JOURNAL OF THE AMERICAN COLLEGE OF CARDIOLOGY VOL. 75, NO.

13, 2020

ª 2020 BY THE AMERICAN COLLEGE OF CARDIOLOGY FOUNDATION

PUBLISHED BY ELSEVIER

Mobile Health Technology to Improve

Care for Patients With Atrial Fibrillation
Yutao Guo, MD, PHD,a Deirdre A. Lane, PHD,b Limin Wang, MD, PHD,c Hui Zhang, MD,a Hao Wang, MD, PHD,a
Wei Zhang, MD,d Jing Wen, MD,e Yunli Xing, MD, PHD,f Fang Wu, MD, PHD,g Yunlong Xia, MD, PHD,h
Tong Liu, MD, PHD,i Fan Wu, MD, PHD,j Zhaoguang Liang, MD, PHD,k Fan Liu, MD, PHD,l Yujie Zhao, MD,m
Rong Li, MD, PHD,n Xin Li, MD, PHD,o Lili Zhang, MD, PHD,p Jun Guo, MD, PHD,a Girvan Burnside, PHD,q
Yundai Chen, MD, PHD,a Gregory Y.H. Lip, MD,a,b on behalf of the mAF-App II Trial Investigators

ABSTRACT

BACKGROUND Current management of patients with atrial fibrillation (AF) is limited by low detection of AF, non-
adherence to guidelines, and lack of consideration of patients’ preferences, thus highlighting the need for a more holistic
and integrated approach to AF management.

OBJECTIVE The objective of this study was to determine whether a mobile health (mHealth) technology-supported AF
integrated management strategy would reduce AF-related adverse events, compared with usual care.

METHODS This is a cluster randomized trial of patients with AF older than 18 years of age who were enrolled in 40 cities
in China. Recruitment began on June 1, 2018 and follow-up ended on August 16, 2019. Patients with AF were randomized
to receive usual care, or integrated care based on a mobile AF Application (mAFA) incorporating the ABC (Atrial Fibril-
lation Better Care) Pathway: A, Avoid stroke; B, Better symptom management; and C, Cardiovascular and other co-
morbidity risk reduction. The primary composite outcome was a composite of stroke/thromboembolism, all-cause death,
and rehospitalization. Rehospitalization alone was a secondary outcome. Cardiovascular events were assessed using Cox
proportional hazard modeling after adjusting for baseline risk.

RESULTS There were 1,646 patients allocated to mAFA intervention (mean age, 67.0 years; 38.0% female) with mean
follow-up of 262 days, whereas 1,678 patients were allocated to usual care (mean age, 70.0 years; 38.0% female) with
mean follow-up of 291 days. Rates of the composite outcome of ‘ischemic stroke/systemic thromboembolism, death, and
rehospitalization’ were lower with the mAFA intervention compared with usual care (1.9% vs. 6.0%; hazard ratio [HR]:
0.39; 95% confidence interval [CI]: 0.22 to 0.67; p < 0.001). Rates of rehospitalization were lower with the mAFA
intervention (1.2% vs. 4.5%; HR: 0.32; 95% CI: 0.17 to 0.60; p < 0.001). Subgroup analyses by sex, age, AF type, risk
score, and comorbidities demonstrated consistently lower HRs for the composite outcome for patients receiving the
mAFA intervention compared with usual care (all p < 0.05).

CONCLUSIONS An integrated care approach to holistic AF care, supported by mHealth technology, reduces the risks of
rehospitalization and clinical adverse events. (Mobile Health [mHealth] technology integrating atrial fibrillation screening
and ABC management approach trial; ChiCTR-OOC-17014138). (J Am Coll Cardiol 2020;75:1523–34).
© 2020 by the American College of Cardiology Foundation.

From the aMedical School of Chinese PLA, Department of Cardiology, Chinese PLA General Hospital, Beijing, China; bLiverpool
Listen to this manuscript’s
Centre for Cardiovascular Sciences, University of Liverpool and Liverpool Heart & Chest Hospital, Liverpool, United Kingdom, and
audio summary by
Aalborg Thrombosis Research Unit, Department of Clinical Medicine, Aalborg University, Aalborg, Denmark; cThe National Center
Editor-in-Chief
for Chronic and Noncommunicable Disease Control and Prevention, Chinese Center for Disease Control and Prevention, Beijing,
Dr. Valentin Fuster on
China; dDepartment of Gerontology and Geriatric Medicine, Seventh Clinical Center, Chinese PLA General Hospital, Beijing, China;
JACC.org.
e
Department of Geriatric Cardiology, Haidian Hospital, Beijing, China; fBeijing Friendship Hospital, Capital Medical University,
Beijing, China; gDepartment of Gerontology and Geriatric Medicine, Ruijin Hospital Affiliated to School of Medicine, Shanghai
Jiaotong University, Shanghai, China; hDepartment of Cardiology, First Affiliated Hospital of Dalian Medical University, Dalian,
China; iTianjin Key Laboratory of Ionic-Molecular Function of Cardiovascular Disease, Tianjin Institute of Cardiology, Second
Hospital of Tianjin Medical University, Tianjin, China; jDepartment of Geriatrics, Tianjin Medical University General Hospital,
Tianjin Geriatrics Institute, Tianjin, China; kDepartment of Cardiology, First Affiliated Hospital of Haerbing Medical University,

