Selecting antithrombotic therapy for stroke prevention in atrial fibrillation: Health professionals’ feedback on a decision support tool

Abstract

A Computerised Antithrombotic Risk Assessment Tool was developed for assisting the selection of antithrombotic therapy based on the risk versus benefit assessment. In view of the recent availability of the novel oral anticoagulants, this tool has been updated to CARATV2.0. To explore health professionals’ perspectives on the tool, semi-structured interviews were conducted in seven pharmacists, seven specialists, six general practitioners and six nurses, who were involved in management of antithrombotic therapy for atrial fibrillation. Three overarching themes emerged: (1) CARATV2.0 provides comprehensive structured assessment of patients and could assist with the prescription and review of antithrombotic therapy, (2) subjective issues such as health professionals’ and patients’ preferences for a particular antithrombotic therapy may affect the usefulness of CARATV2.0 and (3) CARATV2.0 requires integration into existing systems and processes. The majority of health professionals surveyed would like to use CARATV2.0 in practice, believing it would improve antithrombotic use and might reduce stroke incidence.

Keywords

anticoagulant agents atrial fibrillation computer-assisted decision-making decision support

Introduction

The decision-making around antithrombotic therapy (e.g. anticoagulant and antiplatelet therapy) in atrial fibrillation (AF) is complex because it involves assessment of risks versus benefits.¹ For many years, warfarin was the only available OACs, but its unpredictable therapeutic effects, various food and drug interactions and the need for regular monitoring have been associated with great difficulties in its use. Such difficulties have led to a potential increase in adverse events and reluctance by clinicians to prescribe the medication.^2-4 The recently marketed novel oral anticoagulants (NOACs) – dabigatran, rivaroxaban and apixaban – have substantially expanded the treatment armamentarium and are intended to overcome the limitations of warfarin. However, these new anticoagulants are not without risk because some of their so-called advantages can be regarded as potential disadvantages in specific situations.⁵

To optimise the use of antithrombotic therapy in patients with AF, and specifically assist health professionals in selecting appropriate agents, an electronic decision support tool – the original Computerised Antithrombotic Risk Assessment Tool (CARAT) – was developed.⁶ Its decision-making algorithm was computerised by first preparing a prototype in Microsoft Excel^™ and then formatting it as a web-based interface for online access.⁶ The tool generated treatment recommendations (e.g. warfarin vs aspirin therapy) for individual patients based on their risk (bleeding) versus benefit (stroke prevention) estimation, as well as the relevant medication safety considerations (e.g. drug–drug interactions, renal function and medication adherence). The original CARAT has been trialed in real-world hospital patients⁷ and general practice patients⁸ and evaluated by specialist clinicians for its potential clinical application in a vignette-based study.⁶ CARAT has demonstrated its potential utility in practice.^6,7 Although other risk assessment tools have been developed to synthesise the assessment of stroke and bleeding risks – for example, the clinical decision aid developed by LaHaye et al.⁹ and the decision model developed by Casciano et al.¹⁰ – none consider the broader medication safety issues that particularly affect the selection of therapy in the target at-risk patient population.^9,10 Thus, the CARAT provides a novel, more holistic and pragmatic approach to the decision-making around antithrombotic therapy.

Our previous (i.e. original) version of CARAT was designed to address the complexity in decision-making by integrating the relevant assessments around stroke risk, bleeding risk and medication safety for individual patients;⁶ at that time, the tool was able to assist in selecting among two main treatment options – warfarin and aspirin. Now that we have moved forward in time, there are additional issues to consider in the decision-making process which need to be factored into the CARAT. First, the expanded range of treatment options incorporating the NOACs (dabigatran, rivaroxaban and apixaban) has increased the number of parameters (e.g. drug–drug interactions, side effects, doses and frequency of administration) that need to be considered, further complicating decision-making.¹¹ Second, the range of risk assessment tools for stroke risk (e.g. CHADS₂ and CHA₂DS₂VASc¹²) and bleeding risk (HAS-BLED¹³ and HEMORR₂HAGES¹⁴) has evolved. Third, the evidence-base around the effectiveness and safety of available treatment options has grown, for example, aspirin is no longer recommended for stroke prevention in AF.¹⁵ Fourth, a broader range of health professionals is now involved in therapeutic recommendations and decision-making around antithrombotics therapy, whereby hospital specialists and general practitioners (GPs) are able to draw upon the services of nurse practitioners (NPs)¹⁶ and consultant pharmacists (e.g. as part of medicines review process¹⁷). Collectively, these issues have warranted a revision of the original CARAT, into its second version CARATV2.0, which considers the latest clinical evidence (e.g. guidelines^15,18,19 and practice reviews^5,20,21) and available treatments (warfarin and NOACs).

As an initial evaluation of this revised (prototype) CARATV2.0 tool, the aim of this study was to obtain feedback from a wide range of health professionals who are involved in the decision-making around antithrombotics in AF (specialist clinicians, GPs, nurses and pharmacists), to help inform the future implementation of this tool in practice. Specific topics explored were the (1) strengths and weaknesses of this tool; (2) appropriateness and relevance of the content of this tool; (3) usefulness of this tool for selecting appropriate antithrombotics, especially between warfarin and NOACs; (4) feasibility of using this tool in clinical practice; and (5) suggestions for further improvement.

