Application of the simple atrial fibrillation better care pathway for integrated care management in frail patients with atrial fibrillation: A nationwide cohort study

Abstract Background The benefit of integrated care management was unknown in frail atrial fibrillation (AF) patients. This study evaluated whether compliance with the atrial fibrillation Better Care (ABC) pathway for integrated care management would improve clinical outcomes in frail AF patients. Methods From the Korea National Health Insurance Service database, 262,987 nonvalvular AF patients were enrolled between 1 January 2005 and 31 December 2015. For each patient, the Hospital Frailty Risk Score and category were calculated retrospectively using all available ICD‐10 diagnostic codes. Patients were divided into three frailty‐based risk categories: low (<5 points, n = 221,542), intermediate (5‐15 points, n = 37,341), and high risk (>15 points, n = 4,104). Results Over a mean follow‐up of 5.9 (interquartile range 3.2, 9.4) years, in high frailty risk patients, the ABC group had lower rates of all‐cause death (6.5 vs 17.5 per 100 person‐years, P < .001; hazard ratio [HR] 0.74; 95% confidence interval [CI] 0.56‐0.97) but was nonsignificant for the composite outcome (10.5 vs 26.0 per 100 person‐years, P = .101; HR 0.79; 95% CI 0.59‐1.05) compared with the Non‐ABC group. When the three frailty categories were compared, the greatest benefit on mortality was seen in the high frailty group (pint < 0.001), but for the composite outcome, there was no statistical interaction for the three frailty categories (pint = 0.063). Conclusions Compliance with the simple ABC pathway is associated with improved outcomes in AF patients with high frailty risk. Given the high healthcare burden associated with frail AF patients, integrated AF management should be implemented to improve outcomes in these patients.


| INTRODUC TI ON
Atrial fibrillation (AF) is the most common sustained cardiac arrhythmia among elderly individuals, [1][2][3] and it has enormous socioeconomic implications given the risk of mortality and morbidity resulting from stroke, congestive heart failure, dementia, and impaired quality of life. [1][2][3][4][5] Frailty is also associated with more adverse clinical outcomes in elderly individuals admitted to the hospital. AF may be a marker of frailty in elderly individuals and may be related to a loss of independence in performing activities of daily living. 6 AF could worsen the state of frailty, and patients with AF could have 4-fold increased odds of being classified as frail, compared with patients without AF. 7 Recent trials involving AF 8,9 revealed a high (4.6% per year) rate of all-cause death in patients with AF, with only one out of 10 deaths related to stroke and approximately five or six out of 10 deaths related to cardiovascular causes. Therefore, a more integrated and holistic approach beyond anticoagulation therapy for patients with AF has been advocated in guidelines to reduce mortality and adverse outcomes in AF. [10][11][12] One way is to streamline management approaches that would be applicable across the entire AF patient pathway, starting with primary care and linking with secondary care (even for cardiologists and noncardiologists), and be understandable for patients with AF. The ABC (atrial fibrillation better care) pathway has been proposed as a simple, integrated approach. 13 This pathway streamlines the care pathway as follows: "A" Avoid stroke with Anticoagulation; "B" Better symptom management (ie, patient-centered, symptom-directed decisions on rate vs rhythm control); and "C" Cardiovascular and comorbidity management, including lifestyle factors. 13 Application of the simple ABC pathway was associated with a lower risk of adverse outcomes in patients with AF in a post hoc analysis of a clinical trial cohort as well as other AF cohorts. [14][15][16] However, the population-based benefit on clinical outcomes owing to a compliance with the ABC pathway has not been previously evaluated in patients with AF at a high frailty risk. Give the close association between AF and frailty, this study aimed to evaluate whether compliance with the ABC pathway would improve population-based clinical outcomes in patients with AF belonging to different frailty risk categories, using a nationwide AF cohort.

