Clinical Investigations Racial Disparities in the Use of Catheter Ablation for Atrial Fibrillation and Flutter
Address for correspondence: Leonardo Tamariz, MD University of Miami 1120 NW 14th St, Suite 967 Miami, Fl 33136 [email protected]
Leonardo Tamariz, MD, MPH; Alexis Rodriguez, MD; Ana Palacio, MD, MPH; Hua Li, MD, PhD; Robert Myerburg, MD Department of Medicine (Tamariz, Rodriguez, Palacio, Li, Myerburg), Miller School of Medicine, University of Miami, Miami, Florida; Department of Medicine, Veterans Affairs Medical Center (Tamariz, Palacio), Miami, Florida
Background: Atrial fibrillation (AF) is the most common arrhythmia seen in clinical practice. Catheter ablation is an expensive but potentially curable treatment of AF. We explored differences in the use of catheter ablation for AF in the state of Florida and compared the findings to ablation for atrial flutter. Methods: We conducted a cross-sectional analysis of all ambulatory and hospital discharge procedures between 2006 and 2009 in Florida. We identified all subjects with AF and atrial flutter, using International Classification of Diseases, 9th Revision codes along with the race/ethnicity of each individual. We used logistic regression to determine the odds ratio (OR) of having a catheter ablation per disease by race and ethnicity adjusted for Charlson score, insurance status, year of the procedure, and facility location. Results: We identified 923 590 subjects with AF and 28 714 with atrial flutter. Catheter ablations were more commonly used in atrial flutter than in AF. The adjusted OR of having catheter ablation for AF for blacks was 0.67 (95% confidence interval [CI]: 0.60-0.75, P < 0.01), and for Hispanics it was 0.83 (95% CI: 0.75-0.91, P < 0.01) when compared to whites. The adjusted OR of having an ablation for atrial flutter for blacks was 1.08 (95% CI: 0.96-1.21, P = 0.16), and for Hispanics it was 0.90 (95% CI: 0.78-1.08, P = 0.20) when compared to whites. Conclusions: In the state of Florida, black and Hispanic subjects with AF received less catheter ablations, whereas the same minority subjects with atrial flutter received a similar number of ablations compared to white subjects, with the same insurance and comorbidity burden.
Introduction Atrial fibrillaton (AF) and atrial flutter are common arrhythmias generally managed with either restoration and maintenance of sinus rhythm or rate control.1,2 Radio frequency catheter ablation (CA) has emerged as a safe and potentially curative alternative treatment for both AF and atrial flutter.3,4 Originally, only young patients with paroxysmal AF and no structural disease were considered ideal candidates for such procedures, but as the operator experience and ablation technique have dramatically evolved over the years, so has the patient population, to the point that now individuals with persistent and permanent AF,5 heart failure,4 the elderly,6 those with no previous trial of antiarrhythmic medications,7 and even those with implanted devices are considered ablation candidates.8 There is strong evidence supporting that black patients, when compared to whites, have lower rates of referral for invasive procedures like cardiac catheterization, percutaneous coronary intervention, coronary artery bypass graft The authors have no funding, financial relationships, or conflicts of interest to disclose. Received: January 1, 2014 Accepted with revision: August 1, 2014
surgery, and implantation of cardioverter defibrillators.9 – 13 Many have reported these differences, and how they persist even after adjusting for socioeconomic, insurance status, and income.14,15 However, little is known as to whether Hispanics also have lower rates of procedures and if such discrepancies in resource utilization are also present in ablation procedures for AF or atrial flutter. The purpose of this study was to analyze data from the state of Florida to determine whether minority populations have differential referral for CA for the treatment of AF and atrial flutter.
Methods Study Design and Setting We conducted a cross-sectional study using data from the Florida Agency for Healthcare Administration between the years 2006 and 2009. Specifically, we used the hospital inpatient discharge and the ambulatory outpatient datasets. The inpatient hospital discharge dataset contains information on all hospital admissions to acute care hospitals, and the ambulatory outpatient dataset contains on all outpatient encounters in the state of Florida. Both Clin. Cardiol. 37, 12, 733–737 (2014) Published online in Wiley Online Library (wileyonlinelibrary.com) DOI:10.1002/clc.22330 © 2014 Wiley Periodicals, Inc.
