Stroke Risk Screening Scales (SRSS): Identification of Domain

and Item Generation

Suebsarn Ruksakulpiwat, RN, MMed

Background: Stroke is a principal cause of mortality and disability in Thailand and

globally. Early and comprehensive risk identification would be critical to identify
people at high risk for stroke. Therefore, a comprehensive stroke risk screening tool is
needed to assess all possible stroke risks and potential at-risk populations. In the
future, such an instrument would benefit early detection and stroke prevention plan-
ning. Objective: The objective of the Stroke Risk Screening Scales (SRSS) development
is to identify the domains and generating appropriate questions for the new SRSS.
Methods: Using deductive methods suggested by Godfred Boateng and colleagues
(2018), the domains were classified based on the existing literature. The questions
were generated based on a comprehensive analysis of existing stroke risk screening
scales and their representativeness of each domain. Five existing stroke risk screening
tools including 1) the Stroke RiskometerTM, 2) the Framingham 10-Year Risk Score, 3)
the Stroke Risk Screening Tool (The Department of Disease Control of Thailand), 4)
the My Risk Stroke Calculator, and 5) QStroke were included and identified.
Results: Overall, 18 domains were included, and each domain was represented with
at least one or more questions. Eight domains (44.44 %) are consisting of a dichoto-
mous question alone, another eight domains (44.44 %) consist of multiple questions,
which combined between dichotomous, categorical, or fill-in-the-blank questions,
one domain (5.55 %) consists of a fill-in-the-blank question, and another one (5.55 %)
include only one categorical question. Conclusions: Developing a comprehensive tool
to be used for stroke risk screening by extending the knowledge of stroke, stroke risk
factors, and the best practice for tool development is necessitated for further practical
stroke prevention planning.
Key Words: Scale development—Screening scale—Stroke—Stroke risk
© 2021 Elsevier Inc. All rights reserved.

Background is continuously rising.2 In Thailand, stroke is a significant

health burden. It is the principal cause of mortality and
Stroke is a leading cause of death and long-term disabil-
long-term disability in Thai people.3 Every year more than
ity globally.1 In low-income and middle-income countries
250,000 Thai people experience their first stroke, and annu-
combined, there are approximately 6.5 million deaths per
ally nearly 50,000 Thais die because of a stroke.4
year and 113 million associated disability-adjusted life years
Modifiable risk factors such as treating hypertension,
related to the outcomes of stroke.1,2 Moreover, among these
diabetes, dyslipidemia, metabolic syndrome, atrial fibrilla-
countries, the incidence and mortality rate of primary stroke
tion, and smoking are among the main risk factors of stroke
in Thai people. Like many developing countries, Thailand
From the Medical Nursing Department, Faculty of Nursing, Mahi- has changed from a mainly rural society to an urbanized
dol University, 2 Prannok Road, Siriraj, Wanglang, Bangkoknoi, society. This shift has greatly impacted cultural and life-
Bangkok, Thailand. style decisions; for example, people shifted their eating and
Received October 8, 2020; revision received February 25, 2021; physical activity habits from a healthy diet consisting of
accepted March 2, 2021.
mainly rice and vegetables to more westernized foods high
Address correspondence to Suebsarn Ruksakulpiwat, RN, MMed,
Medical Nursing Department, Faculty of Nursing, Mahidol Univer- in fats and sugars.4,5 Consequently, the prevalence of these
sity, 2 Prannok Road, Siriraj, Wanglang, Bangkoknoi, Bangkok, Thai- stroke risk factors has been increasing steadily.3
land E-mail: [email protected] When considering stroke management in Thailand, it
1052-3057/$ - see front matter found that we essentially emphasize therapeutic response
© 2021 Elsevier Inc. All rights reserved.
measures in the early symptoms of the disease after
https://doi.org/10.1016/j.jstrokecerebrovasdis.2021.105740

