Analysis type A decision tree model was used to assess the cost-effectiveness of warfarin intervention versus no antithrombotic therapy for nonrheumatic atrial fibrillation patients over one year. Model Construction of the tree was based on literature findings and modified to include minor bleeding events, which more accurately represents costs by distinguishing between major and minor bleeding events. The model is depicted in Figure 1. An expert panel of four cardiologists and clinical pharmacists from Sunnybrook Health Science Centre in Toronto provided input into the decision tree and the management of patients and patient outcomes. buy asthma inhalers Figure 1) Model for pharmacoeconomic evaluation of warfarin versus no antithrombotic therapy in atrial fibrillation. For pathway (boxed numbers) descriptions, see Appendix A APPENDIX A Explanation of pathways in Figure 1 *Costs from Table 4; fPathway probability from tree; *Cumulative probability multiplied by the cost There were six outcomes measured for each intervention (ie, warfarin and no antithrombotic therapy) including minor bleed, major bleed, fatal major bleed, embolus, fatal embolus and no event. Data collection Probabilities for each outcome were obtained from a metaanalysis to populate the model. This meta-analysis included all randomized, controlled studies published in English literature that assessed the use of warfarin in patients with paroxysmal and chronic nonvalvular atrial fibrillation (nvaf) for the prevention of stroke. A thorough literature search was conducted using the medline and ExcerptaMedica databases. The search was conducted on medline (January 1966 to June 1993) by intersecting the following key words: warfarin, atrial fibril-la tion and stroke (in clude cerebral vascu lar accidents). The search was conducted on Excerpta Medica (January 1988 to December 1992) by intersecting Coumadin (the name used for warfarin in Excerpta Medica), atrial and fibrillation. In addition, original and review articles found from the above search and tertiary references were used to ensure all references pertaining to the topic were found. The manufacturers of war farin in Canada (Dupont-Pharma; Merck-Frosst Canada Inc) were contacted for additional references.