733
datasets contained random patient identification numbers; demographic characteristics; primary and up to 4 secondary diagnoses as classified by the International Classification of Diseases, 9th Revision, Clinical Modification (ICD9); procedure codes based on ICD-9 Current Procedural Terminology; payer type; unique facility identifiers; and facility type. Florida’s Office of Data Collection & Quality Assurance is currently charged with collecting all facility discharge and ambulatory data. Approximately 275 hospitals, 215 emergency departments, 550 freestanding ambulatory surgical centers, and cardiac catheterization laboratories currently submit quarterly reports. A limited dataset is made available to the public once 75% of the reporting facilities have certified the accuracy of their data. Reporting facilities must certify their patient data (the final step of the submission process) within 5 months after the end of the covered quarter. Study Participants The study population consisted of all patients age 18 years and older with an ICD-9 code for AF (427.31) or atrial flutter (427.32) between January 1, 2006 and June 30, 2009. The use of arrhythmia codes has been previously validated.16 – 18 We included the ICD-9 code in any position and excluded subjects who had ICD-9 codes for both AF and atrial flutter (n = 67 745). The purpose of including the atrial flutter population as a control group was to compare the findings of the AF population, because CA has been demonstrated as a curative and safe strategy in the treatment of atrial flutter, and it is a procedure that has been available for several years and requires less technical expertise.19 Evaluation of Covariates Race and ethnicity were reported as a single variable and classified subjects into 8 categories that included black or African American, white, white-Hispanic, and black-Hispanic. For the purpose of this analysis, we merged white and black-Hispanic into a single Hispanic category. To account for confounding, we reported on insurance status recorded as Medicare insurance, uninsured, or other. We also calculated the Charlson score as a measure of disease burden, which is a validated method that records comorbidity status.20 We also included the facility location where the procedure was performed as urban or rural using the 2000 US Census definition of rural regions, which defined rural as counties in Florida with a population density of 100 or less. We included this variable as a surrogate marker of population distribution, because Hispanics and blacks could reside in rural areas and not have access to facilities that provide CA. Definition of Radiofrequency CA We defined CA as having an ICD-9 procedure claim code (37.34) in subjects with AF or atrial flutter. We did not use codes 37.26, because this code reflects electrophysiologic testing and not an intervention procedure.
734
Clin. Cardiol. 37, 12, 733–737 (2014) L. Tamariz et al: Disparities in ablation Published online in Wiley Online Library (wileyonlinelibrary.com) DOI:10.1002/clc.22330 © 2014 Wiley Periodicals, Inc.
Statistical Methods Our primary predictor of interest was race/ethnicity categorized as non-Hispanic white, black, or Hispanic. We compared baseline characteristics at the time of CA by race/ethnicity using the Pearson χ2 statistic for categorical variables and analysis of variance for continuous measures. Because some continuous variables were not normally distributed, we used the Wilcoxon rank sum test. To evaluate for differences, we report the percentage of CA that occurred in each race/ethnicity category with AF or atrial flutter and not the percentage of ablations that were classified as black, white, or Hispanic. To determine if the rate of CA differed by race/ethnicity and account for confounding, we used logistic regression to calculate the odds ratio (OR) of having CA and the corresponding 95% confidence interval (CI). The model included year of the procedure, insurance status, facility location, and the Charlson-Deyo comorbidity score. Analyses were performed using SAS 9.0 (SAS Institute Inc., Cary, NC), and all significance tests were 2-tailed.
