Mumbai stroke registry (2005-2006)--surveillance using WHO steps stroke instrument--challenges and opportunities.

Abstract

India will face enormous socioeconomic burden because life expectancy is increasing placing larger numbers of older people at risk of stroke and other chronic diseases. In order to plan prevention strategies, reliable information on stroke epidemiology is required. For uniform data collection (population based), WHO recommends use of STEPS Stroke instrument.
A well-defined community (H-ward) with verifiable census data, and representative of population structure of Mumbai (Bombay), was selected. The manual on WHO STEPwise approach to stroke surveillance (STEPS; http://www.who.int/chp/steps/Manual.pdf) was the operational protocol.
During the two year study period (Jan 2005 to Dec 2006), 521 new stroke (CVD) cases (males--275 and females--246) were identified; of which 456 (238 males and 218 females) had "first ever stroke"(FES) indicating an annual incidence of 145 per 100,000 persons (CI 95%: 120-170); age adjusted Segi rate: 152/100,000/year (CI 95% 132-172). Two thirds of the FES cases were admitted to health care facilities (Step I: "in-hospital" cases), the remaining 150 (32.8%) either died outside of hospital or were treated at home or nursing homes (Step II: Fatal events in community and Step III: Non-fatal events in community). CVD Diagnosis was supported by CT (Computed Tomography) in 407 (89%) of 456 FES cases: 366 (80.2%) had Ischaemic CVD, 81 (17.7%) had hemorrhagic CVD and 9 (1%) were of unspecified category. The mean age was 66 yrs SD +/- 13.60 and women were older compared with men (mean age 68.9 yrs SD +/- 13.12 versus 63.4 yrs SD +/- 13.53). Hypertension (BP more than 140/90 mm Hg) alone or in various combinations was present in 378 ( 82.8%) cases. Case fatality at 28 days after the FES stroke was 29.8%. Of 320 surviving patients 38.5% had moderate to severe disability.
WHO STEPs stroke surveillance Instrument is simple to use and, practical for community surveys. The data are useful for planning stroke prevention campaigns on public awareness and education with regard to diet, exercise, blood pressure control and early symptoms of minor strokes.

