Note from the National Guideline Clearinghouse (NGC): The National Institute for Health and Clinical Excellence (NICE) commissioned an independent academic centre to perform a systematic literature review on the technology considered in this appraisal and prepare an Evidence Review Group (ERG) report. The ERG report for this technology appraisal was prepared by Liverpool Reviews and Implementation Group, University of Liverpool (see the "Availability of Companion Documents" field).
Clinical Effectiveness
Critique of Manufacturer's Approach
The manufacturer's submission (MS) includes a systematic review (SR) of the clinical evidence available to assess the efficacy and safety of abatacept for the treatment of patients with moderate to severe rheumatoid arthritis (RA) who have failed a tumour necrosis factor alpha inhibitor (TNFi).
Key aspects of the methodological quality of the manufacturer's review of the clinical literature were assessed based on an accepted quality assessment tool and the results are summarised in Table 4-1 of the ERG report (see the "Availability of Companion Documents" field).
Description and Critique of Manufacturers Approach to Validity Assessment
The MS includes a completed validity assessment and a JADAD score of five for the ATTAIN trial, the only randomised controlled trial (RCT) that met the review inclusion criteria. The validity assessment tool used is not referenced but the questions are appropriate and complete.
The ERG agrees that the validity assessment tool used in the MS was appropriate and that all trials were of a good quality. The completed validity assessment tool for ATTAIN as reported in the MS is reproduced in Table 4-5 of the ERG report (see the "Availability of Companion Documents" field).
Describe and Critique the Statistical Approach Used
The ATTAIN trial was powered to 96% to detect a 20% change for the primary outcome of American College of Rheumatology (ACR)20 and 87% to detect an 18% change in Health Assessment Questionnaire (HAQ) scores. For binary measures, Cochran-Mantel-Haenszel chi-square tests with stratification based on baseline history of TNFi treatment (current or prior use) were used. For continuous measures, an analysis of covariance was used, with treatment as the main factor and baseline measures as the covariate. All statistical tests and confidence intervals were two sided. Subgroup analyses were not sufficiently powered to detect a difference. All statistical methods were fully reported for each of the trials.
Refer to Sections 4.1 and 4.2 of the ERG report (see the "Availability of Companion Documents" field) for more information.
Economic Evaluation
Data Extraction
The manufacturer presented summary details of the cost-effectiveness studies (n=10) which described (1) abatacept in any country context or (2) any other biologic used in the UK setting. All of the economic analyses and quality of life studies are also summarised (including details of study, aims, methods, results and comments/relevance) in MS.
Data were extracted into pre-specified tables by one reviewer. A second reviewer conducted independent data abstraction and any discrepancies were discussed.
Quality Assessment
The results of each of the studies were discussed in light of the critical appraisal of its methodology. The specific critical appraisal tool employed was not stated.
Overview of Manufacturer's Economic Evaluation
In the absence of UK-based economic evaluations of abatacept, the manufacturer conducted a de novo economic evaluation. The principal analysis compares abatacept + methotrexate (MTX) versus MTX. An additional analysis compares abatacept versus a cycled TNFi. An economic model was developed to estimate the costs and outcomes of typical RA patients from the beginning of a specific treatment, after having failed a TNFi, until death. The model structure reflects the clinical outcomes of a phase III RCT of abatacept (ATTAIN), published economic evaluations, and expert opinion from clinicians, statisticians and health economists.
The manufacturer constructed a patient-level state simulation model which focuses on a hypothetical cohort of 10,000 patients. Patient disability is simulated over time using six-monthly cycles. Each patient in the hypothetical cohort is "run through" the model, one at a time, to estimate outcomes for the cohort as a whole. The nature of RA is modelled at the patient level in terms of changes in HAQ scores over time. The model estimates the worsening of HAQ scores due to underlying disease progression and treatment discontinuation. The model can be run for different durations up to lifetime duration.
Sensitivity Analyses (SA)
Univariate SA and probabilistic sensitivity analysis (PSA) were conducted by the manufacturer. Univariate SA was performed on a range of key parameters and the results are presented in the MS. In the base-case and additional analyses, the cost-effectiveness results appear to be most sensitive to the following parameters: time horizon, discount rate, annual treatment cost of abatacept and assumption on rebound following treatment discontinuation. In addition, the cost-effectiveness results in the abatacept + MTX versus MTX comparison appear to be sensitive to the annual rate of HAQ progression on MTX.
Refer to Section 5 of the ERG report (see the "Availability of Companion Documents" field) for additional information.