Caregiver-Delivered Stroke Rehabilitation in Rural China

The RECOVER Randomized Controlled Trial

Bo Zhou, PhD*; Jing Zhang, MSc*; Yi Zhao, PhD; Xian Li, PhD; Craig S. Anderson, PhD;
Bin Xie, MD; Ninghua Wang, MD; Yuhong Zhang, PhD; Xiaojun Tang, MSc;
Janet Prvu Bettger, PhD; Shu Chen, MSc; Wanbing Gu, MSc; Rong Luo, MSc;
Qiongrui Zhao, MSc; Xiaoxia Li, MSc; Zhenxing Sun, MSc; Richard I. Lindley, MD;
Sarah E. Lamb, PhD; Yangfeng Wu, PhD; Jingpu Shi, PhD; Lijing L. Yan, PhD

Background and Purpose—Stroke disability is a major health burden in rural China where rehabilitation services are
inadequate. We aimed to determine the effectiveness of a novel nurse-led, caregiver-delivered model of stroke rehabilitation
in rural China.
Methods—A multicenter prospective, randomized open, blinded outcome assessed, controlled trial was conducted in 3 rural
county hospitals in China: Zhangwu, Liaoning Province (Northeast); Qingtongxia, Ningxia Hui Autonomous Region
(Northwest); and Dianjiang, Chongqing Municipality (Southwest). Adult patients (age 18–79 years) with residual
disability (Barthel Index score ≤80/100) after a recent acute stroke were randomized to a new service model or usual
care. The new intervention was multifaceted and was based on a task-shifting / training-the-trainers model, supported by
a custom-designed smartphone application, where patients and caregivers received evidence-based in-hospital education
and stroke rehabilitation training (focus on mobility, self-care, and toileting), delivered by trained nurses before hospital
discharge, and 3 postdischarge support telephone calls. Outcome assessments were undertaken before hospital discharge
and at 3 and 6 months. Primary outcome was physical functioning (Barthel Index scores) at 6 months, assessed by
research staff blind to treatment allocation, adjusted for baseline covariates in an intention-to-treat analysis. Secondary
outcomes included measures of mobility, health-related quality of life, mood, and caregiver burden. The study included a
process evaluation that assessed intervention fidelity.
Results—From November 2014 to December 2016, 246 stroke patients were randomized to intervention (n=118) or control
Downloaded from http://ahajournals.org by on June 21, 2023

(n=128) groups. There was no statistically significant difference in adjusted 6-month Barthel Index scores between groups
(70.1 versus 74.1, mean difference, −4.0 [95% CI, −10.0 to 2.9]), nor any differences across the other outcome measures.
Process evaluation interviews revealed that the intervention was desirable and positively accepted by nurses, caregivers,
and patients but was considered too complex despite efforts to simplify materials for the rural context. Key strategies
identified for future studies included the use of community health workers, smartphone application enhancement, and
simpler and more frequent training for nurses, caregivers, and patients.
Conclusions—A novel nurse-led, digital supported, caregiver-delivered stroke rehabilitation program did not improve
patient physical functioning after stroke in rural China. Further stroke rehabilitation research suitable for resource-poor
settings is required, with several components being suggested through stakeholder interviews in our study.
Clinical Trial Registration—URL: https://www.clinicaltrials.gov. Unique identifier: NCT02247921.    (Stroke.
2019;50:1825-1830. DOI: 10.1161/STROKEAHA.118.021558.)
Key Words: caregivers ◼ China ◼ nurses ◼ rehabilitation ◼ stroke

