OBJECTIVES: AF is the most frequent, clinically significant form of arrhythmia leading to stroke in a substantial proportion of patients. This paper investigates a possible OAC underuse based on a large German health insurance data set. METHODS: Based on data from 2 German statutory medical aid funds (2007-2008), a patient was considered to be in need of OAC if a CHADS2-Score≥2 without any OAC contraindication (CI) had been recorded. We used two CI scenarios; 1) a more extensive CI list based on the German Summary of product characteristics, and 2) a more limited list based on latest publications. In our day-specific analysis covering 2008, we classified each observed day into one of the following 3 categories: OAC use: day covered by OAC prescription (daily required dose defined by daily defined dose (DDD) or within a 180-days grace-period after consumption of last prescription); and Uncertain OAC use: OAC prescription in 2008, but the day was outside the grace-period, or it was covered by other anticoagulants/antiplatelets (DDD+90 days’ grace-period). OAC underuse: no OAC prescription in 2008, or other anticoagulants/antiplatelets did not cover the day (DDD-based+90 days grace-period). RESULTS: Data of 183,448 AF patients were included, mean age 73.2 years (SD: 10.97 years), 55.6% male. Their average CHADS2 was 2.8 (SD 1.58). Considering CI, OAC was recommendable for 54251 (29.6%; scenario 1)/83653 (45.6%; scenario 2) of all patients. Between 28.1% and 31.8% of all days were OAC use days whereas 43.0%-48.7% were OAC underuse days. Relative underuse rates did not differ between the two CI scenarios. Older female patients with a high number of co-morbidities had a greater OAC underuse probability. CONCLUSIONS: Our analysis shows that the OAC underuse problem in AF patients in the German health care system is possibly more widespread than previous observational studies have shown.