Assessment of the attitude, awareness and practice of periprocedural warfarin management among health care professional in Qatar. A cross sectional survey

It is estimated that 10–15% of oral anticoagulant (OAC) patients, would need to hold their OAC for scheduled surgery. Especially for warfarin, this process is complex and requires multi-layer risk assessment and decisions across different specialties. Clinical guidelines deliver broad recommendations in the area of warfarin management before surgery which can lead to different trends and practices among practitioners. To evaluate the current attitude, awareness, and practice among health care providers (HCPs) on warfarin periprocedural management. A multiple-choice questionnaire was developed, containing questions on demographics and professional information and was completed by187 HCPs involved in warfarin periprocedural management. The awareness median (IQR) score was moderate [64.28% (21.43)]. The level of awareness was associated with the practitioner’s specialty and degree of education (P = 0.009, 0.011 respectively). Practice leans to overestimate the need for warfarin discontinuation as well as the need for bridging. Participants expressed interest in using genetic tests to guide periprocedural warfarin management [median (IQR) score (out of 10) = 7 (5)]. In conclusion, the survey presented a wide variation in the clinical practice of warfarin periprocedural management. This study highlights that HCPs in Qatar have moderate awareness. We suggest tailoring an educational campaign or courses towards the identified gaps.

• Warfarin periprocedural management is complex and requires multiple risk assessments and synchronized decisions. • Clinical guidelines deliver broad recommendations in the area of warfarin management before surgery which can lead to different trends and practices among practitioners. • This survey evaluates the current attitude, awareness, and practice among health care providers (HCPs) on warfarin periprocedural management. • The project presented a wide variation in the clinical practice of warfarin periprocedural management and moderate awareness.

Introduction
Oral anticoagulants (OAC) have been used for years in the treatment and prevention of thromboembolism [1,2]. Notably, in Qatar, as well as other parts of the world, warfarin still represents a significant portion of total OAC used [3]. It has been estimated that 10-15% of OAC patients worldwide need to undergo an elective procedure on an annual basis, which may require holding OAC [4]. Periprocedural management of warfarin is a complicated process since it involves multiple steps, each of which must be assessed carefully before making a comprehensive plan. The first step is to decide whether warfarin should be interrupted. While warfarin interruption leads to decreased bleeding risk during and post-procedure, it can also increase the risk of thromboembolism [5]. Second, comes the bridging decision which may be considered to reduce the risk of thromboembolism in patients with moderate to high thromboembolic risk, however, increased risk of bleeding must be put into account [6]. In Perioperative Bridging Anticoagulation in Patients with Atrial Fibrillation trial (the BRIDGE Trial), 1884 warfarin-receiving patients with atrial fibrillation (AF) (mean CHA2DS2-VASc of 2.4) were randomly assigned to receive bridging with low molecularweight heparin (LMWH) or a placebo-controlled bridging perioperatively [7]. The study found that bridging was associated with a more frequent incidence of major bleeding compared to non-bridging (relative risk [RR] = 0.41, 95% confidence interval [CI], 0.2-0.78, P = 0.005). Furthermore, LMWH did not prevent arterial thromboembolism significantly. Similarly, the Outcomes Registry for Better Informed Treatment of Atrial Fibrillation (ORBIT-AF) trial showed that the composite outcome of systemic embolism or stroke, myocardial infarction, bleeding or hospitalization was elevated in the bridging arm significantly [8]. Both studies augment the uncertainty of the need for bridging. Adding to the complexity of the bridging process is that the decision of warfarin interruption according to procedure and patient's bleeding risks are considered another controversy. Most of the guidelines stratify the risk of thromboembolism and procedural bleeding risk into high and low, to facilitate the interruption decision [6]. Unfortunately, these classifications have some drawbacks, such as procedures with a low rate of bleeding, but with severe consequences. Categorizing these procedures as a low bleeding risk instead of a high bleeding risk procedure may be misleading. Moreover, the classification did not consider the level of intermediate bleeding risk category and did not include patients with atrial fibrillation (AF). Besides, there is a disagreement regarding the classification of some procedures such as hip/ knee replacement and prostate biopsy [9].
Collectively, it is evident that the judgment of warfarin holding and periprocedural bridging is not explicit, and decision-makers can be easily misled. This can also create several practices and attitudes among health care professionals. Consequently, a survey on the periprocedural management of warfarin was developed for a better understanding of the current practice, the gap in knowledge and attitude among health care providers in Qatar.