Received March 23, 2018; final revision received April 7, 2019; accepted April 15, 2019.
From the Department of Clinical Epidemiology, First Hospital of China Medical University, Shenyang (B.Z., Q.Z., J.S.); The George Institute
for Global Health at Peking University Health Science Center, Beijing, China (J.Z., Xian Li, C.S.A., R.L., Y.W., L.L.Y.); School of Public Health
and Management, Ningxia Medical University, Yinchuan, China (Y. Zhao, Y. Zhang, Xiaoxia Li); The George Institute for Global Health, Faculty
of Medicine, UNSW, Sydney, Australia (Xian Li, C.S.A.); Department of Rehabilitation Medicine, Peking University First Hospital, Beijing, China
(B.X., N.W.); School of Public Health and Management, Chongqing Medical University, China (X.T., Z.S.); Duke Clinical Research Institutea (J.P.B.)
and Duke Global Health Institute (J.P.B., L.L.Y.), Duke University, Durham, NC; Global Health Research Center, Duke Kunshan University, China
(S.C., W.G., L.L.Y.); Westmead Clinical School, Faculty of Medicine and Health, University of Sydney, Australia (R.I.L.); Oxford Clinical Trials
Research Unit, University of Oxford, United Kingdom (S.E.L.); Peking University Clinical Research Institute, Beijing, China (Y.W.); and School of
Public Health, Peking University Health Science Center, Beijing, China (Y.W.).
*Drs Zhou and J. Zhang are cofirst authors.
The online-only Data Supplement is available with this article at https://www.ahajournals.org/doi/suppl/10.1161/STROKEAHA.118.021558.
Correspondence to Jingpu Shi, PhD, First Hospital of China Medical University, No. 115 Nanjing St, Heping District, Shenyang, Liaoning, China, Email
sjp562013@163.com or Lijing L. Yan, PhD, Global Health Research Center, Duke Kunshan University, No. 8 Duke Ave, Kunshan, Jiangsu Province,
China, Email lijing.yan@dukekunshan.edu.cn
© 2019 The Authors. Stroke is published on behalf of the American Heart Association, Inc., by Wolters Kluwer Health, Inc. This is an open access article
under the terms of the Creative Commons Attribution Non-Commercial License, which permits use, distribution, and reproduction in any medium, provided
that the original work is properly cited and is not used for commercial purposes.
Stroke is available at https://www.ahajournals.org/journal/str DOI: 10.1161/STROKEAHA.118.021558

