Impact of Certified CME in Atrial Fibrillation on Administrative Claims

, , , , , ,
The American Journal of Managed Care, May 2012, Volume 18, Issue 5

Use of administrative claims data is an innovative way of measuring the effect of continuing medical education on physician practice behavior and patient outcomes.

Objective:

To determine whether changes in physician behavior associated with a continuing medical education (CME) activity on atrial fibrillation (AF) can be measured using an administrative claims database.

Study Design:

A retrospective, analytical review of physician practice changes and AF patient— related healthcare utilization and costs derived from an administrative claims database was performed on a cohort of Humana health system physicians.

Methods:

The Humana physicians participated in a specified CME activity on the management of patients with AF. Treatment patterns of these providers and clinical outcomes of a cohort of established AF patients were compared 6 months before and 6 months after physician participation in the AF CME activity.

Results:

Analysis of administrative claims data from Humana providers who participated in an AF CME activity and their patients demonstrated a significant reduction in AF-related healthcare costs and utilization, including decreased length of stay. Humana providers, in addition to the other CME activity participants, demonstrated significant gains in knowledge of evidence-based care strategies when presented with real-world scenarios of patients with AF.

Conclusions:

The use of administrative claims data is an innovative way of measuring the effectiveness of CME. These observations support the need for further investigation into the drivers of change in patient outcomes that may be associated with CME activities, as well as the utility of healthcare claims data as a possible valid measure of the impact of CME on physician performance and patient outcomes.

(Am J Manag Care. 2012;18(5):253-260)Participation in a certified continuing medical education (CME) webcast activity on management of patients with atrial fibrillation (AF) was associated with a significant decrease in AF-related healthcare utilization and costs.

  • CME can improve physician confidence and knowledge.

  • Improved physician competence may translate into healthcare savings through better patient care.

  • CME can help physicians provide quality care, which may affect reimbursements in the new pay-for-performance era of healthcare.

Atrial fibrillation (AF) is associated with significant morbidity and mortality, and affects nearly 2.3 million Americans.1 AF is becoming increasingly prevalent with the aging of the US population; the incidence of AF doubles with each decade of age, and an estimated 5.6 million people are expected to be diagnosed by the year 2050. The economic burden of AF is also significant. Annual healthcare costs exceed 5 times the normal healthcare costs of an average individual, with the majority of these costs driven by interventional procedures and inpatient care.2,3

AF is a complex disease that demands a high degree of individualized patient care. For physicians, this care requires keeping abreast of rapidly changing clinical practice guidelines and an expanding selection of available therapeutic agents. In addition, as the Centers for Medicare & Medicaid Services shifts toward a quality-centric approach to healthcare reimbursements, physician performance and quality patient care are becoming increasingly more important.4 Several definitive measures for quality AF patient care exist, including the assessment for thromboembolic risk factors as well as the provision and appropriate monitoring of anticoagulation therapy.5,6

Despite the existence of quality measures, many challenges remain in the management of this multifaceted disease. Although guidelines continually change to reflect the most recent evidence available, current guidelines do not provide definitive measures on how to provide individualized patient care. Furthermore, treatment challenges can be compounded when comorbid conditions are present.7 Taken together, these factors create a significant need for education on current evidence and best practices for the comprehensive management of AF.

To routinely update, improve, and reinforce practice behaviors, clinicians are required to regularly engage in continuing medical education (CME) activities. Typically, evaluation of CME efforts is limited to self-reporting of participant satisfaction and questions designed to assess the educational objectives of the activity. However, the ever-increasing focus on improving the quality of care demands more stringent and sophisticated methods for measuring changes in clinician performance associated with CME activity teachings. To this end, a new type of outcomes measure was adopted. Administrative claims data have long been used to evaluate the effects of various healthcare practices and interventions; however, the impact of CME teachings on physician behavior has not been commonly evaluated by these means. Here we describe a pilot study that used data from an administrative claims database to explore the potential connection between a CME-certified intervention and improvements in physician knowledge and competence in AF-related healthcare decisions and patient outcomes.

