlled trials, for patients with inadequate response to methotrexate. OBJECTIVE: Compare 2-year cost-effectiveness of two different treatment strategies, from the societal perspective, for rheumatoid arthritis patients with inadequate response to methotrexate: 1) Start patients on methotrexate (MTX) + leflunomide (LEF), 2) Start patients on MTX + tumor necrosis factor a (TNF-a). METHODS: A 2-year decision analysis model with four semiannual cycles was developed to estimate the average cost/QALY, and the incremental cost-effectiveness ratio (ICER) for the two options, for female patients with mean age of 50 years. The model input parameters such as; response rates, dropout rates, costs, and QoL values were obtained either from published literature, or from expert opinion. Univariate sensitivity analysis was conducted to estimate percent changes in ICER from the base case analysis with change in gender, age, response rates, drug costs, and other parameters. RESULTS: The 2-year base case average cost/QALY is $8,551 for patients started on MTX + LEF, and $19,340 for patients started on MTX + TNF-a. The base case ICER for MTX + TNF-a is $36,147. In the univariate sensitivity analysis, ICERs varied from $25,267 to $63,479. The model was extremely sensitive to change in drug costs, and to the method used for conversion of HAQ/QoL scores into QALYs. CONCLUSION: This 2-year cost-effectiveness model suggests that for patients with inadequate response to methotrexate alone, the combination of MTX + TNF-a is a cost-effective strategy compared to MTX + LEF.