Physician decision making in anticoagulating atrial fibrillation: a prospective survey of a physician notification system for atrial fibrillation detected on cardiac implantable electronic devices of patients at increased risk of stroke

Abstract

Objectives:

The objectives of this study were to evaluate the effectiveness of a physician notification system for atrial fibrillation (AF) detected on cardiac devices, and to assess predictors of anticoagulation in patients with device-detected AF.

Methods:

In 2013, a physician notification system for AF detected on a patient’s CIED [including pacemakers, implantable cardioverter defibrillators (ICD) or cardiac resynchronization therapy (CRT) devices] was implemented, with a recommendation to consider oral anticoagulation in high-risk patients. We prospectively investigated the effectiveness of this system, and evaluated both patient and physician predictors of anticoagulation, as well as factors influencing physician decision making in prescribing anticoagulation. Both uni- and multivariable analysis as well as descriptive statistics were used in the analysis.

Results:

We identified 177 patients with device-detected AF, 126 with a CHADS₂ ⩾2. Only 41% were prescribed anticoagulation at any point within 12 months. On multivariable analysis, stroke risk as predicted by CHADS₂ was not a predictor of anticoagulation. ASA use predicted a lower rate of anticoagulation (OR 0.39, 95% CI 0.16–0.97, p = 0.04); physicians in practice for <20 years were more likely to prescribe anticoagulation (OR 3.39, 95% CI 1.28–8.93, p = 0.01); and physicians who believed both cardiologist and family doctor should be involved in managing anticoagulation were more likely to prescribe anticoagulation (OR 3.28, 95% CI 1.02–10.5, p = 0.05).

Conclusions:

Patients on aspirin were less likely to be anticoagulated. Physicians in practice for <20 years and who believed that both the general practitioner and cardiologist should be involved in managing anticoagulants were more likely to prescribe anticoagulation.

Keywords

anticoagulation atrial fibrillation devices subclinical

Introduction

Nearly 800,000 people have a stroke every year in the United States.¹ Approximately 17% of strokes are immediately attributable to atrial fibrillation (AF), while 25% have no apparent cause and are labeled as ‘cryptogenic’.² Strokes related to AF are associated with ~50% increased risk of disability and death within 3 months compared to strokes related to other etiologies.³ It is increasingly recognized that a large proportion of patients labeled with cryptogenic stroke have device-detected AF.^4–6

Asymptomatic AF is frequently detected on pacemakers and defibrillators, and predicts an elevated risk of stroke.⁷ Many patients with clinical and device-detected AF are not anticoagulated and remain at increased risk of stroke.⁸ The decision to anticoagulate these patients should be individualized, considering the patient’s stroke and bleeding risk, among other factors. Perceived risk of anticoagulation, as well as physician inexperience, have previously been associated with a lower rate of initiating anticoagulation of AF.⁹

We implemented a physician notification system for device-detected AF, with a recommendation to consider oral anticoagulation. We sought to assess its effectiveness, and evaluate what patient and physician factors might influence the decision to initiate anticoagulation, through a physician survey. No previous studies have reported on the effectiveness of a physician alert system for device-detected AF, and none have described patient and physician predictors of prescribing anticoagulation to patients with device-detected AF. Our objectives were twofold. First, to evaluate the effectiveness of a physician notification system for anticoagulation of device-detected AF; and second, to assess if patient or physician factors influence the decision to initiate anticoagulation in a patient with device-detected AF.

Methods

The pacemaker and defibrillator clinic at St. Boniface Hospital (Winnipeg, Canada) sees over 6000 patients annually. The clinic is staffed by nurses trained in cardiac device interrogation, with access to a cardiac electrophysiologist for troubleshooting. In 2013, a physician notification system for AF detected on a CIED was implemented. We prospectively evaluated a cohort of patients with AF detected on their CIED between March 2013 and January 2014.

The intracardiac electrograms of all atrial high-rate episodes were reviewed by a cardiac electrophysiologist experienced in device interrogation to confirm the presence of AF lasting ⩾30 s, as per ACC/AHA guidelines.¹⁰ If AF was confirmed, the referring physician was sent a notification via facsimile regarding the presence of device-detected AF, along and a recommendation to consider oral anticoagulation in high-risk patients. Patients were included in the study if they had no previous diagnosis of AF and were not on anticoagulation at presentation.

