Background Evidence on the comparative effectiveness of biologics used for rheumatoid arthritis (RA) is lacking. A validated claims-based algorithm had been developed to estimate the effectiveness of biologics for RA in administrative claims databases.
Objectives The objective was to implement a claims-based algorithm in a US managed care database to estimate the effectiveness of first-line biologics approved to treat moderate-to-severe RA (abatacept [ABA], adalimumab [ADA], certolizumab pegol [CER], etanercept [ETN], golimumab [GOL], infliximab [INF]).
Methods This is a retrospective cohort study using administrative data in the HealthCore Integrated Research Database (HIRDSM). The cohort included patients with ≥1 claim for the first-line biologics used for RA between Jul 1, 2009 and Jan 31, 2013. Patients also had to be aged 18-63, have ≥1 claim for RA, have no prior exposure to biologics for RA (ABA, ADA, anakinra, CER, ETN, GOL, INF, rituximab, or tocilizumab) or other conditions for which the aforementioned biologics were approved to treat during the 6 months pre-index, and be continuously enrolled between 6 months prior to and 12 months post-index. The first biologic use following 6 months of enrollment defined the index event and date. The algorithm defines effectiveness as meeting all of the following criteria: Medication possession ratio ≥80% (or meeting number of infusions/injections as specified on US label), no increase in biologic dose or frequency, no biologic-switching, no addition of new non-biologic Disease Modifying Anti-Rheumatic Drugs (DMARD), no initiation or increase of oral glucocorticoid use, and ≤1 parenteral or intra-articular glucocorticoid injection during the follow-up period. Chi-square or Fisher's exact tests were used to estimate the relative effectiveness of biologics in comparison to ETN. Pearson correlation coefficients were computed to estimate the correlation between each criterion.
Results A total of 4,844 patients were eligible for analysis (mean age 48.6, 76.4% female). Overall, 30.2% patients were classified as effectively treated [ETN (33.3%), ADA (30.7%), ABA (28.7%), GOL (26.2%), CER (21.6%), and INF (19.3%)]. Statistically significant differences (p<0.05) were found in comparison between ETN and CER (p=0.0097) and between ETN and INF (p<0.0001). The majority of patients failed the algorithm due to low adherence (57.8%), followed by biologic switch (17.6%) and adding a non-biologic DMARD (13.3%). With the exception that biologic switch had a “weak” correlation with low adherence (r=0.33), all other correlations between each criterion were “very weak” (r<0.20).
Conclusions Based on a validated claims-based algorithm, ETN users had the greatest likelihood of being effectively treated among first-line biologics users with RA in this managed care population. Despite this, real-world effectiveness appears to be low across biologics.
Acknowledgements Research funded by Immunex Corporation, a wholly owned subsidiary of Amgen Inc.
Disclosure of Interest T. Gu Grant/research support from: Amgen, N. Shah Shareholder of: Amgen, Employee of: Amgen, G. Deshpande Grant/research support from: Amgen, D. Tang Shareholder of: Amgen, Employee of: Amgen, D. Eisenberg Grant/research support from: Amgen, B. Stolshek Shareholder of: Amgen, Employee of: Amgen, D. Harrison Shareholder of: Amgen, Employee of: Amgen