To compare the relative risks and benefits of several clinical strategies for managing patients with chronic atrial fibrillation. Five recent randomized, controlled trials of warfarin in atrial fibrillation, 6 randomized, controlled trials of quinidine, and 13 longitudinal studies of low-dose amiodarone were used. A MEDLINE search was also done (1966 to present). A Markov decision analysis model was used to assess outcomes in large, hypothetical cohorts of patients with atrial fibrillation followed from 65 to 70 years of age within four clinical strategies: 1) no treatment; 2) warfarin; 3) electrical cardioversion followed by quinidine to maintain normal sinus rhythm; and 4) electrical cardioversion followed by low-dose amiodarone. In this hypothetical cohort, fewer patients had disabling events with amiodarone (1.4%) than with quinidine (1.8%), warfarin (2.6%), or no treatment (7.4%). Amiodarone appeared to be associated with the lowest 5-year mortality (13.6%) when compared with warfarin (14.4%), quinidine (15.2%), and no treatment (18.2%). In terms of quality-adjusted life-years, amiodarone had the highest expected value (4.75 years), followed by warfarin (4.72 years), quinidine (4.68 years), and no treatment (4.55 years). Amiodarone remained the preferred strategy using the most plausible scenarios of risks associated with atrial fibrillation. Choices among warfarin, quinidine, and no treatment depended on estimates of bleeding rates with warfarin, stroke rates after discontinuing warfarin, quinidine-related mortality, and the quality of life with warfarin. Cardioversion followed by low-dose amiodarone to maintain normal sinus rhythm appears to be a relatively safe and effective treatment for patients with chronic atrial fibrillation.