Warfarin is a dangerous outpatient medication, by anyone's estimation. It is the second most common cause of adverse drug events in emergency rooms, and the overall risk of major bleeding averages 7–8% per year [1, 2]. Despite the risk, well-established indications for warfarin are increasing in prevalence with aging of the population [3, 4], and new indications for warfarin are regularly recommended [5, 6]. As a result, the proportion of elderly persons taking warfarin has risen to as high as 7% [7].

Increasing a patient's understanding about warfarin is a logical goal. Prior knowledge about warfarin has been associated with a decreased risk of bleeding [8]. Written and verbal information has been shown to improve control of the level of anticoagulation [9]. While past studies suggest that patient education may be associated with better clinical outcomes, doubts remain about the effectiveness of patient education strategies [1012]. As a result, systematic patient education regarding long-term warfarin is not universally implemented.

Our objectives were to (1) identify the published strategies (duration, timing, personnel requirements, content domains) for patient education regarding warfarin anticoagulation and (2) identify published instruments for measuring patient knowledge.


In March 2007, we searched MEDLINE using the MESH terms ("warfarin" or "anticoagulation") AND "patient education". We limited our search to articles published in the English language. We used the related articles link in PubMed and searched the references of identified citations for additional original articles. Similar search terms were used to search Google Scholar. As warfarin is by far the most commonly used oral anticoagulant, we did not seek articles related to other oral anticoagulants.

We sought articles that (a) were original research studies or descriptions of patient education programs that included information on the educational content and strategy related to anticoagulation with warfarin, or (b) contained instruments that measured patient knowledge. Exclusion criteria included studies conducted in pediatric populations, unrelated to patient education, lacking original data or an adequate program description, and those in which the educational effort was focused solely on patient self-testing. Because citations might be excluded for multiple reasons, we used this above mentioned sequence for excluding citations.

An initial search identified 206 citations. Two reviewers (JLW, MDW) reviewed titles and available abstracts to determine relevance to the stated objectives of identifying (1) the optimal educational content and delivery (duration, timing, personnel requirements), and (2) the optimal strategies for measuring patient knowledge. Full text articles were retrieved for citations that met our inclusion criteria and for those where inclusion/exclusion criteria were not determinable by the title and abstract. Two other citations were encountered during the process of reviewing articles that were deemed eligible, raising the number of eligible articles to 208.

A total of 154 citations were initially excluded because patients were of pediatric age (1.9%, 4), the article was not related to patient education (23.1%, 48), did not contain original data or inadequate program description (18.8%, 39), was focused solely on patient self-testing (1), was a duplicate citation (1.4%, 3), or the article was judged otherwise irrelevant (16.8%, 35), or no abstract was available (11.5%, 24) (Figure 1).

Figure 1
figure 1

Search strategy for studies and programs related to patient education about warfarin anticoagulation.

After exclusions, a total of 44 articles qualified for further review. Upon further review, an additional 12 articles were excluded because of inadequate program description, ultimately leaving a total of 32 articles for data extraction (Figure 1). We extracted data on clinical setting, study design, group size, content source, time and personnel involved; and created summary tables. Two reviewers (MDW, JLW) identified the educational topics covered in these reports. Among studies that tested patient knowledge, we extracted information on setting and study population, number and type of questions, and method of administration.


Thirteen articles had a description of the research methods or program that was adequate and consistent with our objectives of identifying the duration, timing and setting, and personnel requirements of the educational program (Table 1) [1325]. Five programs used a nurse or pharmacist (45%), four used a physician, and two studies used other personnel/vehicles (lay educators (1), videotapes (1)). The duration of the educational intervention ranged from one to ten sessions. Patient group size most often averaged three to five patients but ranged from as low as one patient to as much as eleven patients. While the majority of the educational efforts occurred in inpatient settings, most seemed relevant to contemporary outpatient settings.

Table 1 Patient Education Strategies Related to Warfarin and Anticoagulation

Although twelve articles offered information about education content, the wording and lack of detail in the description made it too difficult to accurately assign categories of education topics and to compare articles with one another [2, 11, 12, 15, 19, 2224, 2629]. Nevertheless, we summarized the categories suggested by these studies and listed the potential topics for each category (Table 2).

