The central questions of the study can be summarized as follows:

What stroke / death rates are the result of the use of the following treatment strategies: no anticoagulation with vitamin-k-antagonists (VKAs) or novel oral anticoagulants (NOACs), anticoagulation with VKAs and anticoagulation with NOACs (all). What other thromboembolic complication rates resulting from the use of these treatment strategies? What heavy bleeding rates resulting from the defined management strategies? The analysis is based on a data set of patients with atrial fibrillation (AF)(01/01/2010 - 30/06/2014). 2010 is used as the reference period and 01/01/2011-30/06/2013 as the inclusion period. The minimum observation time per patient is 12 months.

Used is a data base of the AOK PLUS as well as the AOK Baden-Württemberg and AOK Bayern.

The aim of the study is to compare clinical outcomes between the different treatment strategies (frequency of events and time to first event). The influence of alternative treatment strategies is measured on the basis of three different methodological approaches: unadjusted comparison of event council and time to first event between the defined patient groups; matched-pair comparison (propensity score matching) between the different groups of patients (treatment strategies) and multivariate analysis using time to event (Cox proportional hazards model) as the dependent variable.

Inclusion Criteria:

• one inpatient or two outpatient confirmed diagnosis of atrial fibrillation in two different quarters
• Continuously insured by the AOK PLUS, AOK Bayern or AOK Baden-Württemberg

Exclusion Criteria:

• Age < 18 years
• Maximum possible observational period < 12 months (exception: death)

The analysis is based on a data set of patients with atrial fibrillation (01/01/2010 - 30/06/2014) insured by the AOK PLUS, AOK Bayern or AOK Baden-Württemberg.