Journal of Stroke and Cerebrovascular Diseases, Vol. 30, No. 6 (June), 2021: 105740 1
2 S. RUKSAKULPIWAT

people have a stroke. For example, there are more than Human subject protection
110 stroke units in the country, and these are the crucial
In this study, identified and reviewed data came from
component of the improvement of stroke care in this
existing resources and did not involve human partici-
country.6 Moreover, for all Thai citizens, treatment costs
pants. Therefore, Institutional Review Board approval
such as intravenous thrombolysis, antiplatelet, or statin
was exempt.
can be compensated from the Universal Coverage Pro-
gram administered by the National Health Security
Office. Nevertheless, the death rate of stroke is still Research methodology
growing.3
The advancement of the healthcare system in stroke The SRSS development and validation consisted of
management in Thailand has made great progress, imple- three phases; (1) Item development, (2) Scale Develop-
menting adequate stroke screening and primary preven- ment, and (3) Scale evaluation as recommended by the
tion into practice is insufficiently developed.7 According best practice for developing and validating scales devel-
to a review of the literature on stroke risk screening in oped by Godfred Boateng and colleagues.16 In the first
Thailand and internationally found that there are current phase of the scale development (phase 1.1), the researcher
tools that identify some risk factors that increase the classified the domains based on the existing literature and
chance of a stroke but did not include all possible modifi- generated questions based on the deductive method, in
able risk factors.8 12 For example, the Stroke Risk Screen- which each question was classified based on an evalua-
ing Tool commonly used in Thailand is assessing only tion of existing stroke risk screening scales and a literature
eight risk factors, including 1) Having direct relative suf- review (Fig. 1).
fering from ischemic heart disease or paralysis, 2) Current
smoking habits, 3) A history of being diagnosed with Identification of domains
hypertension or current high blood pressure levels, 4) A Following the best practice for developing and validat-
history of being diagnosed with diabetes or current high ing scales, 16 before generating the question, a domain
blood sugar levels, 5) Obtaining information from health- should be identified, 17 and a good identity of domain
care personals that there is an abnormal amount of choles- facilitates question generation and content validity.16
terol in the blood, 6) The current body mass index or waist According to the literature review, stroke can cause by
circumference, 7) History of stroke, and 8) History of heart either non-modifiable stroke risk factors such as age, gen-
disease.10,13 Furthermore, according to stroke risk screen- der, ethnicity, and genetics 18 21 or modifiable stroke risk
ing tools such as the Framingham 10-Year Risk Score, the factors such as high blood pressure, hyperlipidemia, dia-
Stroke RiskometerTM, the QStroke, or the My Risk Stroke betes, atrial fibrillation, smoking, alcohol use, obesity, pre-
Calculator have not covered all possible risks of stroke as vious stroke, inactivity, and stress.3,12,21 23 The current
well.8,9,11,12 research aims to classify the domain and generate an item
The previous research on chronic disease screening and of SRSS that leads to further practical stroke prevention
prevention states that to decrease the disease's burden planning. Therefore, the SRSS domain classification inten-
and prevalence, the procedure that benefits health care tionally covers only modifiable stroke risk factors. Never-
professionals to identify the condition must be develop- theless, some non-modifiable stroke risk factors such as
ing.14,15 Besides, the target population must be those who
have the most significant risk of disease. Consequently,
risk factors can be early discovering, and the implementa-
tion of the preventive program would protect them from
the disease.14
Currently, none of the existing tools is adequately
screening for all risk factors. Hence, developing a compre-
hensive tool for stroke risk screening by expanding the
understanding of stroke risk factors and the best practice
for tool development16 should be implemented. In the
future, a better tool to screen the appropriate target popu-
lation could lead to high gains in future effective disease
prevention.

Objective
The researcher's objective was to identify the domains Fig. 1. The overview of Stroke Risk Screening Scales (SRSS) development
and appropriate questions for the SRSS as the initial phase and validation. Followed the best practice for developing and validating
of the Stroke Risk Screening Scales (SRSS) development. scales.16
STROKE RISK SCREENING SCALES (SRSS) 3