Table 2 Topics for Education of the Anticoagulated Patient

Relevant to our objective of identifying measures of patient knowledge, Table 3 shows the seventeen relevant citations [9, 11, 12, 15, 18, 24, 3040]. Five of the seventeen sites where the surveys were administered were located in anticoagulation clinics/centers. The number of patients included in these studies ranged from as low as 22 to as high as 530. The number of questions ranged from as few as 4 to as many as 28 questions, and were most often of multiple choice format. Three were self-administered, and two were completed over the telephone. Two citations [12, 32] described testing instruments along with formal testing of the validity and reliability of the instrument.

Table 3 Studies Testing Patient Knowledge Regarding Anticoagulation


Patient education has long been thought to be useful for patients receiving long-term anticoagulation. Proposals have been periodically issued suggesting the content of the educational task [2, 23, 41]. However, inadequate attention to health education principles and educational program design have more often been the problem than have issues of content [29, 42]. Despite the practical value of making the patient as knowledgeable as possible, the best strategy for educating patients about anticoagulation is yet to be determined [10].

The variety of strategies shown in Table 1 likely reflect a varying amount of support and resources devoted to this patient education goal. Delegating these educational activities to midlevel practitioners, pharmacists, or designated nurses are strategies well supported by the our literature review. However, in any given clinical setting, local factors such reimbursement and available manpower may determine which health professional(s) is best responsible for managing a population of anticoagulated patients. The advent of warfarin self-monitoring with home coagulometers has sparked renewed interest in improving patient education related to anticoagulation [2, 13]. Government-supported efforts in Germany and Netherlands now devote a significant level of time and manpower to this educational task [21, 43]. However, most clinical settings in the U.S. and elsewhere, may not be able to match that level of support [15]. Because most anticoagulation management still takes place in the offices of clinicians [44, 45], strategies to provide education should be relevant to all clinical settings.

We also found much variability in the content areas reported by educational programs, to the degree that we could not accurately categorize educational domains, let alone make fair comparisons among programs. Some issues (manifestations of bleeding, INR monitoring, etc) were a component of most educational programs, while other issues (Vitamin K, pill color) were present only in a few. Our inability to summarize published efforts likely reflects an underreporting of details rather than extreme variability among programs. Nevertheless, our table of potential educational topics (Table 2) reflects a daunting agenda.

The testing of patient knowledge regarding warfarin and anticoagulation used a variety of instruments. Only two of the sixteen instruments – the Oral Anticoagulation Knowledge (OAK) instrument and the Anticoagulation Knowledge Assessment (AKA) – have been subject to any formal evaluation. The Oral Anticoagulation Knowledge (OAK) investigators evaluated construct and content validity, test-retest reliability, and internal consistency reliability [12]. The Anticoagulation Knowledge Assessment (AKA) investigators used the Rasch model in order to examine validity, and item and person reliability [32]. Both the OAK and AKA are reported to be written at the 7th grade reading level, but neither instrument has been validated in other clinical settings. The best strategy for measuring patient knowledge would depend, in part, on the content of the educational program, but standardization of the testing effort should be a realistic goal.

The limitations of our study deserve acknowledgement. While our study reflected a variety of different strategies for all aspects of the educational process, it is probable that noteworthy and innovative patient education efforts may not be reflected in the medical literature. Second, in reviewing these reports, it is often difficult to separate the management strategy from the educational strategy.

Despite the variability in the content and strategies of educational programs, several important issues should drive future efforts at patient education, in our opinion. Educational programs should focus on topics essential for patient safety, such as what to do when INR is high, rather than the minute details of anticoagulation that overburden the patient. Second, these programs would best be implemented with measures of effectiveness and improvement in patient knowledge, adherence and outcomes using validated instruments. Lastly, educational programs should attempt to maximize office efficiency by delegating this task to physician extenders, nurses, pharmacists, or perhaps an office-based computer.


Patient education is entering a new era where accountability in educational outcomes, interest in literacy/language barriers, and the importance of cost-effectiveness will influence the process of patient education. Prioritizing the educational content and using validated instruments for measuring the outcomes of patient education will be a necessary first step in improving anticoagulation outcomes. This systematic review should guide future efforts.