Full text

Page 1

© JAPI • VOL. 56 • SEPTEMBER 2008 www.japi.org 675
Original Article
Mumbai Stroke Registry (2005-2006) - Surveillance
Using WHO Steps Stroke Instrument - Challenges and
Opportunities#
PM Dalal*, Madhumita Bhattacharjee**, Jaee Vairale***, Priya Bhat****
Abstract
Background: India will face enormous socioeconomic burden because life expectancy is increasing placing larger
numbers of older people at risk of stroke and other chronic diseases. In order to plan prevention strategies, reliable
information on stroke epidemiology is required. For uniform data collection (population based), WHO recommends
use of STEPS Stroke instrument.
Study: A well-defined community (H-ward) with verifiable census data, and representative of population structure
of Mumbai (Bombay), was selected. The manual on WHO STEPwise approach to stroke surveillance (STEPS; http://
www.who.int/chp/steps/Manual.pdf) was the operational protocol.
Results: During the two year study period (Jan 2005 to Dec 2006), 521 new stroke (CVD) cases (males- 275 and
females- 246) were identified; of which 456 (238 males and 218 females) had “first ever stroke”(FES) indicating
an annual incidence of 145 per 100,000 persons (CI 95%: 120-170); age adjusted Segi rate:152 /100,000/year (CI
95% 132-172). Two thirds of the FES cases were admitted to health care facilities (Step I: “in-hospital” cases) , the
remaining 150 (32.8%) either died outside of hospital or were treated at home or nursing homes (Step II: Fatal
events in community and Step III: Non-fatal events in community). CVD Diagnosis was supported by CT (Computed
Tomography) in 407 (89%) of 456 FES cases: 366 (80.2%) had Ischaemic CVD, 81 (17.7%) had hemorrhagic CVD and
9(1%) were of unspecified category. The mean age was 66 yrs SD±13.60 and women were older compared with
men (mean age 68.9rs SD ± 13.12 versus 63.4yrs SD ± 13.53). Hypertension (BP more than140/90 mm Hg) alone or
in various combinations was present in 378( 82.8%) cases. Case fatality at 28 days after the FES stroke was 29.8%.
Of 320 surviving patients 38.5% had moderate to severe disability.
Conclusions: WHO STEPs stroke surveillance Instrument is simple to use and, practical for community surveys. The
data are useful for planning stroke prevention campaigns on public awareness and education with regard to diet,
exercise, blood pressure control and early symptoms of minor strokes. ©
INTRODUCTION
A
of chronic diseases such as stroke.1 For planning prevention
strategies, reliable information on pattern of disease,
exposure to major risk factors and morbidity/mortality
trends for cerebrovascular disease (stroke) in defined
populations are necessary.2 In the literature, there are
several reports on stroke epidemiology from developed
countries using uniform methodology, but there is scant
information on population based data from developing
s life expectancy increases, India will face enormous
socioeconomic burden to meet the costs of management
*Research Director; **Research Associate; ***Senior Research Officer;
****Senior Research Fellow, Lilavati Hospital and L. K. M. M. Trust
Research Centre, A-791, Bandra Reclamation, Mumbai - 400 050,
India.
Received : 6.2.2008; Revised : 21.4.2008; Accepted : 26.6.2008
countries.3,4 There is also great need to monitor these trends
using uniform protocols in simple and reproducible way for
regional comparisons 5. For this purpose, WHO proposes the
use of STEPS Stroke instrument.1,5 We describe observations
made during population based stroke surveillance using
this instrument over a two year period in an urban area of
Mumbai (H-ward), India.
MATERIAL AND METHODS
WHO defines Stroke as “a focal (or at times global)
neurological impairment of sudden onset, and lasting
more than 24 hours (or leading to death) and of presumed
vascular origin.”6 The manual on WHO STEPwise approach to
surveillance (STEPS; http://www.who.int/chp/steps/stroke)7
was the operational protocol. The manual is comprehensive
and describes terminologies and methodologies on case
assessment data collection and management for Step 1(In

Page 2

676 www.japi.org © JAPI • VOL. 56 • SEPTEMBER 2008
–hospital events), Step 2 (fatal events in the community)
and Step 3 (non-fatal events in the community) as shown
in Fig. 1.
In brief, Step 1(in-hospital) questionnaire contains
patient’s demographic data (i.e. name, age, sex, ethnicity,
occupation, contact details), information on acute stroke
event and hospitalization, degree of neurological deficit
(National Institute of Health Stroke Scale-NIHSS), stroke
sub-type by diagnostic tests and information on current
risk factors, data on medical treatment, discharge status
and follow-up at 28 days by Modified Rankin scale. In Step
2 (fatal events in community), apart from above information
specific details on date of death and information on fatal
event by verification of death certificates or medical autopsy
or verbal autopsy are listed. In Step 3 (non-fatal events
in community), all information as described in Step 1 are
collected by interviewing patients at home or nursing home,
and current neurological status by Modified Rankin scale
is assessed. NIHSS, Modified Rankin scale are described in
Appendix.
The duration of study, was twenty four months (January
2005-December 2006). The selected survey population was a
well-defined community (H-ward) which has reliable census
data (2004) and is representative of the population structure
(by sex and mid-decade age bands) of Mumbai (Bombay).8
In view of state and local elections, the electoral roll on
permanent residents is well maintained and periodically
updated. The roster lists full name of the family members
by, sex, age in years and precise residential location. Use of
multiple overlapping source of information confirmed the
validity of entries.
As per the WHO STEPS protocol, out of total 337,391
permanent residents, 156,861 (46%) were 25 years or
older and were eligible if they met the inclusion criteria
for stroke. Census information provides information on
location by street, building, and colony. Information leaflets
on stroke awareness and the purpose of the study were
circulated to local residents. In a special meeting of local
medical practitioners, the WHO protocol was introduced
and discussed. General practitioners, major hospitals,
nursing homes, Computerized Tomography (CT) diagnostic
centres, and municipal health authorities participated or
Appendix : National Institute of Health Stroke Scale (NIHSS)
Fig. 1 : WHO STEP wise approach to stroke surveillance.
No. Category Response/ Score
1A

1B

1C

2.

