Clinical consequences of hospital variation in use of oral anticoagulant therapy after first-time admission for atrial fibrillation

To analyse how hospital factors influence the use of oral anticoagulants (OAC) in atrial fibrillation (AF) patients and address the clinical consequences of hospital variation in OAC use. By linkage of nationwide Danish administrative registers we conducted an observational study including all patients with a first-time hospitalization for AF between 1995 and 2004 as well as prescription claims for OAC. Multivariable logistic regression analysis was used to evaluate hospital factors associated with prescription of OAC therapy. Cox proportional-hazard models were used to estimate the risk of re-hospitalization for thromboembolism and haemorrhagic stroke with respect to discharge from a low, intermediate, or high OAC use hospital. Overall 40 133 (37%) out of 108 504 patients received OAC; ranging from 17% to 50% between the hospitals with the lowest and highest OAC use, respectively. Cardiology departments had the highest use of OAC, but neither tertiary university hospitals nor high volume hospitals had higher OAC use than local community hospitals and low volume hospitals. Risk of a thromboembolic event was significantly increased amongst patients from hospitals with a low OAC use (hazard ratio 1.16, confidence interval 1.10-1.22). Notably, higher OAC use was not associated with a higher risk of haemorrhagic stroke. In Denmark between 1995 and 2004, there was a major hospital variation in AF patients receiving OAC, and consequently, more thromboembolic events were observed amongst patients from low OAC use hospitals. Our study emphasizes the need for a continued vigilance on implementation of international AF management guidelines.