ISSN 0735-1097/$36.00 https://doi.org/10.1016/j.jacc.2020.01.052

1524 Guo et al. JACC VOL. 75, NO. 13, 2020

Mobile Health for Patients With AF APRIL 7, 2020:1523–34

A trial fibrillation (AF), being the most

ABBREVIATIONS secondary care management, as well as allowing pa-
AND ACRONYMS common sustained arrhythmia, re- tient engagement in an integrated care approach,
mains one of the major global causes remain untested.
ACS = acute coronary
syndrome
of stroke, heart failure, dementia, sudden AF clinical integrated care aims to support a
death, and cardiovascular morbidity (1). multidisciplinary approach to optimize stroke pre-
AF = atrial fibrillation
Although there has been substantial progress vention, patient-centered or symptom-directed de-
BP = blood pressure
on the management of AF, doctors’ nonad- cisions on rate or rhythm control, and management of
CAD = coronary artery disease
herence to AF management guidelines and cardiovascular risks and comorbidities, including
CI = confidence interval
lack of incorporation of patients’ preferences lifestyle interventions. Such a holistic approach to AF
DOACs = direct oral
in treatment decisions remain major prob- care is simplified into a practical, simple ABC pathway
anticoagulants
lems (2,3). New approaches to AF manage- (Atrial Fibrillation Better Care pathway, i.e., Avoid
HF = heart failure
ment, including the use of novel stroke; Better symptom management with patient-
HR = hazard ratio
technologies and a more structured inte- centered, symptom directed decisions on rate or
mAFA = mobile atrial
grated or holistic approach to AF care, are rhythm control; Cardiovascular and other comorbid-
fibrillation application
proposed to optimize treatment options for ity risk reduction) (10,11). The simple ABC pathway
mHealth = mobile health
AF (4). has been retrospectively validated in independent
OACs = oral anticoagulants
retrospective or prospective cohort studies, demon-
SEE PAGE 1535
PPG = photoplethysmography
strating a lower risk of adverse outcomes for patients
TTR = time in therapeutic with AF who were managed according to the ABC
The emergence and rapid growth of digital
range
health technology may enable doctors and pathway (12–14).
patients to improve AF management (5). Digital We hypothesized that implementation of a
health refers to the use of mobile computing and mHealth technology-supported integrated manage-
communication technologies (e.g., mobile phones, ment strategy would reduce AF-related adverse
wearable sensors) for health services and information events (stroke/thromboembolism, all-cause death,
(6). Indeed, mobile health (mHealth) tools have been and rehospitalization) compared with usual care. We
studied as an aid to support shared decision-making tested this hypothesis in a prospective cluster ran-
for anticoagulation, to achieve telemonitoring-based domized trial, as part of the mHealth technology for
feedback, and to improve medication adherence improved screening, patient involvement, and opti-
(7,8). In our pilot study, a mHealth technology- mized integrated care in Atrial Fibrillation (mAFA II)
supported AF management model (mobile Atrial study program (15).
Fibrillation Application [mAFA I]) was designed,
including clinical decision support tools (9). We METHODS
demonstrated that the mHealth tool could be used for
AF management to improve knowledge and drug The design and rationale of the mAFA II trial has been
adherence (9). However, strategies to incorporate described previously (15). In brief, we conducted a
such technology effectively into the AF management 2-arm, prospective, cluster-randomized controlled
pathways that can be applicable from primary care to trial that prospectively enrolled patients with AF

Haerbing, China; lDepartment of Cardiology, Second Hospital of Hebei Medical University, Shijiazhuang, China; m
Department of
Cardiology, Henan Cardiovascular Hospital Affiliated to Southern Medical University, Henan, China; nDepartment of Cardiology,
First Affiliated Hospital of Guangzhou University of Chinese Medicine, Guangzhou, China; oDepartment of Cardiology, Benq
Medical Center, Nanjing Medical University, Nanjing, China; pDepartment of Cardiology, Longhua People’s Hospital, Shenzhen,
China; and the qDepartment of Biostatistics, University of Liverpool, Liverpool, United Kingdom. This research project was funded
by the National Natural Science Foundation of China (H2501), National Key Research and Development Project of China
(2018YFC2001200), and the Health and Family Planning Commission of Heilongjiang Province, China (2017-036), and was partly
supported by the National Institute for Health Research (NIHR) Global Health Research Group on Atrial Fibrillation management
at the University of Birmingham, United Kingdom. This study was an investigator-initiated project, with limited funding by in-
dependent research and educational grants. Funders had no role in study design, data collection, data analysis, data interpre-
tation, or writing of the report. Dr. Lip is a consultant for Bayer/Janssen, Bristol-Myers Squibb/Pfizer, Medtronic, Boehringer
Ingelheim, Novartis, Verseon, and Daiichi-Sankyo; and is a speaker for Bayer, Bristol-Myers Squibb/Pfizer, Medtronic, Boehringer
Ingelheim, and Daiichi-Sankyo (no fees are directly received personally). Dr. Lane has received grants from Bristol-Myers Squibb
and Boehringer Ingelheim (paid to the institution); and has received personal fees from Boehringer Ingelheim, Bristol-Myers
Squibb/Pfizer, Bayer, and Daiichii-Sankyo, outside the submitted work. All other authors have reported that they have no
relationships relevant to the contents of this paper to disclose.

Manuscript received November 13, 2019; revised manuscript received January 14, 2020, accepted January 21, 2020.
JACC VOL. 75, NO. 13, 2020 Guo et al. 1525
APRIL 7, 2020:1523–34 Mobile Health for Patients With AF

(Supplemental Figure 1). The study was registered on A user-friendly mAFA was developed for smart
the World Health Organization International Clinical phones based on the Android Operating System
Trials Registry Platform, and the registration number (Google Inc., MountainView, California) and Apple
is ChiCTR-OOC-17014138. iOS (Cupertino, California) for doctors (DmAFA) and
Clusters (sites) were identified by each coordi- patients (PmAFA). In the mAFA intervention group,
nating center based on hospital size, patient volume, the doctors used the mAFA platform (Supplemental
time the doctors could spend with patients after Figures 2 to 4) to manage patients with AF. The
discharge, patient’s smart phone usage, and general mAFA platform provided clinical decision support
education level of the patient population tools (CHA2DS2-VASc, hypertension, abnormal renal/
(Supplemental Table 1). All clusters demonstrated liver function, stroke, bleeding history or predispo-
access to adequate numbers of eligible patients by sition, labile international normalized ratio, elderly,
pre-trial feasibility questionnaires. The sites were drugs/alcohol concomitantly [HAS-BLED], sex, age,
matched based on hospital size and the proportion of medical history, treatment, tobacco use, race
enrolled patients. The hospital sizes were classified as [SAMeTT 2 R2] scores) to facilitate guideline-based
“big” hospitals with enrollment of >20 patients per treatment recommendations, educational materials
month, and “small” hospitals with enrollment of <20 and patient involvement strategies with self-care
patients per month, respectively. The ratio of protocols, and structured follow-up, to support
big:small hospitals was 1:2 based on feasibility implementation of the ABC pathway for integrated or
checks, thus, 142 patients from individual big hospi- holistic AF management, compliant with guidelines
tals and 71 from small hospitals would be needed. on AF management (1), as follows.
Inclusion criteria were as follows: 1) patients $18 A v o i d s t r o k e . Oral anticoagulants (OACs) were cho-
years of age, diagnosed with new-onset, paroxysmal, sen based on clinical decision support tools and pa-
persistent or permanent AF confirmed with electro- tients’ preferences at baseline. The app provided
cardiogram or 24-h Holter monitors; and 2) conges- guidance on appropriate dosing, based on the
tive heart failure, hypertension, $75 years of age, particular drug label or guideline recommendations.
diabetes, stroke, vascular disease, age 65 to 74, and Given the dynamic nature of stroke and bleeding risk
sex category (female) (CHA2 DS2-VASc) score $2. Pa- factors (17), regular clinical risk (re)assessment was
tients were excluded if they met any of the following incorporated into the app, after the initial baseline
criteria: <18 years of age, those with mechanical assessment. For example, dynamic bleeding risks
prosthetic valve or moderate/severe mitral stenosis, were automatically monitored with the HAS-BLED
unable to provide informed consent, or unable to be score by the mAFA platform, once the patient’s data
followed up for 1 year for any reason. (comorbidities, laboratory tests, etc.) were updated.
Suitable patients were enrolled into the mAFA II The trends of bleeding risks over time are shown in
trial from 2 sources: 1) the initial AF screening pro- line charts and personalized modifiable bleeding risk
gram (‘pre-mAFA’) (16); and 2) the outpatient and in- factors were flagged for both doctors and patients to
patient departments of participating centers. Trial help to achieve safe anticoagulant use. The time in
patients were consecutively recruited at each site in therapeutic range (TTR) in patients taking warfarin
China between June 1, 2018 and August 16, 2019. The was automatically calculated for the patients on
study was approved by the Central Medical Ethic warfarin; liver function was assessed with Child-
Committee of the Chinese PLA General Hospital Tucotte-Pugh score; and the dynamic evaluation of
(approval number: S2017-105-02), and by local insti- renal function was calculated with the creatinine
tutional review boards. The study was compliant with clearance (Cockroft-Gault). Optimized dose adjust-
the Declaration of Helsinki. ments of warfarin or direct oral anticoagulants
RANDOMIZATION AND INTERVENTION. In the clus- (DOACs) were proposed based on changes in TTR,
ter randomized parallel intervention trial design, 40 liver, or renal function, being consistent with practice
participating cluster hospitals were randomized in a guidelines (Supplemental Figure 2) (18). A mean TTR
1:1 ratio to the mAFA intervention or usual care. of >65% was defined as good anticoagulation control.
Randomization was done using a computer- Patient-reported thromboembolism or bleeding
generated randomization list. Investigators and site events were captured using the structured question-
personnel were not masked to the intervention. Pa- naire developed by the mAFA platform. Patient-
tients in the usual care group received treatment and reported thromboembolism or bleeding events were
management by local doctors according to local clin- captured using structured questionnaires developed
ical practice. for the mAFA platform. Once patients reported
1526 Guo et al. JACC VOL. 75, NO. 13, 2020

