The adverse effects of oral anticoagulant therapy (OAT) are the main cause of hospitalization for drug accidents, and most of them could be avoided by more rigorous management. We conducted a prospective study based on the analysis of individuals under OAT recruited among the patients admitted to our hospital. The aim was to evaluate the legitimacy of OAT and the quality of its management by referring to general recommendations. Eighty-six patients were included. In 10, the disease justifying OAT was not included in the French recommendations. Contraindications to OAT were observed in 5 patients. Six drug associations were dangerous. The day of admission, the INR value was beyond the therapeutic range in 27 patients and under in 27 patients. Nine patients had been admitted to the hospital for an adverse effect of OAT (hemorrhage or thrombotic event). The risk of adverse effects was higher when the indications of OAT were outside the recommendations, when a contraindication to OAT existed, or when OAT had been prescribed beyond the necessary duration. A low number of patients knew their INR target, the risks of over- and under-anticoagulation, and the dangers of consuming certain drugs and foods. The overall risk linked to OAT would be diminished if the treatment was prescribed within the legitimate indications and the necessary duration, and taking greater account of the contraindications. The awareness of physicians prescribing OAT needs improvement. The lack of knowledge of the treatment by the patient him- or herself, which may be due to a lack of information or to a misunderstanding, should be counterbalanced by a reinforced education, even if it is time-consuming for the physician.