STUDY DESIGN

A 60-minute, CME-certified webcast activity was developed for clinicians who treat patients with AF. The activity incorporated interactive patient case—based questions and an expert panel discussion on current clinical data and guideline-recommended best practices for the optimal management of this patient population. Specific discussion topics included the clinical consequences of AF and the negative effects of the disease on quality of life; the use of risk assessment criteria to optimize antithrombotic treatment decisions for stroke prophylaxis; and the implementation of strategies to individualize care plans that incorporate specific patient needs along with clinical evidence for rate control, rhythm control, and thromboembolic prevention. In accordance with Accreditation Council for Continuing Medical Education standards for commercial support, the content of the webcast activity was examined by an independent, third-party review board; met criteria for objectivity, balance, and scientific rigor; and was determined to be free of commercial bias.8

To assess the impact of the CME activity on clinical practice, a retrospective cohort analysis of physician practices and AF patient outcomes was conducted. The study population consisted of cardiologists, electrophysiologists, and internal medicine physicians who participated in the AF CME activity between October 2009 and May 2010, and who were also identified as being contracted US-based physicians within the Humana, Inc health plan electronic data warehouse. A 12-month study period was identified for each physician participant based upon an index date (ie, the date the individual physician completed the specified CME activity); this period was defined as the 6 months prior to and following the index date (pre-CME period and post-CME period, respectively). Patients of these physicians who had at least 1 International Classification of Diseases, Ninth Revision, Clinical Modification (ICD-9- CM) coded claim for AF (427.3x) in the pre-CME period, who were 18 to 89 years of age, and who were enrolled in a Medicare Advantage or commercial full-insured health plan were also included in the analysis.

Physician and patient data were obtained from 5 electronic databases, including a provider file containing physician practice, specialty, and geographic information; a member file containing demographic and enrollment information for each patient per encounter (age, sex, type of insurance, and geographical region); a medical file containing up to 9 recorded ICD-9-CM codes per encounter and related payment information; a pharmacy file containing all Generic Product Identifier numbers of pharmacy-dispensed medications per claim and related payment information; and a lab file containing Logical Observation Identifiers Names and Codes and test results per encounter.

METHODS

Administrative Claims Data

Therapeutic interventions and AF-related healthcare utilization and costs were obtained from the administrative claims databases for qualifying Humana patients treated by participating physicians. Specifically, AF-related healthcare costs and utilization, rate and rhythm control therapies, stroke prevention treatments and associated monitoring, medical devices and procedures related to AF, cardiovascular comorbidities, CHADS2 scores, Charlson Comorbidity Index scores, and patient demographic information were evaluated. All analyses were conducted using SAS 9.2/SAS Enterprise Guide 4.2 (SAS Institute Inc, Cary, North Carolina). Study exemption was provided by the University of Miami Institutional Review Board prior to data collection.

Administrative claims data corresponding to the predetermined study measures were collected and evaluated on the same group of eligible patients in both the pre-CME and post-CME activity periods of the study. Descriptive analyses were completed on all study variables. Unadjusted preindex and postindex date comparisons on key measures, such as drug utilization and costs, were performed. Wilcoxon matched-pairs, signed-rank tests were used for non—normally distributed continuous outcomes, and McNemar tests were used for binary outcomes. P values for 2-sided tests were calculated with statistical significance set at P <.05.

CME Activity Outcomes

A quantitative analysis of the CME-certified AF webcast was performed to assess changes in participant confidence and knowledge. Participants responded to 1 confidence question on a 4-point Likert scale as well as 5 knowledge questions with only 1 correct answer. Chi-squared tests were conducted to assess immediate gain, retention, and longer-term gains in knowledge. Results were considered statistically significant if the resulting x2 or t test would have occurred by chance less than 10% of the time (P <.10).

