Author + information
- Received June 27, 2016
- Revision received August 9, 2016
- Accepted September 1, 2016
- Published online November 23, 2016.
- Sheldon M. Singh, MDa,∗ (, )
- Lauren Webster, MPHb,
- Dennis T. Ko, MD, MSca,b,c,
- Jack V. Tu, MD, PhDa,b,c and
- Harindra C. Wijeysundera, MD, PhDa,b,c
- aSchulich Heart Centre and Department of Medicine, Sunnybrook Health Sciences Centre, Faculty of Medicine, University of Toronto, Toronto, Ontario, Canada
- bInstitute for Clinical Evaluative Sciences (ICES), Toronto, Ontario, Canada
- cInstitute of Health Policy, Management and Evaluation, University of Toronto, Toronto, Ontario, Canada
- ↵∗Reprint requests and correspondence:
Dr. Sheldon M. Singh, Schulich Heart Centre and Department of Medicine, Sunnybrook Health Sciences Centre, Faculty of Medicine, University of Toronto, 2075 Bayview Avenue, Toronto, Ontario M4N 3M5, Canada.
Objectives This study sought to determine factors associated with cardiac electrophysiologist assessment and atrial fibrillation (AF) ablation in patients with new-onset AF.
Background Factors driving variation in the use of AF ablation have not been well described.
Methods All individuals with new-onset AF in Ontario, Canada, between January 1, 2010, and December 31, 2012, were identified. Survival analysis accounting for the competing risk of death was used to evaluate the association between clinical and nonclinical factors and receipt of an electrophysiologist assessment. Factors associated with AF ablation were then determined in the subgroup of patients who received an electrophysiologist assessment.
Results A total of 22,032 patients with new-onset AF were identified, 8,161 (37%) of whom received an electrophysiology assessment. Prior cardiologist care was associated with electrophysiologist assessment (hazard ratio [HR]: 1.57; p < 0.0001). Rural residence was associated with a decreased incidence of electrophysiology assessment (HR: 0.80; p < 0.0001). A total of 424 (5.2%) patients receiving an electrophysiologist assessment had an AF ablation. Recurrent AF emergency department (ED) visits between the index ED visit and the initial electrophysiologist assessment (HR for ≥2 ED visits: 4.22; p < 0.0001) and rural residence (HR: 1.50; p = 0.002) were both associated with AF ablation. Cardiovascular comorbidities were associated with a decreased incidence of AF ablation.
Conclusions Rural patients with AF have a lower incidence of electrophysiologist assessment but paradoxically a higher incidence of AF ablation compared with their urban counterparts. Clinical factors such as recurrent ED visits for AF and cardiovascular comorbidities are the most important factors associated with of AF ablation.
Atrial fibrillation (AF) is common, with catheter ablation evolving to become an acceptable option for maintaining sinus rhythm. Adequately powered clinical trials assessing the benefits of AF ablation on morbidity and mortality are limited. The paucity of this information coupled with continued debates on the benefits of rate versus rhythm control (1,2) has resulted in highly discretionary use of AF ablation procedures, which has translated to wide practice variation, a phenomenon well reported with other cardiovascular therapies (3–5).
Existing studies highlighting the factors associated with AF ablation are noteworthy but have significant limitations (6–8). First, most have reported on the association with factors noted at the time of an AF hospitalization that frequently occurred remote to the AF ablation. Second, evaluation of the association among clinical comorbidities, AF-related health care utilization, and AF ablation has been limited. Third, a minority of patients with AF in clinical practice are assessed by cardiac electrophysiologists, a mandatory step in the path to undergoing an AF ablation (9). Disentangling of factors associated with the receipt of an electrophysiologist assessment from those specific to undergoing an AF ablation has not been performed to date. Such an analysis is invaluable to gain insight into the source of variation in the use of AF ablation.
