This study investigated the determinants of warfarin use in patients with atrial fibrillation (AF).

Investigators assembled a retrospective cohort of community-dwelling elderly patients (aged ≥66 years) with AF using linked administrative databases.  They identified the physicians responsible for the ambulatory care of these patients using physician service claims and compared patients who did and did not have an identifiable provider.  For those patients with an identifiable provider, investigators assessed the association between patient, physician, and hospital factors and warfarin use.

The cohort consisted of 140,185 patients, of whom 116,200 (83%) had an identifiable cardiac provider.  Patients without a provider were significantly more likely to have comorbid conditions that increase their risk of warfarin-associated bleeding.  After adjustment for clinical factors, patients without a provider were significantly less likely to receive warfarin (odds ratio 0.37, 95% confidence interval: 0.36-0.38).  Of patients with providers, 50,551 patients (43.5%) received warfarin within 180 days after hospital discharge.  Warfarin use was positively associated with AF-associated stroke risk factors (e.g., prior stroke, congestive heart failure) and negatively associated with warfarin-associated bleeding risk factors (e.g., history of intracerebral hemorrhage).  After controlling for patient and hospital factors, patients cared for by non-cardiologist physicians with cardiology consultation were more likely to receive warfarin than patients treated in non-collaborative environments.

Warfarin continues to be substantially underprescribed to patients who are at high risk for AF-associated cardioembolic stroke.  These findings highlight the need for targeted quality improvement interventions and suggest preferred models of AF care involving routine collaboration between cardiologists and other physicians.