3.

4.

5.



6.



7.
8.

9.

10. Dysarthria
11. Extinction


Consciousness

Answers to
questions
Commands

Gaze

Visual fields

Facial palsy

Motor arm

a. left, b. right

Motor leg

a. left, b. right

Ataxia
Sensory

Language

Alert -0 / Arousable -1 /
Obtunded -2 / Unresponsive -3
Answers correctly : Both- 0 /
One – 1 / Neither - 2
Both correctly – 0 / One
correctly – 1 / Neither -2
Normal -0 / Partial palsy -1 /
Total palsy -2
No loss -0 / Partial hemianopsia
-1/ Complete -2/ Bilateral -3
None -0 / Minor – 1 /
Partial – 2 / Complete -3
No drift -0 / Drift <10 sec- 1 /
Falls<10 sec -2/
No effort against gravity – 3 /
No movement -4
No drift -0 / Drift <10 sec- 1 /
Falls<10 sec -2/
No effort against gravity – 3 /
No movement -4
None -0 / One limb- 1 / Two limbs- 2
Normal -0 / Mild loss -1 /
Severe loss -2
Normal -0 / Mild – 1 / Severe -2 /
Mute or global aphasia-3
None-0 / Mild – 1 / Severe -2
None-0 / Mild – 1 / Severe -2
Total NIHSS Score- (00-42)
Legend: NIHSS is a scoring technique defines degree of neurological
deficit ranging from none (0), mild (1) or severe (2 to 4) for eleven
categories of neurologic functions. For practical purpose score of 0 to
5 indicates mild deficit, 6 to 15 denotes moderate deficit and score of
more than 15 is suggestive of severe neurologic deficit.
[Source: Goldstein LB, Samsa GP. Reliability of the National Institutes of
Health Stroke Scale. Extension to non-neurologists in the context of a
clinical trial. Stroke 1997;28:307-10]
Modified rankin scale
Category Score
No symptoms at all
No significant disability despite symptoms
Slight disability
Moderate disability, but able to walk
without assistance
Moderate disability, but unable to walk
without assistance
Severe disability
0
1
2
3
4
5
Legend: Modified Rankin Scale measures independence rather than
performance of specific tasks. Scale consists of six grades from 0 to 5;
0 denotes no symptoms and 5 indicates severe disability. For clinical
purpose, mild disability range is from 0 to 2; moderate disability ranges
from 3 to 4 and 5 indicates severe disability.
[Source: Sulter G, Steen C, Keyser JD. Use of the Barthel Index and
Modified Rankin Scale in acute stroke trials. Stroke 1999;30:1538-41].

