Utilization of Anticoagulation Therapy in Medicare Patients with Nonvalvular Atrial Fibrillation

Atrial fibrillation (AF) is a significant health and cost concern for the Medicare population (age ≥65 years), because of its association with an increased risk for stroke and all-cause mortality.¹ The risk for stroke in patients with AF is almost 5-fold higher than in patients without AF.² One of every 6 strokes in the United States is associated with AF,³ and strokes in patients with AF are more severe and disabling than in patients without AF.⁴ The prevalence of AF in the Medicare population increased from 3.2% in 1992 to 6.0% in 2002.⁵ Recent estimates indicate the prevalence rate is rising.⁶ The prevalence of AF increases with age, currently affecting approximately 3.8% of persons aged ≥60 years and 9.0% of persons aged ≥80 years.⁷ It is expected that by 2050 the prevalence of AF will double from current estimates, when 88% (4.9 million) of those with AF will be aged ≥65 years and 53% (2.9 million) will be aged ≥80 years.⁷

Oral anticoagulation therapy is the cornerstone of stroke prevention for patients with AF. Warfarin reduces the risk for ischemic stroke by 67% compared with placebo and by 38% compared with aspirin.⁸ Recent clinical trials of newer oral anticoagulation agents report a similar or a greater rate of stroke risk reduction with these agents than with warfarin in patients with AF.^9-11 Antithrombotic treatment guidelines from the American College of Cardiology/American Heart Association recommend oral anticoagulation therapy for patients with AF and moderate-to-high stroke risk, assuming that patients are not at high bleeding risk.^12,13

Several evaluation schemes are available to estimate a stroke risk, including CHADS₂¹⁴ and CHA₂DS₂-VASc.¹⁵ Similarly, several bleeding risk schemes are available to assess the risk of bleeding, including HEMORR2HAGES,¹⁶ HAS-BLED,¹⁷ and, more recently, the ATRIA (Anticoagulation and Risk Factors in Atrial Fibrillation) bleeding risk scheme.¹⁸

Despite well-established guideline recommendations, however, oral anticoagulation therapy is underutilized.^19-24   Underutilization has been linked to healthcare system, physician, and patient factors.²⁵ A major factor cited for underutilization involves physician concern with hemorrhage associated with oral anticoagulation.²⁶ For patients who receive warfarin, optimizing anticoagulation and maintaining anticoagulation within the recommended international normalized ratio (INR) therapeutic range can also be a challenge. A large, nationwide study of electronic medical records reported that patients spent only 48% of study days within the recommended INR range.²³ Another study showed that approximately 33% of patients were within the INR therapeutic range less than 20% of the time, and only 19% of patients were within the therapeutic range all or almost all of the time.²⁷

Physician surveys suggest that healthcare system barriers to optimizing anticoagulation include delays in laboratory reports for INR levels, the general inconvenience of INR monitoring, and the lack of consultant services for anticoagulation management.²⁸ Without careful monitoring of INR levels, supratherapeutic INR levels can and do lead to major bleeding events. Bleedings associated with oral anticoagulation therapy are reported as a significant portion of emergency department visits for drug-related adverse events among older adults.²⁹

The objective of this study was to compare actual warfarin utilization with current treatment guideline recommendations, and to assess the effect of warfarin exposure level on ischemic stroke and bleeding rates in Medicare beneficiaries (ie, age ≥65 years) with AF. This analysis used the 5% Sample Medicare Research Identifiable File (RIF) dataset with Medicare Part A and Part B claims data linked to Part D prescription drug claims, unlike previous analyses that were based on 5% Sample Medicare Limited Dataset (LDS), which uses INR monitoring and prothrombin time claims as a proxy for warfarin use.³⁰ This current approach provides a more precise estimate of warfarin use in the Medicare population.

Methods

Study Population and Characteristics

Study participants were patients in the Medicare 5% Sample RIF who were designated as having AF by the Centers for Medicare & Medicaid Services (CMS) Chronic Condition Data Warehouse (CCW) in the years 2006 or 2007. For each patient with AF, the Medicare RIF data included all Medicare Part A and Part B claims incurred between 2005 and 2008, monthly eligibility data for each patient with AF from 2005 to 2008, and all Medicare Part D claims incurred from 2006 to 2008. Unlike LDS data, the actual date of service for each claim was provided in the RIF data.