PACEART^TM (Medtronic, Minneapolis, USA) – a proprietary database designed for cardiac device patients – was used to assess duration of AF, device type, and demographics. We reviewed electronic chart records for patient demographics related to stroke and bleeding risk. The provincial Drug Program Information Network provided patient anticoagulant prescriptions up to 12 months after physician notification. A survey (Appendix A) was sent to all referring providers to identify physician characteristics, including years of training, attendance at continuing medical education seminars on AF, attitudes and practices in anticoagulating AF, among other data.

Primary outcome was initiation of anticoagulation within 12 months. Univariable analysis was performed to identify factors that predicted initiation of anticoagulation. The Fisher exact test was used for categorical variables, Mann–Whitney U test for numeric variables. Multivariable analysis was performed using a stepwise logistic regression to assess whether physician or patient characteristics predicted anticoagulant initiation. Descriptive statistics were used to describe the reasons for not initiating anticoagulation. The sample size was not pre-calculated, but rather all patients over the study period were enrolled to provide maximum statistical power. Ethics approval was obtained through the University of Manitoba Research Ethics Board (REB) (H2014:367). The REB waived the need for informed consent for the present study as it posed minimal risk to the patient, and from the physician perspective completion and return of the survey constituted their consent to participate.

Results

Between 3 March 2013 and 31 January 2014, over 6000 patients had their cardiac devices interrogated. Of these, 217 patients were identified as having device-detected AF. A total of 177 patients from in-province with complete data available were included in the study (Figure 1). Patient demographics are summarized in Table 1. Fifty-four percent of patients were on ASA at baseline. The rate of anticoagulation at any point in 12 months was 41% for patients with CHADS₂ ⩾2 (Figure 2). In January 2015, we sent 284 surveys out to the 239 physicians who had been notified of the presence of device-detected AF in the patient cohort (Figure 3). Some patients had more than one physician. We received 154 responses for a response rate of 54%. Thirty-nine surveys were excluded due to incomplete information. The CHADS₂ score was ⩾1 in 96% of patients on whom surveys were completed, with 6 min or more of device-detected AF.

Figure 1.

Patient identification and CHADS₂ score.

Table 1.

Patient demographics.

Patient variable	N (% or ± SD)
Male	108 (61%)
Age (years)	80 ± 10
Age > 75 years	129 (73%)
CHF/LV dysfunction	42 (24%)
Hypertension	118 (67%)
Diabetes	47 (27%)
Stroke/embolization	24 (14%)
Prior MI	56 (32%)
CHADS₂	2.2 ± 1.1
CHA₂DS₂-VaSc	3.5 ± 1.5
GFR < 30 ml/min/1.73 m²	15 (8%)
End-stage renal disease	8 (5%)
Labile INR	1 (1%)
Prior major bleed	13 (7%)
Liver disease	1 (1%)
Heavy alcohol use	3 (2%)
NSAID use	8 (5%)
HASBLED score	2.0 ± 0.9
Medications
Aspirin	95 (54%)
Statin	82 (46%)
Beta-blocker	97 (55%)
CCB	42 (24%)
ACE inhibitor	106 (60%)
Spironolactone	12 (7%)
Clopidogrel	19 (11%)

ACE, angiotensin converting enzyme; CCB, calcium-channel blocker; CHF, congestive heart failure; GFR, glomerular filtration rate; LV, left ventricular; MI, myocardial infarction; NSAID, non-steroidal anti-inflammatory drug.

Figure 2.

Anticoagulation rate of patients with device-detected atrial fibrillation and CHADS₂ score ⩾2.

Figure 3.

Physician survey responses and rate of anticoagulation stratified by duration of device-detected atrial fibrillation.

Physician survey responses and univariable analyses are summarized in Table 2 for patients with at least 6 min of AF. Most physicians responded by saying they had received the notification regarding the detection of AF on their patient’s CIED (83%, Table 2). Most respondents were family doctors, and most had attended a CME conference on AF in the previous 5 years. A majority of respondents reported managing anticoagulation on their own, and believed that a family physician ought to oversee anticoagulation, rather than a cardiologist. Physicians who had been in practice for >20 years were less likely to start their patient on anticoagulation, compared to those in practice for <20 years (76% were not anticoagulated, compared to 25% of patients with physicians in practice for <20 years (p = 0.03), Table 2).