Study design and population
This study is an observational prospective cross-sectional self-administered questionnaire survey that aims to understand the practice, awareness, and attitude of health care professionals (HCPs) at Hamad Medical Corporation (HMC), Qatar, toward periprocedural management of warfarin patients.
The study was conducted over six months from July 2019 till January 2020. The participants were among physicians and clinical pharmacists from various departments involved in the periprocedural management of warfarin. A hard copy of the survey was delivered by one of the investigators. The first page of the survey contained an introductory invitation informing participants about the purpose and objectives of the survey, and confirming that the contribution to the survey was voluntary and anonymous. Convenience sampling method was used to approach the participants.

Study setting and ethics approval
The study was performed at Al Wakra Hospital (AWH), Hamad General Hospital (HGH), and Heart Hospital (HH). These Three sites are tertiary hospitals and part of HMC, the most prominent medical institution in Qatar. Ethical approval was obtained from the Institutional Review Board (IRB) of HMC in July 2019 (Protocol# MRC-01-19-57).

Sample size calculation
The sample size was calculated using Roasoft online calculator (www.Roaso ft.com) [10], assuming that the HCPs who are involved in warfarin periprocedural management at HMC are 600. To achieve a confidence (power) level of 90% power with a 5% marginal error and taking into consideration 50% response distribution, a sample size of 187 participants was found to be adequate.

Validation and piloting
Content and structure were checked for validity by three senior faculty members at the College of Pharmacy, Qatar University (one with expertise in pharmacy practice research, and two with cardiovascular clinical practice background). Based on their feedback, modifications were performed. A pilot version was created and disseminated to a random sample of (one internal medicine senior consultant, one cardiology specialist, one general resident physician, and one clinical pharmacist). Respondents reported that the questionnaire was well organized, clear, and with a proper sequence of questions. They also completed the survey within 15-20 min, which matched the stated duration at the invitation page of the survey.

Survey development
The survey was designed after performing a thorough literature review using PubMed, Google Scholar, and EMBASE database in January 2019. The search focused on terms related to the HCP's awareness and practice in warfarin periprocedural management.
The survey consisted of four domains. The first domain had 5 questions to assess the attitude of HCPs. The second domain contained 7 questions, and it evaluated the HCP`s practice. The third domain was two case scenarios with 14 questions that assessed the awareness of HCPs. The last domain collected relevant demographic and professional characteristics information of the participants. There was one question with a score ranging from 0 to 10 with one-unit intervals to rate the willingness of HCPs to recommend a genetic test to guide the duration of warfarin discontinuation. The final version of the survey consisted of 31 multiplechoice questions. Survey questions were available only in the English language.

Measured outcome and statistical analysis
All responses were recorded in Excel document and transferred to IBM Statistical Package for Social Science (IBM SPSS 26 software; IBM, New York) for descriptive and inferential statistical analysis. Responses to demographics, professional information practice, and attitude towards periprocedural warfarin management questions, were represented as categorical variables and were expressed in frequencies and percentages. One question was presented as a continuous variable. An awareness score of one point was provided if the participant selected the correct answer for the designated question. For questions with more than one correct answer, a partial score was provided unless the participant selected all the correct answers. The overall score awareness domain was the sum of the scores of all questions under this domain. Percentage Awareness score (PAS) was calculated by dividing the total awareness score by the maximum possible score and multiplying the result by 100. Since data were non-normally distributed, Mann-Whitney U-test and Kruskal-Wallis H test were used to evaluate the effect of participants' demographics and personal information on PAS which was expressed as median and interquartile range (IQR). A Chi-square test was performed to assess the association between different categorical values. Two-tailed P-value of < 0.05 was considered significant.

Participants' characteristics
Over six months, a total of 300 questionnaires were distributed, among which 187 questionnaires were collected (62.3% response rate). The plurality of participants (74.4%) were male, and the majority of them (69.3%) had less than 20 years of experience. Responses were received from 150

Awareness of periprocedural warfarin management
The overall median (IQR) of PAS was moderate 64.28% (21.43). Out of 14 awareness questions, the major deficiency was identified in 5 questions [less than 50% of responders chose the right answer(s)]. Firstly, there is the awareness of the type of surgeries that do not require warfarin interruption (right response rate = 26.2%). Also, there is the awareness regarding the time at which patients must stop warfarin and stop LMWH prior to surgery (right response rate = 42.2%, 47.1%, respectively). Furthermore, bridging decision was another obstacle in both case scenarios (right response rate = 38% & 47.6%). In bridging decision scenarios, we found apparent contrast in response among specialties.  (Fig. 1). Table 3 shows the effect of baseline and professional characteristics on PAS.