Accordingly, the objectives of this study were: 1) to assess clinical and nonclinical factors associated with an assessment by a cardiac electrophysiologist in patients with new-onset AF in Ontario, Canada; and 2) to assess the clinical and nonclinical factors associated with undergoing an AF ablation in the subgroup of patients evaluated by an electrophysiologist.
This retrospective cohort study was approved by the Sunnybrook Hospital Research Ethics Board.
Ontario is Canada’s largest province with more than 13 million inhabitants who receive universal health care coverage through the Ministry of Health and Long Term Care. Population-level administration databases housed within the Institute for Clinical Evaluative Sciences capture details of the care of all Ontarians. These databases are linked using unique encoded identifiers to protect patient confidentiality and allow for the creation of patient cohorts and long-term follow-up.
The Canadian Institute for Health Information Discharge Abstract Database provided data on all hospitalizations, including patient comorbidities. The National Ambulatory Care Reporting Service database provided data on hospital-based ambulatory care, including emergency department (ED) visits. The Ontario Health Insurance Plan database was used to ascertain physician medical claims. The Registered Persons Database was used to ascertain sex, birth, and death dates. Statistics Canada postal code data were used to ascertain socioeconomic status (reported as income quartiles) and rurality (10).
Ontario residents ≥20 years of age possessing a valid Ontario Health Insurance Plan number and presenting to any Ontario ED between January 1, 2010, to December 31, 2012, with a primary diagnosis of AF (International Classification of Disease-10th Revision diagnostic code I480.XX) were identified. This approach has been previously validated with a positive predictive value of 93% and sensitivity of 96.6% for identifying patients with AF (11).
To ensure the cohort included only patients with new-onset AF, those with an AF diagnosis made within the 5 previous years were excluded. Any ablation procedure performed within the previous 5 years was also an exclusion criterion to exclude individuals receiving recent or ongoing electrophysiology care. Because the goal was to create a cohort of individuals likely to be candidates for AF ablation, patients >80 years of age were excluded based on recommendations from the Canadian Cardiovascular Society that AF ablation should be preferentially performed in patients ≤80 years of age (12). Patients who died at the index ED visit were excluded.
As follow-up was available to March 31, 2015, December 31, 2012, was selected as the cohort accrual end date because the median time between an ED visit for AF and AF ablation was previously reported to be approximately 2 years (8). This approach ensured individuals included within the study cohort were afforded a minimum 2-year follow-up to accrue an electrophysiologist assessment and AF ablation. Individuals were followed until receipt of an AF ablation, death, or end of the study period.
Patients in the cohort were divided into 3 mutually exclusive subgroups: 1) those with new-onset AF and who did not receive an assessment by an electrophysiologist or AF ablation; 2) those with new-onset AF who received an electrophysiologist assessment but did not undergo an AF ablation; and 3) those with new-onset AF who both received an electrophysiologist assessment and underwent AF ablation.
Identification of cardiac electrophysiologist assessments
Physicians were identified as a cardiac electrophysiologist if, using validated algorithms (13), they performed any invasive catheter-based electrophysiology procedure during the study period. The physician unique identifier associated with these physicians was determined, and outpatient visits attributed to these physicians were deemed to reflect an electrophysiologist assessment.
Identification of AF ablation procedures
An algorithm consisting of physician procedural codes and accompanying primary hospital diagnostic code was used to identify patients who underwent an AF ablation procedure. This algorithm has been validated using chart review to identify AF ablation procedures involving pulmonary vein isolation in Ontario and has 100% specificity, 87.2% sensitivity, 98.8% positive predictive value, and 100% negative predictive value (13).
Individuals undergoing AF ablation during the study period and not having an ED visit were also identified. The clinical characteristics of these individuals were compared with those of individuals undergoing AF ablation in the study cohort to assess study generalizability.
Two primary analyses were undertaken. First, using the entire patient cohort, clinical and nonclinical factors associated with an electrophysiologist assessment were determined. For this analysis, patient characteristics and prior health care use were determined at the time of the index ED visit.