1825
 1826  Stroke  July 2019
S troke is a major health burden in China, particularly in
rural regions where there are high rates and poor access
to services and optimal care. The economic burden of stroke
is also high, with the average cost of hospital care being esti-
mated at ¥10 489 (US $1311), which is about 92% of annual
median household income for rural families in the absence
of insurance.1 Despite recognition of high levels of disability
among survivors,2 rehabilitation services are limited in many
parts of China, especially in rural regions, where there are
few health professionals specifically trained in recognition
and management of stroke-related complications.3,4 As family
members have prime responsibility for the care of disabled
patients at home in rural China,5 it seems logical to focus the
provision of rehabilitative care to them.
Previous studies on family-based stroke intervention in
countries outside of China have had inconsistent results.6–8
We designed an innovative rehabilitation based on Western
approaches to service delivery involving task-shifting and
training-the-trainer model, where positive results have been
shown for various conditions.9 The task-shifting model, in par-
ticular, seems most suitable in low-resource contexts where
there are few health professionals;10 whereas the training-the-
trainer model is based on a knowledge, skills, and agency
loop: moving from specialist to nurse, nurse to caregiver, and
caregiver to patient. We aimed to determine the effective-
ness of the program to improve the basic self-care (activities
of daily living [ADL]) physical function of patients disabled
after a recent stroke in a multicenter randomized controlled
trial, the RECOVER (Rehabilitation Through Caregiver-
Downloaded from http://ahajournals.org by on June 21, 2023
Delivered Nurse-Organized Service Programs for Disabled
Stroke Patients in Rural China).
Methods
Data from this study are available from the corresponding author on
reasonable request.
Design
Full details of the study protocol for the RECOVER trial have been
published.11 In brief, the study was conducted at 3 county hospitals
in rural China: Zhangwu, Liaoning Province (Northeastern China);
Qingtongxia, Ningxia Hui People Autonomous Region (Northwestern
China); and Dianjiang, Chongqing municipality (Southwestern
China). Hospitalized patients with a clinical diagnosis of acute stroke
(ischemic, hemorrhagic, or undifferentiated) considered suitable for
rehabilitation by their attending physician were eligible for the study
if they were aged 18 to 79 years, required assistance with ADL (score
≤80/100 on the Barthel Index [BI]12) but without severe disability,
were free of serious comorbid diseases and with an estimated life ex-
pectancy >6 months, and able to provide consent. Details of sociode-
mographic characteristics, medical history, and ADL were collected
at baseline. Patients were randomized via a secure, central internet-
based system, with randomly variable block sizes (4, 6, and 8) strati-
fied by ADL status (BI scores ≤40 versus 40–80) either to intervention
(nurse-organized, family caregiver delivered) rehabilitation or control
(usual) care. An integrated Android-based application was installed
on nurses’ smartphones (Red Mi Note 4G, Xiao Mi, Inc) to allow
integrated data collection, randomization, quality control, and oper-
ational guide, with data security ensured by hash algorithm (Secure
Hash Algorithm 512 [SHA512]) and secure socket layer (Hypertext
Transfer Protocol Secure [HTTPS]). The study was approved by the
ethics committees of the China Medical University, Ningxia Medical
University, Chongqing Medical University, and Duke University, and
all patients provided informed consent.
Intervention
Up to 3 nurses from each hospital received training from accred-
ited rehabilitation physicians and therapists at Peking University
First Hospital over 2 days, with a menu-style checklist used to guide
their training and techniques, and case-based methods and teach-
back approaches used to assess their skills. They also received a fur-
ther day of training on returning to their respective hospitals. The
nurses screened for eligible patients and family caregivers, and those
randomized to the intervention group were provided with skills and
guidance, using a culturally appropriate, picture-rich, easy-to-under-
stand manual and several 15- to 30-minute training sessions. They
were also given a videodisk of their training and a “daily reminder”
guide to help motivate patients to complete a routine of daily support
exercises at home (further details of the TIDieR13 template [Template
for Intervention Description and Replication] are provided in Table I
in the online-only Data Supplement). The nurses also offered patients
and caregivers follow-up checks on their status and problem-solving
recommendations through 3 telephone calls scheduled at 2, 4, and
8 weeks after discharge. Patients in the control group received con-
ventional care without any additional recommendations or activities.
The rehabilitation intervention was based on the Motor Relearning
Theory,9 which emphasized task and environmental specific patient
and caregiver training that focused on 3 recovery goals: (1) mobility
(transfer, mobility, and stairs) tailored towards a patient’s level of
function, covering in-bed movements, sitting and standing balance,
and walking; (2) self-care (grooming, feeding, dressing, and bathing);
and (3) continence/toileting (bowel, bladder, and toilet use).14,15 The
training also included information on the usual patterns of poststroke
recovery, importance of cardiovascular risk factor control, awareness
and implications of low mood, importance of task practice, and in the
use of a tailored discharge plan, established jointly with patients and
caregivers, that covered their recovery goals.
Outcomes
The primary outcome was ADL defined by scores on the BI at 6
months postrandomization, assessed in-person by trained research
nurses who were blind to the treatment assignment. Secondary
outcomes included BI scores at predischarge and 3-month postran-
domization; modified Rankin Scale scores16 at the 3 and 6 months;
scores on the Functional Ambulation Category,17 Patient Health
Questionnaire 9,18 and EuroQol-5-dimension19 at 6 months; and a
measure of Caregiver Burden Index20 at 6 months. Other information
collected included hospital length of stay and medical costs obtained
from questionnaires completed by participants and all adverse events
including rehospitalization and death, with details recorded by nurses
reported to the study coordinating center.
Process Evaluation
A process evaluation assessing fidelity to the intervention, involved
standardized, semistructured interviews (20–40 minutes) with nurses
who delivered the rehabilitation model of care and 2 patient/caregiver
pairs at each site, was undertaken during site monitoring visits at 3
and 9 months. Participatory observation of the intervention and focus
groups with nurses, which lasted ≈60 minutes, were also conducted.
All interviews were voice-recorded with consent, and the transcripts
analyzed by independent researchers using a thematic framework21
for coding and descriptions of opinions.
Statistical Analysis
Based on studies indicating considerable variability in BI scores
among stroke survivors, with SD ranging 10 to 3022,23 and scores for
improvement of 10 to 35 points,24 we estimated that a sample size
of 86 patients per group would provide 90% power (2-sided α of
0.05) to detect a 10-point difference in BI scores between groups.
Assuming a SD of 20, this between-group difference corresponds to
a standardized effect size of 0.50. However, taking account of 10%
missing primary outcome, the target sample was 200 patients (100
per group).
 Zhou et al   Stroke Rehabilitation in Rural China   1827