Table 2.

Physician survey responses for patients with six minutes or more of device-detected atrial fibrillation.

	Total surveys (n = 99)	No anticoagulation therapy (n = 55)	Anticoagulation therapy (n = 44)	p-value
Did you receive the communication from the pacemaker clinic regarding the detection of atrial fibrillation in your patient?	82 (82.8%)	44 (80.0%)	38 (86.4%)	0.40
If you received the communication, was the diagnosis of atrial fibrillation and recommendation for considering anticoagulation clearly documented?	77 (77.8%)	40 (72.7%)	37 (84.1%)	0.18
Which of the following best describes your specialty?
Family/general practice	79 (82.3%)	45 (83.3%)	34 (81.0%)	0.94
Cardiology	11 (11.5%)	6 (11.1%)	5 (11.9%)
Other	6 (6.3%)	3 (5.6%)	3 (7.1%)
Have you attended a CME session on atrial fibrillation in the last 5 years?	87 (87.9%)	46 (83.6%)	41 (93.2%)	0.15
How many years have you been in practice?
<20 years	33 (34.0%)	13 (24.5%)	20 (45.5%)	0.03
>20 years	64 (66.0%)	40 (75.5%)	24 (54.5%)	0.03
How many patients with atrial fibrillation have you seen in the last 5 years?
50 patients or less	56 (58.3%)	32 (61.5%)	24 (54.5%)	0.49
More than 50 patients	40 (41.7%)	20 (38.5%)	20 (45.5%)	0.49
Who do you think should be in charge of anticoagulation of atrial fibrillation?
Family physician	77 (79.4%)	47 (88.7%)	30 (68.2%)	0.01
Cardiologist or family physician and cardiologist	20 (20.6%)	6 (11.3%)	14 (31.8%)	0.01
How is anticoagulation managed in your practice?
I manage it myself	83 (85.6%)	48 (90.6%)	35 (79.5%)	0.12
Other*	14 (14.4%)	5 (9.4%)	9 (20.5%)

Univariable analysis of patient factors predicting anticoagulation with at least 6 min of device-detected AF is shown in Table 3. Neither stroke risk as predicted by the CHADS₂ score, nor bleeding risk as estimated by the HASBLED score were associated with anticoagulation. Multivariable predictors of anticoagulation for patients with more than six minutes of AF are shown in Table 4. Aspirin use predicted a lower rate of anticoagulation, with an odds ratio of 0.32 (95% CI 0.13–0.78). Physicians in practice for 20 years or less were more likely to prescribe anticoagulants than those in practice for longer than 20 years (OR 4.1, 95% CI 1.61–10.42, p < 0.01). Physicians who believed the family physician and cardiologist should share the responsibility of prescribing anticoagulants were also more likely to prescribe anticoagulants (OR 3.28, 95% CI 1.01–10.5, p = 0.05).

Table 3.

Patient factors predicting anticoagulation with at least 6 min of device-detected atrial fibrillation.