Patients included in this analysis were required to have ≥2 medical claims (inpatient, emergency department, observation unit, physician evaluation and management [E/M] office visit) coded with an International Classification of Diseases, Ninth Revision (ICD-9) diagnosis code of 427.31 between January 1, 2006, and December 31, 2007. At least 2 of the claims were required to occur at least 30 days apart, with 1 of the claims incurred in an outpatient setting. To ensure that eligible beneficiaries were patients with newly diagnosed AF, patients with ≥1 claim associated with AF in the 12-month period before the index AF diagnosis (first claim associated with the ICD-9 code 427.31 between January 1, 2006, and December 31, 2007) were excluded from the analysis.

Appendix 1

The sample was further limited to patients with nonvalvular AF (NVAF), by excluding AF patients with claims identified by ICD-9 codes for mitral and/or aortic valvular disease, repair or replacement, transient preoperative AF, or hyperthyroidism in the 12 months before and after the index AF diagnosis date (Appendix 1).

Eligible patients were further required to have continuous enrollment in Medicare Part A and Part B benefits for 12 months before the index AF diagnosis date and for the lesser of 12 months after the AF index diagnosis date or the date of mortality, and Part D enrollment for the lesser of 12 months after the AF index date or the date of mortality. Patients who had an ischemic stroke within 7 days before or after the AF index diagnosis date were excluded from the analysis.

Table 1

Appendix 2

CHADS₂ stroke risk14 and ATRIA bleeding risk18 scores were calculated using claim data 12 months before the index AF diagnosis for each eligible patient. The CHADS₂ and ATRIA risk factors (Table 1) of congestive heart failure, hypertension, diabetes, anemia, severe renal disease, prior bleeding, and history of stroke or transient ischemic attack were identified by ≥1 inpatient, emergency department, hospital observation unit, or physician E/M office visit claims associated with specific ICD-9 diagnosis codes (Appendix 2).

Newly diagnosed patients with NVAF and moderate or high CHADS₂ stroke risk (CHADS₂ score ≥2) and low-to-moderate ATRIA bleeding risk (score ≤4) were identified.

Outcome Measures

Warfarin Use

Warfarin use was determined by the presence of ≥1 warfarin prescription claim within 12 months after the index AF diagnosis. Patients who had a warfarin prescription claim within 12 months after the index AF diagnosis but not until after an ischemic stroke claim, and those who had no warfarin prescription claims during the 12-month period after the index AF diagnosis, were considered in this analysis to have not received warfarin.

Warfarin claims were identified using a comprehensive National Drug Code list, which was obtained from the Medi-Span Master Drug Database from Wolters Kluwer Health in May 2011. The algorithm by Go and colleagues³¹ was adopted to determine the duration of warfarin therapy. The duration of warfarin therapy was calculated based on the number of days of supply associated with each prescription claim; continuous warfarin use was assumed for periods of ≤60 days between any 2 consecutive warfarin prescriptions.

For gaps lasting longer than 60 days, continuous warfarin use was assumed if there were intervening outpatient INR measurements at least every 42 days. Otherwise, the patient was considered not to be taking warfarin from day 31 after the end date of the first prescription until the start date of the next prescription. Warfarin exposure level was assessed by the proportion of days covered, by dividing the total duration of warfarin therapy during the 12-month follow-up period by the length of follow-up.

Stroke and Major Hemorrhage Events

Appendix 3

The rate of ischemic strokes and major hemorrhage were identified based on claims data in the 12 months after the index AF diagnosis. Ischemic strokes were defined as an inpatient or emergency department claim associated with a primary ICD-9 diagnosis of 433.xx, 434.xx, or 436.xx. A major hemorrhage was defined as an inpatient admission with an ICD-9 diagnosis code for either an extracranial hemorrhage in the primary diagnosis position or an ICD-9 code for intracranial hemorrhage in any position of the claim (Appendix 3).

Statistical Analysis

The proportion of patients receiving warfarin within 12 months after the index AF diagnosis was evaluated and compared with guideline recommendations by CHADS₂ stroke risk and ATRIA bleeding risk levels. Specifically, this was the cohort of patients with moderate- to-high stroke risk who were not at high risk for bleeding and filled at least 1 warfarin prescription during the 12 months after the index AF diagnosis.