Method

Design and setting

This qualitative study was based on face-to-face interviews. From August to October 2014, health professionals (subgroups: specialist clinicians, GPs, nurses and pharmacists) practising in the Sydney metropolitan area were involved in this study (Figure 1).

Figure 1.

Key stages in eliciting feedback from health professionals about the decision support tool.

Participant recruitment

Purposive sampling was used to identify and recruit health professionals with experience in prescribing antithrombotics and managing antithrombotic therapy for patients with AF.¹¹ Specialist clinicians, hospital-based pharmacists and nurses were recruited via an invitational flyer emailed or faxed to the network of hospitals affiliated with the university. Community-based pharmacists accredited for Home Medicines Review were recruited through an emailed flyer (using contact details from the Australian Association of Consultant Pharmacy). Flyers were also emailed to community-based nurses in community health services affiliated with the university network of hospitals. By visiting family practices and medical centres in the Sydney metropolitan area, the researcher also distributed invitational flyers to GPs. Emails and faxes were also sent to GPs listed on the Internet. Eligible health professionals who agreed to participate provided written consent.

An estimated 24–40 participants (6–10 participants per group) were needed to achieve theme saturation within each subgroup of health professionals (specialist clinicians, GPs, nurses and pharmacists).²² As this tool was developed to support a broad range of health professionals who are involved in the day-to-day management of older persons taking antithrombotics, in particular those who had previously expressed a need for assistance in decision-making,^3,11 the study largely focused on canvassing feedback from GPs, nurses and pharmacists, in preference to experienced cardiologists. For this reason, relatively few cardiologists were recruited to this study.

Data collection

Semi-structured interviews (20–30 min each) were conducted by the researcher at a location convenient to each participant. At the beginning of each interview, demographic data for each participant were collected using a predesigned questionnaire. Then, the researcher presented CARATV2.0 (on the researcher’s laptop) to the participant. After this familiarisation with the tool, the participant was given the opportunity to explore CARATV2.0. Finally, the researcher used a set of open-ended questions to explore the participant’s feedback on the content of the tool and the feasibility of using CARATV2.0 in clinical practice. All questions were pretested in mock individual interviews with nonparticipants.

Prototype of CARATV2.0

The underpinning algorithm of this revised tool has been developed as a Microsoft Excel-based prototype for pretesting, with the intention of subsequently formatting the tool into an online (web-based) platform or mobile application that will enable the integration of this tool into prescribing software and/or electronic health data management systems. The prototype comprises four distinct sections: (1) stroke risk assessment, that is, CHADS₂²³ and CHA₂DS₂VASc;¹² (2) bleeding risk assessment, that is, HAS-BLED¹³ and HEMORR₂HAGES;¹⁴ (3) medication safety issues, for example, mini–mental state examination²⁴ for cognition, estimated glomerular filtration rate (eGFR) (MDRD and Cockcroft and Gault equation²⁵) for renal function and Child-Pugh score²⁶ for liver function; and (4) therapy recommendations and advice. The application of the tool requires the user to input relevant data into the cells, which auto-populates the formulae underpinning the decision-making algorithm, and which, in turn, generates a treatment recommendation. A patient is considered to be eligible for OACs whenever the risk of stroke is equal to or more than the risk of bleeding, otherwise the patient is deemed to be unsuitable for OACs. CARATV2.0 also provides initial advice around any identified medication safety issues which need to be addressed by the clinician. This study explores feedback on the data inputs and usability of CARATV2.0.

Data analysis

The interviews were digitally recorded (audio) and transcribed verbatim by the researcher. The accuracy of the transcripts was confirmed by listening to the digital records and reviewing the transcripts. The transcripts were analysed for themes, using standard thematic analysis techniques (manual inductive coding).²⁷ The two authors independently reviewed the transcripts and identified themes before reaching a consensus through discussion. The accuracy and reliability of the analysis was confirmed by inter-researcher validation (with three other independent researchers).

Results

Overall, 26 participants comprising 7 specialist clinicians, 6 GPs, 7 pharmacists and 6 nurses were interviewed (Table 1). Similar themes were identified among the four subgroups, with three overarching themes emerging (Tables 2 to 4 and Figure 1).

Table 1.

Participant characteristics.

Participant characteristics	Mean years of experience in managing patients with AF ± SD (range)	Mean number of patients with AF managed annually ± SD (range) (self-reported)
Specialist clinicians (n = 7) 3 geriatricians 2 haematologists 1 cardiologist 1 neurologist	23.4 ± 13.1 (5–40)	117.5 ± 109.3 (5–300)
General practitioners (n = 6)	22.3 ± 10.1 (12–40)	21.5 ± 12.4 (4–35)
Pharmacists (n = 7)^a 6 accredited pharmacists 1 hospital pharmacist	20.4 ± 17.7 (5–50)	46.9 ± 39.4 (5–100)
Nurses (n = 6)^b 3 nurse practitioners (NPs) (2 cardiology, 1 neurology) 3 clinical nurse consultants (neurology)	20.2 ± 9.5 (8–30)	145 ± 77.8 (100–300)

AF: atrial fibrillation; SD: standard deviation.