The practice of HCPs in periprocedural warfarin management
Most of the respondents (87.9%) reported that they deal with 1-2 warfarin patients per week undergoing a procedure. There was a statistically significant association between specialty and who is accounted for the direct management of these cases (P < 0.001). Half of the cardiologists (50%) indicated that the anticoagulant clinic is responsible for making plans for the patient, while a similar proportion of internal medicine agreed on warfarin prescriber as the main responsible party. In contrast, 37.5% of surgeon and anesthesia physicians declared that clinician performing the procedure is liable to handle these cases. About a third of the HCP indicated that they encounter a reschedule/ cancellation of the procedure due to elevation in INR some or most of the time.
In terms of warfarin interruption, 85.2% of respondents indicated that around 75% of patients need warfarin discontinuation before elective surgery, and that about half of those patients (56.1%) will require bridging.
When the respondents were asked to indicate which criteria are used to assess patient's stroke risk, just under 70% reported that they use the CHA2DS2-VASc score, while fewer (17%) reported the use of CHADS 2 score.

Attitude towards periprocedural warfarin management
A chi-square test for association was conducted between demographics and warfarin periprocedural management attitude. Females significantly perceived more than males that warfarin interruption, and heparin bridging are overused (34% vs 22.6%, P = 0.003, P = 0.034 respectively). More emphasis on the difference in the attitude of physicians and pharmacists; whereby, more physicians believed that the cost of bridging is very important (38.5%vs 24.3%, P = 0.042). Participants expressed a good level of interest in using genetic tests to guide periprocedural warfarin management [median (IQR) score (out of 10) = 7 (5)].