Mobile Health for Patients With AF APRIL 7, 2020:1523–34

thromboembolism or bleeding events, they were caused by ST-segment elevation myocardial

required to upload the supporting files (e.g., pictures infarction/non–ST-segment elevation myocardial
of the bleeding, medical material). Doctors could also infarction, heart failure (HF), arrhythmia, cardiac
communicate with patients to confirm the reported perforation/tamponade, and other deaths of cardiac
events and decide if they needed to be further diag- origin. Vascular death included death ascribed to
nosed and treated in the hospital or could be ischemic stroke, hemorrhagic stroke, systemic hem-
managed remotely through the mAFA app. orrhage, peripheral embolism, and pulmonary em-
Better symptom m a n a g e m e n t . Cardiac rhythm bolism. We also recorded rehospitalization for any
monitoring was available in the mAFA platform, us- cause for AF, thromboembolism, major bleeding, HF,
ing photoplethysmography (PPG)-smart devices, as ACS, and admission for other cardiovascular disease.
previously reported (16). AF symptoms were evalu- Other cardiovascular outcomes included recurrent
ated using the European Heart Rhythm Association AF, which was defined as recurrent onset of AF for
classification and assessment, once AF episodes were patients with paroxysmal AF or with AF ablation.
detected using PPG. Other reported symptoms, such A secondary outcome included event rates for the
as headache, dizziness, shortness of breath, and chest components of the primary endpoint, and the change
pain, could be monitored and recorded. Once such in the proportion of patients able to continue
data were updated, the instant message function anticoagulation.
could be used to communicate information to doc- STATISTICAL ANALYSIS. Sample size and power for
tors. Doctors could also communicate with the pa- the primary outcome were calculated according to
tients on the mAFA in a timely manner and regulate Eldridge’s method (19). Study feasibility was investi-
use of antiarrhythmic drugs or rate control therapies gated among 52 researchers from 42 hospitals be-
according to guidelines on AF management (1) tween September 1, 2017 and March 31, 2018, in
(Supplemental Figure 3). relation to hospital size, ability of enrollment per
Cardiovascular and other comorbidities risk month, age distribution of patients, possible time
m a n a g e m e n t . Associated comorbidities could be doctors would like to spend with patients after
proactively managed, for example, blood pressure discharge, mobile phone usage, educational levels,
(BP) could be monitored, and treatment optimized, and so on (Figure 1).
aiming for BP <140/85 mm Hg (and ideally, 130/80 Assuming the coefficient of variation of the cluster
mm Hg); statins were used in association with size of 2.60 based on the hospital size and enrollment
vascular disease. Lifestyle factors were recorded, ability, with an intra-cluster correlation of 0.02, 20
such as alcohol intake (with education and recom- clusters per group with an average cluster size of 71 to
mendations for reduction, as appropriate). 142 patients, and a 2-sided type 1 error of 0.05, the
The app also encouraged patient engagement by study was powered >90% to detect a 5% absolute
encouraging their participation in educational pro- difference in the composite of stroke/thromboembo-
grams, provision of informative articles, videos, game lism, all-cause death, and rehospitalization at 1 year,
playing, and so on. Educational materials were about and to achieve 80% power to detect 3% absolute
AF, hypertension, acute coronary syndrome (ACS), difference of such an improvement between mAFA
heart failure, valvular disease, self-care, and so on intervention and usual care. Therefore, a total of
(Supplemental Figure 4). 3,294 patients needed to be enrolled in this study.
Supplemental Figures 2D and 3C show how the All analyses followed the intention-to-treat prin-
doctors with DmAFA and patients with PmAFA ach- ciple. Baseline characteristics for continuous vari-
ieved integrated care management. ables are summarized as mean (standard deviation)
OUTCOMES. All patients were followed up in the and median (interquartile range). Frequencies and
outpatient clinics at 6 and 12 months for clinical percentages per group as well as hazard ratios (HRs)
events. The primary endpoint was the composite of with 95% confidence interval (CI) are reported for
stroke/thromboembolism, all-cause death, and reho- binary outcomes. Cox proportional hazards models,
spitalization. The thromboembolism endpoint with shared frailties to account for the effect of
included ischemic stroke, pulmonary embolism, deep clustering, adjusted for baseline risk factors, were
vein thromboembolism, and other thromboemboli used to analyze the primary composite outcome of
(peripheral embolism, atrial thrombus, and left atrial stroke/thromboembolism, all-cause death, and reho-
appendage thrombus, and so on). All-cause death spitalization. Additionally, the impact of the mAFA
included cardiac death, vascular death, and non- intervention on clinical outcomes was investigated
cardiovascular death. Cardiac death included death with the frailty Cox model, including the time to first
JACC VOL. 75, NO. 13, 2020 Guo et al. 1527
APRIL 7, 2020:1523–34 Mobile Health for Patients With AF