RESULTS

Administrative Claims Data

Provider Characteristics. In an effort to explore possible changes in physician practice patterns and AF patient outcomes after heightened awareness of guideline measures and current evidence, 395 physicians were identified who completed both the presurvey and postsurvey of the AF CME activity. Of these physicians, 204 were excluded due to incomplete name information, missing activity completion date, foreign residence, or completion of the activity after June 1, 2010. The remaining 191 providers were merged with the Humana electronic data warehouse. A total of 114 providers encountered Humana member patients during the

study period. Of those providers, a total of 84 encountered 932 Humana member patients with a diagnosis of AF in the pre-CME period.

Table 1

All study participants were physicians, and the majority specialized in cardiology (). Of the study participants who provided practice information, more than one-half practiced in community or private practice, and nearly one-fourth were associated with hospital practices. Sixteen percent of study physicians did not provide practice details.

Patient Characteristics. The majority of patients included in the study were male (57%) and patients had a mean age of 74 years. Overall, 24% of patients were in their 60s, 39% of patients were in their 70s, and 32% were in their 80s. The mean Charlson comorbidity score was 2.7 (SD 1.9), and the most common comorbidities were hypertension (88%), heart failure (53%), cardiovascular disease (42%), and diabetes (14%). The mean CHADS2 score was 2.3.

Prescription Patterns. Evaluation of prescriptions for rhythm and rate control therapies revealed that 83% of patients were being treated with at least 1 of these approaches; 78% were treated with rate control agents, and 28% were treated with rhythm control agents. A significant increase in the use of dronedarone was observed during the study period, with nearly a 3-fold increase in new users in the post-CME activity period (7 vs 20, P = .0004). A trend toward decreased use of flecainide was also observed after the CME intervention, although these results were not statistically significant.

According to American College of Cardiology/American Heart Association AF guideline recommendations, patients with AF and a CHADS2 score of 2 or more should receive combination therapy with a rate or rhythm control agent and an anticoagulant.6 At the time of this study, warfarin was the only guideline-recommended anticoagulant agent for stroke prevention in patients with AF.7 Overall, 70% of patients had a CHADS2 score of 2 or more and should have received some form of combination therapy (ie, anticoagulant and rate control therapy, anticoagulant and rhythm control therapy, or all 3 therapies together). Of these patients, 60% in the pre-CME period and 57% in the post-CME period received some form of combination therapy. Sixty-two percent of patients were on warfarin for stroke prevention; however, no significant changes in anticoagulant use were observed in the overall cohort between the pre-CME and post-CME activity periods. Althoug 53 patients in the post-CME activity period were started on warfarin therapy, a slightly greater number of patients (n = 56) discontinued this anticoagulant for a net loss of 3 warfarin users.

Table 2

Table 3

AF-Related Healthcare Utilization and Costs. Onefourth of patients were hospitalized for AF over the course of the study period. Significant decreases in the number and duration of AF-related hospitalizations were observed in the post-CME period (and ). Additionally, declines in the absolute number of AF-related 30-day readmission rates during the study period were observed, as well as decreases in hospitalizations for common AF-related complications (stroke and gastrointestinal bleeding).

A significant decline in AF-related healthcare spending was observed in the post-CME activity period (Table 3), primarily driven by significant savings in medical costs (inpatient, outpatient, and emergency department). Healthcare utilization also decreased significantly; the number of inpatient and outpatient visits dropped.

CME Activity Outcomes

Of the overall webcast participants (Humana and non-Humana providers) who redeemed CME credit, 285 participants completed the pretest, and 280 completed the posttest. The majority of webcast participants were medical doctors who specialized in cardiology and/or electrophysiology, and most practiced in community hospitals or private practices. In addition, 4% of participants were nurse practitioners and physician assistants. Of the 84 Humana providers who participated, 79 had completed the webcast at the time the initial outcomes analysis was performed. For comparison purposes, outcomes data from these 79 participants are reported alongside data from the 206 non-Humana providers who completed the webcast activity.