Results Baseline Characteristics of the AF Population Table 1 reports the baseline characteristics of 923 590 patients with a diagnosis of AF in Florida. Whites represented the majority of the patients with AF, followed by Hispanics and blacks. Blacks were younger, more likely to be female, less likely to have Medicare insurance, and had higher comorbidity scores when compared to non-Hispanic whites. The use of CA was low in all ethnicities, but blacks were less likely to have a CA for AF when compared to whites and Hispanics. Baseline Characteristics of the Atrial Flutter Population Table 2 reports the baseline characteristics of 28 714 patients with a diagnosis of atrial flutter in Florida. Whites represented the majority of the patients with atrial flutter, followed by blacks and then Hispanics. Blacks were younger, more likely to be female, less likely to have Medicare insurance, and had higher comorbidity scores when compared to non-Hispanic whites. Blacks were as likely as Hispanics and whites to have CA. Use of CA in the AF Population Figure 1 reports the use of CA in the AF cohort by year and ethnicity. In 2006, the use of CA was highest in non-Hispanic whites, followed by Hispanics and then blacks. In 2009, the use of CA in whites was higher than the prior years, whereas there was a sharp decrease in the use of CA in Hispanics and no change in blacks. Table 3 reports the univariate and multivariate ORs for CA of AF patients. In the univariate analysis, the more recent year when the procedure was done and being insured were associated with a statistically significant increase in the use of CA. Being black, Hispanic, or having higher comorbidity scores were associated with a statistically significant decrease in the use of CA. The multivariate estimates were similar when compared to the univariate. The adjusted OR of having CA for AF in blacks
Table 1. Baseline Characteristics of 923 590 Patients With Atrial Fibrillation in the State of Florida Characteristic
White
Black
Hispanic
799 479
61 037
63 074
77.1 ± 11.5
68.8 ± 14.8
74.9 ± 13.0
48
54
50
Medicare, %
87
82
84
Catheter ablation, %a
1.0
0.81
0.99
Hypertension, %a
71
79
75
Diabetes, %a
29
43
39
5
6
5
2.49 ± 2.14
3.07 ± 2.25
2.17 ± 2.12
94
93
99
No. Agea a
Female gender, % a
a
Heart failure, %
a
Charlson score
Urban facility, %a a
P < 0.01 for comparisons between ethnicities.
Table 2. Baseline Characteristics of 28 714 Subjects With Atrial Flutter in the State of Florida
Figure 1. Percentage of the use of catheter ablation by ethnicity (y-axis) per year (x-axis) in patients with atrial fibrillation.
Table 3. Univariate and Multivariate Odds Ratios of Catheter Ablation for Atrial Fibrillation Univariate OR (95% CI) P Value
Multivariate OR (95% CI) P Value
Characteristic
White
Black
Hispanic
No.
23 253
3 257
1934
Characteristic
Agea
71.2 ± 14.0
61.4 ± 16.8
65.5 ± 19.7
Black vs white
0.59 (0.52-0.66)
Address for correspondence: Leonardo Tamariz, MD University of Miami 1120 NW 14th St, Suite 967 Miami, Fl 33136 [email protected]
Leonardo Tamariz, MD, MPH; Alexis Rodriguez, MD; Ana Palacio, MD, MPH; Hua Li, MD, PhD; Robert Myerburg, MD Department of Medicine (Tamariz, Rodriguez, Palacio, Li, Myerburg), Miller School of Medicine, University of Miami, Miami, Florida; Department of Medicine, Veterans Affairs Medical Center (Tamariz, Palacio), Miami, Florida
Background: Atrial fibrillation (AF) is the most common arrhythmia seen in clinical practice. Catheter ablation is an expensive but potentially curable treatment of AF. We explored differences in the use of catheter ablation for AF in the state of Florida and compared the findings to ablation for atrial flutter. Methods: We conducted a cross-sectional analysis of all ambulatory and hospital discharge procedures between 2006 and 2009 in Florida. We identified all subjects with AF and atrial flutter, using International Classification of Diseases, 9th Revision codes along with the race/ethnicity of each individual. We used logistic regression to determine the odds ratio (OR) of having a catheter ablation per disease by race and ethnicity adjusted for Charlson score, insurance status, year of the procedure, and facility location. Results: We identified 923 590 subjects with AF and 28 714 with atrial flutter. Catheter ablations were more commonly used in atrial flutter than in AF. The adjusted OR of having catheter ablation for AF for blacks was 0.67 (95% confidence interval [CI]: 0.60-0.75, P < 0.01), and for Hispanics it was 0.83 (95% CI: 0.75-0.91, P < 0.01) when compared to whites. The adjusted OR of having an ablation for atrial flutter for blacks was 1.08 (95% CI: 0.96-1.21, P = 0.16), and for Hispanics it was 0.90 (95% CI: 0.78-1.08, P = 0.20) when compared to whites. Conclusions: In the state of Florida, black and Hispanic subjects with AF received less catheter ablations, whereas the same minority subjects with atrial flutter received a similar number of ablations compared to white subjects, with the same insurance and comorbidity burden.