F I G U R E 1 Flow Chart of Patients Included Into the mAFA-II Trial

52 researchers from 42 sites assessed for eligibility with regard

to hospital size and patient catchment
(Sep 1, 2017-March 31, 2018)

Sites Researchers
40 sites participated in the study Willingness to manage patients discharged from the hospitals
2 sites dropped 51 researcher’s willingness
1 site for the patient catchment 1 researcher without response
1 site for operational reasons Concerns in AF management
28 on safe drug use, treatment persistence and adherence
24 on syndrome improvement cardiac rhythm/heart rate
20 on stroke/thromboembolism prevention, bleeding etc
16 on blood pressure, cardiac function, comorbidities control
4 on biochemical test (liver, renal function), coagulation test

40 cites randomized
(June 1, 2018 to August 16, 2019)

20 sites with mAFA intervention

Received integrated care ABC pathway 20 sites in the control group
Avoid stroke: Better symptom management: Cardiovascular Received usual care
and other comorbidity risk reduction

1,786 patients with mAFA intervention 1,842 patients with usual care
140 patients lost to follow-up 164 patients lost to follow-up
131 cannot contact patients 151 cannot contact patients
9 refused follow-up 13 refused follow-up

1,646 patients (92%) 1,678 patients (91%)

entered into the present analysis entered into the present analysis

AF ¼ atrial fibrillation; mAFA ¼ mobile atrial fibrillation application.

occurrence of ischemic stroke, systemic thromboem- adjusting for baseline risk factors. The change in
bolism, ACS, HF, rehospitalization, or all-cause death, proportion of patients on anticoagulation was evalu-
also adjusting for baseline risk factors. Adjusted ated with Mantel-Haenszel statistics and adjusted for
model included age, sex, AF type, prior AF treatment the effect of clustering. All statistical tests were done
(cardioversion, AF ablation, rhythm control), and at the nominal 0.05 (2-sided) significance level. All
comorbidities (hypertension, diabetes, coronary ar- statistical analyses were conducted using IBM SPSS
tery disease [CAD], obstructive sleep apnea Statistics, version 22.0 (SPSS Inc., Chicago, Illinois),
syndrome, HF, hyperthyroidism, prior ischemic MedCalc version 19.0.4 (MedCalc Software bvba,
stroke, dilated cardiomyopathy, hypertrophic Ostend Belgium), and SAS software, version 9.4 (SAS
cardiomyopathy). Institute, Cary, North Carolina) for frailty Cox model.
Subgroup analyses for the primary composite
outcome and secondary outcomes (rehospitaliza- RESULTS
tions) were conducted by age strata, sex, CHA2DS2 -
VASc score, HAS-BLED score, hypertension, CAD, We enrolled 3,324 patients in 40 centers between
paroxysmal AF, and persistent/permanent AF, after June 1, 2018 and August 16, 2019 (Figure 1). Most of
1528 Guo et al. JACC VOL. 75, NO. 13, 2020

Mobile Health for Patients With AF APRIL 7, 2020:1523–34

The treatments, with respect to the ABC pathway,

T A B L E 1 Baseline Characteristics of the mAFA Intervention and Usual Care
with mAFA intervention and usual care are summa-
mAFA Intervention Usual Care rized in Supplemental Table 3. For the patients with
(n ¼ 1,646) (n ¼ 1,678)
mAFA, there were 1,260 patients using PPG-based
Age, yrs 67.0  15.0 70.0  12.0
smart devices to monitor pulse rhythm, and they
Female 625 (38.0) 637 (38.0)
Current smoking 159 (9.5) 168 (10.2) achieved adequate symptom control using mAFA (B
Medical history criterion of ABC pathway, i.e., Better symptom con-
Hypertension 908 (55.2) 962 (57.3) trol; case shown in Supplemental Figure 3).
CAD 635 (38.6) 724 (43.1)
Diabetes mellitus 381 (23.1) 366 (21.8) CARDIOVASCULAR OUTCOMES. The rates of the
Heart failure 360 (21.9) 354 (21.1) composite outcome of ischemic stroke/systemic
Prior ischemic stroke 191 (11.6) 232 (13.8) thromboembolism, death, and rehospitalization were
PAD 172 (10.4) 172 (10.3)
significantly lower in those allocated to the mAFA
Renal dysfunction 138 (8.4) 172 (10.3)
intervention group compared with the usual care
Pulmonary hypertension 87 (5.3) 83 (4.9)
Liver dysfunction 55 (3.3) 48 (2.9)
group (1.9% vs. 6.0%, respectively; HR: 0.39; 95% CI:
Prior TE 54 (3.3) 59 (3.5) 0.22 to 0.67; p < 0.001). The rate of rehospitalization
Prior other bleeding 54 (3.3) 67 (4.0) was significantly lower in the mAFA group compared
Dilated cardiomyopathy 44 (2.7) 61 (3.6) with the usual care group (1.2% and 4.5%, respec-
Hyperthyroidism 37 (2.2) 51 (3.0) tively; HR: 0.32; 95% CI: 0.17 to 0.60; p < 0.001)
Hypertrophic cardiomyopathy 25 (1.5) 29 (1.7)
(Table 2). Detailed reasons for rehospitalization are
Prior brain bleeding 24 (1.5) 38 (2.3)
summarized in Supplemental Table 4.
AF type
The cumulative risk of cardiovascular events over
New-onset AF 195 (11.9) 232 (13.8)
Paroxysmal AF 673 (40.9) 660 (39.3) time, adjusting for age, AF type, prior AF treatment,
Persistent AF 380 (23.1) 448 (26.7) and comorbidities, are shown in Figures 2 and 3.
Long-standing AF 56 (3.4) 101 (6.0)
Permanent AF 48 (2.9) 123 (7.3) CHANGES OF OACs IN mAFA AND USUAL CARE
Unknown AF type 281 (17.1) 113 (6.7) GROUPS. The baseline proportions of OAC use in the
Prior AF treatment usual care and mAFA arms was 48.4% (812 of 1,678
Pharmacy cardioversion 213 (12.9) 155 (9.2) patients) and 66.1% (1,088 of 1,646 patients),
Electrical cardioversion 30 (1.8) 35 (2.1)
respectively (p < 0.001). Patients with mAFA more
AF ablation 183 (11.1) 173 (10.3)
commonly received DOACs than usual care
Dual chamber pacemaker 76 (4.6) 85 (5.1)
(Supplemental Table 3). The change in the use of
LAAO 33 (2.0) 30 (1.8)
CHA2DS2-VASc 3 (2-4) 3 (2-4) OACs among high risk patients (with CHA 2DS 2-
HAS-BLED 1 (1-2) 1 (1-2) VASc $2 in males, $3 in females) between baseline
SAMe-TT2R2 4 (3-4) 4 (3-4) and 12 months are shown in Supplemental Figure 5.