Figure B

Humana webcast participants demonstrated notable gains in knowledge, with the average percentage of correct responses to knowledge-based questions increasing from 63% to 81% (P = .004; Figure A). Similarly, all 283 webcast participants were found to have increased knowledge after completing the webcast (56% vs 78%, P <.001). Significant gains in confidence were also observed in both groups, particularly with those who felt “extremely” confident in selecting an individualized, evidence-based treatment regimen after completing the activity (). Although more Humana participants reported high levels of confidence before and after the webcast, both groups had significant gains in confidence after completing the webcast.

Table 4

Comparing the Humana participants with the remaining 206 non-Humana participants who completed the webcast activity revealed that outcomes results were very similar. Humana participants demonstrated greater knowledge at pretest and posttest on only 1 question, which asked participants to assess a patient’s stroke risk using the CHADS2 risk score. Overall, there were no differences in knowledge outcomes between the 2 groups ().

DISCUSSION

This pilot study, which involved a relatively new approach to establishing a partnership between a CME provider and a health services research center to perform a retrospective cohort analysis of physician practices and corresponding patient outcomes in the treatment of AF, represented an effort to move beyond traditional CME outcomes assessments to determine the real-world patient impact of an educational activity. From this partnership, we report that participation in the CME-certified intervention was associated with a 33% reduction in AF-related healthcare costs accrued by the 932 eligible patients with established AF who were treated by the 84 participating physicians. Furthermore, patients treated by these physicians had a 40% reduction in the number of inpatient days in the 6 months following the activity.

A strength of this study is the demographic characteristics of the patient population. The AF patients included in our study were fairly representative of the AF population as a whole. In the overall population of AF patients, the median age is approximately 75 years, and the distribution of men and women is roughly equal.7 Furthermore, comorbidities are common in this patient population. In an analysis of the National Hospital Discharge Survey, hypertension, congestive heart failure, ischemic heart disease, and diabetes mellitus were commonly reported comorbid conditions among patients hospitalized for AF.9

Although the analyses presented here were unable to identify the specific factors driving the reductions in healthcare costs and utilization, we hypothesize that participation in the AF CME activity was associated with heightened awareness and better implementation of evidence-based AF care, which may have translated into the selection of more appropriate rate and/or rhythm control therapies for individual patients and, in turn, improved overall disease management and patient outcomes. Recent findings from ATHENA (a placebo-controlled, double-blind, parallel-arm trial to assess the efficacy of dronedarone 400 mg twice a day for prevention of cardiovascular hospitalization or death from any cause in patients with AF/ atrial flutter) showed that patients receiving dronedarone had an approximate 20% lower risk of first cardiovascular hospitalization related to AF.10 Those patients who were hospitalized due to either AF or other cardiovascular-related conditions also had significantly shorter stays. Although the data from the present study are generally in agreement with the results from the ATHENA trial (ie, dronedarone use was higher and AFrelated healthcare utilization and costs were lower in the post- CME period), our analysis was not designed to examine the relationship between any specific rhythm control therapy and hospitalizations. Furthermore, the overall number of patients treated with this agent represented less than 3% of the overall patient cohort and was unlikely a major driver of the decrease in healthcare costs observed in the present analysis. However, the increased use of dronedarone and shorter hospital stays were associated with healthcare cost savings in select patients with AF, and indicates that the observed positive association between the CME intervention and true improvements in patient outcomes is valid.