The second analysis was limited to patients who received an electrophysiologist assessment. Clinical and nonclinical factors associated with the receipt of an AF ablation were determined. The clinical characteristics of patients in this subgroup were updated to account for the fact that a nontrivial period of time may intervene between the index ED visit and the electrophysiologist assessment, during which time a patient’s health may change.
Covariates of interest in our model included age, gender, and presence of cardiovascular comorbidities. Covariates reflecting the severity of the index AF episode included ambulance use and cardioversion during the index ED visit. For patients seeing an electrophysiologist, additional covariates reflecting AF health care use from the time of the index ED visit to the time of the electrophysiologist visit (all cardioversions, ED visits, and inpatient admissions) were included. Covariates reflecting nonclinical factors included rural residence and socioeconomic status.
Descriptive statistics were used to summarize the study cohort. A Fine-Gray subdistribution model accounting for the competing risk of all-cause death was used to evaluate the association between clinical and nonclinical characteristics determined at the time of the index ED visit and receipt of an electrophysiologist assessment. A similar approach was applied to the cohort receiving an electrophysiologist assessment to evaluate the association between clinical and nonclinical characteristics and AF ablation.
Cumulative incidence functions, accounting for the competing risk of death, were calculated to estimate the rates of: 1) electrophysiology assessment within the entire cohort; and 2) AF ablation in the cohort receiving an electrophysiology assessment. All analyses were performed using SAS Enterprise Guide Version 6.1 (SAS Institute, Cary, North Carolina). Statistical significance was defined by a 2-tailed p < 0.05.
A total of 45,013 patients presented to the ED with a primary diagnosis of AF between January 1, 2010, and December 31, 2012 (Figure 1). Exclusions resulted in a final cohort size of 22,032 individuals. Death occurred in 13% (n = 2,900) of the cohort.
Frequency of and Time to Electrophysiologist Assessment and AF Ablation
Thirty-seven percent of the cohort (n = 8,161) received an electrophysiology assessment. The median time to initial electrophysiology assessment was 87 days (25th percentile to 75th percentile: 5 to 409 days) (Figure 2). Of the entire cohort, 1.9% (n = 424), or 5.2% of those who received an electrophysiology assessment, had an AF ablation procedure during the follow-up period. The median time to ablation was 300 days (25th percentile to 75th percentile: 159 to 627 days) from the time of the initial electrophysiology assessment (Figure 2).
Table 1 summarizes the characteristics of patients with AF who: 1) neither received an electrophysiology assessment nor underwent AF ablation; 2) received an electrophysiology assessment but did not undergo an AF ablation; and 3) received an electrophysiology assessment and underwent AF ablation procedure. Individuals undergoing AF ablation were younger, with a lower frequency of cardiovascular and noncardiovascular comorbidities. The clinical characteristics of the individuals undergoing AF ablation in our study cohort were similar to those of all others undergoing AF ablation in Ontario during the study period (Online Table 1).
Predictors of Electrophysiology Assessment
Cardiology assessment (hazard ratio [HR]: 1.56; p < 0.001) and presence of congestive heart failure (HR: 1.18; p < 0.0001) were associated with an increase in the incidence of electrophysiology assessment (Table 2). In contrast, rural residence (HR: 0.80; p < 0.0001) was associated with a decrease in the incidence of electrophysiology assessment.
Predictors of AF Ablation After an Electrophysiology Assessment
Increasing age (HR: 0.80; p < 0.0001), presence of congestive heart failure (HR: 0.60; p = 0.001), hypertension (HR: 0.74; p = 0.01), and diabetes (HR: 0.74; p = 0.06) were all associated with a decreased incidence of AF ablation.