age, gender, race, and genetic appear in the SRSS domain The Stroke RiskometerTM, which developed from the
as they are great potential stroke risk factors. Framingham 10-Year Risk Score, has added stroke risk
factors in the measurement such as a family history of
Question generation stroke or heart attack, alcohol consumption, stress, low
physical activity, waist to hip ratio (WHR), non-Cauca-
After the domain was drawn, an author aimed to gener-
sian, inadequate diet, cognitive problems or dementia,
ate items (questions) in this process. In this SRSS develop-
poor memory, previous traumatic brain injury, body
ment, the researcher applied the deductive approach to
mass index (BMI), and waist circumference to make a tool
identify relevant questions.24 Each question in SRSS was
more generalized. However, this has not yet incorporated
generated from literature review and five existing stroke
all potential stroke risk factors.11 Moreover, this study has
risk screening tools, including (1) the Stroke Riskome-
suggested that additional research should improve the
terTM, (2) the Framingham 10-Year Risk Score, (3) the
stroke risk scoring system for accurate prediction,11 as the
Stroke Risk Screening Tool (The Department of Disease
Stroke RiskometerTM performed poorly in predicting
Control of Thailand), (4) the My Risk Stroke Calculator,
stroke events.
and (5) QStroke were included and identified.8 12 Fur-
Hippisley et al. have developed and validated the
thermore, the researcher analyzed five critical characteris-
QStroke to predict the risk of stroke or transient ischemic
tics of questions; (1) Consistently recognized, (2) Well
attack (TIA) in 451 primary health care units in England
communicate to respondents, (3) Constitutes sufficient
and Wales by comparing efficacy with the Framingham
answers, (4) Presented correct information for respond-
10-Year Risk Score. According to measurement analysis
ents, and (5) The hospitality for participants to give an
found that if compare to a three tool above (the Depart-
accurate answer by the question at all times.25
ment of Disease Control of Thailand, the Framingham 10-
The Stroke Risk Screening Tool (The Department of Dis-
Year Risk Score, and The Stroke RiskometerTM), Hippisley
ease Control of Thailand) was developed by the Bureau of
and teams have added more possible stroke risk factors to
Non-Communicable Disease, Department of Disease Con-
QStroke such as ethnicity, congestive cardiac failure, heart
trol, Ministry of Health of Thailand.13 This tool is widely
attack or angina, rheumatoid arthritis, valvular heart dis-
used in Thailand for initial stroke risk screening, origi-
ease, cholesterol/high-density lipoprotein (HDL) ratio,
nally in the Thai language. The author (SR) translated this
and BMI.8 Regarding the validity, QStroke presents a
tool into English. It was then given to two anonymous
valid measure of absolute stroke risk in the general popu-
translators and asked them to translate it into Thai. These
lation of patients free of stroke or transient ischaemic
anonymous translators were not previously involved in
attack.8
the study and did not know its original content, objec-
Furthermore, the longitudinal study of the My Risk
tives, or context. Ultimately, the author compared the
Stroke Calculator states that this tool has developed from
back translation to the original text to verify whether or
the Framingham 10-Year Risk Score, and the author has
not the original translation was accurate.
added more modifiable and non-modifiable stroke risk
factors such as gender, level of education, diagnosis with
Result
renal or kidney disease, diagnosis with peripheral arterial
Deductively, five stroke risk screening tools were iden- disease, physical activity, indicators of anger, depression,
tified. The results showed that none of these tools had and anxiety into the tool.9 Nevertheless, this has not yet
covered all potential stroke risk factors. Table 1 shows the included all possible stroke risk factors, although the My
existing stroke risk screening tools covering the instru- Risk Stroke Calculator is a simplistic approach of distrib-
ment's structure, composition, and risk factor coverage. uting knowledge to the general population about their
The stroke risk screening tool (The Department of Disease stroke risk.9
Control of Thailand), the Framingham 10-Year Risk Score, In this stage of the SRSS development, the researcher
The Stroke RiskometerTM, QStroke, and the My Risk has identified 18 domains; (1) Age, (2) Gender, (3) Race,
Stroke Calculator were identified. (4) Obesity, (5) Hypertension, (6) Hyperlipidemia, (7) Dia-
The Stroke Risk Screening Tool developed by the betes mellitus, (8) Genetic, (9) Previous stroke, (10) Tran-
Department of Disease Control of Thailand only assesses sient Ischemic Attack, (11) Heart disease, (12) Peripheral
eight risk factors, which cover modifiable and non-modifi- arterial disease, (13) Rheumatoid arthritis, (14) Physical
able stroke risk factor, and there is no reliability or valid- exercise, (15) Smoking, (16) Alcohol, (17) Traumatic brain
ity reported for the scale.10,13 injury, and (18) Stress and emotion (Supplementary
For the Framingham 10-Year Risk Score, the measure- Table 1). For domain 1 5, 7 8, 10 12, and 14 18 were
ment that predicts stroke incidence in 10 years also does identified base on The Stroke RiskometerTM, domain 1 5,
not cover all possible stroke risk factors.12 Although, three 7 8, and 10 18 were identified base on the Framingham
studies have reported the Framingham Risk Score's valid- 10-Year Risk Score, domain 4, and 6 8 were identified
ity with overall satisfactory accuracy in the risk predic- according to the stroke risk screening tool (The Depart-
tions for future coronary heart disease events.26 28 ment of Disease Control of Thailand), the My Risk Stroke
 4
Table 1. Stroke risk factors for each stroke risk screening tool.