To assess the effect of warfarin exposure level on the rates of ischemic stroke and major bleeding events for those receiving and not receiving warfarin within 12 months of the index AF diagnosis, incidence rates for ischemic stroke and major bleeding events were separately analyzed, using multivariate logistic regression by controlling for differences in baseline characteristics. Ischemic stroke rates were also compared in patients with any exposure to warfarin (proportion of days covered >0%), in those with warfarin proportion of days covered exposure ≥80%, and in patients who did not receive warfarin during the 12-month follow-up period.

Covariates included in the ischemic stroke model were age, dual-eligibility (Medicare and Medicaid eligibility) status, and CHADS₂ score. Major bleeding event rates were compared for patients who had any exposure to warfarin versus those who did not receive warfarin during the 12-month follow-up period. Covariates included in the major bleeding event model were age, dual-eligibility status (Medicare and Medicaid eligibility), and ATRIA score.

Results

A total of 189,835 Medicare beneficiaries with AF were identified by CMS CCW as having AF in 2006 and in 2007. Limiting the sample to AF patients with ≥2 medical claims associated with an ICD-9 diagnosis code of 427.31 from January 1, 2006, to December 31, 2007, reduced the initial cohort to 165,120 beneficiaries with AF. Limiting this sample to patients with newly diagnosed AF reduced the sample size to 65,438, and further limiting the sample to patients with NVAF reduced the sample size to 54,937.

Figure 1

Requiring Medicare Parts A and B eligibility for the 12 months before the AF index and requiring Medicare Parts A, B, and D eligibility for the lesser of 12 months after the AF index date or date of death reduced the sample size to 14,555 patients with newly diagnosed NVAF. Removing patients with a stroke on or within 7 days before or after the AF index date reduced the sample size to a final total of 14,149 patients with newly diagnosed NVAF (Figure 1).

Baseline characteristics for patients who received warfarin were generally similar to those who did not receive warfarin within 12 months after the index NVAF diagnosis, with a few noteworthy exceptions. There was a higher proportion of patients who had dual eligibility for Medicare and Medicaid (12.6% vs 7.6%, respectively), as well as a higher proportion of patients aged ≥85 years (37.1% vs 21.1%, respectively) among those who did not receive warfarin.

Table 2

The mean CHADS₂ score was significantly lower in patients receiving warfarin, and patients not receiving warfarin had a higher mean ATRIA bleeding risk score (3.7 vs 3.2, respectively; Table 2).

Figure 2

Of the 14,149 patients with newly diagnosed NVAF, 75.8% had moderate or high stroke risk (CHADS₂ score ≥2), and 24.5% had high bleeding risk (ATRIA score ≥5). The proportion of patients with high bleeding risk increased from 8% of those with low stroke risk to 43% in the group with high stroke risk (Figure 2).

Figure 3

Warfarin Use

Table 3

Overall, 6981 (49.3%) patients received warfarin within 12 months of the index AF diagnosis. Figure 3 presents the number and proportion of patients who received warfarin within 12 months of the index AF date at each stroke and bleeding risk level. Warfarin use was lower in the group with a high bleeding risk than in the groups with low/moderate bleeding risk across all stroke risk levels.

However, warfarin use did not seem to differ across stroke risk levels. Of the 14,149 patients with NVAF included in the study, 10,725 patients (75.8%) had moderate (CHADS₂, 2 or 3) or high (CHADS₂ ≥4) stroke risk; of these, 7524 patients did not have high bleeding risk (ATRIA ≥5) and would meet the recommendations for oral anticoagulation therapy. Those recommended for oral anticoagulation therapy (moderate-to-high stroke risk and low-to-moderate bleeding risk) comprised 53.2% (7524/14,149) of the study sample of patients with newly diagnosed NVAF. Among patients recommended for oral anticoagulation therapy, 41.3% (3110/7524) did not receive warfarin within 12 months of the index NVAF diagnosis (Table 3).