All six accredited pharmacists (home medication review and/or residential medication management review) were community-based pharmacists.

Among them, the two cardiology nurses were community-based, while the others were hospital-based.

Table 2.

Theme 1: need for comprehensive structured assessment of patients to assist with the prescription and review of antithrombotic therapy.

General practitioners	Pharmacists	Nurses	Specialist clinicians
It (CARATV2.0) is very thorough. It has everything there that we should definitely take into account. When I see a patient … there may be aspects that I wouldn’t have thought about that I should have thought about that wasn’t there. So it is a very good tool that has everything on it. (G03) You (CARATV2.0) have all renal liver function, gastrointestinal problem and other illnesses. You seem to cover everything that a good clinician would have to take into consideration anyway … As a doctor you think along these lines anyway, so all you (CARATV2.0) are doing is to put them into a chart or table or tool that gives you a visual and a check list, I suppose. (G05) The strength is that it is very important that as clinicians whether we still use the warfarin or the newer agents. I think if the prescribers have got a tool that would help them … Having it as a rigid module, I think it is a good idea. (G06) I can see it (CARATV2.0) would be useful for GPs that don’t have access to a cardiologist … It is good for doctors who are not certain or don’t have access to cardiologist … The new doctors love to have something like that, because it gives them the confidence to manage the patient and to be able to assess the patient to know what they should be doing. (G02)	I will definitely use it (CARATV2.0) … if I have a patient that I wasn’t sure whether or not they should be on say warfarin … All of these criteria that you have got listed here, taken into account with the choices basically they will go. Ok this person had an event and he need to be on kind of anticoagulant. Ok then let’s put him on warfarin. Maybe there are CHADS₂ score and HES-BLAD score … but they (CHADS2 score and HES-BLAD) are not looking at such comprehensive criteria, so this (CARATV2.0) is much more specific. (P02) The strength would be that it (CARATV2.0) considers all the factors of a patient which I usually consider when I am recommending an anticoagulant or checking what they are already taking, so I will look at the safety issues I look at the drugs that they are taking. I look at their actual medical condition as well so it has got all those things covered. And having that universal tool would be a big strength because as I said I can refer to a doctor what I am referring to. (P06) I think it (CARATV2.0) is good for pharmacist. It looks good to show this to doctor to say this is my recommendation. You have got something to back it up. So decision-making tool to back it up. (P04) Comprehensive. All the factors that I would consider starting someone on anticoagulant is in this tool. (P07)	I think that (CARATV2.0) offers a global assessment around anticoagulation not just simply risk stratification tool but a global assessment tool. It takes into consideration other factors that are above and beyond stroke and bleeding risk, such as adherence. (N01) … we might be in a situation especially with patients in nursing homes things like that. I would not feel confident without a pharmacist’s recommendation, but with that tool (CARATV2.0) I can possibly. (N02) I think it (CARATV2.0) is good because one again it makes you look at the patient as a whole … It (CARATV2.0) brings in all those other factors like their cognition their function and those sorts of things. I think that gets forgotten when people are prescribing. They forget the whole patient … I think that is what is good about being able to select or being able to choose between warfarin and other OACs. (N05) I think it (CARATV2.0) is good because it would make you also think about things that you might not think about when prescribing these medications that you might forget about the drug interactions and things like that. (N06)	For me it (CARATV2.0) might be useful but I imagine that I would most likely be using it in patients where it is not certain either way. Like I probably wouldn’t use it if it clear in my mind that there is high risk of stroke and low risk of bleeding … Whereas for patients who a bit equivocal then I may use this tool to help me choose one or the other. (S01) The benefit is that you (CARATV2.0) are assessing both the risk of stroke and the risk of haemorrhage, and you have also got some aged care risk factors in there, which a lot of other tools don’t have. (S03) One of the first thing that is good is that if a doctor who has not treated many patients with AF is that by going through this (CARATV2.0) they see a lot to consider and it also makes them to think the patient a little bit more thoroughly. I think that is good … It is good that these scores (CHADS₂ and CHA₂DS₂-VASc, and HAS-BLED and HEMORR₂HAGE) are included and the scores are also given at the end. (S05) All the components (in CARATV2.0) are the same as the experienced clinicians would use and make a judgement about whether to use or not to use anticoagulant therapy. So the use of CHA₂DS₂-VASc and HAS-BLED scores, the differentiation between warfarin and NOACs are all quite appropriate. (S04)

CHADS₂: congestive heart failure, hypertension, age = 75 years, diabetes mellitus, stroke; CHA₂DS₂-VASc: congestive heart failure, hypertension, age ⩾ 75, diabetes, prior stroke or transient ischaemic attack, vascular disease, age 65−74 and sex; HAS-BLED: hypertension, abnormal renal/liver function, stroke, bleeding history or predisposition, labile international normalised ratio, elderly, drugs/alcohol; HEMORR₂HAGE: hepatic or renal disease, ethanol abuse, malignancy, older age, reduced platelet count or function, re-bleeding, hypertension, anaemia, genetic factors, excessive fall risk and stroke; OACs: oral anticoagulants; NOACs: novel oral anticoagulants.