The increased use of dronedarone we observed may be due to the timing of the study coinciding with the US Food and Drug Administration (FDA) approval of this agent. Dronedarone was approved by the FDA for treatment of AF shortly before the launch of the webcast activity on July 2, 2009.11 It is likely that physician participants sought to learn about and understand dronedarone’s place in the armamentarium ofavailable therapeutic options for AF. Importantly, although dronedarone was not formally adopted in clinical practice guidelines until January 2011, faculty experts were able to provide insight into how this drug should be incorporated into practice based on the available clinical evidence.12

The fact that our data did not show an increase in the use of guideline-recommended anticoagulation therapy may be due to the well-documented patient adherence challenges associated with warfarin—monitoring requirements, dose adjustments, bleeding risks, and dietary restrictions—coupled with the nature of pharmacy claims data.13 Pharmacy claims represent the action of patients filling prescriptions, rather than physicians’ actual prescribing behavior. It is possible, therefore, that any improvements in guideline-recommended prescribing of anticoagulant therapy for patients with a CHADS2 score of 2 or more were obscured by poor patient adherence to therapy, as approximately 48% of all eligible AF patients take warfarin.14 Additionally, clopidogrel was a possible therapeutic alternative, although not guideline recommended at the time. Fifteen percent of all study patients were receiving this therapy, and it is possible that a portion of those patients may have been prescribed clopidogrel instead of warfarin for stroke prevention.

Initial analysis of outcomes from all CME-certified webcast activity participants demonstrated significant gains in confidence and knowledge. When activity outcomes were limited to only Humana participants, similar findings for gains in knowledge were found. This finding was validated in the comparison of knowledge data between the Humana cohort and the non- Humana webcast participants. Humana participants were more likely to have a high degree of confidence in their abilities to individualize treatment regimens for their patients with AF. A likely explanation for this result relates to the stringent study criteria for the Humana cohort, which was limited to physicians, while the webcast activity was open to participants of any degree type, including nurse practitioners and physician assistants.

Given the similarities between the knowledge gains by both the Humana and non-Humana participants, it is possible there was a greater impact on health outcomes than that observed with information from the Humana database. Differences between the Humana cohort and remaining CME activity participants in degree type, specialty, and practice type were insignificant. Assuming similar knowledge and behaviors, patients treated by the non-Humana webcast

participants had the potential to reduce overall AF-related healthcare spending and utilization as well. Unfortunately, limitations with collecting health outcomes data on all participants hinder the verification of this hypothesis.

To our knowledge, only 1 other group has used administrative claims data as a method of analysis of physician performance improvement. Walden and colleagues demonstrated a reduced time to diagnosis of myelodysplastic syndromes after physicians participated in a traditional medical educational forum compared with nonparticipating physicians.15 Like ours, this study was retrospective in nature, using a sliding scale window to examine physician behavior 6 months prior to and 6 months after CME activity participation. In contrast, our analysis tracked performance changes and outcomes within the same group of physicians and their patients, and provides a more global assessment of physician behaviors and healthcare system changes after a CME activity.

Several limitations are associated with this study that may restrict the generalizability of the results. Physicians within the Humana cohort may have participated in additional CME activities on the topic of AF patient management during the study period, thereby influencing their practice patterns. The inclusion of a comparator group of physicians who practiced in the same health system and were matched to participating physicians by demographic and practice criteria would potentially have provided more insight on the influence of outside factors. Although this group was not included as part of the current study, this design has been integrated into a future retrospective claims database analysis that is currently under way.

Additionally, the time frame of the study was limited to 12 months, and changes in practice behavior may have occurred more gradually, over a longer period of time. Limitations common to administrative claims data include the nonrandomization of subjects, the potential for selection bias, and errors in claims coding. In addition, patients who use sample medications or pay for therapies out of pocket cannot be tracked with claims and could influence study results. Also, patients of participating physicians were not assessed for adherence to medical therapies. As previously noted, lack of patient adherence to medical treatments and therapies is well known.16 This fact, combined with a limited study period, may have resulted in the initial decrease in healthcare utilization and cost. If this hypothesis were true, then an increase in healthcare utilization and costs after an extended period of time would be expected.

