Search

Objective: To estimate the comparative lifetime cost-effectiveness of sequenced therapy with tumor necrosis factor (TNF) antagonists as the initial therapeutic intervention for patients with early rheumatoid arthritis (RA)., Methods: Because patients with RA switch regimens many times throughout the course of disease, sequenced therapeutic interventions were modeled, continuing until the last effective agent failed or death occurred. The model used published clinical outcomes from short-term, randomized controlled trials. Direct treatment costs and costs of lost productivity were modeled for each of 5 alternative treatment sequences. Incremental cost-effectiveness ratios are expressed as quality-adjusted lifeyears (QALY) gained., Results: Treatment sequences that included TNF antagonists produced a greater number of QALY than conventional disease modifying antirheumatic drug regimens alone. The cost-effectiveness of sequenced therapy initiated with adalimumab plus methotrexate (MTX) extendedly dominated both infliximab-plus-MTX-initiated and etanercept sequences. The cost of adalimumab plus MTX per QALY was US $47,157 excluding productivity losses, and $19,663 including productivity losses. A supplementary sequence that incorporated adalimumab-plus-MTX-initiated first-line therapy followed by another TNF antagonist as second-line therapy was modeled; this sequence resulted in additional QALY gained and extendedly dominated all single-TNF strategies., Conclusion: Of the 3 single-TNF antagonist sequences, the adalimumab-plus-MTX-initiated sequence was cost-effective in producing the greatest number of QALY. Multiple TNF strategies, such as the supplementary sequence modeled in this analysis, may be cost-effective in producing even greater health gain.

Objective: To estimate relative efficiencies of the 7 rheumatoid arthritis (RA) Core Data Set measures to distinguish adalimumab from control treatments in 4 clinical trials., Methods: Four adalimumab clinical trials were analyzed for arithmetic and percentage changes for each Core Data Set measure from baseline to endpoint: 3 assessor/physician measures -- swollen joints, tender joints, and global estimate; 1 laboratory test -- C-reactive protein; and 3 patient measures -- physical function, pain, and global estimate. Relative efficiencies of each measure to distinguish adalimumab from control group responses were assessed, with tender joint count as the referent measure., Results: Relative efficiencies were in a similar range for physician/assessor, patient, and laboratory measures, with some variation between trials. Among physician/assessor measures, relative efficiencies for global estimates were greater than for swollen and tender joint counts in 8/8 comparisons. Among patient measures, relative efficiencies for global estimates were greater than for physical function and pain scores in at least 6/8 comparisons. Among all measures, relative efficiencies for patient global estimates were greater than for swollen joint counts in 5/8 comparisons, and for tender joint counts in 8/8 comparisons., Conclusion: Patient and physician/assessor measures distinguished adalimumab from control treatment groups in similar ranges. Among all measures, physician/assessor global estimate was most efficient, and tender joint count least efficient, in all 4 trials. This information suggests that while joint counts are the most specific measure to assess RA, their sensitivity to detect treatment effects in patients with RA is generally no greater, and usually less, than other measures.

Objective: To analyze 2 indices composed of the 3 patient reported outcomes (PRO) in the American College of Rheumatology (ACR) Core Data Set--physical function, pain, and global estimate--without joint count or laboratory data, for capacities to distinguish active from control treatments in 4 pivotal clinical trials., Methods: Data from 4 clinical trials involving adalimumab, in combination with methotrexate or other disease-modifying antirheumatic drugs (DMARD) or as monotherapy, versus control treatment were made available to analyze properties of various indices. A categorical PRO-Index M was defined as "majority" improvement in 2 of the 3 PRO measures at 20%, 50%, and 70% levels; results were evaluated to analyze agreement with ACR20, ACR50, ACR70 responses and an "all Core Data Set measures" index based on 4 of the 7 measures having such levels of improvement. A continuous PRO-Index C was defined as the median or 2nd highest of 3 percentage differences from baseline to endpoint; results were evaluated to analyze agreement with a continuous ACR-N, "all Core Data Set measures" index, and Disease Activity Score 28 (DAS28)., Results: All indices distinguished active versus control treatment at similar levels, including PRO-Index M versus ACR20, ACR50, and ACR70 responses, and PRO-Index C versus DAS28., Conclusion: PRO indices based only on patient questionnaire data, without joint counts or laboratory tests, may be useful quantitative measures of therapeutic efficacy for use in standard rheumatology clinical care.