Introduction Atrial fibrillaton (AF) and atrial flutter are common arrhythmias generally managed with either restoration and maintenance of sinus rhythm or rate control.1,2 Radio frequency catheter ablation (CA) has emerged as a safe and potentially curative alternative treatment for both AF and atrial flutter.3,4 Originally, only young patients with paroxysmal AF and no structural disease were considered ideal candidates for such procedures, but as the operator experience and ablation technique have dramatically evolved over the years, so has the patient population, to the point that now individuals with persistent and permanent AF,5 heart failure,4 the elderly,6 those with no previous trial of antiarrhythmic medications,7 and even those with implanted devices are considered ablation candidates.8 There is strong evidence supporting that black patients, when compared to whites, have lower rates of referral for invasive procedures like cardiac catheterization, percutaneous coronary intervention, coronary artery bypass graft The authors have no funding, financial relationships, or conflicts of interest to disclose. Received: January 1, 2014 Accepted with revision: August 1, 2014
surgery, and implantation of cardioverter defibrillators.9 – 13 Many have reported these differences, and how they persist even after adjusting for socioeconomic, insurance status, and income.14,15 However, little is known as to whether Hispanics also have lower rates of procedures and if such discrepancies in resource utilization are also present in ablation procedures for AF or atrial flutter. The purpose of this study was to analyze data from the state of Florida to determine whether minority populations have differential referral for CA for the treatment of AF and atrial flutter.
Methods Study Design and Setting We conducted a cross-sectional study using data from the Florida Agency for Healthcare Administration between the years 2006 and 2009. Specifically, we used the hospital inpatient discharge and the ambulatory outpatient datasets. The inpatient hospital discharge dataset contains information on all hospital admissions to acute care hospitals, and the ambulatory outpatient dataset contains on all outpatient encounters in the state of Florida. Both Clin. Cardiol. 37, 12, 733–737 (2014) Published online in Wiley Online Library (wileyonlinelibrary.com) DOI:10.1002/clc.22330 © 2014 Wiley Periodicals, Inc.
733
datasets contained random patient identification numbers; demographic characteristics; primary and up to 4 secondary diagnoses as classified by the International Classification of Diseases, 9th Revision, Clinical Modification (ICD9); procedure codes based on ICD-9 Current Procedural Terminology; payer type; unique facility identifiers; and facility type. Florida’s Office of Data Collection & Quality Assurance is currently charged with collecting all facility discharge and ambulatory data. Approximately 275 hospitals, 215 emergency departments, 550 freestanding ambulatory surgical centers, and cardiac catheterization laboratories currently submit quarterly reports. A limited dataset is made available to the public once 75% of the reporting facilities have certified the accuracy of their data. Reporting facilities must certify their patient data (the final step of the submission process) within 5 months after the end of the covered quarter. Study Participants The study population consisted of all patients age 18 years and older with an ICD-9 code for AF (427.31) or atrial flutter (427.32) between January 1, 2006 and June 30, 2009. The use of arrhythmia codes has been previously validated.16 – 18 We included the ICD-9 code in any position and excluded subjects who had ICD-9 codes for both AF and atrial flutter (n = 67 745). The purpose of including the atrial flutter population as a control group was to compare the findings of the AF population, because CA has been demonstrated as a curative and safe strategy in the treatment of atrial flutter, and it is a procedure that has been available for several years and requires less technical expertise.19 Evaluation of Covariates Race and ethnicity were reported as a single variable and classified subjects into 8 categories that included black or African American, white, white-Hispanic, and black-Hispanic. For the purpose of this analysis, we merged white and black-Hispanic into a single Hispanic category. To account for confounding, we reported on insurance status recorded as Medicare insurance, uninsured, or other. We also calculated the Charlson score as a measure of disease burden, which is a validated method that records comorbidity status.20 We also included the facility location where the procedure was performed as urban or rural using the 2000 US Census definition of rural regions, which defined rural as counties in Florida with a population density of 100 or less. We included this variable as a surrogate marker of population distribution, because Hispanics and blacks could reside in rural areas and not have access to facilities that provide CA. Definition of Radiofrequency CA We defined CA as having an ICD-9 procedure claim code (37.34) in subjects with AF or atrial flutter. We did not use codes 37.26, because this code reflects electrophysiologic testing and not an intervention procedure.