Page 3

© JAPI • VOL. 56 • SEPTEMBER 2008 www.japi.org 677
cooperated in the study. The clinical syndrome of limb
weakness accompanied by speech or visual disturbance
and/ or imbalance posed no major difficulty in identification
of CVD case by the patient or caregiver.
The medical research officers received intensive training
on definitions/terminologies in respect to questionnaire
and responses, including independent assessment of
neurological deficit at onset and disability status at 28
days by Modified Rankin scale (MRS). They identified
“first ever stroke” (FES) as defined in the WHO manual.
Where information was not possible to obtain the item
was coded as “incomplete data”. Stroke diagnosis was
cross-checked by medical practitioner/consultant and was
supported in the majority of cases by diagnostic tests (e.g.
CT scan). All suspected stroke cases were traced by “hot
pursuit” (prospective case registration) or “cold pursuit”
(retrospective case registration) or a combination of both.
The stroke team made periodic (daily/weekly) visits to
health care facilities to ascertain completeness of case
assessment and to minimize “under-reporting” of events
especially those cared for out of H Ward hospitals.
Ethics Committee recommended that names of the
subjects in the study and their contacts were not divulged
without permission. After completing the defined responses
to questionnaire by itemized codes, the data on each new
stroke case were transferred to a Central registry.
Statistics
Statistical analysis was carried out using “True Epistat
(version:5.3)” on following variables : age, sex, previous
stroke, stroke sub type, delay from onset to admission, NIH
score on admission, timing of CT scanning, vital status at day
28 (alive/dead status), MRS score at day 28 and associated
risk factors. Gender differences on above variables were also
assessed by chi-square test and log likelihood ratio test. To
calculate age standardized incidence rates by direct method
with Segi 1996 world population,9 True Epistat package
(version 5.3, Epistat services, USA) was used and this enabled
us to compare similar age standardized incidence rates as
reported for different studies.
RESULTS
During the study period 521 cases with new stroke
events were identified: 456 patients had “first ever stroke”
(males- 238 and females- 218) and 65 had a history of
previous stroke. The overall crude annual incidence rate for
“first ever stroke” (FES) is 145 per 100,000 persons, (CI 95%:
120-170); this rate for men and women is 149 and 141.5 per
100,000 persons respectively. Age standardized rate of total
FES by direct method with Segi 1996 world population is:
152/100,000 persons/year (CI 95% 132-1729); similar age
standardized rates for men is 162/ 100,000 persons /year
(95% CI- 133-192) and for women is 141 /100,000 persons
/year (95% CI- 114-167).
The overall mean age was 66 yrs, SD± 13.60; women
were older (68.9 yrs) as compared to men (63.4 yrs). Table
1 and Fig. 2 show percentage distribution of population by
age and sex and incidence of “first ever stroke” cases per
100,000 per year.
Two-thirds (67.2%) of FES cases were seen at health
care (“in-hospital”) facilities (Step I) and the remaining
150(32.8%) of 456 FES were treated at home or nursing
homes (Step II and Step III).Clinical diagnosis on new stroke
event (FES) was supported by CT (Computed Tomography)
data in 407(89.2%) of 456 cases Diagnostic tests indicated
that, 366(80.2%) had ischaemic stroke, 81(17.7%) had
hemorrhagic stroke and 9(1.9%) were of unspecified type.
There was no significant gender difference with regard to
neurological deficit on admission, timing of first scan, stroke
subtype, alive or dead status at day 28. For risk factor analysis
reliable data for hypertension (by history, by value-BP more
than 140/90 mm Hg, supported by anti-hypertensive drug
therapy or physician’s prescription) was available in 378
(82.8%) cases; however, verifiable data for other risk factors
like diabetes mellitus, ischaemic heart disease, tobacco use,
lipid levels were not available for analysis in every case.
Disability status (by MRS) and Case fatality rate: At day 28,
Table 1 : Crude annual incidence of “first ever stroke” per 100,000 by age groups
Age Groups
Total Population
1,56,861
FES cases
Cases/100,000/yr
95% CI*
120-170
25-34
47,298
35-44
38,245
45-54
29,806
55-64
20,232
65-74
14,401
75-84
5,593
85-94+
1,286
Total
6
6.3
2-13
26
33.9
23-47
55
92.2
72-110
96
237.2
210-270
140
486
92
822.4
770-880
41
1594
456
145
440-530 1520-1670
Age standardized rate by direct method with Segi 1996 world population9: 152 / 100,000 persons/year (95% CI- 132-172). for males it is 162/ 100,000
Fig. 2 : Crude annual incidence of “First ever Stroke” per 100,000 persons
and Population distribution (%) by age groups.