The Humana study cohort of 84 physicians is relatively small compared with the number of physicians who treat AF patients within the Humana network. More dramatic changes may have been observed with a larger sample of physicians completing the CME activity. Finally, quality healthcare is provided by a team of clinicians. Although the physician is the recognized team leader, it takes the knowledge of all team members, including patients themselves, to provide demonstrable and sustained changes in patient outcomes.17

As healthcare costs continue to increase, focus is shifting away from fee-for-service reimbursements and is moving toward a performance-based, quality care platform. Physicians are being assessed on the quality of care they provide and the outcomes of their patients in an effort to reduce patient complications and, in turn, healthcare costs. With greater knowledge comes a greater ability to make treatment decisions that can optimize patient health. CME is a valuable tool in enhancing physician knowledge, competence, and performance and is influential in driving effective changes in healthcare utilization and costs.Acknowledgments

The AF CME activity was developed by Med-IQ. The retrospective claims data analysis was conducted by Med-IQ in collaboration with Competitive Health Analytics, a subsidiary of Humana Inc. The authors acknowledge Carolyn Berry, PhD, Kenny Khoo, MBA, MS, and Victoria Lawson, MA, for assistance with data analysis; Rebecca Julian, MS, for editorial assistance; Reshma Carter, PharmD, and Amy Sison, BA, for study coordination; Whitney Stevens Dollar, BA, for project management; and Sara Zimmer, BA, for production assistance.

Author Affiliations: Med-IQ (SAS, AJG), Baltimore, MD; Department of Medicine (JSA), University of Arizona Health Science Center, Tucson, AZ; Division of Cardiology (GVN), Penn State University College of Medicine, Hershey, PA; Competitive Health Analytics, Inc (TH, AML), Miramar, FL; Department of Medicine (LT), Miller School of Medicine, University of Miami, Miami, FL.

Funding Source: Supported by an educational grant (#29556) from sanofiaventis US.

Author Disclosures: Dr Alpert reports receiving paid consultancies from Elsevier Publications, Inc, sanofi-aventis, Merck, Bristol-Myers Squibb, Pfizer, AstraZeneca, McNeil, Organon, Berlex, Novartis, Ciba-Geigy, Servier, Boehringer-Ingleheim, Bayer, Johnson & Johnson, Exeter CME, North American Center for Consulting Medical Education, and FRANCE foundation CME. Dr Naccarelli reports receiving paid consultancies from GlaxoSmithKline, Medtronic, Pfizer, Xention, sanofi-aventis, Gilead, Novartis, Bristol-Myers Squibb, Merck, Biosense-Webster, Ortho-McNeil-Janssen, Otsuka, Blue Ash Pharmaceuticals, Forest, and Daiichi-Sankyo. The other authors (SAS, AJG, AML, LT, TPH) report no relationship or financial interest with any entity that would pose a conflict of interest with the subject matter of this article.

Authorship Information: Concept and design (AJG, JSA, GVN, AML, LT, TPH); acquisition of data (AML, TPH); analysis and interpretation of data (SAS, AJG, JSA, GVN, AML, LT, TPH); drafting of the manuscript (SAS, AJG, JSA, GVN); critical revision of the manuscript for important intellectual content (SAS, AJG, JSA, GVN, LT, TPH); statistical analysis (SAS, AML, LT); obtaining funding (AJG); administrative, technical, or logistic support (SAS, TPH); and supervision (JSA, GVN, LT).

Address correspondence to: Stephanie A. Stowell, MPhil, Med-IQ, 5523 Research Park Dr, Ste 210, Baltimore, MD 21223. E-mail: sstowell@med-iq.com.1. Sanoski CA. Prevalence, pathogenesis, and impact of atrial fibrillation. Am J Health Syst Pharm. 2010;67(9)(suppl 5):S11-S16.