	Total population (n = 90)	No anticoagulation (n = 51)	Anticoagulation therapy (n = 39)	p-value
Gender (female)	37 (41%)	25 (49.0%)	12 (30.8%)	0.08
Valvular disease	1 (1%)	0 (0.0%)	1 (2.6%)	0.43
Congestive heart failure	20 (22%)	13 (25.5%)	7 (18.0%)	0.39
Hypertension	58 (64%)	31 (60.8%)	27 (69.2%)	0.41
Age ⩾ 75	65 (72%)	36 (70.6%)	29 (74.4%)	0.69
Age ⩾ 65	84 (93%)	47 (92.2%)	37 (94.9%)	0.69
Type 2 diabetes	20 (22%)	9 (17.7%)	11 (28.2%)	0.23
Smoker	8 (9%)	6 (11.8%)	2 (5.1%)	0.27
Cerebrovascular accident	9 (10%)	6 (11.8%)	3 (7.7%)	0.73
Vascular disease	1 (1%)	0 (0.0%)	1 (2.6%)	0.43
Myocardial infarction/coronary artery disease	25 (28%)	15 (29.4%)	10 (25.6%)	0.69
eGFR ⩽ 30	7 (8%)	6 (11.8%)	1 (2.6%)	0.13
End-stage renal disease	2 (2%)	2 (3.9%)	0 (0.0%)	0.50
Liver disease	0 (0%)	0 (0.0%)	0 (0.0%)	1.00
Prior bleeding	7 (8%)	4 (7.8%)	3 (7.7%)	1.00
Labile INR	0 (0%)	0 (0.0%)	0 (0.0%)	1.00
Alcohol use	1 (1%)	0 (0.0%)	1 (2.6%)	0.43
NSAID use	4 (4%)	4 (7.8%)	0 (0.0%)	0.13
ASA	45 (50%)	30 (58.8%)	15 (38.5%)	0.06
Clopidogrel	8 (9%)	5 (9.8%)	3 (7.7%)	1.00
Statin	33 (37%)	20 (39.2%)	13 (33.3%)	0.57
Beta-blocker	44 (49%)	24 (47.1%)	20 (51.3%)	0.69
CCB	25 (28%)	13 (25.5%)	12 (30.8%)	0.58
ACE-i/ARB	51 (57%)	29 (56.9%)	22 (56.4%)	0.97
MRA	3 (3%)	2 (3.9%)	1 (2.6%)	1.00
CHADS₂	2 (IQR 1–3)	2 (1–3)	2 (1–3)	0.60
CHADS₂ VASC	3 (IQR 2–4)	3 (2–4)	3 (2–4)	0.85
HASBLED	2 (IQR 2–3)	2 (2–3)	2 (2–3)	0.20

ACE-I, angiotensin converting enzyme inhibitor; ARB, angiotensin receptor blocker; ASA, acetylsalicylic acid; CCB, calcium-channel blocker; INR, international normalized ratio; MRA, mineralcorticoid receptor antagonist; NSAID, non-steroidal anti-inflammatory drug.

Table 4.

Predictors of anticoagulation in patients with 6 min or more of device-detected atrial fibrillation (n = 99 surveys).

Variable	Odds ratio	95% confidence interval	p-value
CHADS₂ score	0.88	0.57–1.36	0.55
ASA	0.39	0.16–0.97	0.04
Physician in practice 20 years or less	3.39	1.28–8.93	0.01
Opinion that combination of family doctor and cardiologist should be involved in managing anticoagulation	3.28	1.02–10.5	0.05

ASA, acetylsalicylic acid.

Discussion

We evaluated the effectiveness of a system alerting physicians to the presence of device-detected AF in a cohort of patients with elevated stroke risk. We found that our notification was widely received by the primary care providers, and that fewer than half of patients were started on anticoagulation. Stroke risk as indicated by CHADS₂ was not associated with a higher rate of anticoagulation. Aspirin use predicted a lower rate of anticoagulation, while physicians who were of a collaborative belief regarding anticoagulation (involving both cardiologist and family physician), and physicians in practice <20 years, were more than three times more likely to prescribe anticoagulation to patients with device-detected AF. Our results suggest that there remains a belief in primary care that aspirin is adequate in preventing stroke in AF, despite data showing that it carries minimal to no benefit, with similar bleeding risk.^11–13

Many patients worldwide with AF who have an increased risk of stroke and low risk of bleeding are not anticoagulated and remain at excessive risk of stroke.^8,14–16 Reports have been published assessing factors that influence physician decision making in anticoagulating AF through case vignettes or general attitudes,^17–30 with far fewer publications examining real-world practices and decision making at the individual patient level.^31–33 Patient factors such as previous stroke,^16,18,31 hypertension,¹⁶ valvular disease,¹⁶ and CHF¹⁶ have been associated with a higher rate of anticoagulation in real-world studies. Previous bleed or risk of bleeding,^18,31–33 paroxysmal AF,^16,32 non-white ethnicity,^16,31 younger^16,31 and older age,³² dementia,³¹ lower functional status,³¹ recurrent falls,²⁸ expected poor patient compliance,¹⁸ perception of a high risk–benefit, patient contraindications,¹⁷ and perceived low risk of stroke¹⁸ have been associated with a lower rate of anticoagulation. In case vignettes and surveys of general attitudes,²⁰ bleeding risk,^20,21,24 fall risk,^18,24 monitoring^18,20 and irreversibility²⁰ have been associated with a lower rate of anticoagulation use. With regards to physician factors, studies have identified physician experience with previous stroke¹⁸ with a higher likelihood of prescribing anticoagulants. Primary care physicians have previously identified lack of time, laboratory result delays, medication interactions, fear of litigation and space constraints as barriers to anticoagulation.²³

