PURPOSE: The aim of this explorative analysis was to estimate the incremental cost-effectiveness of a treatment strategy that involves adding rituximab as another option to the sequence of disease modifying anti-rheumatic drugs (DMARDs) for 10,000 (hypothetical) patients with rheumatoid arthritis (RA), over a 10-year time horizon. Rituximab is not currently licensed for treatment of RA in Canada but is already marketed for other indications.

METHODS: Two DMARD strategies were compared – with and without rituximab. A Markov model using patient cohorts was used to simulate patients through different treatment regimens for both strategies using Microsoft Excel, reflecting the course of RA disease progression. Clinical data (ACR20 response rates not adjusted for placebo responses, drug withdrawal rates and mortality data) were obtained from literature searches. Response data on rituximab + methotrexate as observed during the first 6 months of a phase II trial (Edwards et al, NEJM 2004) was used for the model. Drug costs were obtained from published list prices. Monitoring costs were calculated from recommendations provided in products monographs for each drug, in conjunction with procedure costs provided by the Ontario Schedule of Benefits for the year 2004. Costs and outcomes were discounted by 5% annually in accordance with provincial guidelines and univariate sensitivity analyses were conducted to validate the results.

RESULTS: In all scenarios analysed, the incremental cost-effectiveness ratio (ICER) was in the range of $5,000-$17,000 per additional person-year of ACR20 response. Sensitivity analyses showed that the model was most sensitive to the rituximab re-treatment rate.

CONCLUSION: Based on this explorative model, the additional use of using rituximab is expected to be in a range that would be considered cost-effective in the Canadian healthcare setting. This analysis suggests that rituximab offers extended clinical benefit (at a reasonable cost) when incorporated into a therapeutic management plan for patients with RA. This initial pharmacoeconomic model should be expanded to include long-term data when such data becomes available in the future.