This is a critical abstract of an economic evaluation that meets the criteria for inclusion on NHS EED. Each abstract contains a brief summary of the methods, the results and conclusions followed by a detailed critical assessment on the reliability of the study and the conclusions drawn.

This study assessed the cost-effectiveness of dabigatran etexilate and warfarin in patients with non-valvular atrial fibrillation and a moderate-to-high risk of stroke. Dabigatran 110mg was not cost-effective. Dabigatran 150mg could be cost-effective, compared with warfarin, but this was uncertain. It was more cost-effective for patients at increased risk of stroke or whose international normalised ratio was less well controlled. The methods were valid and satisfactorily presented. The authors' conclusions appear to be robust.

Cost-utility analysis

This study assessed the cost-effectiveness of dabigatran etexilate 110mg or 150mg twice daily and warfarin in patients with non-valvular atrial fibrillation who were at a moderate-to-high risk of stroke.

The three preventive interventions were dabigatran 110mg twice daily, dabigatran 150mg twice daily, and warfarin. Patients who discontinued dabigatran because of a bleed or who discontinued warfarin for any reason were switched to aspirin. Patients who discontinued dabigatran for reasons other than bleeds were switched to warfarin.

UK/primary or secondary care.

Analytical approach:

The analysis was based on a discrete event model that considered various subgroups of patients over a lifetime. The authors stated that the perspective of the UK NHS was adopted.

Effectiveness data:

A literature review was undertaken in commonly used electronic databases to identify the sources for the clinical inputs. Most of the evidence for the treatment effect and safety came from the multinational, phase III, Randomized Evaluation of Long-Term Anticoagulation Therapy (RE-LY) trial, which included 18,113 patients who were each allocated to one of the three treatments. Additional data were from official statistics and a published meta-analysis of comparative trials. The reduction in the stroke event rate was a key input to the model. Assumptions on the long-term benefits of dabigatran etexilate had to be made, as there was a lack of available data.

Monetary benefit and utility valuations:

The utility values were from published sources, including the European Stroke Prevention Study, which used the European Quality of life (EQ-5D) questionnaire, the US Medical Expenditure Panel Survey, and a study of 83 patients with atrial fibrillation.

Measure of benefit:

Quality-adjusted life-years (QALYs) were used as the summary benefit measure and they were discounted at an annual rate of 3.5%.

Cost data:

The analysis included the costs of drugs, stroke, myocardial infarction, bleeds, pulmonary embolism, and transient ischaemic attack. These included ward care (staffing, equipment, consumables, and overheads) procedures (including hospital drugs), in-patient and out-patient care, visits by general practitioners and district nurses, and other drugs. Drug costs were from the British National Formulary. The costs of pulmonary emboli, transient ischaemic attacks, and bleeds were NHS reference costs. The cost of dyspepsia was included and was from a published study. The cost of warfarin and monitoring was from a micro-costing analysis of 165 patients. All costs were in UK pounds sterling (£) and were discounted at an annual rate of 3.5%. The price year was 2009.

Analysis of uncertainty:

One-way sensitivity analyses were carried out on all inputs, using published 95% confidence intervals or plausible ranges. Alternative scenarios were modelled: dabigatran 150mg was given until 80 years with dabigatran 110mg after 80 years old; a higher cost of managing intracranial haemorrhage and gastrointestinal bleeding with dabigatran than with warfarin; the benefit of drugs persisting for two years; and the benefit of drugs decreasing linearly to zero over 10 years. A probabilistic sensitivity analysis was performed, using Monte Carlo simulation, and cost-effectiveness acceptability curves were generated. Several subgroup analyses were carried out.

The total lifetime costs were £10,529 with dabigatran 110mg, £9,850 with dabigatran 150mg, and £6,480 with warfarin.

Compared with warfarin, dabigatran 110mg led to a gain of 0.094 (95% CI -0.083 to 0.267) QALYs and dabigatran 150mg led to a gain of 0.146 (95% CI -0.029 to 0.322) QALYs.

Dabigatran 110mg was dominated by dabigatran 150mg, as 150mg was more effective and less expensive. Compared with warfarin, the incremental cost per QALY gained was £43,074 with dabigatran 110mg and £23,082 with dabigatran 150mg.

In the scenario with dabigatran 150mg until 80 years then 110mg, the cost per QALY gained was £24,340 or £27,940 (depending on the data used), compared with warfarin. Dabigatran 150mg dominated this scenario. The sensitivity analyses confirmed the base-case findings; the incremental cost-utility ratio for dabigatran 110mg exceeded £32,415, compared with warfarin, in all sensitivity analyses.

The probabilistic sensitivity analysis showed that warfarin was the preferred treatment at cost-effectiveness thresholds of £24,400 or lower. Above this value, dabigatran 150mg was the optimal strategy.

In all the subgroups analysed, dabigatran 110mg was dominated by 150mg. The cost-utility ratio for dabigatran 150mg compared with warfarin was within the £30,000 threshold, except for patients in centres where the mean time that the international normalised ratio was within the therapeutic range was at least 65.5%. Dabigatran 150mg was particularly cost-effective in patients at a high risk of stroke.

The authors concluded that dabigatran 150mg could be cost-effective compared with warfarin, but this was uncertain. Low-dose dabigatran (110mg) was not cost-effective. High-dose dabigatran (150mg) was more cost-effective for patients at increased risk of stroke or for whom the international normalised ratio was less well controlled.

Interventions:

The rationale for the selection of the interventions was clear, as warfarin was the main oral anticoagulation treatment to prevent thrombosis, while dabigatran etexilate was a new oral direct thrombin inhibitor.

Effectiveness/benefits:

An appropriate approach was used to identify the relevant sources of evidence. The key methods of the literature review and the reasons for excluding some phase II trials were clearly reported. Most of the evidence was from a pivotal, phase III, head-to-head, clinical trial, which should have ensured its validity. It was acknowledged that the population included in this trial might not have been representative of the UK population, and several subgroup analyses were conducted to address this issue. Valid sources of data, such as a meta-analysis of clinical trials, were used for other parameters. Extensive sensitivity analysis was conducted on all clinical parameters. QALYs were a valid benefit measure and permit comparisons to be made with the benefits of other health care interventions. They capture the impact of the disease on both survival and quality of life, which are key dimensions of health for these patients. Some of the utility weights were from US patients as there was a the lack of reliable UK data.

Costs:

The cost categories and their sources were consistent with the perspective of the NHS. A clear description of each cost item was reported. Typical UK sources were used for the unit costs, and the resource quantities were from the pivotal clinical trial and other UK data. Except for dyspepsia, the adverse event rates did not differ between groups and their costs were not included; a sensitivity analysis was conducted to consider potential extra costs for dabigatran. Details, such as the price year and the discount rate, were considered.

Analysis and results:

The projected costs and benefits were reported and were synthesised, using an incremental approach, which allowed the identification of the best strategy. Extensive deterministic and probabilistic sensitivity analyses were carried out to investigate the uncertainty, and the methods and results were clearly reported and discussed. Appropriate subgroup analyses were performed. The discrete simulation model was a good design, but it was only partly described. The authors compared their results with those of other published studies that had shown that dabigatran was cost-effective. The transferability of the results was not discussed, but the extensive sensitivity analysis and subgroup analyses simulated various changes in parameters.

Concluding remarks:

The methods were valid and satisfactorily presented. The authors' conclusions appear to be robust.

Funded by the Medical Research Council, UK.