You are on page 1of 3


International Journal of Stroke

0(0) 1–3
Statistical analysis plan for the ! 2016 World Stroke Organization
Reprints and permissions:
family-led rehabilitation after stroke
DOI: 10.1177/1747493016674956

in India (ATTEND) trial: A multicenter

randomized controlled trial of a

new model of stroke rehabilitation
compared to usual care

Laurent Billot1,2, Richard I Lindley1,2, Lisa A Harvey2,

Pallab K Maulik3,4, Maree L Hackett1,5, Gudlavalleti VS Murthy6,7,
Craig S Anderson1,2, Bindiganavale R Shamanna8, Stephen Jan1,
Marion Walker9, Anne Forster10, Peter Langhorne11,
Shweta J Verma12, Cynthia Felix3, Mohammed Alim3,
Dorcas BC Gandhi12 and Jeyaraj Durai Pandian12

Background: In low- and middle-income countries, few patients receive organized rehabilitation after stroke, yet the
burden of chronic diseases such as stroke is increasing in these countries. Affordable models of effective rehabilitation could
have a major impact. The ATTEND trial is evaluating a family-led caregiver delivered rehabilitation program after stroke.
Objective: To publish the detailed statistical analysis plan for the ATTEND trial prior to trial unblinding.
Methods: Based upon the published registration and protocol, the blinded steering committee and management team,
led by the trial statistician, have developed a statistical analysis plan. The plan has been informed by the chosen outcome
measures, the data collection forms and knowledge of key baseline data.
Results: The resulting statistical analysis plan is consistent with best practice and will allow open and transparent
Conclusions: Publication of the trial statistical analysis plan reduces potential bias in trial reporting, and clearly outlines
pre-specified analyses.
Clinical Trial Registrations: India CTRI/2013/04/003557; Australian New Zealand Clinical Trials Registry
ACTRN1261000078752; Universal Trial Number U1111-1138-6707.

Rehabilitation, statistical analysis plan, cost factors, India, clinical trial, disability, caregivers, developing countries

Received: 11 August 2016; accepted: 11 September 2016

School of Medical Sciences, University of Hyderabad, Hyderabad,
Telangana, India
The George Institute for Global Health, University of Sydney, Sydney, School of Medicine, University of Nottingham, Nottingham, UK
NSW, Australia Academic Unit of Elderly Care and Rehabilitation, Bradford Teaching
Sydney Medical School, University of Sydney, Sydney, NSW, Australia Hospitals NHS Foundation Trust, University of Leeds, Leeds, UK
Research and Development, George Institute for Global Health India, Academic Section of Geriatric Medicine, Institute of Cardiovascular &
Hyderabad, Telangana, India Medical Sciences, University of Glasgow, Glasgow, UK
The George Institute for Global Health, Oxford University, Oxford, UK Department of Neurology, Christian Medical College, Ludhiana, Punjab,
College of Health and Wellbeing, University of Central Lancashire, India
Preston, UK Corresponding author:
Indian Institute of Public Health, Hyderabad, India Richard I Lindley, The George Institute for Global Health, University of
Clinical Research Department, London School of Hygiene & Tropical Sydney, Level 3, 50 Bridge St., Sydney, NSW 2000, Australia.
Medicine, London, UK Email:

International Journal of Stroke, 0(0)

