OBJECTIVES: To conduct a systematic review of economic models of pharmacologic stroke prevention in atrial fibrillation (SPAF). METHODS: We searched Medline, Embase, NHSEED and the Tuft,s Registry through May 2012. Included models assessed pharmacologic SPAF using a Markov process or discrete event simulation (DES), calculated both costs and effectiveness, and was published in English. Two investigators independently screened models and extracted data. RESULTS: Twenty- two models, published between 1995 and 2012, were identified. One model was a DES, and the remainder Markov models. Eleven models used a structure similar to Gage et al. (1995); five were derivatives of Sorensen et al. (2009), with the remainder using unique structures. Only 5 models had a non-CNS systemic embolism health state. Models typically started at 65 or 70 years and followed patients for their lifetime (e.g., ≥75 years of age). Inaccuracies in reporting of perspective existed; however, no model included indirect costs and all but one calculated qualityadjusted life-years (QALYs). Twenty models included warfarin; however, only 50% assessed the impact of INR control on conclusions. Most models included aspirin alone (73%), ten evaluated newer anticoagulants, and three evaluated clopidogrel+aspirin. Comparative efficacy and safety data for warfarin vs. aspirin/ control models were often derived from meta-analyses; whereas, data for newer agents came from a lone randomized trial. Models otherwise used similar sources of non-drug dependent inputs. Eighty-two percent of reported base-case incremental cost-effectiveness ratios (ICERs) were cost-effective (<$50,000/QALY). Models typically found warfarin (vs. aspirin/no therapy), dabigatran and rivaroxaban (vs. warfarin), and apixaban (vs. aspirin) to be cost-effective; data on clopidogrel+aspirin (vs. aspirin) to be conflicting, and genotyped-warfarin and ximelagtran not cost-effective. CONCLUSIONS: Cost-effectiveness models of pharmacologic SPAF have been extensively published; but none have estimated the comparative cost-effectiveness of newer agents. Models used similar structures and non-drug-specific inputs, and commonly find innovator strategies to be cost-effective.
Limone, B., Baker, W. L., Roberts, M., & Coleman, C. I. (2012). PCV48 Systematic Review of Cost-Effectiveness Models for Pharmacologic Stroke Prevention in Atrial Fibrillation. Value in Health, 15(7), A370–A371. https://doi.org/10.1016/j.jval.2012.08.991