| Data source
We performed a retrospective analysis of data from the national health claims database (NHIS-2016-4-

| Study cohort
From the Korean NHIS database, a total of 955 111 patients with prevalent AF who were aged 18 years or older were identified from 1 January 2005 to 31 December 2015. Patients with valvular AF, such as those with any mechanical or bioprosthetic heart valves, mitral valve repair, or rheumatic mitral stenosis (n = 59 189); those without baseline health check-up data up to 1 year before enrolment (n = 571 585); and those who had ischemic stroke (n = 61 350) were excluded. Finally, a total of 262,987 patients with nonvalvular AF were enrolled in the study to evaluate the impact of the ABC pathway on the long-term clinical outcomes of these patients ( Figure 1).
For each patient, the Hospital Frailty Risk Score was calculated retrospectively using all available ICD-10 diagnostic codes that were documented for the particular admission, as recommended by Gilbert et al. 17 The score is an aggregate of 109 ICD-10 diagnostic codes found to be associated with frailty-based risk (Supplementary   Table S1). Each diagnostic code was assigned a specific value proportional to how strongly it predicted frailty. According to the aggregate score, patients were divided into three frailty-based risk categories: low risk (<5 points, n = 221 542), intermediate risk (5-15 points, n = 37 341), and high risk (>15 points, n = 4104). 17

| Definition of the ABC pathwaycompliant group
The integrated care group (ABC group) was defined according to the criteria summarized in Supplementary Figure S1. "A" was defined as the use of oral anticoagulants, in accordance with guidelines, with high adherence (prescription covering ≥80% of days); "B" was defined in relation to visits requiring medical contact with outpatient clinics (<5 visits per year during the follow-up period); "C" was defined as optimal management of the main cardiovascular comorbidities (hypertension, heart failure, myocardial infarction, peripheral artery disease, stroke/transient ischemic attack [TIA], diabetes mellitus, and obesity). Optimal management of hypertension was defined as baseline blood pressure values <140/90 mmHg. For obesity, body mass index less than 30 kg/m 2 was considered optimal management. For other comorbidities, appropriate use of cardiovascular prevention medications according to current guidelines was considered optimal management. Patients who fulfilled all criteria were defined as the "ABC" group, and those who did not fulfill all criteria were defined as the "non-ABC" group.

| Comorbidities and endpoints
AF was identified using International Classification of Disease, Tenth Revision (ICD-10) codes: I48. To ensure accuracy, the diagnosis was established based on more than one or two outpatient records of ICD-10 codes in the database. The diagnosis of AF has previously been validated in the NHIS database, with a positive predictive value of 94.1%. 1,4,18,19 Comorbidities were identified from ICD-10 codes and prescription, as in previous studies (Supplementary Table S2). 18 The primary clinical outcomes of this study were all-cause death, ischemic stroke, heart failure admission, acute myocardial infarction, major bleeding, and a composite outcome of these five outcomes.
Any diagnosis of ischemic stroke in the emergency room or inpatient clinic with concomitant brain imaging studies, including computed tomography or magnetic resonance imaging, was defined as incident ischemic stroke. The accuracy of the diagnosis of ischemic stroke in the NHIS claims data has been previously validated. 4 The other definitions of clinical outcomes are presented in Supplementary   Table S2. Patients were followed from the index date until the study outcomes occurred or up to the end of follow-up, whichever occurred first.

| Statistical analysis
Categorical data are reported as proportions, while continuous data are reported as medians with interquartile ranges (IQRs). The categorical variables were compared using Fisher's exact test or Pearson chi-square test, and continuous variables were compared using Student's t test. The main analyses compared the clinical outcomes between the ABC (ie, integrated care) and non-ABC groups.
Incidence rates were defined as events per 100 person-years at risk but expressed as annualized percentage rates for comprehensiveness. The relationships between the total number of ABC criteria fulfilled and the clinical outcomes were also investigated.
The cumulative incidences of adverse outcomes were presented using Kaplan-Meier curves and compared across the groups using the log-rank test. Using Cox proportional hazard regression model, the hazard ratios (HRs) for adverse outcomes according to the use of integrated care (ABC) were analyzed. Clinical variables including age, gender, heart failure, hypertension, diabetes mellitus, previous myocardial infarction, peripheral artery disease, economic status, CHA 2 DS 2 -VASc, and HAS-BLED score were adjusted for HR. Pvalues < 05 were considered significant. Statistical analyses were conducted using SAS version 9.3 (SAS Institute) and R version 3.3.2 (The R Foundation, www.R-proje ct.org).