Stroke and Bleeding Events

The unadjusted rate of ischemic stroke during the 12-month period after the index NVAF diagnosis was 2.6% for those receiving warfarin within 12 months of the index NVAF versus 5.5% for those not receiving warfarin within 12 months of the index NVAF. After adjusting for age, dual-eligibility status, and CHADS₂ score, the risk for ischemic stroke was 49% lower (odds ratio [OR], 0.51; 95% confidence interval [CI], 0.43-0.61; P <.001) for patients receiving warfarin compared with those not receiving warfarin within 12 months.

Among patients who received warfarin within 12 months after the index AF date, 66.74% had warfarin exposure proportion of days covered ≥80% during the follow-up period. After adjusting for age, dual-eligibility status, and CHADS₂ score using regression, the risk of ischemic stroke for patients with warfarin exposure proportion of days covered ≥80% was significantly lower (unadjusted stroke rate of 2.9% vs 4.8%; adjusted OR, 0.586; 95% CI, 0.48-0.72; P <.001) compared with those who had proportion of days covered <80% (including patients who did not receive any warfarin) during the 12-month postindex AF date. However, among warfarin users, no significant difference was seen in ischemic stroke risk between patients with warfarin exposure proportion of days covered ≥80% (unadjusted stroke rate, 2.9%) and patients with warfarin exposure proportion of days covered <80% (unadjusted stroke rate, 1.7%; adjusted OR, 0.95; 95% CI, 0.71-1.27).

The unadjusted major bleeding event rate during the 12-month period after the index AF diagnosis was 6.8% for those receiving warfarin within 12 months of index NVAF versus 6.5% for those not receiving warfarin within 12 months of the index NVAF. After adjusting for age, dual-eligibility status, and ATRIA score using logistic regression, the major bleeding rate for patients who received warfarin was statistically higher (adjusted OR, 1.19; 95% CI, 1.04-1.36; P = .013) than patients who did not receive warfarin within 12 months of the index NVAF. Subgroup analysis found that warfarin exposure was significantly associated with increased bleeding risk in patients aged ≥65 years (OR, 1.17; 95% CI, 1.02-1.35; P = .024) but not for patients aged <65 years (OR, 1.32; 95% CI, 0.45-3.91; P = .613).

Discussion

Our results confirm previous studies regarding the underuse of warfarin for patients with NVAF, and the lower rate of stroke for those receiving warfarin. Using INR monitoring claims as a proxy for warfarin use, Mercaldi and colleagues found that 58.5% of patients with NVAF were receiving warfarin, and the incidence of ischemic stroke was 27% lower in patients taking warfarin than in those not taking warfarin.³⁰ Using a similar methodology, Lakshminarayan and colleagues reported a 26% lower risk of ischemic stroke in patients with NVAF taking warfarin.⁵

We found that 41.3% of patients at moderate-to-high stroke risk and not at high risk of bleeding were not receiving warfarin within 12 months of the index diagnosis, which amounts to 22% of the total patients with AF study sample. For those taking warfarin within 12 months of the index diagnosis, the relative risk of ischemic stroke was 49% lower than for those not taking warfarin within 12 months. The level of ischemic stroke reduction observed in our study fell within similar ranges reported in previous clinical trials,⁹ where the relative risk reductions of stroke for adjusted-dose warfarin versus aspirin and placebo were 38% and 67%, respectively. Similar to previous clinical trials, we found that absolute increases in major hemorrhage were less than the absolute reductions in stroke.

These findings highlight quality-of-care issues for patients with AF and the need to improve compliance with anticoagulation guidelines in the Medicare population, which would be associated with fewer stroke-related deaths and lower stroke-related healthcare costs for Medicare. Several studies have emphasized the overestimation of bleeding risk and underestimation of stroke risk as reasons for underprescribing anticoagulants in patients with AF,^28,32,33 and our study results corroborate these reports.

We further found that warfarin use was sensitive to bleeding risk, but it did not vary across stroke risk levels. Although the CHADS₂ score has been available and validated for stroke risk stratification for the past decade, and was cited by national treatment guidelines as a tool to determine when oral anticoagulation is warranted, it remains unclear to what extent it has been routinely used in clinical practice to guide decisions on anticoagulation therapy.³⁴ It has been suggested that the CHADS₂ score may not be sensitive in stratifying patients clearly into low-, intermediate-, and high-risk groups in clinical practice.³⁴

The more recently developed stroke stratification scheme, CHA₂DS₂-VASc,¹⁵ which was endorsed by European guidelines,³⁵ may provide more specificity on anticoagulation decision, especially in patients with moderate stroke risk based on CHADS₂. However, the performance of this new scheme in predicting strokes in clinical practice remains to be confirmed.³⁶ These results suggest the need for action by health insurers and policymakers to develop programs to aid clinicians to overcome the clinical decision challenges in prescribing oral anticoagulation therapy for the population with NVAF.