There have been no published reports on physician attitudes toward anticoagulation of device-detected AF. We received a good response rate to our survey, and had a captive sample of patients in a large academic center with a catchment area of over one million patients. Limitations of our study include using pharmacy dispensary records as a surrogate for anticoagulant use, which would not account for medication samples. Furthermore, our study is a single-center experience, and the sample size was small, which may further bias results. Additionally, dispensary records were audited over a 12-month period, which might underestimate the rate of anticoagulation if the patient was prescribed anticoagulants after 12 months.

Although patients with device-detected AF are at increased risk of stroke, there is a lack of conclusive evidence to support widespread anticoagulation of device-detected AF, though randomized trials are underway.³⁴ Despite clinical equipoise, anticoagulation of patients with device-related AF and a high CHADS₂ score is common practice.^35,36

Current recommendations for anticoagulation of symptomatic AF do not differ based on duration. In patients with high stroke risk, it is reasonable to prescribe anticoagulation to patients with device-detected AF.³⁷ One recent randomized trial³⁸ evaluated the efficacy of using device-detected atrial arrhythmias to guide anticoagulation compared to routine clinical practice. They found no difference in their composite endpoint of stroke, major bleeding and systemic embolism.³⁸ However, the patients in this cohort were of relatively low stroke risk, anticoagulation was discontinued after freedom from device-detected AF for 30–90 days, and many of the subjects in the control group were started on oral anticoagulation.³⁸ On the other hand, a recent subgroup analysis from the ASSERT trial showed that only episodes of device-detected AF longer than 24 h were associated with increased risk of stroke.³⁹ Individualizing a patient’s risk–benefit ratio when considering anticoagulation in this patient population is a reasonable practice, given the often-devastating outcome of stroke.

In conclusion, we found that a large proportion of patients with device-detected AF were not started on oral anticoagulation and remain at increased stroke risk. A belief that aspirin is adequate in reducing stroke risk in AF persists. Physicians in practice longer than 20 years and who believe that anticoagulants should be managed without the input of a cardiologist were less likely to prescribe anticoagulation. As we learn more about the benefits of anticoagulation of device-detected AF, these results can guide future education programs and policy to increase uptake of anticoagulation in device-detected AF.

Footnotes

Appendix A. Physician Survey

1. Did you receive the communication from the pacemaker clinic at St Boniface Hospital regarding the detection of atrial fibrillation on patient X’s device?

____________________________________________________________________________________________________________________

□ Yes □ No

2. If you received the communication, was the diagnosis of atrial fibrillation and recommendations for considering anticoagulation clearly documented?

____________________________________________________________________________________________________________________

□ Yes □ No

3. Physician information

Which of the following best describes your area of specialty? (Circle): Family/general practice Internal Medicine Nurse practitioner Cardiology (nonEP vs EP) Other:______________

4. Have you attended a CME session on atrial fibrillation in the last 5 years?

____________________________________________________________________________________________________________________

□ Yes □ No

4a. How many years have you been in practice

____________________________________________________________________________________________________________________

□ Less than 5 years □ 5-10 years □ 10-20 years □ More than 20 years

4b. How many patients with atrial fibrillation have you seen in the past year?

____________________________________________________________________________________________________________________

□ Less than 10 □ 10-50 □ 50-100 □ More than 100

4c. Who do you think should be in charge of anticoagulation of atrial fibrillation?

____________________________________________________________________________________________________________________

□ Family physician □ Cardiologist □ Other (please describe): ________________

4d. How is anticoagulation monitoring handled in your practice?

