1 Centre for Applied Health Services Research, Department of Public Health, Det Sundhedsvidenskabelige Fakultet, SDU2 Det Sundhedsvidenskabelige Fakultet, SDU3 Health Economics, Faculty of Business and Social Sciences, SDU4 Faculty of Business and Social Sciences, SDU5 COHERE - Center for Health Economic Research, Department of Public Health, Det Sundhedsvidenskabelige Fakultet, SDU6 unknown7 Health Economics, Faculty of Business and Social Sciences, SDU8 COHERE - Center for Health Economic Research, Department of Public Health, Det Sundhedsvidenskabelige Fakultet, SDU
Background: When estimating the cost of biological treatment many analyses rely on cross sectional data or standard consumption patterns indicated in the manufacturers' instruction leaflet. Unless such consumption patterns truly reflect routine clinical practice they may result in wrong assumptions about the true long run cost. Taking the actual medication practice into account is important for the evaluation of the costs and optimal sequencing of new and existing biological treatments. Objectives: To investigate the drug cost of TNF-inhibitors in the treatment of RA using real-life data from a nationwide clinical database (DANBIO) and use an incidence approach to compare the expected annual dosages and costs of the three most common treatments. Methods: 2438 RA patients from the DANBIO database who started TNF-Inhibitor therapy between January 2003 and August 2008 were included in the study. Consumption of TNF-Inhibitors and the associated costs were accumulated for each patient and assessed at the last observed visit. Missing data on dosage and frequency were either extrapolated from individual specific data, if available, otherwise standard dosages and frequencies were assumed. For infliximab an average dosage per kg body weight ratio was calculated from non-missing observations for reduced, standard and increased treatment regimes respectively and applied to patients with missing data on dosage. The annual dosage, duration and treatment costs were analysed at the end of each treatment line using ordinary least square (OLS) or Cox regression analysis with the type of biological drug, age, gender, disease duration, functional status and disease activity at treatment start as independent variables. Comparisons between TNF-inhibitor therapies were analysed using t-tests. Results: During the first year the true average annual consumption of TNF-inhibitors was higher than indicated by the standard treatment for all three TNF-alpha inhibitors. Therefore the associated costs exceeded the expected annual costs. For infliximab the average costs remained significantly above the expected standard treatment regime costs after the initiation period and increased with the number of treatment lines. Conclusion: The current consumption patterns of TNF-Inhibitors in Denmark indicate that the drug costs for adalimumab and etanercept are similar but exceed the drug costs for infliximab after the first year of treatment. Cost estimates based on short term observational data or on instruction leaflets from manufacturers may provide wrong cost assessments of TNF-alpha therapy. It is important to take the long term cost structure into account to arrive at unbiased treatment cost estimates.
Annals of the Rheumatic Diseases, 2009, Issue suppl(3)
Main Research Area:
EULAR 2009 European League against Rheumatoism Congress