The strength of this study is the use of Medicare Part D data to evaluate warfarin use among patients with NVAF who are recommended for oral anticoagulation therapy and, in particular, the identification of patients with newly diagnosed NVAF and their use of warfarin after the index diagnosis. Past studies have used a proxy of INR testing to establish warfarin use, and dates of service have been unavailable to precisely identify an index diagnosis date and stroke incidence date after the diagnosis of NVAF.

Studies that do not have precise start dates of warfarin and stroke incidence dates may inadvertently assign individuals to the group receiving warfarin when in fact they were prescribed warfarin after their stroke occurred. This will understate the outcomes differences between the groups taking and not taking warfarin. We have tried to control for this by identifying patients who are newly diagnosed with NVAF and excluding strokes that occur 7 days before or after the index diagnosis, and by excluding from the cohort receiving warfarin within 12 months, those receiving warfarin within 12 months but only after a stroke.

Novel oral anticoagulants, including dabigatran and rivaroxaban, have recently been approved for stroke prevention in US patients with AF. Clinical studies have shown that these new agents have either similar or better efficacy in preventing stroke while having a lower risk for intracranial bleeding than warfarin. Future studies are needed to determine whether these newer agents would reduce underutilization of anticoagulation therapy for stroke prevention in patients with AF. Additional studies to assess compliance of warfarin and novel oral anticoagulants and their implications on patient outcomes are also warranted.

Limitations

The present study has several limitations. First, it shares the limitations associated with all administrative claims studies, and it depends on accurate and complete coding practices by providers. Because of data availability, we were limited to a 12-month period preceding the index AF diagnosis to determine whether a patient was newly diagnosed with AF. This methodology cannot rule out the exclusion of all existing patients with AF, because it is possible that a patient with AF may not get coded with AF during a 12-month preindex period, especially if other conditions become more prominent in the patient’s care.

An insufficient preindex period may also underestimate the prevalence of the risk factors for the CHADS₂ stroke and ATRIA bleeding risk schemes, leading to lower than actual stroke and bleeding risk levels. In addition, clinical information not available in claims data and inaccurate coding practice may introduce bias when determining stroke and bleeding risk levels in this study.

Because we required Medicare patients to have 12 months of eligibility before the index diagnosis, we inadvertently excluded 65-year-old Medicare beneficiaries who are new to Medicare through aging. Therefore, our patient selection methodology may have resulted in a slightly older population, which can impact risk factor prevalence. Warfarin use and level of warfarin exposure were identified by filled warfarin prescription claims, which may not accurately represent patient adherence to warfarin therapy.

In this study, we used the ATRIA bleeding risk scheme as a tool to stratify patients into different bleeding risk levels. Different bleeding risk assessment schemes have been developed for the past decade. Current treatment guidelines for anticoagulation do not specify a standard instrument that should be used for bleeding risk assessment. We arbitrarily chose to use the ATRIA bleeding risk scheme because it was developed and validated based on administrative claims data of patients receiving warfarin. The ATRIA bleeding risk scheme has been shown to identify a higher proportion of patients at high bleeding risk and higher major bleeding rates than other bleeding risk schemes.¹⁸ Therefore, our estimate that 53.2% of patients with NVAF are eligible for oral anticoagulation therapy (ie, at moderate-to-high stroke risk and not at high bleeding risk) is a conservative estimate, because using other bleeding risk schemes may lead to a higher proportion of patients with NVAF being eligible for oral anticoagulation therapy. Further studies are needed to confirm our study findings.

Conclusion

AF is a prevalent disease in the Medicare population, with patients with AF having a 5-fold higher risk of stroke than those without AF. Strokes in the AF population are more severe and disabling than in patients without AF. Oral anticoagulation significantly reduces the risk of stroke for patients with AF, yet there is continued underuse of anticoagulation therapy for this population. Notwithstanding its limitation, this study corroborates results from the previous literature and highlights the need for quality improvement initiatives to reduce underuse of oral anticoagulation in patients with clinically indicated AF to reduce stroke-related mortality and healthcare resource use.