5. Decision regarding oral anticoagulation

5a. What was the primary reason that you did not start oral anticoagulation on this patient

5b. Did you consider using a novel oral anticoagulant on this patient such as rivaroxaban (Xarelto), dabigatran (Pradaxa), or apixaban (Eliquis)? Why not?

Funding

This research received no specific grant from any funding agency in the public, commercial or not-for-profit sectors.

Conflict of interest statement

The authors declare that there is no conflict of interest.

References

Mozaffarian

Benjamin

et al . Heart disease and stroke statistics – 2016 update: a report from the American Heart Association. Circulation 2016; 133: e38–e48.

Hart

Diener

H-C

Coutts

et al . Embolic strokes of undetermined source: the case for a new clinical construct. Lancet Neurol 2014; 13: 429–438.

Lamassa

Di Carlo

Pracucci

et al . Characteristics, outcome, and care of stroke associated with atrial fibrillation in Europe: data from a multicenter multinational hospital-based registry (The European Community Stroke Project). Stroke 2001; 32: 392–398.

Sanna

Diener

H-C

Passman

et al . Cryptogenic stroke and underlying atrial fibrillation. N Engl J Med 2014; 370: 2478–2486.

Gladstone

Spring

Dorian

et al . Atrial fibrillation in patients with cryptogenic stroke. N Engl J Med 2014; 370: 2467–2477.

Seet

RCS

Friedman

Rabinstein

AA.

Prolonged rhythm monitoring for the detection of occult paroxysmal atrial fibrillation in ischemic stroke of unknown cause. Circulation 2011; 124: 477–486.

Healy

Connolly

Gold

et al . Subclinical atrial fibrillation and the risk of stroke. N Engl J Med 2012; 366: 120–129.

Gladstone

Bui

Fang

et al . Potentially preventable strokes in high-risk patients with atrial fibrillation who are not adequately anticoagulated. Stroke 2009; 40: 235–240.

Gross

Vogel

Dhond

et al . Factors influencing physicians’ reported use of anticoagulation therapy in nonvalvular atrial fibrillation: a cross-sectional survey. Clin Ther 2003; 25: 1750–1764.

10.

Fuster

Ryden

Asinger

et al . ACC/AHA/ESC guidelines for the management of patients with atrial fibrillation: executive summary – 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 and Policy Conferences (Committee to Develop Guidelines for the Management of Patients With Atrial Fibrillation). Circulation 2001; 104: 2118–2150.

11.

Mant

Hobbs

FDR

Fletcher

et al . Warfarin versus aspirin for stroke prevention in an elderly community population with atrial fibrillation (the Birmingham Atrial Fibrillation Treatment of the Aged Study, BAFTA): a randomised controlled trial. Lancet 2007; 370: 493–503.

12.

Sato

Ishikawa

Kitabatake

et al . Low-dose aspirin for prevention of stroke in low-risk patients with atrial fibrillation: Japan atrial fibrillation stroke trial. Stroke 2006; 37: 447–451.

13.

Connolly

Eikelboom

Joyner

et al . Apixaban in patients with atrial fibrillation. N Engl J Med 2011; 364: 806–817.

14.

Zimetbaum

Thosani

H-T

et al . Are atrial fibrillation patients receiving warfarin in accordance with stroke risk? Am J Med 2010; 123: 446–453.

15.

Ogilvie

Newton

Welner

et al . Underuse of oral anticoagulants in atrial fibrillation: a systematic review. Am J Med 2010; 123: 638–645.e4.

16.

Reynolds

Shah

Essebag

et al . Patterns and predictors of warfarin use in patients with new-onset atrial fibrillation from the FRACTAL Registry. Am J Cardiol 2006; 97: 538–543.

17.

Gattellari

Worthington

Zwar

et al . The management of non-valvular atrial fibrillation (NVAF) in Australian general practice: bridging the evidence–practice gap. A national, representative postal survey. BMC Fam Pract 2008; 9: 62.

18.

Nicholls

Brehaut

Arim

et al . Impact of stated barriers on proposed warfarin prescription for atrial fibrillation: a survey of Canadian physicians. Thromb J 2014; 12: 13.

19.