Page 4

678 www.japi.org © JAPI • VOL. 56 • SEPTEMBER 2008
320(70%) of 456 FES cases were alive and here 176 (38.5%)
had moderate to severe disability pointing to burden of
stroke; whereas 29.3% (134 cases) had minimal disability;
and in 2.1% (10 cases) information was missing. By 28 days,
136 (29.8%) of 456 FES died. Here, 82 (17.9%) were stroke
related deaths, and 54 (11.8%) deaths appeared related to
co-morbid conditions (eg cardiopulmonary illness etc).
Recurrent Strokes: During the two year study period, a
total of 521 new stroke events were registered, of which
456 were “first ever strokes” and 65 had had a previous
stroke which brought them into the study. Among the
recurrent strokes there was no significant gender difference
with regard to age groups, stroke subtypes (57 had ICVD,
7 had ICH and 1 was unspecified category), or alive/dead
status at 28 days (55 alive and 10 dead). For the purposes
of this paper, these patients have been excluded from the
analyses.
DISCUSSION
To plan effective intervention and stroke preventive
strategies, one must know the annual incidence rates; its
clinical profile based on supportive diagnostic tests, as well
as mortality and morbidity trends. Only epidemiological
surveys in defined communities can determine the overall
burden on health services, communities and family
members. The purpose of WHO STEPS Stroke surveillance
(6 modules in 3 steps) is to (i) assess the magnitude of
stroke, (ii) describe population at risk, (iii) identify associated
risk factors, (iv) monitor trends over time, (v) implement
intervention strategies and (vi) evaluate the outcome of
intervention. WHO STEPS encourages the collection of
standardised core data; options are available for more
extended data collection to suit local circumstances or
needs. The three “steps” represent different, but overlapping
case finding sources with each step representing increasing
difficulty in achieving comprehensive coverage.
Step-1: Hospitalized cases: The main aim of Step 1 is to
collect core data (demography, information on acute event,
evaluation and management, identifying stroke subtype,
associated risk factors and outcome status by Modified
Rankin scale on admission and at 28 days or at discharge)
on patients who come to health care facilities enumerated
within the defined population. We found the WHO Stroke
Surveillance protocol using STEPs questionnaire (Version 2)7
is simple and practical to administer. Earlier, (2003-2004) we
had tested STEPs questionnaire (Version 1.1) in hospitalized
cases and had practical experience in administering the
basic protocol1. Data collection posed no major difficulty;
but complete information in some cases was not possible
due to short stay or discharge against medical advice or
unwillingness for costly tests (e.g. CT scan). The information
once coded was easily transferred by “Data Entry Tool”
(DET) to a central registry. Stroke awareness leaflets, media
reports, meeting with community doctors, to explain stroke
as “burden of disease” and guide on comprehensive care
played a significant role in successful implementation. It
was also shown to be useful to take everyone (patients
and family, referring physicians, community and resident
doctors, nursing and paramedical staff, administrative and
municipal health authorities) in confidence at periodic
intervals.
Step-2: Fatal events in community In Mumbai, it is
mandatory to have “medical certification of cause of death”
before disposal of a dead body. This certificate (FORM NO.
4 and 4A) is attested by a registered medical practitioner.
It contains information on “Name of deceased” age, sex,
residential location, cause of death – immediate cause
as well as antecedent and the manner of death (natural,
accidental or otherwise) Regular perusal of death registry
records for residents of the H Ward, together with follow
up and verification of the diagnosis, suggested that these
data were by and large properly recorded. For cross-
verification of medical illness, attempts were made for a
“verbal autopsy” but this was not always successful. The
reasons for non-cooperation included the following: (a)
family bereavement; (b) resistance to part with copies of
documents which sometimes had been filed for insurance /
medical claims; and (c) privacy related to illness. The natural
calamity (monsoon deluge and post-monsoon havoc) with
loss of life and property was also a contributing factor. It is
also possible that “sudden deaths” due to stroke may not be
correctly classified. However, review of death certificates
with the certifying doctor together with “verbal autopsy”
proved helpful in reclassifying 36 cases of “encephalopathy”
as recent strokes.
Step-3: Non-fatal events in the community. The most
difficult cases to identify are those occurring in the
community and potentially represent “missed cases” will
result in under- reporting of incidence rates. For example,
subjects with uncontrolled hypertension and/or diabetes
mellitus often attribute their neurological deficit to
underlying co morbid condition. Denial of minor stroke like
symptoms by patients or not reporting to family physician or
health care facility is another possibility. Furthermore these
subjects with early neurological deficit may be taking local
herbal or non-allopathic treatment and might not seek or
attend allopathic health care facility. In addition, patients
having prolonged TIA where deficit lasts around 24 hours
may be misclassified can also result in “under reporting”.
The current trend to advocate thrombolytic therapy
within 6 hours of an acute event could result in more rapid
reference of “minor stroke” cases (NIH score 5 or less) for
diagnostic tests. Unless the subject is hospitalized, medical
insurance does not reimburse the cost of CT scanning. This
situation may result in hospitalization of acute stroke cases
with mild deficit who would have been otherwise treated
at home. We believe that our contact with practicing
physician’s and community leaders helped us to identify all
eligible acute stroke events that occurred in the community
of H Ward. Unless there is intensive tracking (mixed pursuit)
of community on weekly basis, by health visitors, there
is possibility of “under- reporting” of events and thereby