Values are mean  SD, n (%) or median (interquartile range). SUBGROUP ANALYSES. Subgroup analyses, by age,
AF ¼ atrial fibrillation; CAD ¼ coronary artery disease; CHA2DS2-VASc ¼ chronic heart failure,
hypertension, age >75 years, diabetes, stroke, vascular disease, age 65–74 years, sex; HAS- sex, AF type, risk scores (CHA 2DS2-VASc and HAS-
BLED ¼ hypertension, abnormal renal/liver function, stroke, bleeding history or predisposition, BLED scores), and comorbidities, demonstrated
labile international normalized ratio, elderly, drugs/alcohol concomitantly; IQR ¼ interquartile;
LAAO ¼ left atrial appendage occlusion; mAFA ¼ mobile Atrial Fibrillation Application; consistently lower HRs for the primary composite
PAD ¼ peripheral arterial disease; SAMe-TT2R2 ¼ sex, age, medical history, treatment, tobacco
outcome and rehospitalization for patients allocated
use, race; TE ¼ thromboembolism.
to mAFA intervention when compared with patients
receiving usual care (all p < 0.05 vs. usual
care) (Figure 4).

the patients, both mAFA intervention and usual care, DISCUSSION

were enrolled from an inpatient department
(Supplemental Table 2). In this study, 1,646 patients First, in the present prospective, multicenter, cluster
were allocated to mAFA intervention (mean age, 67.0 randomized trial of subjects with AF, an integrated
years; 38.0% female) with mean follow-up of care pathway approach based on the ABC pathway,
262 days, whereas 1,678 patients were allocated to supported by mHealth technology, significantly
usual care (mean age, 70.0 years; 38.0% female), with reduced the composite outcome of ischemic stroke/
mean follow-up of 291 days. The baseline character- systemic thromboembolism, death, and rehospitali-
istics of the mAFA intervention group and usual care zation (Central Illustration). Second, lower rehospi-
are shown in Table 1. talization rates in the mAFA intervention group were
JACC VOL. 75, NO. 13, 2020 Guo et al. 1529
APRIL 7, 2020:1523–34 Mobile Health for Patients With AF

observed, even after adjusting baseline age, AF type,

T A B L E 2 Clinical Outcomes in the mAFA and Usual Care Groups
treatment, and comorbidities, compared with usual
care. Third, the impact of mAFA intervention on the mAFA Usual Care HR* (mAFA vs.
(n ¼ 1,646) (n ¼ 1,678) Usual Care) 95% CI p Value
composite outcome was consistent irrespective of
Thromboembolism
age, sex, AF type, risk scores (CHA2DS2 -VASc and
Ischemic stroke 3 (0.2) 3 (0.2) 1.31 0.18–9.31 0.78
HAS-BLED scores), and comorbidities. Other systemic 4 (0.2) 3 (0.2) 1.02 0.18–5.93 0.97
The integrated care approach to disease manage- thromboembolism

ment originates in the chronic care model, which Bleeding events

Intracranial bleeding 0 (0.0) 2 (0.1) – – –
aimed to redesign daily practice to facilitate treat-
Extracranial bleeding 31 (1.9) 36 (2.1) 0.95 0.54–1.66 0.85
ment optimization and enhance patient outcomes
Cardiovascular outcomes
(20). An integrated care approach to AF management Recurrent atrial fibrillation 23 (1.4) 56 (3.3) 0.48 0.29–0.79 0.004
was proposed in the 2016 European Society of Cardi- Heart failure 24 (1.5) 33 (2.0) 0.99 0.51–1.92 0.97
ology guidelines on AF management (1), but how best Acute coronary syndrome 2 (0.1) 9 (0.5) 0.21 0.04–1.21 0.08
to operationalize an integrated, structured approach All-cause death 12 (0.7) 25 (1.5) 0.71 0.26–1.91 0.49

to AF care remained unclear. A nurse-led integrated Rehospitalization 20 (1.2) 75 (4.5) 0.32 0.17–0.60 <0.001
Composite outcome of IS/TE, 32 (1.9) 101 (6.0) 0.39 0.22–0.67 <0.001
outpatient AF care approach demonstrated improved
death, and rehospitalization
adherence to guidelines, with psychosocial support
and educational interventions during follow-up (21). Values are n (%) unless otherwise indicated. Extracranial bleeding included gastrointestinal, urogenital, skin, eye
bleeding, and other non-major bleeding. Recurrent AF: recurrent onset of AF for patients with paroxysmal AF or
Another post-discharge integrated care plan, con- with AF ablation. Reasons for rehospitalization included any cause for AF, heart failure, thromboembolism, major
sisting of home visits and 7- to 14-day Holter moni- bleeding, CAD, and other cardiovascular disease (Supplemental Table 4). The causes of death are shown in
Supplemental Table 5. *The frailty Cox model, adjusted for cluster effect, age, comorbidities, AF type, and prior
toring by a cardiac nurse and multidisciplinary AF treatment, was used to assess the effect of mAFA intervention on the clinical events.
support showed improvements in survival compared CI ¼ confidence interval; HR ¼ hazard ratio; IS ¼ ischemic stroke; TE ¼ thromboembolism.

with usual care (22).

Other studies have aimed to explore improvements
in (anticoagulant) drug adherence in primary care
F I G U R E 2 Cumulative Risk of Composite Outcome of Ischemic Stroke/TE,
(23,24). Educational interventions such as IMPACT-
Death, and Rehospitalization
AF (IMProve treatment with AntiCoagulanTs in pa-
tients with Atrial Fibrillation) have led to improved
rates of OAC use, which would translate into stroke mAFA vs. usual care
reduction in developing countries (24). Although
HR: 0.39 (95% CI: 0.22-0.67) P < 0.001
some progress on integrated care approaches are 0.06
implied, prior studies on integrated care have not
Cumulative Hazard Ratio

covered all main treatment targets for AF and avoid-

ance of AF-related complications.
In the present trial, we tested implementation of
the ABC pathway for the holistic, integrated man- 0.04
agement of patients with AF: 1) “A” Avoid stroke with
anticoagulation with dynamic monitoring of stroke/
bleeding risks, clinical decision support with dosage
adjustments of warfarin based on changes in TTR, or 0.02
label-adherent use of DOACs; 2) “B” Better symptom
management with patient-centered, symptom-
directed decisions on rate or rhythm control; and 3)
“C” Cardiovascular and comorbidity risk reduction,
with the timely monitoring (and treatment) of BP, 0.00
optimization of cardiovascular prevention strategies
(e.g., statins for vascular disease, angiontensin-con- 0 100 200 300 400
verting enzyme inhibitors or angiotensin receptor Time to Ischemic Stroke/TE and Death and
blockers for HF). The ABC pathway simplifies and Rehospitalization
streamlines the patient journey, is uniformly appli- Groups Usual Care mAFA
cable across the whole AF patient pathway, starting
with primary care and linking with secondary care CI ¼ confidential interval; HR ¼ hazard ratio; TE ¼ thromboembolism; other abbreviation
(including cardiologists and noncardiologists), and as in Figure 1.
understandable for the patients with AF per se to
1530 Guo et al. JACC VOL. 75, NO. 13, 2020