Table 1. Demographic Characteristics of Subjects and Their Caregivers in

Intention-to-Treat Dataset

Intervention Control Total

N=116 N=128 N=244
Patients
 Age, yr 64.3 (9.2) 66.2 (9.4) 65.4 (9.2)
 Male 54 (46.6) 56 (43.8) 110 (45.1)
 Years of schooling
  
Illiterate 32 (27.6) 41 (32.0) 73 (29.9)
  
Primary school 44 (37.9) 39 (30.5) 83 (34.0)
  
Middle school 32 (27.6) 36 (28.1) 68 (27.9)
  Senior high school 8 (6.9) 12 (9.4) 20 (8.2)
 First-ever stroke 93 (81.6) 105 (82.0) 198 (81.1)
 Ischemic stroke 85 (73.3) 104 (81.3) 189 (77.5)
 Household income, Chinese Yuan (CNY)
  
<5,000 29 (25.0) 22 (17.2) 51 (20.9)
Figure 1. Flow of patients through rehabilitation services programs in rural
  
5,000–10,000 29 (25.0) 31 (24.2) 60 (24.6)
China. BI indicates Barthel Index; and ITT, intention-to-treat.
  
10,000–20,000 19 (16.4) 19 (14.8) 38 (15.6)
The primary analysis was based on a modified intention-to-   
20,000–30,000 22 (19.0) 30 (23.4) 52 (21.3)
treat population that included all randomized patients who com-
pleted at least one BI assessment; a complete data analysis was   
>30,000 17 (14.6) 26 (20.4) 43 (17.6)
used in those with a primary outcome at 6 months. Multiple linear
 Smoker, n (%) 38 (32.8) 36 (28.1) 74 (30.1)
regression was used to adjust for baseline imbalances, and mixed
models were used to account for repeated measures and missing  Hypertension, n (%) 85 (73.3) 97 (75.8) 182 (74.4)
values. Predefined subgroups included age, sex, stroke type and
sequence (first versus recurrent), severity of 6-month BI score, and  Diabetes mellitus, n (%) 15 (12.9) 17 (13.3) 32 (13.0)
Downloaded from http://ahajournals.org by on June 21, 2023

by site. All analysis was conducted using SAS 9.4 software (SAS  Cardiovascular disease, n (%) 27 (23.3) 26 (20.3) 53 (21.7)
Institute, Inc, Cary, NC).
Caregivers

Results  Age, yr 51.4 (12.3) 52.3 (12.8) 51.9 (12.5)

Between November 2014 and December 2016, 447 acute  Sex, male 57 (49.1) 63 (48.4) 120 (49.2)
stroke patients were screened for inclusion, but 168 were  Years of schooling
excluded for clinical reasons and 4 for other reasons, and
  