Table 3.

Theme 2: health professionals’ and patients’ preferences for a particular antithrombotic therapy.

General practitioners	Pharmacists	Nurses	Specialist clinicians
I think traditionally that people with high risk score of CHADS₂ should be a candidate to warfarin. That is the decision that I always made, so I would need be shown why I should be changing my decision. Why somebody might not be given warfarin but given NOACs instead. (G01) My only concern is that I am always very aware of warfarin, so if it (CARATV2.0) does recommend warfarin I may not take the recommendation, I would have to looking closely to see why it recommends warfarin instead of NOACs. (G03) The difficulty is to negotiate with the patients about their degree of risk. The meaning behind not being on an anticoagulant. What it would mean in terms of follow-up for it to be safe, if they are on anticoagulant. And the management around that … A lot of patient, where warfarin is appropriate have bad stories. Maybe the parents have been on it or the relatives, then they had a bleeding and regular blood test. So for lots of GPs, to address these issues is in fact the main task … So to manage the people’s expectation of health is the most important part of the consultation about anticoagulants. So I suppose with this tool it is really more about the technical decision. (G04) One of the difficulties in general practice is the cardiologist does something different from what you do. That becomes quite complicated, because I have been quite surprised that some of my patients are under cardiology care who are not on medication. (G02)	Well, you (CARATV2.0) have got all these you have listed all possible risk factors. As much as you can, of course sometimes these risk factors might vary or might be severe than in other situations. Like anaemia for instance, that might not be present might later be present. So it is just you would have to constantly measure these from time to time. (P03) Because some people do not like to go to their doctor to get the blood check, I don’t think it (CARATV2.0) addresses this problem, but otherwise it is pretty good. (P04) It is good tool. A lot depends on the patient … the doctor may evaluate the patient according to this tool. But the outcome does not depend on that they recommend warfarin or NOACs, it really depends on the patient’s cognitive function. Because sometimes, they (patients) don’t understand this situation very well and they don’t know how to ask questions. And even Webster pack is not always suitable. (P05) If there is an option to include all the drugs that the patient is on or if they want to do a little drug interactions … put the drug that they (patients) are on you might come up with some extra interactions. But that is probably going take away the simpleness of this tool, which is a big plus to it. So obviously it (CARATV2.0) is very very good but there still be a few little drugs that might interact that might be missed. (P06)	I think the nurses have limited input into choices and decision makings. I think the decision of anticoagulant actually come from the cardiologist in the end. Nurses may fill it (CARATV2.0) in and the pharmacist can fill it in. Is cardiologist going to look at it and follow? … I think still a lot of clinicians are hesitant to use NOACs … if you have cardiologist who have been using warfarin for the last 40 yrs and why they are going to change to NOACs? No. If they have someone who is on warfarin and adherent why change it … So it (CARATV2.0) doesn’t really consider the clinicians preference … Well it is nice to give a recommendation based on international guidelines validated risk assessment tools, I think it doesn’t consider the clinicians previous experience of the agents and that is going to influence their decision making. But I think this is still useful tool. (N01) I think it is good that it (CARATV2.0) makes you think about other things from a nursing point of view. I think it makes you think very much the cognition the falls risk visual deficit … that is good. Because all these things can be difficult for patients to actually overcome. So by taking that into consideration I think that is good … I like that. (N04) I think that is what is good about being able to select or being able to choose between warfarin and the NOACs. Because warfarin is a really hard medication to manage whether we can actually say this patient would definitely benefit from one the NOACs, because of their cognition bla bla then I think this tool is helpful in that respect. (N05)	I do not believe you should use warfarin as first line. It is a choice and some patients who we would prefer them to be on warfarin typically with renal impairment and maybe noncompliance. And certainly people who are well established on warfarin, they do not need to switch them. But for anyone who is new, I do not see there is any justification for saying warfarin first and then the new drugs. (S06) The other thing that I am not totally convinced by it yet, is whether there is enough data in older patients for the NOACs. That is the other thing that I am not sure about given that they have been safety signals for things like gastrointestinal bleeding. You know I probably want more data in older patients. (S01) One of the clinical concerns that we all face now with the NOACs is what happen to people who are elderly and who are around NOACs because of AF and happened to get gastrointestinal disturbance get diarrheal, vomiting, dehydration, renal impairment. And suddenly get an eGFR that drops down to 30 and they are on a NOACs. Unless they are aware of the risk. They are going to have increased risk of bleeding. So I think as an educational tool it is useful and that is going to be one of the main clinical concerns that I have in the use of NOACs. (S04)

CHADS₂: congestive heart failure, hypertension, age = 75 years, diabetes mellitus, stroke; GPs: general practitioners; AF: atrial fibrillation; NOACs: novel oral anticoagulants; eGFR: estimated glomerular filtration rate.

Table 4.

Theme 3: integration into existing systems and processes.

