[The author responds:]
Malcolm Man-Son-Hing draws attention to 2 interesting factors related to underuse of anticoagulant therapy in patients with atrial fibrillation. Anticoagulant therapy is inconvenient both for physicians and for patients. To my knowledge, there is little information about the factors that influence physician decisions to prescribe (or not prescribe) anticoagulant therapy for atrial fibrillation. Large anticoagulation services are available in many urban centres, and many of these accept referrals from physicians. This reduces the burden of caring for these patients. It would be interesting to know whether anticoagulant therapy is more readily prescribed in such areas. Would the wider availability of such a service increase anticoagulant use in atrial fibrillation?
Man-Son-Hing also draws attention to his recent analysis of the relation between anticoagulant therapy for atrial fibrillation and risk of falling among elderly people. His Markov decision analytic model suggests that the benefit-risk ratio favours using anticoagulant therapy even in elderly patients who are at high risk for falling: The reliability of decision modelling (based on literature review) is only modest. In the absence of randomized trial data pertaining to this issue, I would remain cautious about prescribing anticoagulant therapy in elderly patients with a history of repetitive falling.