Objective: The goal of this study was to assess the differences between an ex ante and an ex post cost-effectiveness analysis of Dabigatran etexilate vs VKAs for the prevention of thromboembolic events in non-valvular atrial fibrillation patients and to draw lessons on the design and use of real-world data for decision making.
Methods: The same model was used to calculate the cost-effectiveness ratio using two sets of parameters. One set included the efficacy and safety outcomes data from RE-LY, the pivotal trial comparing Dabigatran to warfarin; cost data came from an ex ante publication. Outcomes data for the second set came from real-world data studies. Cost data were a mix of real-world data and other sources. Two treatment strategies were compared: treatment initiation by either Dabigatran or VKAs, followed by either VKAs or Dabigatran. A crude comparison of results was performed; the impact of data differences was then assessed. Probabilistic sensitivity results of the two analyses were compared.
Results: With real-world evidence, Dabigatran at both dosages was more effective for the prevention of ischemic strokes, intra-cranial haemorrhages, with less major extra-cranial haemorrhages and a similar risk of myocardial infarction. Using clinical trial data, Dabigatran150 mg (resp. Dabigatran110 mg) as a first-line treatment vs VKAs yielded an ICER of € 8077/QALY (resp. € 13,116/QALY). Real-world evidence scenarios were cost-saving and more effective for both dosages.
Conclusion: The reassessment of outcomes and cost data had an impact on results, improving the efficiency of Dabigatran. We identify methodological issues which should be discussed if post-launch RWE based cost-effectiveness data become a standard in HTA decision making.
Keywords: Atrial fibrillation; Cost-effectiveness; Direct oral anticoagulants; Real-world evidence.