Illiterate 26 (22.4) 20 (15.6) 46 (18.9)
29 refused participation (Figure 1). Overall, 246 patients
were randomized into intervention (118) and control (128)   
Primary school 35 (30.2) 40 (31.2) 75 (30.7)
groups, but 2 patients died before receiving the intervention   
Middle school 36 (31.0) 41 (32.0) 77 (31.6)
(Figure 1).
  Senior high school 19 (16.4) 27 (21.1) 46 (18.8)
Table 1 outlines the baseline characteristics of participat-
Data are n (%) or mean (SD).
ing patients and their caregivers. Most patients (mean age 65.5
years; 45.1% male) had suffered their first (81.2%) primarily
Table 3 provides details surrounding the training associ-
ischemic (77.2%) stroke. Caregivers (mean age 52.1 years)
ated with the intervention. Training was completed by most
had either primary school (30.3%) or middle school (31.5%)
patient/caregivers (92%) over 2 or 3 sessions, for a duration
education.
Table 2 shows that at 6 months, there was no signifi- of >1 hour (91%), and most (99%) received 3 postdischarge
cant between-group difference in BI scores (70.1±25.5 and follow-up telephone calls.
74.1±23.0, mean difference, −4.0 [95% CI, −10.0 to 2.9]; Table II in the online-only Data Supplement provides a
crude P=0.23 and adjusted P=0.27). A mixed model indi- summary of the key themes identified from the in-depth indi-
cates that BI scores similarly increased in both groups dur- vidual interviews with 7 nurses and 6 patient/caregiver pairs
ing follow-up (baseline, before discharge, and at 3 and 6 from participating hospitals, indicating not only high accept-
months), but did not significantly differ between the inter- ability of the intervention but also certain challenges, such as
vention and control groups (Figure 2). Secondary outcomes extra workload and complexity of the training.
were also similar between groups, except for Functional
Ambulation Category scores being higher in the usual care Discussion
group (P=0.04) after adjustment confounders. The neutral Our novel nurse-led, caregiver-delivered, rehabilitation model
treatment effect on the primary outcome was consistent of care did not improve physical recovery, according to basic
across subgroups (Figure 3). ADL measured by the BI, over usual care for patients recently
 1828  Stroke  July 2019
Table 2. Effectiveness of Intervention at 6-Month Target Date After
Rehabilitation Services Based on Intention-to-Treat Analysis (Multiple Linear
Regression)
Outcomes
Primary
 BI (N=233) at 6 mo
Secondary
 FAC (N=233) at 6 mo
 mRS scores 0–1
(N=233) at 6 mo†
 PHQ-9 (N=230) at 6 mo
 EQ-5D (N=226) at 6 mo
Exploratory
 CBI (N=231) at 6 mo
 Total expenses in
hospital (N=238), CNY
 Average length of
hospital stay, d‡
(N=238), median (IQR)
Data are n (%), mean (SD) or median (IQR). BI indicates Barthel Index; CBI,
Caregiver Burden Index; CNY, Chinese Yuan; CVD, cardiovascular disease; EQ-
5D, EuroQol-5-dimension; FAC, Functional Ambulation Category; mRS, modified
Rankin Scale; and PHQ-9 Patient Health Questionnaire 9.
*Adjusting different hospitals, baseline BI, age, sex, education level, stroke
type, no. of previous strokes, hypertension, CVD, and diabetes mellitus.
†χ2 test.
‡Wilcoxon test, median (Q1, Q3).
Downloaded from http://ahajournals.org by on June 21, 2023
disabled from acute stroke in rural China. Although there is
evidence of effectiveness of task-shifting and training-the-
trainers in other contexts, the use of this approach to stroke re-
habilitation in low-resource settings cannot be recommended,
despite being coveted and enthusiastically accepted by hos-
pital executives, nurses, family caregivers, patients, and other
stakeholders.