Mobile Health for Patients With AF APRIL 7, 2020:1523–34

managed with the ABC pathway had a lower risk of

F I G U R E 3 Cumulative Risk of Rehospitalization
mortality, stroke/major bleeding/cardiovascular
death, and hospitalization. In the ATHERO-AF
0.05 mAFA vs. usual care (Atherosclerosis in Atrial Fibrillation) study, Pastori
HR: 0.32 (95% CI: 0.17-0.60) P < 0.001 et al. (13) reported that ABC pathway–adherent man-
agement resulted in a significantly lower rate of car-

0.04 diovascular events (13) as well as lower health

care costs (25). Indeed, adherence with the ABC
Cumulative Hazard Ratio

pathway also had an impact on a nationwide scale,

leading to reduced clinically relevant outcomes in a
0.03 nationwide cohort of 204,842 nonvalvular AF pa-
tients (14).
We found a significant reduction in hospitaliza-
0.02 tions in those allocated to the mAFA intervention.
Comorbid chronic diseases are increased in associa-
tion with increased hospitalization rates for AF (26).
Indeed, hospitalizations for AF itself and related
0.01
complications are the main driver for health care
costs (26), and may be preventable with appropriate
guideline-adherent care delivery to enhance out-
0.00 comes in this population.
High-risk patients with CHA 2DS2-VASc score $3 in
0 100 200 300 400
females and $2 in males using mAFA had a high rate
Time to Rehospitalization
of OAC use (>80%), which generally persisted over
Groups Usual Care mAFA time (Supplemental Figure 5). In various registries
from China, OAC use was substantially lower, being
Rehospitalization included any cause of AF, heart failure, thromboembolism, major 25% to 31% (27–29). In the IMPACT-AF trial including
bleeding, artery coronary disease, and other cardiovascular disease. Abbreviations as in
China, OAC use was increased from 68% at baseline to
Figure 1 and 2.
80% at 1 year, following intervention of an educa-
tional program (24). The high rate of OAC use in the
present study could have been facilitated by the
mAFA app providing clinical decision support, dy-
enable them to engage with their care. Indeed, the namic risk monitoring, and timely communication
ABC pathway covers all main treatment options for between doctors and patients, as well as educational
patients with AF relevant to primary care physicians, and interactive programs. Bleeding risk was low,
secondary care (whether cardiology or noncardiology approximately 2% in both intervention and usual care
clinics), and patient-centered approaches. Patients groups, perhaps related to the app assessing bleeding
could monitor their pulse rhythm with smart devices, risk based on dynamic calculation of HAS-BLED score
assess their symptoms, and communicate to their in a timely manner, thus facilitating optimal man-
doctors in a timely manner, using mAFA, and with agement guided by the doctors (even remotely) using
better symptom management delivered and tailored the mAFA app. This could also be one reason for the
to the patient’s needs and values (Supplemental lower rate of OAC discontinuation among these pa-
Figure 3). Moreover, patients with AF could take tients. Finally, mHealth technology in the present
part in educational programs via different methods, study provided a clinical decision support tool, which
for example, articles, videos, and games, as well as improved guideline adherence for anticoagulant
mAFA providing encouragement for self-care and therapy (30), as well as a clinical management
lifestyle changes, through the Question and Answer pathway with mHealth technology to streamline and
sections (Supplemental Figure 4). simplify clinical practice.
Our results are consistent with prior studies testing
the impact of ABC pathway adherence on clinical STUDY STRENGTHS AND LIMITATIONS. To the best
outcomes (12–14). For example, Proietti et al. (12) of our knowledge, mAFA is the first integrated pro-
conducted a post hoc ancillary analysis of the AFFIRM gram that links AF screening (‘pre-mAFA’) (16) with
(Atrial Fibrillation Follow-up Investigation of Rhythm eligible patients entered into a structured care
Management) trial, which showed patients with AF pathway. Another integrated care trial of chronic
JACC VOL. 75, NO. 13, 2020 Guo et al. 1531
APRIL 7, 2020:1523–34 Mobile Health for Patients With AF

F I G U R E 4 HR of Primary Composite Outcome of Ischemic Stroke/TE, Death, and Rehospitalization, and Secondary Endpoint (Rehospitalization), by Sex, Age,
AF Type, Risk Score, and Comorbidities, Adjusting for Cluster Effect, and Baseline Risk Factors

A Composite Outcome of Ischemic Stroke/TE, Death, and Rehospitalization

HR (95% CI) P Test for Interaction

Sex 0.27
Female 0.48 (0.22-1.04) 0.06
Male 0.34 (0.18-0.67) 0.002
Age 0.004
Age <75 years 0.17 (0.08-0.36) < 0.001
Age ≥75 years 0.63 (0.29-1.38) 0.25
AF type 0.15
Paroxysmal AF 0.49 (0.25-0.94) 0.03
Persistent and permanent AF 0.40 (0.17-0.94) 0.04
CHA2DS2-VASc 0.01
CHA2DS2-VASc ≥2 in males, ≥3 in females 0.57 (0.31-1.03) 0.06
CHA2DS2-VASc 0-1 in males, 1-2 in females 0.04 (0.01-0.27) 0.001
HAS-BLED 0.005
HAS-BLED ≥3 0.86 (0.35-2.16) 0.75
HAS-BLED 0-2 0.21 (0.12-0.37) < 0.001
Hypertension 0.01
Yes 0.52 (0.26-1.03) 0.06
No 0.11 (0.03-0.36) < 0.001
CAD 0.04
Yes 0.53 (0.26-1.11) 0.09
No 0.22 (0.11-0.44) < 0.001

0.0 0.5 1.0 1.5 2.0 2.5

HRs
Favors mAFA Favors Usual Care

B Rehospitalization
HR (95% CI) P Test for Interaction

Sex 0.87
Female 0.27 (0.10-0.72) 0.009
Male 0.31 (0.15-0.64) 0.002
Age 0.08
Age <75 years 0.17 (0.07-0.40) < 0.001
Age ≥75 years 0.46 (0.19-1.12) 0.09
AF type 0.06
Paroxysmal AF 0.43 (0.19-0.94) 0.035
Persistent and permanent AF 0.34 (0.13-0.86) 0.02
CHA2DS2-VASc 0.10
CHA2DS2-VASc ≥2 in males, ≥3 in females 0.41 (0.21-0.80) 0.009
CHA2DS2-VASc 0-1 in males, 1-2 in females 0.07 (0.01-0.55) 0.011
HAS-BLED 0.02
HAS-BLED ≥3 0.78 (0.24-2.56) 0.68
HAS-BLED 0-2 0.18 (0.09-0.38) < 0.001
Hypertension 0.25
Yes 0.33 (0.15-0.75) 0.008
No 0.17 (0.05-0.58) 0.005
CAD 0.03
Yes 0.45 (0.21-1.00) 0.05
No 0.13 (0.04-0.38) < 0.001