No. Factors The stroke risk screening tool The Framingham The Stroke RiskometerTM QStroke The My Risk Stroke
(The Department of Disease 10-Year Risk Score Calculator
Control of Thailand)
1 Sex ‘ @ @ @ @
2 Age @ @ @ @ @
(> 35 years) (35 74 years) (20 90 years) (25 84 years) (For each year above the
age of 20, get 1 point)
3 Systolic Blood pressure @ @ @ @ ‘
4 BMI @ ‘ @ @ ‘
(> 25 kg/m2) (Weight, Height) (Weight, Height)
5 Waist circumference @ ‘ @ ‘ ‘
(cm.) (If > 90 cm. in male or > 80 cm.
in female answer yes)
6 Waist-to-hip ratio ‘ ‘ @ ‘ ‘
7 HDL- Cholesterol (mg/dl) ‘ @ ‘ @ ‘
(Leave blank if unknown)
8 Total Cholesterol (mg/dl) @ @ ‘ @ ‘
(yes, no) (Leave blank if unknown)
9 On hypertensive @ @ @ @ @
medication
10 Smoking @ @ @ @ @
(yes, no) (yes, no) (yes, no) (Non-smoker, ex-smoker, (Currently, smoke,
light smoker (less than smoked in the past,
ten scores), moderate never smoked)
smoker (10 to 19 score),
heavy smoker (20 scores
or over))
11 Diabetes @ @ @ @ @
(Yes, no) (Types 1 or Types 2 (Types 1 or Types 2 (Types 1 or Types
combined) combined) 2)
12 Atrial Fibrillation ‘ @ @ @ ‘
13 Left ventricular ‘ @ @ @ ‘
hypertrophy
14 Valvular heart disease ‘ ‘ ‘ @ ‘

S. RUKSAKULPIWAT
15 Congestive cardiac failure ‘ ‘ ‘ @ @
16 History of CVD @ @ @ @ ‘
17 Family history of stroke/ @ @ @ ‘ ‘
heart attack
18 Peripheral arterial disease ‘ ‘ ‘ @
 STROKE RISK SCREENING SCALES (SRSS)
Table 1 (Continued)

No. Factors The stroke risk screening tool The Framingham The Stroke RiskometerTM QStroke The My Risk Stroke
(The Department of Disease 10-Year Risk Score Calculator
Control of Thailand)
@
(Heart/peripheral arterial
disease)
19 Heart attack or angina @ ‘ ‘ @ @
(Heart attack/angina/ MI)
20 Previous stroke or TIA @ ‘ @ ‘ ‘
21 Rheumatoid arthritis ‘ ‘ ‘ @ ‘
22 Chronic kidney disease ‘ ‘ ‘ @ @
(Renal/kidney disease)
23 Cognitive problems or ‘ ‘ @ ‘ ‘
dementia
24 Poor memory ‘ ‘ @ ‘ ‘
25 Previous TBI ‘ ‘ @ ‘ ‘
26 Non-European ‘ @ ‘ @ ‘