Funding Source

The study and the preparation of this manuscript were funded by Daiichi Sankyo, Inc.

Author Disclosure Statement

Ms Fitch, Mr Broulette, Mr Pyenson, and Mr Iwasaki have reported no conflicts of interest. Dr Kwong is an employee of Daiichi Sankyo, Inc.

Author Affiliations

Ms Fitch is a Principal and Healthcare Management Consultant, Mr Broulette is an Actuarial Analyst, Mr Pyenson is a Principal and Consulting Actuary, Mr Iwasaki is a Consulting Actuary, all at Milliman, Inc, New York, NY; Dr Kwong is Senior Director, Health Economics and Outcomes Research, Daiichi Sankyo, Inc, Parsippany, NJ. Part of the study findings was presented at the Academy of Managed Care Pharmacy 24th Annual Meeting and Exposition, April 20, 2012, San Francisco, CA.

References

1. Stewart S, Hart CL, Hole DJ, McMurray JJ. A population-based study of the long-term risks associated with atrial fibrillation: 20-year follow-up of the Renfrew/Paisley study. Am J Med. 2002;113:359-364.
2. Wolf PA, Abbott RD, Kannel WB. Atrial fibrillation as an independent risk factor for stroke: the Framingham Study. Stroke. 1991;22:983-988.
3. Hart RG, Halperin JL. Atrial fibrillation and thromboembolism: a decade of progress in stroke prevention. Ann Intern Med. 1999;131:688-695.
4. Lin HJ, Wolf PA, Kelly-Hayes M, et al. Stroke severity in atrial fibrillation. The Framingham Study. Stroke. 1996;27:1760-1764.
5. Lakshminarayan K, Solid CA, Collins AJ, et al. Atrial fibrillation and stroke in the general medicare population: a 10-year perspective (1992 to 2002). Stroke. 2006;37:1969-1974.
6. Naccarelli GV, Varker H, Lin J, Schulman KL. Increasing prevalence of atrial fibrillation and flutter in the United States. Am J Cardiol. 2009;104:1534-1539.
7. Go AS, Hylek EM, Phillips KA, et al. Prevalence of diagnosed atrial fibrillation in adults: national implications for rhythm management and stroke prevention: the Anticoagulation and Risk Factors in Atrial Fibrillation (ATRIA) study. JAMA. 2001;285:2370-2375.
8. Hart RG, Pearce LA, Aguilar MI. Meta-analysis: antithrombotic therapy to prevent stroke in patients who have nonvalvular atrial fibrillation. Ann Intern Med. 2007;146:857-867.
9. Patel MR, Mahaffey KW, Garg J, et al. Rivaroxaban versus warfarin in nonvalvular atrial fibrillation. N Engl J Med. 2011;365:883-891.
10. Connolly SJ, Eikelboom J, Joyner C, et al. Apixaban in patients with atrial fibrillation. N Engl J Med. 2011;364:806-817.
11. Connolly SJ, Ezekowitz MD, Yusuf S, et al. Dabigatran versus warfarin in patients with atrial fibrillation. N Engl J Med. 2009;361:1139-1151.
12. Fuster V, Rydén LE, Cannom DS, et al. ACC/AHA/ESC 2006 Guidelines 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 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. Circulation. 2006;114:e257-e354.
13. Wann LS, Curtis AB, Ellenbogen KA, et al. 2011 ACCF/AHA/HRS focused update on the management of patients with atrial fibrillation (update on dabigatran): a report of the American College of Cardiology Foundation/American Heart Association Task Force on practice guidelines. Circulation. 2011;123:1144-1150.
14. Gage BF, Waterman AD, Shannon W, et al. Validation of clinical classification schemes for predicting stroke: results from the National Registry of Atrial Fibrillation. JAMA. 2001;285:2864-2870.
15. Lip GY, Nieuwlaat R, Pisters R, et al. 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.
16. Gage BF, Yan Y, Milligan PE, et al. Clinical classification schemes for predicting hemorrhage: results from the National Registry of Atrial Fibrillation (NRAF). Am Heart J. 2006;151:713-719.
17. Pisters R, Lane DA, Nieuwlaat R, et al. 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.