2. Wu EQ, Birnbaum HG, Mareva M, et al. Economic burden and comorbidities of atrial fibrillation in a privately insured population. Curr Med Res Opin. 2005;21(10):1693-1699.

3. Ringborg A, Nieuwlatt R, Lindgren P, et al. Costs of atrial fibrillation in five European countries: results from the Euro Heart Survey on atrial fibrillation. Europace. 2008;10(4):403-411.

4. Centers for Medicare & Medicaid Services. Quality initiatives—general information: overview. https://www.cms.gov/QualityInitiatives- GenInfo/. Accessed June 16, 2011.

5. National Quality Forum. NQF-Endorsed® Standards. http://qualityforum.org/Measures_List.aspx#. Accessed June 15, 2011.

6. Estes NA 3rd, Halperin JL, Calkins H, et al; American College of Cardiology/American Heart Association Task Force on Performance Measures; Physician Consortium for Performance Improvement (Writing Committee to Develop Clinical Performance Measures for Atrial Fibrillation); Heart Rhythm Society. ACC/AHA/Physician Consortium 2008 clinical performance measures for adults with nonvalvular atrial fibrillation or atrial flutter: a report of the American College of Cardiology/ American Heart Association Task Force on Performance Measures and the Physician Consortium for Performance Improvement (Writing Committee to Develop Clinical Performance Measures for Atrial Fibrillation): developed in collaboration with the Heart Rhythm Society. Circulation. 2008;117(8):1101-1120.

7. Fuster V, Rydén LE, Cannom DS, et al; American College of Cardiology/American Heart Association Task Force on Practice Guidelines; European Society of Cardiology Committee for Practice Guidelines; European Heart Rhythm Association; Heart Rhythm Society. 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 [published correction appears in Circulation. 2007;116(6):e138]. Circulation. 2006;114(7):e257-354.

8. Accreditation Council for Continuing Medical Education. ACCME Standards for Commercial Support. http://www.accme.org/about-us/ accountability-to-the-public/standards-commercial-support. Accessed November 22, 2011.

9. Wattigney WA, Mensah GA, Croft JB. Increasing trends in hospitalization for atrial fibrillation in the United States, 1985 through 1999: implications for primary prevention. Circulation. 2003;108(6):711-716.

10. Torp-Pedersen C, Crijns HJ, Gaudin C, Page RL, Connolly SJ, Hohnloser SH; ATHENA Investigators. Impact of dronedarone on hospitalization burden in patients with atrial fibrillation: results from the ATHENA study. Europace. 2011;13(8):1118-1126.

11. US Food and Drug Administration. FDA approves Multaq to treat heart rhythm disorder [FDA news release]. http://www.fda.gov/ NewsEvents/Newsroom/PressAnnouncements/ucm170276.htm. Published July 2, 2009. Accessed November 22, 2011.

12. Wann LS, Curtis AB, January CT, et al; ACCF/AHA Task Force Members. 2011 ACCF/AHA/HRS focused update on the management of patients with atrial fibrillation (updating the 2006 guideline): a report of the American College of Cardiology Foundation/American Heart Association Task Force on Practice Guidelines [published correction appears in Circulation. 2011;124(5):e173]. Circulation. 2011;123(1):104-123.

13. 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.

14. Baker WL, Cios DA, Sander SD, Coleman CI. Meta-analysis to assess the quality of warfarin control in atrial fibrillation patients in the United States. J Manag Care Pharm. 2009;15(3):244-252.

15. Walden PD, Dennison B, Jame C, et al. Administrative health data to assess performance in a myelodysplastic syndromes CME initiative. CE Measure. 2010;4(2):26-33.

16. Osterberg L, Blaschke T. Adherence to medication. N Engl J Med. 2005;353(5):487-497.

17. Zeltser MV, Nash DB. Approaching the evidence basis for aviationderivedteamwork training in medicine. Am J Med Qual. 2010;25(1): 13-23.