27 Race-ethnicity ‘ ‘ @ @ ‘
(White/Caucasian, (White or not stated,
African, Arabian/Persian, Indian, Pakistani, Ban-
Chinese, Indian, gladeshi, Other Asian,
Japanese, Latin Ameri- the Black Caribbean,
can/Hispanic, Malay/ Black African, Chinese,
Indonesian/ Other Others)
Southeast Asian, Maori,
Other Asian, Pacific
Islander, Not specified)
28 Low diet (unhealthy diet) ‘ ‘ @ ‘ ‘
29 Low physical activity ‘ ‘ @ ‘ @
30 High alcohol consumption ‘ ‘ @ ‘ @
31 Experienced stress ‘ ‘ ‘ @ @
(Mental ill health)
32 Cry easily ‘ ‘ ‘ ‘ @
33 Feel fearful ‘ ‘ ‘ ‘ @
34 Education ‘ ‘ ‘ ‘ @
Note. BMI Body Mass Index, HDL high-density lipoprotein, CVD cardiovascular disease, TIA transient ischemic attack, MI myocardial infarction, TBI Traumatic Brain Injury.

5
6 S. RUKSAKULPIWAT

Calculator was evaluated for domain 12, and domain 11 In this initial phase of SRSS development, 18 domains
and 13 were identified base on QStroke (Supplementary were identified (Supplementary Table 1). Some domain
Table 1). contains more than one question because only one ques-
Additionally, each domain consists of at least one or tion may not comprehend, constitutes sufficient answers,
more than one question in ascending order without over- presented correct information for respondents, and assists
lap. Overall, eight domains (44.44 %) are consisting of a participants in giving an accurate answer. This approach
dichotomous question alone (e.g., have you ever been diag- support by literature that recommended that the initial
nosed with high blood pressure? (yes/no)), another eight pool of items development can be at minimum twice as
domains (44.44 %) consist of multiple questions, which long as the desired final scale or five times as large as the
combined between dichotomous, categorical, or fill-in- final version,16,37,38 as further evaluation such as content
the-blank questions (e.g., please indicate: weight. . ..kg, validity, pretest the question, survey, and sampling, or
height. . ..cm, and waist circumference (> 90 cm in male or > question reduction will decrease unwanted parts from the
80 cm in female or
90 cm in male or
80 cm in female)), 1 primary pool.
(5.55 %) is a fill-in-the-blank question (e.g., how old are
you?....year), and another 1 (5.55 %) is a categorical ques-
Limitation
tion (e.g., what race/ethnicity do you belong to? (Caucasian/
African-American/Asian/Hispanic/Other)), which the word There are two methods to identify domains and gener-
used for each question is simple to avoid ambiguity (Sup- ate items: deductive and inductive methods.16 The deduc-
plementary Table 1). tive method, which the current study applied to identify
domains and generate items of SRSS, is based on the
description of the related domain and items' identifica-
tion. It has been done through literature review and
Discussion
assessment of existing scales. In comparison, the inductive
There are five existing stroke risk screening instruments method comprises the generation of items from individu-
were included in this study, including the Stroke Risk als’ responses, qualitative data collected through direct
Screening Tool (The Department of Disease Control of observations—for example, focus groups and personal
Thailand), the Framingham 10-Year Risk Score, the Stroke interviews, which can be used to identify domain items
RiskometerTM, the QStroke, and the My Risk Stroke Cal- inductively. While the deductive method presents the the-
culator. However, these are not incorporate all potential oretical basis for determining the domain, qualitative
stroke risk factors. For example, previous hypertensive techniques like the inductive method move the domain
studies reveal that high blood pressure, particularly sys- from an abstract point to identifying its visible forms.
tolic blood pressure, significantly increases stroke chance Therefore, consolidate both approaches should be consid-
compared to people without high blood pressure.29 33 ered for future scale development to make more practical
Likewise, people with high blood pressure have an oppor- judgments about domains and items. Furthermore, since
tunity to get stroke 3.8 times compared to people with the results emerged from solely five existing stroke risk
normal blood pressure, same-sex, and age.34 Accordingly, screening tools, there might be other tools to identify and
if we use the My Risk Stroke Calculator to measure stroke benefits the development of new SRSS.
risk from a population, the blood pressure (factor 3 in
Table 1) will be underrepresented. Similarly, when the
Conclusion and implication of the study
Framingham 10-Year Risk Score and The Stroke Riskome-
terTM are using, the heart attack or angina (factor 19 in The knowledge of stroke and stroke risk factors has
Table 1) will also be underrepresented, but the My Risk been improving. Hence, the investigation and confirm sig-
Stroke Calculator will not since it covers this factor. There- nificant factors that need to be part of instrument develop-
fore, improving a more broad stroke risk screening scales ment and developing a comprehensive tool used for
that consolidate all potential stroke risk factors through stroke risk screening by expanding the knowledge of
literature review and assessments of existing scales stroke, stroke risk factors, and the best practice for tool
appear to be an appropriate strategy.17,35 development should be addressed. We can then use the
The best practice for scale development and validation tool to screen the target population to further effective dis-
for health science suggested that the first step for develop- ease prevention planning in the future. In the further step,
ing the scale, domain, and question should be identified.16 phase 1.2 of the Stroke Risk Screening Scales (SRSS) devel-
In the current SRSS development, the deductive method opment (content validity). The author proposed to assess
was used to determine the domain and generate the ques- each of the questions composing the domain for content
tion. Likewise, the previous study of self-efficacy scale relevance, representativeness, and scientific quality.
development also applies this approach to create the ques- Experts from the stroke or related field will be invited to
tion, and it provided initial support for the validity and evaluate each question to determine whether they repre-
reliability.36 sent the domain of interest. Moreover, the content validity
STROKE RISK SCREENING SCALES (SRSS) 7