Rural residence (HR: 1.50; p = 0.001) was significantly associated with an increase in the incidence of AF ablation. Health care use after the index ED visit had the largest impact on the incidence of AF ablation (HR: 2.3; p < 0.001 for individuals with a single additional ED visit; HR: 4.2; p < 0.001 for those with ≥2 additional ED visits). Although a trend was present, increasing income was not significantly associated with AF ablation.
In this population-level analysis of patients with new-onset AF, a minority of patients with new-onset AF received specialized electrophysiology assessment. Moreover, an even smaller number of patients with AF assessed by an electrophysiologist underwent an AF ablation. Recurrent AF ED visits was the most important factor associated with undergoing an AF ablation, suggesting that increased health care utilization plays a strong role in clinical decision making regarding subsequent ablation. Our findings provide substantial insight into previous work and will be of significant interest to practitioners and policymakers developing AF care pathways (6–8).
AF is increasing in prevalence worldwide. The morbidity associated with this condition has resulted in significant direct health care costs for AF hospitalizations, drug treatment, and outpatient testing (14,15). Costs due to lost productivity are also not trivial (16). AF ablation has emerged as an important therapeutic option for symptomatic AF. Randomized trials have consistently demonstrated a reduction in arrhythmia burden with AF ablation comparable to antiarrhythmic drug therapy for both paroxysmal and persistent AF (17–20). Furthermore, a reduction in cardiovascular hospitalizations after AF ablation has been highlighted (20). However, AF ablation does have a high initial cost and upfront procedural risk, and it is not widely available in all jurisdictions. In addition, whether AF ablation is cost-effective in all patient subgroups compared with drug therapy is not clear (21). Controversy also continues to persist as to whether we should strive for rhythm control in patients with AF (1,2). These factors, and the fact that AF ablation is elective, have resulted in wide variation in the use of this procedure. Our work draws attention to this finding and, more importantly, provides an explanation for the observed variation.
Thirty-seven percent of patients with new-onset AF in Ontario, Canada, received an electrophysiology assessment, a mandatory step before undergoing an AF ablation. This finding is not unique; a similarly low rate of electrophysiology assessment was reported in the ORBIT-AF (Outcomes Registry for Better Informed Treatment of AF) study (9). In this registry, patients cared for by an electrophysiologist had a higher likelihood of undergoing treatment with a rhythm control strategy, a phenomenon similar to other disease states such as congestive heart failure, for which specialist care was associated with more resource-intensive care (22,23). Whether the differential care afforded by an electrophysiology assessment can result in improved outcomes in patients with AF remains to be determined.
Rural primary residence was associated with a lower incidence of electrophysiology assessment. Policymakers should be aware of, and care pathways designed to account for, this finding. Mobile or telehealth solutions may be necessary to improve access of rural patients with AF to electrophysiology assessment (24). Improving access of rural patients with AF to electrophysiology assessment is important as, paradoxically, rural patients assessed by electrophysiologists had a higher incidence of AF ablation compared to their urban counterparts. The increased incidence of AF ablation in rural dwellers may reflect the fact that they may have more severe symptoms of AF or may be better candidates for AF ablation, a decision best made with an electrophysiologist. Alternatively, rural patients may be more agreeable to undergoing an AF ablation given perceptions that AF ablation may reduce future health care needs, which may be challenging to access in rural environments. Additional study in this area is warranted to confirm these findings and determine the underlying driving factors.
Recurrent AF ED visits was the most impactful factor associated with AF ablation. This finding intuitively makes sense because health care use typically reflects the burden and severity of AF symptoms. This association is in accordance with current guidelines that recommend AF ablation in patients with symptomatic AF (12,25). Furthermore, the absolute and increasing magnitude of this association with increasing numbers of ED visits suggests the doselike importance of this factor in electrophysiologists’ decision making when assessing patients with AF for ablation procedures. Further study to determine the impact of shorter wait times for electrophysiology assessment on recurrent ED visits and AF health care use in general is warranted. Such work will quantify the benefits of early electrophysiology assessment, which will be important when creating care pathways for patients with AF after an ED visit.