Our results are consistent with other recently completed
trials, including one that used a similar task-shifting family-
delivered stroke rehabilitation intervention in India,6 and an-
other cluster clinical trial of training caregivers in the United
Kingdom.7 Moreover, a meta-analysis8 of several trials involv-
ing 333 patients showed that caregiver-mediated rehabilitation
did not improve ADL function after stroke. Our trial differed
from these studies in several aspects but most notably in
regards to the approach to training and support which was free
Intervention Control P Value* of direct input from rehabilitation physicians and therapists;
rather, nurses were trained to deliver rehabilitation to patients/
caregivers in hospital and at home.
70.1 (25.5) 74.1 (23.0) 0.27
Our imbedded process evaluation showed that the stroke
rehabilitation program was well implemented by the nurses
4.6 (1.7) 5.0 (1.4) 0.04 (dose delivered), and nearly all caregivers and patients received
79 (32.1) 88 (35.8) 0.21 training in hospital and at home (dose received). Even so, the
qualitative interviews disclosed several negative responses,
highlighting challenges in delivering this model of care which
5.5 (5.5) 5.0 (4.9) 0.44
affected fidelity of the intervention. Despite our efforts to en-
0.7 (0.3) 0.8 (0.3) 0.15 sure that the training could be integrated into routine care, the
trainer nurses expressed concern over an increase in workload
73.0 (23.1) 74.5 (19.6) 0.56 and that their other duties assumed priority over those associated
with the intervention. Although increased in-hospital rehabilita-
2861.4 (4205.7) 2765.8 (3780.9) 0.84
tion may assist patients, it may not be feasible in a task-shifting
model without freeing up nursing time from other tasks. Another
14 (11–17) 14 (10–16) 0.35 concern of nurses was inconsistency in the delivery of care by
caregivers, whereas patients expressed reluctance to exercise
without appropriate supervision, which led to the level of partici-
pation of patients in exercises fell off upon their discharge home.
Thus, future task-shifting models of care need to ensure there
are sufficient human resources (ie, nurses and caregivers), with
the potential to integrate a trained community health worker, to
facilitate the development of skills by patients at home.
Another potential limitation of our study was that the nurses
were insufficiently skilled to deliver an appropriate standard of
intensity of training to caregivers and patients to affect a signif-
icant change. The multidisciplinary intervention design team
aimed to develop contextually appropriate training materials
and innovative easy-to-use tools for training. Nevertheless,
many nurses and caregivers/patients expressed difficulties in
acquiring certain rehabilitation skills, raising the issue as to
whether the nurses would have benefited from a longer pe-
riod of training and for caregivers/patients to have had greater
frequency of engagement. Conversely, the rehabilitation tools
and training materials might have been too complex. Greater
thought is required to develop task-shifting or training-the-
trainer interventions that balance simplicity with effectiveness.
The RECOVER trial indicates that rigorous evalua-
tions of rehabilitation services are possible in low-resource
Figure 2. Barthel Index (BI) as measured at
baseline, predischarge, and 3 and 6 mo post-
randomization, in the modified intention-to-treat
population with adjustment for hospital, base-
line BI, age, sex, education level, stroke type,
number of previous strokes and chronic disease
history (hypertension, cardiovascular, and dia-
betes mellitus).

