We performed a systematic review of the available evidence on the relationship between the individual clinical, echocardiographic and laboratory characteristics of patients with atrial fibrillation (AF) and the risk of stroke. A systematic review was also performed of all published stroke risk stratification models, as well as the accuracy of their discriminative ability between risk strata. Third, we reviewed the literature on cost-effectiveness analyses with oral anticoagulation in AF. From the systematic review on stroke risk factors, a prior stroke or transient ischemic attack (15/16 studies positive, risk ratio [RR] 2.86), hypertension (11/20 studies positive, RR 2.27), aging (9/13 studies positive, RR 1.46 per decade increase), structural heart disease (9/13 studies positive, RR 2.0) and diabetes (9/14 studies positive, RR 1.62) were found to be good independent predictors of stroke. Supportive evidence was found for sex (8/22 studies positive, RR 1.67), vascular disease (6/17 studies positive, RR 2.61) and heart failure (7/18 studies positive, RR 1.85). The various risk stratification schemes classified variable proportions as low, moderate and high risk, but the CHA(2)DS(2)-VASc score classified the smallest proportion of patients as 'low risk'. Anticoagulation with vitamin K antagonists and dabigatran is cost-effective in patients at high risk of stroke, but not in patients without any other stroke risk factor beside AF. Continued efforts are warranted to improve the antithrombotic management of AF patients to identify, and challenge, risk factors and refine risk stratification models in order to realize an individualized tailored, risk factor-based approach.