An important association between the presence of patient comorbidities and a reduced incidence of AF ablation was noted, a finding likely related to greater AF recurrence post-ablation in patients with cardiovascular comorbidities compared with those without similar comorbidities (26,27). Although AF ablation may be perceived to be less effective in patients with comorbidities compared with those without, it may be more effective than pharmacological therapy in similar patient groups. This is exemplified by the recently published Ablation versus Amiodarone for Treatment of Atrial Fibrillation in Patients with Congestive Heart Failure study, which clearly demonstrated a benefit of AF ablation using contemporary ablation strategies on arrhythmia burden, hospitalizations, and death in patients with persistent AF compared with drug therapy with amiodarone (28). Electrophysiologists should be aware of and encouraged to perform investigation into the efficacy of contemporary AF ablation strategies in patients with cardiovascular comorbidities to ensure an appropriate rate of AF ablation in these subgroups of patients.
Limitations of our work include the lack of clinical details, such as the type of AF (paroxysmal, persistent, or permanent) and patient preference, which are essential factors affecting the use of AF ablation. Second, because the average age of our cohort was <65 years and drug data in Ontario are available only for patients ≥65 years of age, we were unable to evaluate the association between antiarrhythmic drug therapy and AF ablation, which is an important consideration because drug therapy is considered first-line therapy in patients with AF (12,25). Third, the small number of AF ablation patients in this study may not have provided sufficient statistical power to detect all factors associated with AF ablation. For example, a nonsignificant but plausible trend toward increasing incidence of AF ablation with increasing income quintile was noted in our analysis. It is possible that this association may be real but was not demonstrated to be statistically significant because of the small numbers of AF ablation patients. Alternatively, this association may be less prominent in Ontario, where universal health care is present. Indeed, prior work in the United States demonstrated an association between AF ablation and socioeconomic status (6). Additional work to better assess the association between socioeconomic status and AF ablation is prudent. Fourth, our approach evaluated clinical and nonclinical factors determined at landmark time points (i.e., the index ED visit and initial electrophysiology assessment) and does not account for the impact of health care utilization or development of new clinical conditions subsequent to these events, which may influence the receipt of an electrophysiology assessment and AF ablation. Finally, as with all administrative data studies, there is the potential for inaccuracies in coding of individual patient diagnoses and health care interventions received.
Unique features of our work must be highlighted. First, our cohort evaluated highly symptomatic patients with AF who utilize health care resources, a group who arguably should be aggressively managed. Furthermore, because AF symptoms are the primary indication for AF ablation, identifying a cohort of symptomatic patients with AF is vital when describing factors associated with AF ablation (12,25). We do appreciate that discrepancies in the care of these patients may be magnified in less symptomatic patients with AF. Second, the large cohort studied with clinical characteristics similar to a previously reported population-level cohort of outpatients with AF (29) and the similarity between patients undergoing AF ablation within the ED cohort and patients who never had an ED visit (Online Appendix) speak to the generalizability of our findings. Furthermore, our findings will be of importance to primary care physicians given their role in providing care to the significant majority of patients with new-onset AF discharged from the ED (30). Third, unlike previous work in this field, validated algorithms with high accuracy to identify patients with AF (11) and AF ablation procedures (13) within administrative databases were used. Fourth, a competing risk model allowed for more appropriate estimates of the incidence of AF ablation (31).
As the prevalence of AF and its associated health care costs will continue to increase, efforts to determine the optimal care of patients with AF are necessary. Our work identifies sources of variation in the care of patients with AF that may be targeted with knowledge translation activities. We encourage future studies to determine the optimal rate and timing of electrophysiologist assessments in patients with AF as well as the optimal population rate for AF ablation procedures. Such work will be important to stem the bourgeoning costs associated with this condition.