ratio, content validity index, Cohen's coefficient alpha, or 13. Risk Assessment Guide for Ischemic Heart Disease and
Delphi method will be used as appropriate.16 Cerebrovascular Disease: The Bureau of Non-Communi-
cable Disease, Department of Disease Control, Ministry
of Health of Thailand; 2017 [cited 2020. Available from:
Funding http://thaincd.com/document/file/info/non-communi-
cable-disease/%E0%B8%84%E0%B8%B9%E0%B9%88
No funding was received for this study. %E0%B8%A1%E0%B8%B7%E0%B8%AD%E0%B8%81
%E0%B8%B2%E0%B8%A3%E0%B8%9B%E0%B8%A3
%E0%B8%B0%E0%B9%80%E0%B8%A1%E0%B8%B4%E0
Declaration of Competing Interest %B8%99%E0%B9%82%E0%B8%AD%E0%B8%81%E0
There are no conflicts of interest to declare. %B8%B2%E0%B8%AA%E0%B9%80%E0%B8%AA%E0
%B8%B5%E0%B9%88%E0%B8%A2%E0%B8%87%E0
%B8%95%E0%B9%88%E0%B8%AD%E0%B8%81%E0
Supplementary materials %B8%B2%E0%B8%A3%E0%B9%80%E0%B8%81%E0
%B8%B4%E0%B8%94%E0%B9%82%E0%B8%A3%E0
Supplementary material associated with this article can %B8%84%E0%B8%AB%E0%B8%B1%E0%B8%A7%E0
be found in the online version at doi:10.1016/j.jstrokecere %B9 %83%E0%B8%88_%E0%B8%AA%E0 %B8%A1%E0
brovasdis.2021.105740. %B8%AD%E0%B8%87%E0%B8%AF.pdf.
14. Crook ED, Washington DO, Flack JM. Screening and pre-
vention of chronic kidney disease. J Natl Med Assoc
References 2002;94:55S.
15. Grunfeld E, Manca D, Moineddin R, et al. Improving
1. Chen Y, Wright N, Guo Y, et al. Mortality and recurrent chronic disease prevention and screening in primary
vascular events after first incident stroke: a 9-year com- care: results of the BETTER pragmatic cluster random-
munity-based study of 0¢5 million Chinese adults. The ized controlled trial. BMC Family Pract 2013;14:175.
Lancet Glob Health 2020;8:e580-e590. 16. Boateng GO, Neilands TB, Frongillo EA, Melgar-Qui-
2. Feigin VL, Krishnamurthi RV, Parmar P, et al. Update on ~ onez HR, Young SL. Best practices for developing and
n
the global burden of ischemic and hemorrhagic stroke in validating scales for health, social, and behavioral
1990 2013: the GBD 2013 study. Neuroepidemiology research: a primer. Front Public Health 2018;6:149.
2015;45:161-176. 17. Raykov T, Marcoulides GA. Introduction to Psychometric
3. Suwanwela NC. Stroke epidemiology in Thailand. J Theory. Routledge; 2011.
Stroke 2014;16:1. 18. Brown R, Whisnant JP, Sicks J, et al. Stroke incidence,
4. Jitnarin N, Kosulwat V, Rojroongwasinkul N, et al. Risk prevalence, and survival: secular trends in Rochester,
factors for overweight and obesity among Thai adults: Minnesota, through 1989. Stroke 1996;27:373-380.
results of the National Thai Food Consumption Survey. 19. Jerntorp P, Berglund G. Stroke registry in Malm€ o, Swe-
Nutrients 2010;2:60-74. den. Stroke 1992;23:357-361.
5. Kosulwat V. The nutrition and health transition in Thai- 20. Horner RD, Matchar DB, Divine G, et al. Racial variations