| Baseline characteristics
Comparisons between the ABC and non-ABC groups are presented in Supplementary Table S3 Hospital Frailty Risk Score

| Death and composite outcomes
Patients with all three types of frailty risk in the ABC group had significantly lower cumulative incidences of all-cause death

| Other outcomes
The Kaplan-Meier cumulative incidence curves for other outcomes are presented in Figure 4. Among AF patients with low and intermediate frailty risk, lower cumulative incidences of ischemic stroke, heart failure admission, acute myocardial infarction, and major bleeding were observed in the ABC group compared with the non-ABC group (all log-rank P < .001). Among AF patients with high frailty risk, the lower cumulative incidence of clinical outcomes was also observed in the ABC group than in the non-ABC group, but statistical significance was on the border.
The event rates and risks for other outcomes according to the ABC and non-ABC groups are presented in Figure 5.

TA B L E 1 (Continued)
interaction (P < .001). But among AF patients with overall and low frailty risk group, there was no significant difference in the risk of major bleeding between ABC and non-ABC groups.

| D ISCUSS I ON
In this largest nationwide analysis of patients with AF according to frailty risk, the event rates and risks of the all-cause death, ischemic stroke, heart failure admission, acute myocardial infarction, major bleeding, and composite of these outcomes were significantly lower in the ABC group than in the non-ABC group. In addition, among patients with high frailty risk, compared with patients in the non-ABC group, those in the ABC group had lower rates of all-cause death; however, the composite outcome was nonsignificantly lower in the ABC group. When the three frailty categories were compared, the greatest benefit on mortality was observed in the high frailty group; however, with regard to the composite outcome, there was no statistical interaction for the three frailty categories. Given the close association between AF and frailty and the high healthcare burden associated with AF, a streamlined holistic approach to the management of AF would improve outcomes in such patients.

| Mortality and outcomes according to frailty risk
The use of an integrated care approach to AF management has been associated with reduced cardiovascular hospitalization and all-cause mortality. 20 Nevertheless, approaches for providing integrated care Abbreviations: CI, confidence interval; HR, hazard ratio; NOAC nonvitamin K antagonist oral anticoagulant. Other abbreviations are same as in Table 1.
have varying complexity. 11 There is a need to have a simple, practical, and easily operational method to streamline the decision-making management approaches to allow uniform applicability across the entire AF patient pathway, linking primary care and secondary care (including cardiologist and noncardiologists), and to be understandable for patients with AF, facilitating their engagement.
The ABC pathway was proposed to streamline the interventions and decision-making, and to optimize the patient management pathway, providing simple guidance for the main components of integrated care. 13 Recent AF management guidelines have incorporated the ABC pathway. 12,21 Nevertheless, there are limited data on the value of the ABC pathway in high-risk patient groups. In this study, we show that the ABC pathway was related to reduced mortality and composite outcomes in frail patients with AF. The strong impact of the ABC pathway on overall mortality substantiates and strengthens the concept that a holistic approach for integrated management is associated with a significant clinical benefit for patients with AF.
Indeed, compliance to the ABC pathway was also associated with a lower risk of ischemic stroke, heart failure admission, and acute myocardial infarction, as well as major bleeding, in patients with AF. A greater benefit in terms of major bleeding was observed in patients with AF and high frailty risk.
Although the risk of thromboembolic events is high, the rate of adequate oral anticoagulation is lower in frail patients with AF compared to nonfrail patients. 8  Despite these limitations, to the best of our knowledge, this study presents the largest nationwide population dataset available in the literature to investigate the relationship between frailty and cardiovascular outcomes in patients with AF.

| CON CLUS IONS
Compliance with the simple ABC pathway is associated with improved outcomes in patients with AF who have a high frailty risk.
Given the high healthcare burden associated with AF, such a streamlined holistic approach to the management of AF should be implemented to improve outcomes in such patients.

ACK N OWLED G M ENTS
The National Health Information Database was provided by the National Health Insurance Service of Korea. We thank the National Health Insurance Service for its cooperation.
F I G U R E 5 Events, event rates, and risks of other adverse outcomes according to the use of integrated care (ABC) in patients with different frailty risks. CI, confidence interval; HR, hazard ratio; PYRs, person-years