COMPETENCY IN MEDICAL KNOWLEDGE: Clinical and nonclinical factors associated with referral for an initial electrophysiology assessment are not the same as those factors associated with AF ablation in the subgroup of patients who receive an electrophysiology assessment. This knowledge is important to ensure optimal care is provided to patients with AF.
TRANSLATIONAL OUTLOOK: More research to define the impact of electrophysiology assessment and AF ablation on the overall care of patients with AF is important to help determine the optimal population rate and timing of these interventions in patients with AF.
For a supplemental table, please see the online version of this article.
This study was supported by a Heart & Stroke Foundation/University of Toronto Polo Chair in Cardiology Young Investigator Award and a generous donation from the Tambakis family. Additional support was obtained from the Institute for ICES, which is funded by an annual grant from the Ontario Ministry of Health and Long-Term Care (MOHLTC). Dr. Ko is supported by a Mid-Career Scientist Award from the Heart and Stroke Foundation of Canada, Ontario Office. Dr. Tu is supported by a Tier 1 Canada Research Chair in Health Services Research and an Eaton Scholar Award. Dr. Wijeysundera is supported by a Distinguished Clinical Scientist Award from the Heart and Stroke Foundation of Canada. The funding organizations did not have any role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; and preparation, review, or approval of the manuscript. No endorsement by ICES or the Ontario MOHLTC is intended or should be inferred. Parts of this material are based on data and information compiled and provided by Canadian Institutes for Health Information (CIHI). However, the analyses, conclusions, opinions, and statements expressed herein are those of the authors and not necessarily those of CIHI. All other authors have reported that they have no relationships relevant to the contents of this paper to disclose.
- Abbreviations and Acronyms
- atrial fibrillation
- emergency department
- Received June 27, 2016.
- Revision received August 9, 2016.
- Accepted September 1, 2016.
- American College of Cardiology Foundation
- Tu J.V.,
- Ko D.T.,
- Guo H.,
- et al.
- Matlock D.D.,
- Groeneveld P.W.,
- Sidney S.,
- et al.
- Matlock D.D.,
- Peterson P.M.,
- Heidenreich P.A.,
- et al.
- Kummer B.R.,
- Bhave P.D.,
- Merkler A.E.,
- Gialdini G.,
- Okin P.M.,
- Kamel H.
- Avigil Tsadok M.,
- Gagnon J.,
- Joza J.,
- et al.
- Patel N.,
- Deshmukh A.,
- Takkar B.,
- et al.
- Fosbol E.L.,
- Holmes D.N.,
- Piccini J.P.,
- et al.
- ↵Kralj B. Measuring “rurality” for purposes of health-care planning. Ont Med Rev 200;10:33–52.
- Verma A.,
- Cairns J.A.,
- Mitchell L.B.,
- et al.
- Singh S.M.,
- Webster L.,
- Calzavara A.,
- Wijeysundera H.C.
- Kim M.H.,
- Klingman D.,
- Lin J.,
- Pathak P.,
- Battleman D.S.
- Packer D.L.,
- Kowal R.C.,
- Wheelan K.R.,
- et al.
- Mont L.,
- Bisbal F.,
- Hernandez-Madrid P.,
- et al.
- Piccini J.P.,
- Lopes R.D.,
- Kong M.H.,
- et al.
- Aronsson M.,
- Walfridsson H.,
- Janzon M.,
- et al.
- Jong P.,
- Gong Y.,
- Liu P.P.,
- Austin P.C.,
- Lee D.S.,
- Tu J.V.
- January C.T.,
- Wann L.S.,
- Alpert J.S.,
- et al.
- Mohanty S.,
- Mohanty P.,
- Di Biase L.,
- et al.
- Di Biase L.,
- Mohanty P.,
- Mohanty S.,
- et al.
- Atzema C.,
- Yu B.,
- Ivers N.,
- et al.
- Austin P.C.,
- Lee D.S.,
- Fine J.P.