4.1 Main Results and Comparison with the Literature
Several studies analysed the attitudes of AF patients towards OAC treatment in general, especially the view of AF patients towards the trade-off between a lower stroke and higher bleeding risk. Most of the literature shows that patients are willing to accept higher bleeding risks if a certain threshold in reduced stroke risk can be reached. Nevertheless, the majority of the publications also showed that many AF patients may weigh bleeding risk and stroke risk differently from both physicians and clinical guidelines. Therefore, based on the preferences of AF patients only, fewer patients would receive anticoagulation treatment than may be expected on the basis of recommendations found in clinical guidelines. Consequently, preferences of AF patients may be one important and potentially modifiable explanation for the often observed OAC undertreatment of patients [
50,
51]. Therefore, it is of utmost importance to not only identify but also to understand patient preferences with regard to anticoagulation treatment in order to improve adherence to guideline recommendations. Besides that, it is also important that the treating physician educates and informs the patient about stroke and bleeding risks since, in that way, adherence to guideline recommendations can be further improved. All the publications we analysed reported a high variability of AF patient preferences towards anticoagulation treatment; some of the analyses even identified specific AF patient segments, defined by different degrees of bleeding risk aversion.
The above-mentioned results were confirmed by two previous systematic reviews. The authors of the first review concluded that higher patient values and preferences regarding thromboprophylaxis treatment may depend on patients’ prior experience with the treatments, as well as on the methods used for preference elicitation [
52]. The second review argued that AF patient preferences may indicate that fewer patients would take VKAs compared with the recommendations of the guidelines. Based on this review, at a stroke rate of 1 % with aspirin, half of the participants would prefer VKAs and, at a rate of 2 % with aspirin, two thirds would prefer VKA treatment [
53].
As shown in our review, a second part of the scientific literature analysed the preferences of AF patients towards OAC treatment by including ‘convenience’ attributes, which characterise the alternative OAC treatment options, VKAs and NOACs. Generally, the published data show that AF patients, in accordance with clinical guidelines, weigh clinical attributes such as stroke or bleeding risk more heavily than convenience attributes. Therefore, it is in line with the preferences of AF patients that a treating physician first investigates the clinical effectiveness and safety of the recommended anticoagulant before suggesting alternative treatment choices to the patient. However, this review also showed that if alternative OAC treatments are similar in terms of efficacy and safety, as is the case with many anticoagulation options in AF, convenience attributes such as mode of application, interactions with food or drugs, availability of an antidote, need for bridging, or frequency of application may matter to patients. Furthermore, patients may not only have a preference for a more convenient mode of application but the adherence of patients may also depend on the convenience of medication therapy. Thus, it has been shown that a less frequent dosing schedule, such as once daily on chronic cardiovascular disease medication, is associated with higher treatment adherence [
54].
We also identified a variability of results with regard to preferences towards these convenience attributes, which could be explained by several reasons. First, patients analysed in the included studies were different from each other in terms of sociodemographic characteristics, study site characteristics (inpatient versus outpatient treatment, general practitioners [GPs] versus cardiologists) and current anticoagulation treatment, or in terms of the treatment experience of surveyed patients. For example, one German study found that AF patients currently treated with VKAs for at least 3 months (which may be an indicator of stable anticoagulation) did not show any clear preference for or against monthly blood checks, whereas patients treated with an NOAC agent had a clear preference against such checks [
41]. Thus, it is important to inform the patient about the differences between NOACs and VKAs, which are apparent in the need for regular blood checks, amongst others. With VKAs, regular checks are necessary, whereas with NOACs, no regular checks are needed.
Second, study methodologies differed markedly between the different studies in terms of the number of included study sites, patient inclusion criteria and patient selection, and the preference elicitation technique used. Because of this, it cannot be ruled out that the selected questions in the descriptive questionnaires used in several studies also influenced reported results. In some studies, interviews were conducted face-to-face, whereas others used phone, written or online interviews, therefore interviewer bias cannot be excluded in these studies [
29]. On the other hand, when questionnaires were sent out to participants, there was the risk of non-response bias, potentially leading to skewed results of the study [
25].
Three studies analysing preferences towards NOACs versus VKAs used trade-off/standard gamble techniques. These comprise a method that offers the opportunity to compare therapy preference between two different options [
45]. The hypothetical efficacy of the intervention is systematically varied until the lowest risk reduction, the point at which patients are willing to take the therapy, is found [
28]. A disadvantage of this method is that the consideration of more than two options might generate confusion or fatigue in respondents, resulting in a notable proportion of respondents indicating initial choices that were internally inconsistent [
55].
Finally, four studies applied a conjoint analysis/discrete choice experiment (DCE) to analyse preferences towards either NOACs or VKAs. DCE was introduced into health economics as a technique to identify the key characteristics of alternative treatments because patients were concerned with aspects of healthcare other than only clinical outcomes. This method has been used to elicit preferences for health and healthcare in a range of contexts and is now, to a certain degree, seen as a gold-standard technique. The method assumes that the value of medical treatments depends on a number of characteristics. DCE allows one to simultaneously weigh various characteristics of different therapeutic options and to establish the relative importance of each characteristic in the implementation of that therapeutic option. It can also be used to estimate how individuals trade these characteristics; for instance, the rate at which they are willing to give up one characteristic for an increase in another [
56]. The main reason for applying a DCE is that simply asking patients to rate treatment-related attributes generally yields no substantial information since patients will state in such a survey that they want all the benefits and none of the indirect/direct costs [
57]. One advantage of applying a DCE is that patients are forced to make a trade-off between two or more options and that they have to choose, as is the case in reality, between options that may be associated with utility-increasing and utility-decreasing attribute levels [
58].
The studies included in our review that used a DCE technique came to similar conclusions. An Australian study that included both clinical and convenience attributes concluded that the overall profiles of NOACs, compared with VKA treatment, are preferred by patients, especially if an antidote exists and if there are reasonable costs for NOAC treatment [
18]. Similarly, in an Italian study that included bleeding risk as an attribute, patients preferred once-daily tablet treatment without regular monitoring [
56]. In a German DCE study, AF patients also favoured attribute levels that are best presented by a once-daily NOAC treatment [
41].
In addition to the above-mentioned difficulties in comparing the identified studies, due to differences in their methodology, we acknowledge two additional limitations. First, our research was based on a review of publications available in the selected databanks only and, second, due to methodological differences between the studies, we did not conduct a quantitative meta-analysis.