Influence of decision aids on patient preferences for anticoagulant therapy: a randomized trial.

BACKGROUND Decision aids have been shown to be useful in selected situations to assist patients in making treatment decisions. Important features such as the format of decision aids and their graphic presentation of data on benefits and harms of treatment options have not been well studied. METHODS In a randomized trial with a 3 x 2 factorial design, we investigated the effects of decision aid format (decision board, decision booklet with audiotape, or interactive computer program) and graphic presentation of data (pie graph or pictogram) on patients' comprehension and choices of 3 treatments for anticoagulation, identified initially as "treatment A" (warfarin), "treatment B" (acetylsalicylic acid) and "treatment C" (no treatment). Patients aged 65 years or older without known atrial fibrillation and not currently taking warfarin were included. The effect of blinding to the treatment name was tested in a before-after comparison. The primary outcome was change in comprehension score, as assessed by the Atrial Fibrillation Information Questionnaire. Secondary outcomes were treatment choice, level of satisfaction with the decision aid, and decisional conflict. RESULTS Of 102 eligible patients, 98 completed the study. Comprehension scores (maximum score 10) increased by an absolute mean of 3.1 (p < 0.01) after exposure to the decision aid regardless of the format or graphic presentation. Overall, 96% of the participants felt that the decision aid helped them make their treatment choice. Unblinding of the treatment name resulted in 36% of the participants changing their initial choice (p < 0.001). INTERPRETATION The decision aid led to significant improvement in patients' knowledge regardless of the format or graphic representation of data. Revealing the name of the treatment options led to significant shifts in declared treatment preferences.