Andrade

Krahn

Skanes

et al . Values and preferences of physicians and patients with nonvalvular atrial fibrillation who receive oral anticoagulation therapy for stroke prevention. Can J Cardiol 2015; 32: 747–753.

20.

Frankel

Parker

Rosenfeld

et al . HRS/NSA 2014 survey of atrial fibrillation and stroke: gaps in knowledge and perspective, opportunities for improvement. J Stroke Cerebrovasc Dis 2015; 24: 1691–1700.

21.

Sen

Dahlberg

KW.

Physician’s fear of anticoagulant therapy in nonvalvular atrial fibrillation. Am J Med Sci 2014; 348: 513–521.

22.

Gattellari

Worthington

Zwar

et al . Barriers to using warfarin in non-valvular atrial fibrillation. BMJ 2006; 332: 303–304.

23.

Rodgers

Sudlow

Dobson

et al . Warfarin anticoagulation in primary care: a regional survey of present practice and clinicians’ views. Br J Gen Pract 1997; 47: 309–310.

24.

Monette

Gurwitz

Rochon

et al . 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.

25.

McCrory

Matchar

Samsa

et al . Physician attitudes about anticoagulation for nonvalvular atrial fibrillation in the elderly. Arch Intern Med 1995; 155: 277–281.

26.

Lip

Zarifis

Watson

et al . Physician variation in the management of patients with atrial fibrillation. Heart 1996; 75: 200–205.

27.

Kutner

Nixon

Silverstone

Physicians’ attitudes toward oral anticoagulants and antiplatelet agents for stroke prevention in elderly patients with atrial fibrillation. Arch Intern Med 1991; 151: 1950–1953.

28.

Chang

Bell

Deroo

et al . Physician variation in anticoagulating patients with atrial fibrillation. Dartmouth Primary Care COOP Project. Arch Intern Med 1990; 150: 83–86.

29.

Brodsky

Chun

Podrid

et al . Regional attitudes of generalists, specialists, and subspecialists about management of atrial fibrillation. Arch Intern Med 1996; 156: 2553–2562.

30.

Beyth

Antani

Covinsky

et al . Why isn’t warfarin prescribed to patients with nonrheumatic atrial fibrillation? J Gen Intern Med 1996; 11: 721–728.

31.

Abdel-Latif

Latif

AKA

Peng

et al . Predictors of anticoagulation prescription in nursing home residents with atrial fibrillation. J Am Med Dir Assoc 2005; 6: 128–131.

32.

Bahri

Roca

Lechani

et al . Underuse of oral anticoagulation for individuals with atrial fibrillation in a nursing home setting in France: comparisons of resident characteristics and physician attitude. J Am Geriatr Soc 2015; 63: 71–76.

33.

Choudhry

Anderson

Laupacis

et al . Impact of adverse events on prescribing warfarin in patients with atrial fibrillation: matched pair analysis. BMJ 2006; 332: 141–145.

34.

Institute PHR. Apixaban for the reduction of thrombo-embolism in patients with device-detected sub-clinical atrial fibrillation (ARTESiA). Bethesda, MD: National Library of Medicine, 2015.

35.

Tomson

Passman

Management of device-detected atrial high-rate episodes. Card Electrophysiol Clin 2015; 7: 515–525.

36.

Boriani

Pettorelli

Atrial fibrillation burden and atrial fibrillation type: clinical significance and impact on the risk of stroke and decision making for long-term anticoagulation. Vascul Pharmacol 2016; 83: 26–35.

37.

Kallinen

Hauser

Lee

et al . Failure of impedance monitoring to prevent adverse clinical events caused by fracture of a recalled high-voltage implantable cardioverter-defibrillator lead. Heart Rhythm 2008; 5: 775–779.

38.

Martin

Bersohn

Waldo

et al . Randomized trial of atrial arrhythmia monitoring to guide anticoagulation in patients with implanted defibrillator and cardiac resynchronization devices. Eur Heart J 2015; 36: 1660–1668.

39.

Van Gelder

Healey

Crijns

HJGM

et al . Duration of device-detected subclinical atrial fibrillation and occurrence of stroke in ASSERT. Eur Heart J 2017; 38: 1339–1344.