Figure 3. Average Barthel Index (BI) scores for intervention and control groups with stroke in rural China by predefined subgroups.
Downloaded from http://ahajournals.org by on June 21, 2023
settings, in particular for task-shifting and training-the-train-
ers approaches being successfully adapted to rural China. The
study team’s evidence-based rehabilitation training with pic-
ture-rich manuals for both nurses and patients/family caregiv-
ers was acceptable to all parties and achieved clear, practical
benefits across the intervention program. In this context, sev-
eral important future research directions were defined: training
time and intensity reconfigurations, training manual develop-
ment, community healthcare worker uptake, and smartphone
application enhancement.
In summary, our evaluation of task-shifting stroke reha-
bilitation by nurses who trained family caregivers to provide
Table 3. Adherence to Intervention Protocol in the Intervention Group
Zhangwu
County
N=31
No. of completed training sessions in hospital
 ≥3× 26 (83.9)
Total duration of training session in hospital
 ≥1 h, n (%) 27 (81.7)
 Training time in 2.3 (1.2)
hospital, h
No. of telephonic follow-up calls postdischarge
 3× 30 (99.9)
Data are n (%), mean (SD).
Zhou et al   Stroke Rehabilitation in Rural China   1829
care, though desired and well-accepted by stakeholders, did
not produce significant improvements over usual care in the
recovery of ADL among participants in rural China. In this
foundational study, future directions for research have been
highlighted that include the utilization of community health-
care workers, enhancement of smartphone applications for
communication, education and data collection, and review
of the appropriate format and duration of training. More
efforts and different models are required to develop effective
approaches to improve recovery from stroke in low-resource
setting of China and other countries.‍‍
Acknowledgments
R. Luo and J. Zhang searched the literature; Drs Xie, Wang, Y. Zhang,
Prvu Bettger, Lindley, Anderson, Lamb, Wu, Shi, and Yan were re-
Qingtongxia Dianjing sponsible for study design; Dr Xie and Wang developed the inter-
County County Total vention protocol; Q. Zhao, X. Li, and Z. Sun collected data; XL, J.
N=54 N=29 N=114 Zhang, and Dr Yan analyzed and interpreted the data; Drs Zhou, Y.
Zhang, Shi, YZ, and X. Tang were responsible for project implemen-
tation; J. Zhang, Drs Zhou, YZ, Chen, and W. Gu undertook project
management; Q. Zhao, X. Li, Z. Sun, J. Zhang, Drs Zhou, YZ, X.
47 (87.0) 26 (89.7) 99 (86.8)
Tang, and Yan were responsible for quality control; and J. Zhang and
Dr Zhou wrote the first draft of the article. All authors made crit-
ical revisions of the article. We thank all directors, nurses, patients,
51 (94.4) 27 (93.1) 105 (92.1) and caregivers of participating hospitals for their cooperation in the
1.7 (0.7) 1.9 (0.7) 1.9 (0.9) study and Yu Liu of Beihang University for developing the smart-
phone application.
Sources of Funding
53 (99.9) 29 (100.0) 112 (99.9) This study had principle funding from the China Medical Board
(United States) and Amsterdam Health and Technology Institute,
 1830  Stroke  July 2019
the Netherlands and additional funding from The George Institute
China and Duke Kunshan University. None of these organizations
had any role in writing of the article or the decision to submit it for
publication. The corresponding authors had full access to all the data
in the study and final responsibility for the decision to submit for
publication.
Disclosures
There is no support from any organizations for the submitted work.
No financial relationships with any organizations that might have an
interest in the submitted work in the previous 3 years. Dr Anderson
reports receiving Advisory Board fees from Amgen and Boehringer
Ingelheim and grants and speaker fees and travel reimbursement from
Takeda. No other relationships or activities that could appear to have
influenced the submitted work. The other authors report no conflicts.
References
1. Yang L, Wu M, Cui B, Xu J. Economic burden of cardiovascular diseases
in China. Expert Rev Pharmacoecon Outcomes Res. 2008;8:349–356.
doi: 10.1586/14737167.8.4.349
2. Yang Y, Shi YZ, Zhang N, Wang S, Ungvari GS, Ng CH, et al.
The disability rate of 5-year post-stroke and its correlation fac-
tors: a national survey in China. PLoS One. 2016;11:e0165341. doi:
10.1371/journal.pone.0165341
3. Jianjun Y, Yongshan H, Wu Y, Wenhua C, Yulian Z, Xiao C, et al. The
effects of community-based rehabilitation on stroke patients in china:
