OBJECTIVES: The aim of this research was to assess the quality of anticoagulation therapy of patients with atrial fibrillation (AF) in a real life setting and to identify the causal factors explaining anticoagulation quality deficits. Furthermore, clinical consequences of suboptimal anticoagulation therapy (strokes, TIA, bleedings, embolism, myocardial infarcts) are identified. METHODS: The INR values as well as other clinical events concerning AF patients recruited into a prospective cohort study (observation period of 12 month) were documented at every visit to the treating doctor. Using the Rosendaal linear trend method, the time in therapeutic range (TTR) of 2.0-3.0 was estimated. Additionally, the squared INR deviation was investigated. To identify causes of INR-values below/above 2.0-3.0, a logistic regression on the basis of a TTR < 60% as the dichotomous outcome was conducted. RESULTS: For 525 patients from 71 study centers, at least two INR values were available over a median observational period of 228.9 days (SD: 106.1 days). The average TTR was 68.1 % (SD: 26.3 %). The average deviation of INR value from the mean of the INR target range (2.5) was 0.44 (SD: 1.29). The results of the multivariate regression (R2=0,179) show that the most important factors explaining a poor quality of anticoagulation therapy are bridging periods and patients self-reported need of help regarding medication therapy without getting that help. In the group of patients with a TTR<60 %, the occurrence rate of clinical outcomes was higher (p=0.031) than in the group of patients with a TTR>60 %. CONCLUSIONS: Labile INR values lead to negative clinical outcomes. In order to improve the situation, the main identified causes of poor anticoagulation quality should be addressed.