Discussion
In this study, we attempted to assess the attitude, knowledge, and practice of HCPs in Qatar on periprocedural management of warfarin patients undergoing a procedure. The main finding of the study was that participants' awareness is moderate. In a recent study in Qatar, a similar level of awareness was achieved among HCPs on direct oral anticoagulants (DOACs) [13]. Three areas of knowledge deficiency were the driver of the decline in awareness level in the current study. Firstly, conflicting ability to determine the duration of discontinuation of warfarin prior to the procedure. This is surprising given the fact that a clear recommendation in the 2017 American College of Cardiology (ACC) guideline states that, warfarin should be held 5-7 days before an elective procedure [9]. A second area of deficiency was the inconsistencies between HCPs on who bridge warfarin patient and the duration of preoperative parenteral anticoagulation when a decision to bridge is made. Thirdly, the majority of participants were lacking awareness of the type of procedures that do not require warfarin interruption, such as cataract and tooth extraction due to their low-risk of bleeding [14,15]. Whether the low score achieved in these elements is due to true lack of awareness or judgement from clinical practice and experience is hard to assess. Regardless, we believe that applying inappropriate timing, duration of warfarin interruption or bridging can yield significant risk of bleeding and thromboembolic events. It was also found that cardiologists were the best in continuing warfarin in procedures with low risk of bleeding, while most of the surgeons still stopped warfarin. This is potentially due to the cardiologists' attention to patient's thromboembolic risk, while surgeons give more attention to the procedure's bleeding risk. Results from a survey that evaluated the practice patterns in the Unites States for bridging AC showed that 25% and 45% decided not to interrupt warfarin during dental extraction and cataract surgery, respectively [16]. Bridge or Continue Coumadin for Device Surgery Randomized Controlled trial (BRUISE CONTROL) has shown that maintaining warfarin with an INR of ≤ 3 on the day of the procedure in patients undergoing implantation of pacemakers or cardioverter defibrillators was associated with significantly less bleeding than warfarin discontinuation along with bridging with heparin (Odds ratio:0.19; p < 0.001) [17]. Another critical observation in the survey is that clinical pharmacists had better awareness scores compared to physicians. A possible explanation for this might be that clinical pharmacists have a reasonable knowledge of pharmacokinetics and pharmacology of warfarin, and are frequently involved with warfarin dosing and periprocedural management through anticoagulation clinics and in-patient services [12]. A significant difference was also noted among the physician's specialties, where cardiologists and internalists achieved the highest scores. This result is likely related to these specialties being more involved in the management of warfarin patients.
As expected, HCPs holding professional degrees had a superior awareness than fresh graduate HCPs holding a bachelor's degree. Surprisingly, HCPs with PhD got a lower awareness score than HCPs with a professional degree. It is possible that practical training plays a significant factor in determining the awareness level. We also observed that the position or rank was positively associates with the awareness of periprocedural warfarin management (highest in consultants/senior consultants). While one may expect from recent graduates to have better awareness, extensive clinical practice appears to have a vital role in augmenting awareness levels. These results are also in alignment with the previous survey on DOACs awareness in Qatar [13].
Response to the involvement in periprocedural warfarin management was another interesting finding. The majority of each specialty were biased towards their own practice. For instance, cardiologists, being the specialty running jointly or in close relation to the anticoagulation clinics in Qatar agreed on the anticoagulant service as the main responsible party for periprocedural management. Similarly, surgeons and anesthesiologists referred to the clinician performing the procedure as the responsible, while internalist referred to the warfarin prescriber as the responsible party. These findings are consistent with data from a recent survey in which respondents distributed the responsibility among cardiologists, surgeons, internalists and anticoagulant services to manage warfarin periprocedural (56%, 36%, 28%,and 27%, respectively) [16].
In addition to our main findings above, respondents revealed that warfarin is discontinued in the majority of patients who will undergo elective surgery. This was reflected when most of the participants chose to stop warfarin in cataract and tooth extraction surgeries in separate questions. Similar trends were expressed by participants in this survey and those described by Starks et al. [18], Krahan et al. [19], and Balbino et al. [20] (75%, 83%, and 83% interrupted warfarin preoperatively correspondingly). We believe that this clinical practice leans towards fear of bleeding events from warfarin much more than thromboembolic events. However, HCPs in our study stated that almost half of those patients undergoing warfarin discontinuation would require bridging to protect them from thromboembolic events. Both of these practice behaviors (exaggerated discontinuation and bridging) may put the patients at higher risk of thromboembolism and bleeding, respectively. This comes also against the recent expert call to reduce the use of bridging during preoperative management due to the increased risk of bleeding from heparin use [21]. In this report, it was estimated that over 90% of patients receiving warfarin therapy should not receive bridging anticoagulation during periprocedural management. This conclusion was based on accumulating evidence that rated overall and major bleeding significantly higher in bridged rather than non-bridged patients by 2-5 folds while there was no difference in the risk of thromboembolism between both arms [22].
As an area of future research and possible clinical translation we asked HCPs on their opinion to use a genetic test as a tool to help in personalizing the duration of warfarin interruption before surgery. Remarkably, the survey articulated the interest of HCPs (especially pharmacists) in recommending this tool to their patient in the future. These results are in agreement with Elewa et al. [23] findings in 2015, which showed that pharmacists had more willingness and positive attitude towards the application of pharmacogenetics in practice when compared to physicians in Qatar.
A key strength of the current survey is that it investigated different domains (attitude, knowledge, and practice) of various specialties involved in warfarin periprocedural management. On the other hand, this study had some limitations. First, there is a potential for sampling bias since we surveyed a governmental hospital only, i.e. HMC, which could affect the generalizability of the results. Despite a high response rate in this survey (62.3%), some HCPs did not agree to participate possibly due to lack of knowledge or interest which may have had an impact on the generalizability of the results. To overcome that, we intentionally used a paper-based survey instead of an online version to increase the response rate. In addition to the above limitations, survey fatigue, and lack of required time to answer the survey are obstacles that could have affected the response quality. We tried to solve this issue by limiting the number of case scenarios. Moreover, validation of the questionnaire helped to ensure it had appropriate time and clarity. Lastly, and similar to other survey-based studies, our findings may be distinct from what applies in practice.
In conclusion, this research highlights that HCPs in Qatar have moderate awareness of warfarin periprocedural management with a lack of standardized practice. Practice leans to overestimate the need for warfarin discontinuation due to fear of bleeding risk. Besides, it overestimates the need for bridging to overcome thromboembolic risk. Additionally, HCPs are interested in applying pharmacogenetics to their practice to gage the duration of warfarin discontinuation. Future work should focus on reassessing practitioners' knowledge after providing well-designed education campaigns.