Page 5

© JAPI • VOL. 56 • SEPTEMBER 2008 www.japi.org 679
resulting in lower incidence rates. For developing countries,
this situation needs to be taken into consideration.
The Mumbai stroke registry based on reliable census
population describes epidemiology of stroke (incidence
rates, stroke subtypes, and alive /dead status at 28 days).
Experience suggests that the WHO STEPS protocol is feasible
and gives useful information to plan intervention strategies.
Another approach for studying stroke epidemiology
on smaller populations (30,000 or more) has been
suggested.10
For Mumbai registry, the age standardized annual
incidence rate by direct method with Segi 1996 world
population for FES has been 152/100,000; (95% CI 132-1729)
and this is not significantly different from that reported
from China (Jiang),11 Auckland (New Zealand),3 Martinique
(ERMANCIA study)12 and Chile (PISCIS study).13 On the other
hand, the WHO Rohtak (India) study (1971-74)14,15 quoted an
annual incidence of 33/100,000 population for all ages and
356/100,000 for those above the age of 70 years. However,
precise details on numerator (for cases) and denominator
(for population) are not available for interpretation.
In analysis of proportional frequency of stroke types in
selected population (ten studies) around the world, Feigin
et al3 reported, that ischaemic strokes were present in about
67.3-80.5%, whereas primary intracerebral haemorrhage
was responsible for 6.5-19.6% of cases, subarachnoid
haemorrhage accounted for 0.8%-7.0% and 2.0-14.5% were
undefined type. In the Mumbai Stroke Registry (2005-06),
80.2% had ischaemic strokes, 14.6% had intracerebral
haemorrhage, 3% had subarachnoid haemorrhage and 1.9%
were unspecified type. These findings are not significantly
different from those reported from above studies.3
The 28 day case-fatality rate in thirteen selected studies
has ranged from 17% to 33%, average being 22.9%.3 In the
Mumbai Stroke Registry (2005-06), the overall case-fatality
rate for FES was 29.8% (136 cases) but the stroke related
case-fatality rate was 17.9% (82 cases) and deaths unrelated
to stroke were 11.8% (54 cases). The latter were related to
concurrent co-morbid conditions like congestive cardiac
failure, respiratory infections, septicemia etc.
It is emphasized that there are difficulties in this type of
comparative analysis because of regional variations, differing
ethnic groups etc but it does give general information with
regard to incidence rates, stroke subtypes and case fatality.
In absence of time trend studies, future projections to define
stroke burden and interventions are not possible.
SUMMARy
The World Health Organization (WHO) STEPwise
approach to stroke surveillance (WHO STEPS) designed
to collect in a standardized manner basic epidemiological
data on incidence, major risk factors, morbidity and
mortality trends, intervention strategies in recent (acute)
stroke. The present report describes the testing of WHO
Steps Stroke Instrument (Question guide-Version: 2.0)7 on
data collection and management in a defined segment