18. Fang MC, Go AS, Chang Y, et al. A new risk scheme to predict warfarin-associated hemorrhage: the ATRIA (Anticoagulation and Risk Factors in Atrial Fibrillation) Study. J Am Coll Cardiol. 2011;58:395-401.
19. Fang MC, Stafford RS, Ruskin JN, Singer DE. National trends in antiarrhythmic and antithrombotic medication use in atrial fibrillation. Arch Intern Med. 2004;164:55-60.
20. Waldo AL, Becker RC, Tapson VF, et al. Hospitalized patients with atrial fibrillation and a high risk of stroke are not being provided with adequate anticoagulation. J Am Coll Cardiol. 2005;46:1729-1736.
21. Samsa GP, Matchar DB, Goldstein LB, et al. Quality of anticoagulation management among patients with atrial fibrillation: results of a review of medical records from 2 communities. Arch Intern Med. 2000;160:967-973.
22. Glazer N, Dublin S, Smith NL, et al. Newly detected atrial fibrillation and compliance with antithrombotic guidelines. Arch Intern Med. 2007;167:246-252.
23. Boulanger L, Kim J, Friedman M, et al. Patterns of use of antithrombotic therapy and quality of anticoagulation among patients with non-valvular atrial fibrillation in clinical practice. Int J Clin Pract. 2006;60:258-264.
24. Harley CR, Riedel AA, Hauch O, et al. Anticoagulation therapy in patients with chronic atrial fibrillation: a retrospective claims data analysis. Curr Med Res Opin. 2005;21:215-222.
25. Kneeland PP, Fang MC. Current issues in patient adherence and persistence: focus on anticoagulants for the treatment and prevention of thromboembolism. Patient Prefer Adherence. 2010;4:51-60.
26. Gross CP, Vogel EW, Dhond AJ, et al. Factors influencing physicians’ reported use of anticoagulation therapy in nonvalvular atrial fibrillation: a cross-sectional survey. Clin Ther. 2003;25:1750-1764.
27. Walker AM, Bennett D. Epidemiology and outcomes in patients with atrial fibrillation in the United States. Heart Rhythm. 2008;5:1365-1372.
28. Bungard TJ, Ghali WA, Teo KK, et al. Why do patients with atrial fibrillation not receive warfarin? Arch Intern Med. 2000;160:41-46.
29. Budnitz DS, Shehab N, Kegler SR, Richards CL. Medication use leading to emergency department visits for adverse drug events in older adults. Ann Intern Med. 2007;147:755-765.
30. Mercaldi CJ, Ciarametaro M, Hahn B, et al. Cost efficiency of anticoagulation with warfarin to prevent stroke in medicare beneficiaries with nonvalvular atrial fibrillation. Stroke. 2011;42:112-118.
31. Go AS, Hylek EM, Chang Y, et al. Anticoagulation therapy for stroke prevention in atrial fibrillation: how well do randomized trials translate into clinical practice? JAMA. 2003;290:2685-2692.
32. Bungard TJ, Ghali WA, McAlister FA, et al. The relative importance of barriers to the prescription of warfarin for nonvalvular atrial fibrillation. Can J Cardiol. 2003; 19:280-284.
33. Beyth RJ, Antani MR, Covinsky KE, et al. Why isn’t warfarin prescribed to patients with nonrheumatic atrial fibrillation? J Gen Intern Med. 1996;11:721-728.
34. Klein D, Levine M. Are family physicians using the CHADS₂ score? Is it useful for assessing risk of stroke in patients with atrial fibrillation? Can Fam Physician. 2011;57:e305-e309.
35. Camm AJ, Kirchhof P, Lip GY, et al, for the European Heart Rhythm Association, European Association for Cardio-Thoracic Surgery. Guidelines for the management of atrial fibrillation: the Task Force for the Management of Atrial Fibrillation of the European Society of Cardiology (ESC). Eur Heart J. 2010;31:2369-2429.
36. Singer DE, Fang MC, Go AS. CHA₂DS₂-VASc risk scheme not ready for clinical use. Chest. 2010;138:1020; author reply 1020-1021.