General practitioners	Pharmacists	Nurses	Specialist clinicians
I think if you can incorporate it (CARATV2.0) into the (practice) software and the software would populate as much as the data from our file as much as possible that would be very good. (G03) The weakness will be if this tool is not in cooperated in the practice software, so it must be integrated. Because it would be very difficult for a doctor to go somewhere else or another website. It is going be there as you are prescribing and also in a hospital situation it has to be accessible. So you know in the patient’s file it is going to be part of that medical record and accessible to people. (G06) I think if you can incorporate it (CARATV2.0) into the software and the software would populate as much as the data from our file as much as possible that would be very good. Also, we have got a chronic disease nurse, and she could administer it and fill the answer then that could also make it quite easier for us. (G02)	If you are doing a medication review, you look at the pathology first before you went out … It will be better your renal function and liver function altogether, because that will be the pathology that the patients wouldn’t know those answers. So it will be better if they are together at the beginning or at the end (of CARATV2.0) … I think it (CARATV2.0) will work really well within HMR or RMMR … mean a lot of questions you going to ask anyway. (P01) I usually not initiating treatment, so in my practice it would be basically just check what the patient is taking is best for them. And particularly if they are recently being changed or initiated a drug, I will be checking that it is the right drug that they should be on. So for me it (CARATV2.0) is a good checking mechanism. (P06) I think it is good as well. In a community would be useful. For annual review or something, they (patients) can not reach therapeutic level they have a labile INR … using this tool to determine whether warfarin still the therapeutic choice for this patient or not. (P07)	I think the mobility thing is important to have. Like to have it (CARATV2.0) with the patient. Like explaining to the patient about the things. And have it portable not on a desk top. I guess there is access issues around if it is something like internet based. (N01) Maybe you can get an app. If you can get it (CARATV2.0) into an app, maybe everyone would use it because it would be easy to carry around and ask patients questions. (N06) I think that (integration into electronic health system) will be useful. Because there is a lot of things there that you don’t need to do it again … A lot of that stuff whether the patient has AF or not could then go into that system. (N05)	I think easy to use is the main thing. If you do turn it (CARATV2.0) into an app make it simple to enter things so you don’t have to enter a lot of text more than that you have tick box to make it bit easier. Might be useful to have section of patient information, so you know based on this tool your doctor has recommended XYZ, these are precaution whenever using this agent is. So if you use a website or something that it can generate patient information sheet. (S01) I think we probably need it (CARATV2.0) as an app and need it to be electronic and need to be touch screen if we can have that. And then we just need to practice. (S03) It (CARATV2.0) will be much more acceptable on an app … Because you can just have it in your phone and download the things and much quicker than finding a computer and calling it up in computer and excel file to type it in for me. (S02)

HMR: Home Medicines Review; RMMR: Residential Medication Management Review; INR: international normalised ratio.

Theme 1: need for comprehensive structured assessment of patients to assist with the prescription and review of antithrombotic therapy

The most highly appreciated feature of CARATV2.0 was that it provides comprehensive assessment of a patient’s risk versus benefit of using antithrombotics. Perhaps more importantly, the ability of CARATV2.0 to provide guidance and assistance in selecting among OACs, especially between warfarin and the NOACs, was highlighted by health professionals. Overall, health professionals considered this tool helpful in the decision-making for antithrombotic therapy and hoped that it could help reduce the incidence of strokes.

Specifically, many GPs and specialists felt that CARATV2.0 validated or organised their own decision-making process. Interested in using this tool for the prescription of antithrombotics, GPs and specialists tended to see the tool as most useful in those cases in which there are clinical dilemmas (i.e. where the risk versus benefit of using OACs is not clear-cut). With regard to selection among OACs, both GPs and specialists appreciated that the tool offered a specific recommendation among the OACs (especially either warfarin or an NOAC). The doctors considered this useful because they perceived that the differences in the benefits and risks of individual anticoagulant agents were not clear to many doctors. One GP mentioned that this tool could be useful for the initiation of therapy.

Similarly, the senior accredited and hospital-based pharmacists (with ⩾40 years of experience) also stated that CARATV2.0 validated or organised their own decision-making process. In contrast, nurses and the junior accredited pharmacists felt that they could use this tool as a reference for their medication reviews and patient assessments, especially when choosing among OACs. Pharmacists and nurses also emphasised that patients’ risk factors associated with antithrombotic therapy were not static; therefore, they tended to see this tool as most useful for regular reviews of patients. This aspect of tool use seemed to be overlooked by the GPs and specialists.

Theme 2: health professionals’ and patients’ preferences for a particular antithrombotic therapy

Health professionals’ opinions on CARATV2.0’s recommendations were underpinned by whether they perceived the tool as preferring any particular antithrombotic therapy and whether this therapy was the one they preferred to use. While warfarin was preferred by the majority of health professionals, one neurologist, two haematologists, one GP and one nurse stated that they preferred using NOACs. Some pro-NOAC health professionals perceived that CARATV2.0 was biased towards warfarin and thus distrusted CARATV2.0’s recommendation when it did not recommend their preferred therapy. Similarly, several pro-warfarin health professionals questioned and disliked CARATV2.0’s recommendations because it did not allow negotiation with their preference. Because patients were routinely referred to either GPs (in remote and regional areas) or specialists (in metropolitan areas) for the prescription of antithrombotic therapy, many health professionals believed that the usefulness of CARATV2.0 in improving antithrombotic selection would depend on whether the tool’s recommendations are followed by the GPs and hospital doctors.

