Background Biologic disease modifying antirheumatic drugs (DMARDs) entail considerable improvement in patient outcomes for patients with rheumatoid arthritis (RA), but the treatment is costly. Several model studies have explored the cost-effectiveness of TNF inhibitors (TNFi), but the results have been conflicting in that some studies conclude that biologics are cost-effective while others do not. Most models are based on clinical trial data with limited generalizability to real life outcomes.

Objectives To use real life register data to estimate the additional costs and additional health benefits of replacing synthetic DMARDs alone with a combination of TNF inhibitors (TNFi) and synthetic DMARDs.

Methods We developed a Markov model to explore two treatment strategies for patients with RA: TNFi plus synthetic DMARD (TNFi-strategy) and synthetic DMARD alone (synthetic-strategy). Patients in both strategies started in one of seven health states defined on the basis of SF-6D utilities (0.30-0.39, 0.40-0.49 …. 0.90-1.00). The patients could subsequently move to better or worse utility levels according to transition probabilities which varied over time. The model had a 10-year perspective and a cycle length of three months. In the TNFi-strategy, patients could stay on TNFi (incl. TNFi cycling), or switch to non-TNFi-biologics, synthetic DMARDs or no DMARDs. In the synthetic-strategy, patients remained on the same therapy, but could change utility levels. In both strategies we assumed a standardized mortality ratio of 1.54 using mortality statistics from Norway.

Transition probabilities for the TNFi-strategy were derived from the Norwegian longitudinal observational study on disease-modifying antirheumatic drugs (NOR-DMARD) using 810 RA-patients who started TNFi-treatment after failure on 2.1 synthetic DMARDs on average. Patients were followed from 2000 to 2011 with a mean follow-up time of 3.6 years. Data from Oslo Rheumatoid Arthritis register (ORAR) were used for the synthetic-strategy. All cost data (pharmaceuticals, hospital care, GP and physiotherapist visits, indirect costs, etc.) were taken from NOR-DMARD[1]. SF-36 data were taken from NOR-DMARD and ORAR and translated into SF-6D utilities. Future cost and outcomes were discounted at the nationally recommend rate of 4% per year. For the sensitivity analyses, Dirichlet distributions were used to express uncertainty in the Markov transition probabilities.

Results Ten-year discounted costs of a patient on the TNFi-strategy totaled €129,090 (€470,709 including indirect costs) while the synthetic-strategy led to costs of €52,765 (€415,434). The strategies entailed 5.42 and 4.78 quality adjusted life years (QALYs) (6.51 and 5.73 undiscounted), respectively, representing €118,802 per gained QALY (€86,079 when indirect costs were included). The latter ratio is below the suggested willingness-to-pay for a QALY (€40,300 to €107,600) in Norway. In probabilistic sensitivity analysis, the probability that the combined TNFi and synthetic DMARD is cost-effective (< €107,600 per QALY) was 100% when including indirect costs.

Conclusions TNFi treatment of RA in a 10-year perspective appears to be cost-effective when indirect costs are accounted for in a setting of real life data.

References

1. Kvamme MK et al. Rheumatology (Oxford).2012;51:1618-27.

Disclosure of Interest M. Kvamme Grant/research support from: A course in Paris, France 12-14 of September 2012 payed by Pfizer and organized by the International value coalition., E. Lie Consultant for: Speaker/consulting honoraria Pfizer, BMS, Abbott og Roche, T. Uhlig Consultant for: honorary/Speaker/consulting: Abbott, BMS, MSD, Pfizer, Roche, UCB, T. Kvien Grant/research support from: Research grants received to Diakonhjemmet Hospital from: Abbott, BMS, MSD/Schering-Plough, Pfizer/Wyeth, Roche, UCB, Consultant for: Consultancy fees: Abbott, BMS, MSD/Schering-Plough, Nicox, Pfizer/Wyeth, Roche, UCB, Schering-Plough, UCB, Wyeth, Speakers bureau: Speaker honoraria: Abbott, AstraZeneka, MSD/Schering-Plough, Nicox, Pfizer/Wyeth, Roche, UCB, Schering-Plough, UCB, Wyeth, T. Moger: None Declared, I. Kristiansen: None Declared