a single-blind, randomized controlled multicentre trial. Clin Rehabil.
2009;23:408–417
4. Xianghong Meng WW, Hong Dai. Investigation on current status of re-
habilitation of stroke at Chongwen district in Beijing. Chin J Rehabil
Med. 2009;24:4.
5. Sit JW, Wong TK, Clinton M, Li LS, Fong YM. Stroke care in the home:
the impact of social support on the general health of family caregivers.
J Clin Nurs. 2004;13:816–824. doi: 10.1111/j.1365-2702.2004.00943.x
Downloaded from http://ahajournals.org by on June 21, 2023
6. ATTEND Collaborative Group. Family-led rehabilitation after stroke in
India (ATTEND): a randomised controlled trial. Lancet. 2017;390:588–
599. doi: 10.1016/S0140-6736(17)31447-2
7. Forster A, Dickerson J, Young J, Patel A, Kalra L, Nixon J, et al; TRACS
Trial Collaboration. A cluster randomised controlled trial and economic
evaluation of a structured training programme for caregivers of inpatients
after stroke: the TRACS trial. Health Technol Assess. 2013;17:1–216.
doi: 10.3310/hta17460
8. Vloothuis JD, Mulder M, Veerbeek JM, Konijnenbelt M, Visser-Meily JM,
Ket JC, et al. Caregiver-mediated exercises for improving outcomes
after stroke. Cochrane Database Syst Rev. 2016;12:CD011058. doi:
10.1002/14651858.CD011058.pub2
9. Gustavsen M, Jansen R, Kjendahl A, Lorentzen A. Motor Relearning
Program approach improves short-term motor outcomes and reduces
hospital stay after stroke. Aust J Physiother. 2002;48:59.
10. Ogedegbe G, Gyamfi J, Plange-Rhule J, Surkis A, Rosenthal DM,
Airhihenbuwa C, et al. Task shifting interventions for cardiovascular risk
reduction in low-income and middle-income countries: a systematic re-
view of randomised controlled trials. BMJ Open. 2014;4:e005983. doi:
10.1136/bmjopen-2014-005983
11. Yan LL, Chen S, Zhou B, Zhang J, Xie B, Luo R, et al. A randomized
controlled trial on rehabilitation through caregiver-delivered nurse-orga-
nized service programs for disabled stroke patients in rural china (the
RECOVER trial): design and rationale. Int J Stroke. 2016;11:823–830.
doi: 10.1177/1747493016654290
12. Mahoney FI, Barthel DW. Functional evaluation: the barthel index. Md
State Med J. 1965;14:61–65.
13. Hoffmann TC, Glasziou PP, Boutron I, Milne R, Perera R, Moher D, et
al. Better reporting of interventions: template for intervention description
and replication (TIDieR) checklist and guide. BMJ. 2014;348:g1687.
doi: 10.1136/bmj.g1687
14. Sulter G, Steen C, De Keyser J. Use of the Barthel index and modified
Rankin scale in acute stroke trials. Stroke. 1999;30:1538–1541.
15. Kwon S, Hartzema AG, Duncan PW, Min-Lai S. Disability meas-
ures in stroke: relationship among the Barthel Index, the Functional
Independence Measure, and the modified Rankin Scale. Stroke.
2004;35:918–923. doi: 10.1161/01.STR.0000119385.56094.32
16. Nunn A, Bath PM, Gray LJ. Analysis of the modified Rankin Scale in
randomised controlled trials of acute ischaemic stroke: a systematic re-
view. Stroke Res Treat. 2016;2016:9482876. doi: 10.1155/2016/9482876
17. Mehrholz J, Wagner K, Rutte K, Meissner D, Pohl M. Predictive validity
and responsiveness of the functional ambulation category in hemiparetic
patients after stroke. Arch Phys Med Rehabil. 2007;88:1314–1319. doi:
10.1016/j.apmr.2007.06.764
18. Zhang YL, Liang W, Chen ZM, Zhang HM, Zhang JH, Weng XQ,
et al. Validity and reliability of patient health questionnaire-9 and
patient health questionnaire-2 to screen for depression among col-
lege students in China. Asia Pac Psychiatry. 2013;5:268–275. doi:
10.1111/appy.12103
19. Gusi N, Olivares PR, Rajendram R. The eq-5d health-related quality of
life questionnaire. In: Preedy VR, Watson RR, eds. Handbook of Disease
Burdens and Quality of Life Measures. New York, NY: Springer; 2010.
20. Liu PC, Gau BS, Hung CC. Development and psychometric test-
ing of a chinese version of the caregiver burden index for parents
of children with allergies. J Pediatr Nurs. 2015;30:208–218. doi:
10.1016/j.pedn.2014.04.006
21. Pope C, Ziebland S, Mays N. Qualitative research in health care.
Analysing qualitative data. BMJ. 2000;320:114–116. doi: 10.1136/bmj.
320.7227.114
22. Reding MJ, Potes E. Rehabilitation outcome following initial unilat-
eral hemispheric stroke. Life table analysis approach. Stroke. 1988;19:
1354–1358.
23. Yu J, Yu Y, Fan JT, Xu Y, Hao FC. Impact of different treatment and
evaluation methods on stroke rehabilitation effect. Tianjin Med J.
2010;11:2.
24. Shah S, Vanclay F, Cooper B. Efficiency, effectiveness, and duration of
stroke rehabilitation. Stroke. 1990;21:241–246.