Patients with mechanical heart valves must follow lifelong warfarin therapy. Warfarin, however, is a difficult drug to manage because it has a narrow therapeutic window and potentially serious side effects. Successful anticoagulation treatment is dependent upon the patient's knowledge of this drug; however, little is known regarding the determinants of such knowledge. Therefore, the purpose of this study was to determine the influence of both in-hospital teaching practices as well as socioeconomic status and demographic variables on patients' knowledge of warfarin therapy. A telephone survey was conducted among 100 patients 3 to 6 months after mechanical heart valve replacement. A previously validated 20-item questionnaire was used to measure the patient's knowledge of warfarin, its side effects, and vitamin K food sources. Demographic information, socioeconomic status data, and medical education information were also collected. Knowledge scores were compared using the Student t test or one-way analysis of variance. Variables with P < or = .2 on univariate analysis were entered in multiple stepwise regression analysis. Sixty-one percent of participants had scores indicative of insufficient knowledge of warfarin therapy (score < or = 80%). Age was negatively related to warfarin knowledge scores (r = 0.27, P = .007). Patients with family incomes greater than $25,000, who had greater than a grade 8 education, and who were employed or self-employed had significantly higher warfarin knowledge scores (P = .007, P = .002, and P = .001, respectively). Gender, ethnicity, and warfarin therapy before surgery were not related to warfarin knowledge scores. Furthermore, none of the in-hospital teaching practices significantly influenced knowledge scores, whereas receiving postdischarge community counseling significantly improved knowledge scores (P = .001). Multivariate regression analysis revealed that understanding the concept of International Normalized Ratio, knowing the acronym, age, and receiving community counseling after discharge were the strongest predictors of warfarin knowledge. Accessing postdischarge counseling resulted in significantly improved warfarin knowledge scores. Because improved knowledge has been associated with improved compliance and control, our findings support the need to develop a comprehensive postdischarge education program or at least to ensure that patients have access to a community counselor to compliment the in-hospital education program.