Some GPs believed that CARATV2.0 might be able to assist in negotiations with patients by providing evidence (e.g. stroke risk score) for explanations. However, one GP argued that CARATV2.0 could not help in persuading patients to take certain OACs because the negotiation to persuade or convince patients to take antithrombotic therapy involves managing individualised health expectations rather than only presenting scientific evidence about this form of therapy.

Pharmacists’ and nurses’ perspectives on the usefulness of CARATV2.0 was largely determined by whether they thought this tool considered important issues in medication management when selecting antithrombotic therapy for individual patients. While pharmacists focused more on the medication safety issues (e.g. drug–drug interactions, adherence and international normalised ratio (INR)) when using this tool, nurses paid more attention to the tool’s assessment of patients’ capability to manage their medications (e.g. mobility, cognitive function and lifestyle). Although pharmacists and nurses believed that CARATV2.0 comprehensively assessed the major medication management issues, they also pointed out that CARATV2.0 did not consider every issue, for example, use of fish oil supplements or binge drinking. But they also admitted that the selective inclusion of the most important medication management issues ensured the simplicity and ease of use of this tool.

Theme 3: integration into existing systems and processes

Hospital-based health professionals (including specialists, nurses and pharmacists) and GPs suggested integration of CARATV2.0 into existing systems and processes due to the limited time available for making clinical decisions. Most hospital-based health professionals and GPs recommended that the tool’s usefulness could be improved if it was integrated into or linked with electronic medical records or the electronic prescribing software used in hospitals and general practices. They recommended it be accessible through mobile phones, computers and tablets in order to self-populate the medical records and databases and to make the tool easily accessible and portable. However, both GPs and hospital-based health professionals were worried that some of the practice computer software might not interact well with CARATV2.0. They were also concerned that some of the information required by CARATV2.0 might not be available in the electronic health system, which would mean that such information would require manual entering. To solve this problem, some suggested that pharmacists, junior medical residents, medical students or practice staff (e.g. nurses) could populate CARATV2.0 manually, allowing senior clinicians more time to review CARATV2.0’s recommendations.

In contrast to hospital-based health professionals, the time needed for populating CARATV2.0 was not raised as a major issue by community-based pharmacists and community-based nurses. Some of these practitioners actually thought that CARATV2.0 would save them time because it considers all the antithrombotics indicated for AF and integrates many relevant risk assessment tools into one tool. They paid more attention to how the tool’s usability could be improved by incorporation into their medication review and patient assessment processes within an electronic format (e.g. ‘apps’, websites or software).

Discussion

The results from this study show that CARATV2.0 is generally welcomed by health professionals and that they consider it can potentially improve prescription of OACs and clinical outcomes of patients. This is consistent with the evidence that computerised decision support tools can significantly improve prescription among clinicians and can improve the quality and safety of care provided.^28,29 Although decision support tools that focus on the assessment of stroke and bleeding risk are widely available,³⁰ so far CARATV2.0 is the only tool that integrates stroke risk assessment, bleeding risk assessment and medication safety assessment and that considers both the traditional antithrombotic agents, warfarin and the NOACs.

Since decision-making is an emotive process, comprehensive risk versus benefit assessment, systematic documentation and communication of decisions can assist in the selection of optimal therapy for individual patients.³ However, due to limited experience with the use of newly available NOACs, especially NOAC use in elderly patients, the risk versus benefit assessment of using OACs in these, and other, patients is a complex task for many health professionals. CARATV2.0’s comprehensive risk versus benefit assessment of individual patients provides guidance and a reference for, and confidence in, not only the decision on whether a patient should be treated with antithrombotics but also choosing the appropriate therapy among various OACs. Furthermore, because CARATV2.0 is based on the latest clinical guidelines, the tool can also reinforce the use of clinical guidelines by health professionals.

The study found that subjective issues, including clinician and patient preferences for particular antithrombotic therapies, can have a substantial impact on the clinical decision-making process.³¹ Studies have shown that clinicians tend to override recommendations made by a decision support tool if they have a strong preference for a particular medication.³² Also, patients’ preference of therapy has been reported to substantially affect the clinical decision-making for therapy.³³ Although it is widely recognised that computerised decision support tools have the potential to improve the behaviour of clinicians in terms of prescription and consistency of decision-making, evidence supporting the long-term impact of decision support tools on clinicians’ prescribing behaviour is lacking.³⁴ Given these subjective issues, the impact of CARATV2.0 on decision-making for antithrombotic therapy in AF needs to be further explored.

The suggestion of integrating CARATV2.0 into existing systems and processes shows that the health professionals valued the tool as an effective support for clinical decision-making.³² According to a systematic review by Kawamoto et al.,³⁵ a successful decision support tool needs to be computer-based, to have automatic provision of decision support as part of clinician workflow, to provide recommendations rather than just assessments and to provide decision support at the time and location of decision-making. However, this suggestion also reflects that clinicians are reluctant to prioritise and allocate time for the initial decision-making around antithrombotics, which contrasts with the time spent in managing the adverse outcomes of poor or suboptimal prescription. Also, ‘pharmacotherapy’ as an intervention follows a less structured decision-making process than other interventions such as surgery.