2 International Journal of Stroke 0(0)

As infectious diseases have been brought under increas- received usual care. Baseline, demographic, and initial
ing control, non-communicable diseases are on the rise, follow-up data were collected at hospital discharge.
with stroke, in particular, increasing in incidence and Outcomes were collected at three months and six
subsequent burden. The majority of those with stroke months post-randomization by blinded assessors
live in low- and middle-income countries, and these employed at each site. The primary outcome is the
countries have few organized stroke services.1 In modified Rankin Scale (mRS) at six months to reflect
India, for example, most stroke patients do not have the lasting effects of the intervention on disability. The
access to formal rehabilitation.2 In rural areas of India, three-month data will be used to explore trajectories of
few patients get to hospital, and most do not receive recovery and to minimize dropouts (e.g., if a patient
appropriate secondary preventative treatment.3 cannot be found at three-months, there will be another
Economic drivers are clearly important as the per three months to try and locate the patient for the pri-
capita Gross Domestic Product is more than 30 times mary outcome). The main outcome was a dichotomous
greater in countries such as the USA or Australia, as outcome on the mRS score of 0–2 versus 3–6, with an
compared to India.4 Affordable healthcare should ordinal analysis as a key secondary outcome. Other
therefore be a public health priority. The ATTEND outcomes include basic activities of daily living, quality
trial of family-led rehabilitation after stroke is an inter- of life, carer burden, anxiety, and depression and
national attempt to address this priority as recom- extended activities of daily living. Length of stay,
mended by the World Bank and World Health place of residence, and return to work will also be col-
Organization report on Disability.5 ATTEND is an lected, as part of a larger planned economic analysis.
individual subject randomized controlled trial involving The trial is funded by the National Health and
the training of the patient and nominated carer in key Medical Research Council of Australia and recruitment
aspects of stroke rehabilitation, with an emphasis on commenced in January 2014 and was successfully com-
repeated performance in task-specific activities. The pleted (1250 participants) in early 2016. In keeping with
intervention was piloted,6 and then refined with input best practice, the trial is registered and the protocol
from the steering committee and project research team published.7 Prior to unblinding, we present the statis-
in Indian. tical analysis plan (see supplementary material avail-
ATTEND uses a pragmatic design, and the protocol able online with this article) which helps prevent
has been published.7 In brief, it is a randomized con- undue emphasis on data dependent analyses and
trolled trial across 14 sites in India. Patients are eligible reduces potential bias in future reporting. The prelim-
if they are adults (aged 18 years or older) who were inary results are expected to be announced at the World
admitted to hospital with a recent (<1 month) stroke Stroke Congress in Hyderabad in October 2016.
(ischemic, hemorrhagic, or undifferentiated) and had Results from ATTEND trial have the potential to
residual disability with a reasonable expectation of sur- have major impact as the costs are likely to be modest,
vival. Participants need to have had a nominated care- and therefore affordable to those not only in India but
giver and be willing to adhere to the follow-up also in other low- and middle-income countries. If suc-
arrangements. After consent has been obtained from cessful in the context of stroke, further work exploring
both the patient and caregiver, they were randomized similar models (a ‘‘polypill’’ of rehabilitation) would be
through a secure web-based randomization system to important for other acutely disabling conditions such
either intervention or control. The intervention was as hip fracture, spinal cord injury, burns, and brain
started in hospital, with an aim of participants receiving injury given the burden of disease in low- and middle-
it for 1 h each day until discharge, delivered by a trained income countries.
stroke coordinator (usually with a physiotherapy back-
ground). Part of the intervention involved discharge Declaration of conflicting interests
planning. After discharge, the patient and carer were The author(s) declared no potential conflicts of interest with
supported by up to five home visits by a stroke coordi- respect to the research, authorship, and/or publication of this
nator who provided them with an illustrated ATTEND article.
manual outlining key activities. The details of the inter-
vention have been kept confidential until the trial Funding
follow-up period has been completed to reduce the
The author(s) disclosed receipt of the following financial sup-
risk of contamination between intervention and control port for the research, authorship, and/or publication of this
patients. Overall, the intervention was developed from article: This study is funded by the National Health and
new models of care emerging in India, incorporating Medical Research Council of Australia (Project grant no.
aspects of the best evidence for stroke rehabilitation APP1045391). Pallab K Maulik is a recipient of an
and good practice summarized in an intervention Intermediate Career Fellowship of Wellcome Trust-
guide for the stroke coordinators. Control patients Department of Biotechnology India Alliance. Maree L

International Journal of Stroke, 0(0)

Billot et al. 3

Hackett is a recipient of a National Heart Foundation Future 3. Kalkonde Y, Deshmukh M, Sahane V, Nilas S and
Leader Fellowship, Level 2 (100034, 2014–2017). Stephen Jan Bang A. Awareness and healthcare seeking for stroke in
is the recipient of an NHMRC Senior Research Fellowship. rural Gadchiroli, India: a community-based study.
Craig Anderson holds an NHMRC Senior Principal Research Neuroepidemiology 2015; 45: 325.
Fellowship. 4. Dieleman JL, Templin T, Sadat N, et al. National spend-
ing on health by source for 184 countries between 2013 and
2040. Lancet 2016; 387: 2521–2535.
References 5. World Health Organisation. World report on disability.
1. Feigin VL, Fourouzanfar MH, Krishnamurthi R, et al. Geneva: World Health Organisation, 2011.
Global and regional burden of stroke during 1990-2010: 6. Pandian JD, Felix C, Kaur P, et al. FAmily-Led
findings from the Global Burden of Disease Study 2010. RehabiliTaTion aftEr Stroke in INDia: the ATTEND
Lancet 2014; 383: 245–255. pilot study. Int J Stroke 2015; 10: 609–614.
2. Langhorne P, de Villiers L and Pandian JD. Applicability 7. Alim M, Lindley R, Felix C, et al. Family-led rehabilita-
of stroke-unit care to low-income and middle-income tion after stroke in India: the ATTEND trial, study proto-
countries. Lancet Neurol 2012; 11: 341–348. col for a randomized controlled trial. Trials 2016; 17: 1–8.

International Journal of Stroke, 0(0)