FRI0528 Using a validated algorithm to evaluate the effectiveness and cost per effectively treated patient for biologics for rheumatoid arthritis (RA) in patients with employer provided health insurance

Background Administrative claims contain detailed medication, diagnosis, and procedure data, but their lack of clinical outcomes for RA has limited their use in comparative effectiveness research. A validated claims-based algorithm uses a combination of adherence, dosing, and treatment modifications to estimate biologics’ clinical effectiveness as proxies for low disease or remission for RA.1 In the validation study, “effectiveness” was defined as reaching DAS 28 low disease activity or remission one year after initiating treatment. Objectives To implement this algorithm in a US managed-care database and calculate the cost per effectively treated patient among biologics approved for moderate to severe RA (etanercept, adalimumab, infliximab, golimumab, and abatacept). Methods Data were obtained from the commercially insured cohort in the Truven Marketscan Database. The cohort included patients with RA aged 18-63, initiating treatment between January 2007 and December 2010, without biologics 6 months before their first (index) treatment in the database, and continuously enrolled 6 months before and 12 months after their index biologic. Other disease indications for TNF therapy were excluded. The algorithm defines lack of effectiveness as: medication possession ratio < 80% (or fewer infusions/injections than specified on US label), increase in biologic dose or frequency interval, switching biologics, adding new non-biologic Disease Modifying Anti-Rheumatic Drugs, glucocorticoid, dose increase or initiation, or more than one parenteral or intra-articular glucocorticoid injection during follow-up. Drug and administration costs were obtained from allowed amounts on claims. Cost per effectively treated patient was calculated as total drug cost for all patients initiating treatment on a given agent divided by the number of patients in whom that agent was classified as “effective” by the algorithm. Results The cohort included 15,351 patients, with a mean age of 49.7 (SD 9.5) years and 78.3% were female. Algorithm effectiveness criteria were met in 30% of etanercept (n=6,374), 30% of adalimumab (n=4,661), 20% of infliximab (n=2,765), 27% of abatacept (n=1,338), and 29% of golimumab (n=213) patients in the first 12 months of treatment. Mean first year cost per “effectively treated patient” was lowest for etanercept ($49,952), followed by golimumab ($50,189), adalimumab ($52,858), abatacept ($71,866), and infliximab ($104,333). Conclusions Algorithm-defined effectiven