0.0 0.5 1.0 1.5 2.0 2.5 3.0

HRs
Favors mAFA Favors Usual Care

(A) Composite outcome of ischemic stroke/TE, death, and rehospitalization. (B) Rehospitalization. CAD ¼ coronary artery disease; CHA2DS2-VASc ¼ congestive heart
failure, hypertension, age $75, diabetes, stroke, vascular disease, age 65–74, and sex category (female); HAS-BLED ¼ hypertension, abnormal renal/liver function,
stroke, bleeding history or predisposition, labile international normalized ratio, elderly, drugs/alcohol concomitantly; mAFa ¼ mobile atrial fibrillation application;
other abbreviations as in Figure 2.
1532 Guo et al. JACC VOL. 75, NO. 13, 2020

Mobile Health for Patients With AF APRIL 7, 2020:1523–34

C E NT R AL IL L U STR AT IO N Mobile Atrial Fibrillation Application

Guo, Y. et al. J Am Coll Cardiol. 2020;75(13):1523–34.

AF ¼ atrial fibrillation; CAD ¼ coronary artery disease; HF ¼ heart failure; DOAC ¼ direct oral anticoagulants; TTR ¼ time in therapeutic range.

cardiovascular diseases (AF, HF, hypertension) using appropriate for analyses of cluster randomized trials.
a mobile tool (phone, tablet) is ongoing, providing There was also no statistical interaction in relation to
vital sign measurements and on-remote communica- the beneficial impact of mAFA intervention on the
tions between patients and physicians, but not pro- composite outcome, even among the elderly ($75
active management options (24). Nevertheless, there years of age), which constituted approximately 32%
were some limitations. Although the present study of the study cohort. The low mean age overall may
was a cluster randomized control trial, the patients in also reflect that mobile health devices are more likely
the mAFA arm were younger than the usual care arm to be used in the younger population in a real-
and compared with the general AF population. Some world setting.
differences in baseline comorbidities were also We are testing a package of care with the mAFA
evident given the cluster randomized trial design, intervention in a holistic approach to AF manage-
rather than individual randomization. Nevertheless, ment, and it was never the intention to investigate
the outcomes were fully adjusted for age and other the individual components of the ABC pathway, nor
comorbidities, using statistical methodology the factors within the A, B, or C components, e.g.,
JACC VOL. 75, NO. 13, 2020 Guo et al. 1533
APRIL 7, 2020:1523–34 Mobile Health for Patients With AF

warfarin or DOAC, BP control or not. Indeed, adher- members include Jiabing Yan, Wenjuan Chen, Qin
ence with therapy (for example, oral anticoagulation) Chen, Jie Zhang, Xi Huang, and Hongbao Li. We are
may reduce over time—but our holistic package of grateful to Tong Xinyuan, from the Statistics Teaching
care includes more than only the ‘A’ (anticoagulation) and Research Office, Medical School of Chinese Peo-
component per se. We did not observe a significant ple’s Liberation Army, Beijing, China, to Liu Miao, from
difference in the individual endpoints of stroke/ the Centre for Epidemiology and Biostatistics, Na-
thromboembolism and major bleeding, which may tional Center for Geriatrics Clinical Research Center,
perhaps be related to the low rate of events during the Beijing, China, for statistical analysis, and for Luo
mean follow-up of 286 days, and some were lost to Yanxia for her contribution to frailty Cox model anal-
follow-up. Further analyses of adherence and long- ysis, who works for Capital Medical University, China.
term outcomes in the mAFA II trial are planned.
Most research centers in present study are Grade 3 ADDRESSES FOR CORRESPONDENCE: Prof. Gregory
level in China (Supplemental Table 1), representing Y.H. Lip, Liverpool Centre for Cardiovascular Science,
the top medical level hospitals. The finding of the William Henry Duncan Building, 6 West Derby St,
present study may not necessarily be generalized to Liverpool, Merseyside L7 8TX, United Kingdom.
Grade 2 level hospitals or Community Service Cen- E-mail: [email protected]. OR Prof. Yundai
ters, which will be explored in future studies. Finally, Chen, Chinese PLA General Hospital, Department of
this trial was conducted in China, and further studies Cardiology, No. 28, Fuxin Rd, Beijing 100853, China.
in non-Asian cohorts from different health care sys- E-mail: [email protected].
tems are in progress or planned.

PERSPECTIVES
CONCLUSIONS

An integrated care approach to holistic AF care, sup- COMPETENCY IN PATIENT CARE AND PROCEDURAL
ported by mHealth technology, significantly reduced SKILLS: Application of a mHealth technology to facilitate inte-
clinical adverse outcomes in patients with AF. grated care for patients with AF by integrating clinical decision
Implementation of the ABC pathway using an app- support, risk monitoring, education, and patient self-
based mHealth approach may improve clinical out- management can reduce hospitalizations and clinical adverse
comes in patients with AF. events.
ACKNOWLEDGMENTS The authors thank the Huawei
TRANSLATIONAL OUTLOOK: Further studies are needed to
Heart Health Research Team for the
optimize the mobile technology algorithms and assess their
photoplethysmography-based pulse rhythm moni-
utility across a broader range of patient populations and care
toring for Better symptom management with patient-
settings.
centered, symptom-directed decisions on rate or
rhythm control, headed by Mr. Xiaoxiang He. Team