Some limitations of the study need to be acknowledged. The participating specialist clinicians, GPs, pharmacists and nurses in this study were volunteers who showed interest in the study. This could have biased their feedback on CARATV2.0. Also, the sampling strategy affects to some extent the generalisability of the study findings beyond these participants. Furthermore, CARATV2.0 inputs are based on available evidence from guidelines and reviews, which may not be relevant to all patient populations and may change over time as new evidence emerges. Since the findings from this research are restricted to the content and feasibility of this tool, the potential clinical and economic impact of the tool and the feasibility of using the tool in real-world clinical practice require further evaluation.

Overall, the feedback from health professionals identifies that the only drawback of this tool is the time needed to complete the assessment (i.e. input the relevant data). To address this issue for future application in practice, CARATV2.0 may be integrated into other systems (e.g. electronic medical records) to enable the auto-population of patient data into the tool. Furthermore, the tool may be used by other health professionals (e.g. NPs and consultant pharmacists) where comprehensive patient assessment and medication review are part of their targeted services.^17,36

Conclusion

CARATV2.0 was regarded by a variety of health professionals as a potentially useful tool that provided a systematic assessment around the decision-making for antithrombotic therapy in patients with AF. The tool also shows potential for rationalising the use of antithrombotics and for improving the clinical outcomes of patients with AF. Future research should evaluate the impact of this tool on the prescription of antithrombotics in clinical practice. The main drawback of this prototype tool is that it requires the manual input of data, which may not be time-efficient for busy health professionals. Therefore, processes for the auto-population of the tool with relevant patient data need to be explored, for example, the integration of CARATV2.0 into electronic databases or prescribing software or/and re-formatting it into a mobile online application.

Footnotes

Acknowledgements

The authors would like to thank Ekta Pandya, Shamsher Singh, Riana Rahmawati and Leigh Findlay.

Authors’ contribution

All authors have made substantial contributions to all of the following: (1) the conception and design of the study, acquisition of data, analysis and interpretation of data; (2) drafting the article or revising it critically for important intellectual content; and (3) final approval of the version to be submitted.

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship and/or publication of this article.

Ethical approval

Ethical approval was given by University of Technology Sydney (UTS HREC REF NO. 2013000338).

Funding

The author(s) received no financial support for the research, authorship and/or publication of this article.

References

Bajorek

A review of the safety of anticoagulants in older people using the medicines management pathway: weighing the benefits against the risks. Ther Adv Drug Saf 2011; 2: 45–58.

Bajorek

Krass

Ogle

. The impact of age on antithrombotic use in elderly patients with non-valvular atrial fibrillation. Australas J Ageing 2002; 21: 36–41.

Bajorek

Ogle

Duguid

. Management of warfarin in atrial fibrillation: views of health professionals, older patients and their carers (Research Support, Non-US Gov’t). Med J Aust 2007; 186: 175–180.

De Breucker

Herzog

Pepersack

. Could geriatric characteristics explain the under-prescription of anticoagulation therapy for older patients admitted with atrial fibrillation? A retrospective observational study. Drugs Aging 2010; 27: 807–813.

Wang

Bajorek

New oral anticoagulants in practice: pharmacological and practical considerations. Am J Cardiovasc Drugs 2014; 14: 1–15.

Bajorek

Masood

Krass

Development of a Computerised Antithrombotic Risk Assessment Tool (CARAT) to optimise therapy in older persons with atrial fibrillation. Australas J Ageing 2012; 31: 102–109.

Bajorek

Krass

Ogle

. Optimizing the use of antithrombotic therapy for atrial fibrillation in older people: a pharmacist-led multidisciplinary intervention. J Am Geriatr Soc 2005; 53: 1912–1920.

Bajorek

Magin

Hilmer

. A cluster-randomized controlled trial of a computerized antithrombotic risk assessment tool to optimize stroke prevention in general practice: a study protocol. BMC Health Serv Res 2014; 14: 55.

LaHaye

Gibbens

Ball

DGA

. A clinical decision aid for the selection of antithrombotic therapy for the prevention of stroke due to atrial fibrillation. Eur Heart J 2012; 33: 2163–2171.

10.

Casciano

Singer

Kwong

. Anticoagulation therapy for patients with non-valvular atrial fibrillation. Am J Cardiovasc Drugs 2012; 12: 313–323.

11.

Wang

Bajorek

Decision-making around antithrombotics for stroke prevention in atrial fibrillation: the health professionals’ views. Int J Clin Pharm 2016; 38: 985–995.

12.

Lip

GYH

Nieuwlaat

Pisters

. Refining clinical risk stratification for predicting stroke and thromboembolism in atrial fibrillation using a novel risk factor-based approach: the Euro heart survey on atrial fibrillation. Chest 2010; 137: 263–272.

13.

Pisters

Lane

Nieuwlaat

. A novel user-friendly score (HAS-BLED) to assess 1-year risk of major bleeding in patients with atrial fibrillation: the Euro heart survey. Chest 2010; 138: 1093–1100.

14.