734
Clin. Cardiol. 37, 12, 733–737 (2014) L. Tamariz et al: Disparities in ablation Published online in Wiley Online Library (wileyonlinelibrary.com) DOI:10.1002/clc.22330 © 2014 Wiley Periodicals, Inc.
Statistical Methods Our primary predictor of interest was race/ethnicity categorized as non-Hispanic white, black, or Hispanic. We compared baseline characteristics at the time of CA by race/ethnicity using the Pearson χ2 statistic for categorical variables and analysis of variance for continuous measures. Because some continuous variables were not normally distributed, we used the Wilcoxon rank sum test. To evaluate for differences, we report the percentage of CA that occurred in each race/ethnicity category with AF or atrial flutter and not the percentage of ablations that were classified as black, white, or Hispanic. To determine if the rate of CA differed by race/ethnicity and account for confounding, we used logistic regression to calculate the odds ratio (OR) of having CA and the corresponding 95% confidence interval (CI). The model included year of the procedure, insurance status, facility location, and the Charlson-Deyo comorbidity score. Analyses were performed using SAS 9.0 (SAS Institute Inc., Cary, NC), and all significance tests were 2-tailed.
Results Baseline Characteristics of the AF Population Table 1 reports the baseline characteristics of 923 590 patients with a diagnosis of AF in Florida. Whites represented the majority of the patients with AF, followed by Hispanics and blacks. Blacks were younger, more likely to be female, less likely to have Medicare insurance, and had higher comorbidity scores when compared to non-Hispanic whites. The use of CA was low in all ethnicities, but blacks were less likely to have a CA for AF when compared to whites and Hispanics. Baseline Characteristics of the Atrial Flutter Population Table 2 reports the baseline characteristics of 28 714 patients with a diagnosis of atrial flutter in Florida. Whites represented the majority of the patients with atrial flutter, followed by blacks and then Hispanics. Blacks were younger, more likely to be female, less likely to have Medicare insurance, and had higher comorbidity scores when compared to non-Hispanic whites. Blacks were as likely as Hispanics and whites to have CA. Use of CA in the AF Population Figure 1 reports the use of CA in the AF cohort by year and ethnicity. In 2006, the use of CA was highest in non-Hispanic whites, followed by Hispanics and then blacks. In 2009, the use of CA in whites was higher than the prior years, whereas there was a sharp decrease in the use of CA in Hispanics and no change in blacks. Table 3 reports the univariate and multivariate ORs for CA of AF patients. In the univariate analysis, the more recent year when the procedure was done and being insured were associated with a statistically significant increase in the use of CA. Being black, Hispanic, or having higher comorbidity scores were associated with a statistically significant decrease in the use of CA. The multivariate estimates were similar when compared to the univariate. The adjusted OR of having CA for AF in blacks
Table 1. Baseline Characteristics of 923 590 Patients With Atrial Fibrillation in the State of Florida Characteristic
White
Black
Hispanic
799 479
61 037
63 074
77.1 ± 11.5
68.8 ± 14.8
74.9 ± 13.0
48
54
50
Medicare, %
87
82
84
Catheter ablation, %a
1.0
0.81
0.99
Hypertension, %a
71
79
75
Diabetes, %a
29
43
39
5
6
5
2.49 ± 2.14
3.07 ± 2.25
2.17 ± 2.12
94
93
99
No. Agea a
Female gender, % a
a
Heart failure, %
a
Charlson score
Urban facility, %a a
P < 0.01 for comparisons between ethnicities.
Table 2. Baseline Characteristics of 28 714 Subjects With Atrial Flutter in the State of Florida
Figure 1. Percentage of the use of catheter ablation by ethnicity (y-axis) per year (x-axis) in patients with atrial fibrillation.
Table 3. Univariate and Multivariate Odds Ratios of Catheter Ablation for Atrial Fibrillation Univariate OR (95% CI) P Value
Multivariate OR (95% CI) P Value
Characteristic
White
Black
Hispanic
No.
23 253
3 257
1934
Characteristic
Agea
71.2 ± 14.0
61.4 ± 16.8
65.5 ± 19.7
Black vs white
0.59 (0.52-0.66)