REFERENCES

1. January CT, Wann LS, Calkins H, et al. 2019 management: the 6th AFNET/EHRA Consensus 8. Ahmed I, Ahmad NS, Ali S, et al. Medication
AHA/ACC/HRS Focused Update of the 2014 AHA/ Conference. Europace 2018;20:395–407. adherence apps: review and content analysis. JMIR
ACC/HRS Guideline for the Management of Pa- Mhealth Uhealth 2018;6:e62.
5. McConnell MV, Turakhia MP, Harrington RA,
tients With Atrial Fibrillation: A Report of the
King AC, Ashley EA. Mobile health advances in 9. Guo Y, Chen Y, Lane DA, Liu L, Wang Y, Lip GYH.
American College of Cardiology/American Heart
physical activity, fitness, and atrial fibrillation: Mobile health technology for atrial fibrillation
Association Task Force on Clinical Practice Guide-
moving hearts. J Am Coll Cardiol 2018;71:2691–701. management integrating decision support, edu-
lines and the Heart Rhythm Society. J Am Coll
cation, and patient involvement: mAF App Trial.
Cardiol 2019;74:104–32. 6. World Health Organisation. Geneva: World
Am J Med 2017;130:1388–96 e6.
Health Organisation; 2019 Apr 17. WHO Releases
2. Miyazawa K, Ogawa H, Mazurek M, et al. 10. GYH L. The ABC pathway: an integrated
First Guideline on Digital Health Interventions.
Guideline-adherent treatment for stroke and approach to improve AF management. Nature re-
Available at: https://www.who.int/news-room/
death in atrial fibrillation patients from UK and views. Cardiology 2017;14:2.
detail/17-04-2019-who-releases-first-guideline-
Japanese AF registries. Circ J 2019;83:2434–42.
on-digital-health-interventions? Accessed 11. Lip GYH, Banerjee A, Boriani G, et al. Antith-
3. Lane DA, Meyerhoff J, Rohner U, Lip GYH. Atrial November 19, 2019. rombotic therapy for atrial fibrillation: CHEST
fibrillation patient preferences for oral anti- guideline and expert panel report. Chest 2018;154:
7. Zeballos-Palacios CL, Hargraves IG,
coagulation and stroke knowledge: results of a 1121–201.
Noseworthy PA, et al. Developing a conversation
conjoint analysis. Clin Cardiol 2018;41:855–61.
aid to support shared decision making: reflections 12. Proietti M, Romiti GF, Olshansky B, Lane DA,
4. Kotecha D, Breithardt G, Camm AJ, et al. Inte- on designing anticoagulation choice. Mayo Clin Lip GYH. Improved outcomes by integrated care of
grating new approaches to atrial fibrillation Proc 2019;94:686–96. anticoagulated patients with atrial fibrillation
1534 Guo et al. JACC VOL. 75, NO. 13, 2020

Mobile Health for Patients With AF APRIL 7, 2020:1523–34

using the simple ABC (atrial fibrillation better 19. Eldridge SM, Ashby D, Kerry S. Sample size for 26. Patel NJ, Deshmukh A, Pant S, et al.
care) pathway. Am J Med 2018;131:1359–66.e6. cluster randomized trials: effect of coefficient of Contemporary trends of hospitalization for atrial
variation of cluster size and analysis method. Int J fibrillation in the United States, 2000 through
13. Pastori D, Pignatelli P, Menichelli D, Violi F,
Epidemiol 2006;35:1292–300. 2010: implications for healthcare planning. Circu-
Lip GYH. Integrated care management of patients
lation 2014;129:2371–9.
with atrial fibrillation and risk of cardiovascular 20. Wagner EH. Organizing care for patients with
events: the ABC (atrial fibrillation better care) chronic illness revisited. Milbank Q 2019;97: 27. Liu C, Du X, Jiang C, et al. Long-term persistence
pathway in the ATHERO-AF study cohort. Mayo 659–64. with newly-initiated warfarin or non-VKA oral anti-
Clin Proc 2019;94:1261–7. coagulant (NOAC) in patients with non-valvular
21. Hendriks JM, de Wit R, Crijns HJ, et al. Nurse-
atrial fibrillation: insights from the Prospective
14. Yoon M, Yang PS, Jang E, et al. Improved led care vs. usual care for patients with atrial
China-AF Registry. Med Sci Monit 2019;25:2649–57.
population-based clinical outcomes of patients fibrillation: results of a randomized trial of inte-
with atrial fibrillation by compliance with the simple grated chronic care vs. routine clinical care in 28. Liang HF, Du X, Zhou YC, et al. Control of
ABC (atrial fibrillation better care) pathway for in- ambulatory patients with atrial fibrillation. Eur anticoagulation therapy in patients with atrial
tegrated care management: a nationwide cohort Heart J 2012;33:2692–9. fibrillation treated with warfarin: a study from the
study. Thromb Haemost 2019;19:1695–703. Chinese Atrial Fibrillation Registry. Med Sci Monit
22. Stewart S, Ball J, Horowitz JD, et al. Standard
2019;25:4691–8.
15. Guo Y, Lane DA, Wang L, Chen Y, Lip GYH. versus atrial fibrillation-specific management
Mobile health (mHealth) technology for improved strategy (SAFETY) to reduce recurrent admission 29. Steinberg BA, Gao H, Shrader P, et al. Inter-
screening, patient involvement and optimising and prolong survival: pragmatic, multicentre, national trends in clinical characteristics and oral
integrated care in atrial fibrillation: the mAFA randomised controlled trial. Lancet 2015;385: anticoagulation treatment for patients with atrial
(mAF-App) II randomised trial. Int J Clin Pract 775–84. fibrillation: results from the GARFIELD-AF, ORBIT-
2019;73:e13352. AF I, and ORBIT-AF II registries. Am Heart J 2017;
23. van den Dries CJ, Oudega R, Elvan A, et al.
16. Guo Y, Wang H, Zhang H, et al. Mobile 194:132–40.
Integrated management of atrial fibrillation
photoplethysmographic technology to detect including tailoring of anticoagulation in primary 30. Schmidt C, Oner A, Mann M, et al. A novel
atrial fibrillation. J Am Coll Cardiol 2019;74: care: study design of the ALL-IN cluster rando- integrated care concept (NICC) versus standard
2365–75. mised trial. BMJ Open 2017;7:e015510. care in the treatment of chronic cardiovascular
17. Chang TY, Lip GYH, Chen SA, Chao TF. impor- diseases: protocol for the randomized controlled
24. Vinereanu D, Lopes RD, Bahit MC, et al.
tance of risk reassessment in patients with atrial trial CardioCare MV. Trials 2018;19:120.
A multifaceted intervention to improve treatment
fibrillation in guidelines: assessing risk as a dy- with oral anticoagulants in atrial fibrillation
namic process. Can J Cardiol 2019;35:611–8. (IMPACT-AF): an international, cluster- KEY WORDS adverse events, atrial
randomised trial. Lancet 2017;390:1737–46. fibrillation, integrated care, mobile health
18. Steffel J, Verhamme P, Potpara TS, et al. The
2018 European Heart Rhythm Association Prac- 25. Pastori D, Farcomeni A, Pignatelli P, Violi F,
tical Guide on the use of non-vitamin K antagonist Lip GY. ABC (atrial fibrillation better care) pathway A PPE NDI X For supplemental investigators,
oral anticoagulants in patients with atrial fibrilla- and healthcare costs in atrial fibrillation: the results, figures, and tables, please see the on-
tion. Eur Heart J 2018;39:1330–93. ATHERO-AF study. Am J Med 2019;132:856–61. line version of this paper.