Gage

Yan

Milligan

. Clinical classification schemes for predicting hemorrhage: results from the National Registry of Atrial Fibrillation (NRAF). Am Heart J 2006; 151: 713–719.

15.

January

Wann

Alpert

. AHA/ACC/HRS guideline for the management of patients with atrial fibrillation: a report of the American College of Cardiology/American Heart Association Task Force on practice guidelines and the heart rhythm society. J Am Coll Cardiol 2014; 64: e1–e76.

16.

Becker

DeMong

Kaplan

. Anticoagulation therapy and primary care internal medicine: a nurse practitioner model for combined clinical science. J Gen Intern Med 1994; 9: 525–527.

17.

Stafford

Peterson

Bereznicki

. Clinical outcomes of a collaborative, home-based postdischarge warfarin management service. Ann Pharmacother 2011; 45: 325–334.

18.

You

Singer

Howard

. Antithrombotic therapy for atrial fibrillation: antithrombotic therapy and prevention of thrombosis, 9th ed: American College of Chest Physicians evidence-based clinical practice guidelines. Chest 2012; 141: e531S–e575S.

19.

Wann

Curtis

Ellenbogen

. Management of patients with atrial fibrillation (compilation of 2006 ACCF/AHA/ESC and 2011 ACCF/AHA/HRS recommendations): a report of the American College of Cardiology/American Heart Association Task Force on practice guidelines. Circulation 2013; 127: 1916–1926.

20.

Miller

Grandi

Shimony

. Meta-analysis of efficacy and safety of new oral anticoagulants (Dabigatran, Rivaroxaban, Apixaban) versus warfarin in patients with atrial fibrillation. Am J Cardiol 2012; 110: 453–460.

21.

Australian Government Department of Health and Ageing. Review of anticoagulation therapies in atrial fibrillation, http://www.pbs.govau/reviews/atrial-fibrillation-files/report-anticoagulation.pdf (2012, accessed 2 February 2015).

22.

Trusler

Well-managed warfarin is superior to NOACs. Can Fam Physician 2015; 61: 23–24.

23.

Gage

Waterman

Shannon

. Validation of clinical classification schemes for predicting stroke: results from the National Registry of Atrial Fibrillation. JAMA 2001; 285: 2864–2870.

24.

Pangman

Sloan

Guse

An examination of psychometric properties of the mini-mental state examination and the standardized mini-mental state examination: implications for clinical practice. Appl Nurs Res 2000; 13: 209–213.

25.

Johnson

Jones

Mathew

. Chronic kidney disease and automatic reporting of estimated glomerular filtration rate: new developments and revised recommendations. Med J Aust 2012; 197: 224–225.

26.

Pugh

Murray-Lyon

Dawson

. Transection of the oesophagus for bleeding oesophageal varices. Br J Surg 1973; 60: 646–649.

27.

Pirmohamed

Warfarin: almost 60 years old and still causing problems. Br J Clin Pharmacol 2006; 62: 509–511.

28.

Hylek

Borowsky

. Warfarin use among ambulatory patients with nonvalvular atrial fibrillation: the anticoagulation and risk factors in atrial fibrillation (ATRIA) Study. Ann Intern Med 1999; 131: 927–934.

29.

Fuster

Rydén

Cannom

. ACC/AHA/ESC 2006 guidelines for the management of patients with atrial fibrillation: full text A report of the American College of Cardiology/American Heart Association Task Force on practice guidelines and the European Society of Cardiology Committee for Practice Guidelines (Writing committee to revise the 2001 guidelines for the management of patients with atrial fibrillation): developed in collaboration with the European Heart Rhythm Association and the Heart Rhythm Society. Europace 2006; 8: 651–745.

30.

Wang

Bajorek

Safe use of antithrombotics for stroke prevention in atrial fibrillation: consideration of risk assessment tools to support decision-making. Ther Adv Drug Saf 2014; 5: 21–37.

31.

Man-Son-Hing

Gage

Montgomery

. Preference-based antithrombotic therapy in atrial fibrillation: implications for clinical decision making. Med Decis Making 2005; 25: 548–559.

32.

Monette

Gurwitz

Rochon

. Physician attitudes concerning warfarin for stroke prevention in atrial fibrillation: results of a survey of long-term care practitioners. J Am Geriatr Soc 1997; 45: 1060–1065.

33.

Zhu

W-G

Xiong

Q-M

Hong

Meta-analysis of CHADS2 versus CHA2DS2-VASc for predicting stroke and thromboembolism in atrial fibrillation patients independent of anticoagulation. Tex Heart Inst J 2015; 42: 6–15.

34.

Singer

Albers

Dalen

. Antithrombotic therapy in atrial fibrillation: American College of Chest Physicians evidence-based clinical practice guidelines. Chest 2008; 133: 546S–592S.

35.

Kawamoto

Houlihan

Balas

. Improving clinical practice using clinical decision support systems: a systematic review of trials to identify features critical to success. BMJ 2005; 330: 765.

36.

Hendriks

JML

De Wit

Crijns

HJGM

. Nurse-led care vs. usual care for patients with atrial fibrillation: results of a randomized trial of integrated chronic care vs. routine clinical care in ambulatory patients with atrial fibrillation. Eur Heart J 2012; 33: 2692–2699.