of community (H-ward) in Mumbai. The data collection,
case finding methods and reporting templates should be
uniform for inter-regional comparisons. The WHO Stepwise
approach to Stroke Surveillance is a practical instrument to
administer to “in-hospital” cases as well as those occurring
in community set up trained health workers will have
to follow “mixed pursuit” methods in identifying “minor
strokes”. It is proposed that similar surveys should be
initiated on representative samples of regional population
for comparative evaluation, to plan effective intervention
and prevention strategies. Building on the populations
enumerated for cancer surveillance purposes, might provide
an excellent opportunity in meeting the challenge of
understanding the population impact of this increasingly
important chronic disease.
Acknowledgements
#Presented at ICMR - WHO Workshop on Stroke
Surveillance, New Delhi, 13-15th Nov, 2006.
We are thankful to Shri. Vijaybhai Mehta, President,
Lilavati Hospital and Research Center and to Dr. Narendra
Trivedi, Vice President, LKMMT Research Centre for their
unstinted support. We are ever grateful to Dr Ruth Bonita
and Dr Thomas Truelsen for constant help and guidance at
all stages of the study. We would also grateful to Dr Seema
Malik (Chief Medical Superintendent Peripheral Hospitals,
Mumbai), Dr S. Deshmukh (Medical Officer Health –H
ward Mumbai), Dr K. Khandelwal (Consultant Radiologist
P H Medical Centre, Mumbai), Dr V. D. Mathur (Medical
Superintendent Mahavir Medical Centre, Mumbai), Swami
Rudra Maharajji (Administrator, Ramakrishna Mission
Hospital Mumbai)for their unstinted support at all stages of
this study, Dr. A. Nanivadekar, Consultant Biostatistician, for
guiding us in statistical analysis and Dr. P. Mehrotra, Dr. Hetal
Rathod, (Research Officers), Dr. Nandini Arora (MSW) from
the stroke team and Ms Shilpa Phansalkar for secretarial
help. The study was supported by an unrestricted grant from
the Global Stroke Fund of International Stroke Society via
SEARO and LKMMT Research Centre at Mumbai.
REFERENCES
1. Dalal PM. Burden of Stroke: Indian Perspective. International J Stroke
2006;1:164-66.
Bonita R, Mendis S, Truelsen T, Bogousslavsky J, Toole J,
yatsu F. The Global Stroke Initiative. Lancet Neurol 2004;3:
391-93.
Feigin VL, Lawes CM, Bennett DA, Anderson CS. Stroke epidemiology:
a review of population based studies of incidence, prevalence, and
case-fatality in the late 20th century. Lancet Neurol 2003;2:43-53.
Das SK, Banerjee TK, Biswas A, Roy T, Raut DK, Mukherjee
CS, Chaudhuri A, Hazra A, Roy J. A prospective community-
based study of stroke in Kolkata, India. Stroke 2007;38:
906-10.
Truelsen T, Heuschman PU, Bonita R, et al. Standard method for
developing stroke registers in low-income and middle- income
countries: experiences from a feasibility study of a stepwise approach
to stroke surveillance (STEPS Stroke). Lancet Neurol 2007;2:134-39.
Hatano S. Experience from a multicentre stroke register: a preliminary
report. Bull WHO 1976;54:541-53.
World Health Organization. STEPwise approach to stroke surveillance.
2.
3.
4.
5.
6.
7.