land. Public Health Nutr 2002;5:183-189. in ischemic stroke-related physical and functional impair-
6. Nijasri Charnnarong Suwanwela NE, Phanthumchinda ments. Stroke 1991;22:1497-1501.
K, Piravej K, Locharoenkul C. Combination of acute 21. Singh JK, Ranjan P, Kumari A, et al. Types, outcome and
stroke unit and short-term stroke ward with early sup- risk factors of stroke in tribal patients. Int J Stroke
ported discharge decreases mortality and complications 2013;8:675-680.
after acute ischemic stroke. J Med Assoc Thail 22. Lee CD, Folsom AR, Blair SN. Physical activity and
2007;90:1089. Chotmaihet thangphaet. stroke risk: a meta-analysis. Stroke 2003;34:2475-2481.
7. Artinuch Benjarak PK, Naraongat P, Ruksakulpiwat S, 23. Truelsen T, Nielsen N, Boysen G, et al. Self-reported
Singha-Dong N. Epidemiology and Health care of stroke stress and risk of stroke: the Copenhagen City Heart
in Nakhon Ratchasima, Thailand. In: Proceedings of the Study. Stroke 2003;34:856-862.
International Conference: Moving Towards the New Era 24. Hunt SD. Modern Marketing Theory: Critical Issues in
of NCDs and Global Health; Phuket, Thailand 2016. the Philosophy of Marketing Science. South-Western
8. Hippisley-Cox J, Coupland C, Brindle P. Derivation and Pub; 1991.
validation of QStroke score for predicting risk of ischaemic 25. Fowler Jr FJ, Fowler FJ. Improving Survey Questions:
stroke in primary care and comparison with other risk Design and Evaluation. Sage; 1995.
scores: a prospective open cohort study. BMJ 2013:346. 26. Morris S. A comparison of economic modelling and
9. Nobel L, Mayo NE, Hanley J, et al. MyRisk_Stroke calcu- clinical trials in the economic evaluation of choles-
lator: a personalized stroke risk assessment tool for the terol-modifying pharmacotherapy. Health Econ
general population. J Clin Neurol 2014;10:1-9. 1997;6:589-601.
10. Srikha D, Yothathoon SA, Nualnetr N. Situation of risk of 27. Roze S, Liens D, Palmer A, et al. A health economic
stroke in Samliam community, Muang District, Khon model to determine the long-term costs and clinical out-
Kaen Province. J Med Technol Phys Ther 2011;23: comes of raising low HDL-cholesterol in the prevention
159-164. of coronary heart disease. Curr Med Res Opin
11. Parmar P, Krishnamurthi R, Ikram MA, et al. The S troke 2006;22:2549-2556.
riskometer TM app: validation of a data collection tool 28. Stollenwerk B, Gerber A, Lauterbach KW, et al. The Ger-
and stroke risk predictor. Int J Stroke 2015;10:231-244. man coronary artery disease risk screening model: devel-
12. Wolf PA, D'Agostino RB, Belanger AJ, et al. Probability of opment, validation, and application of a decision-
stroke: a risk profile from the Framingham Study. Stroke analytic model for coronary artery disease prevention
1991;22:312-318. with statins. Med Decis Mak 2009;29:619-633.
8 S. RUKSAKULPIWAT

29. Georgakis MK, Gill D, Webb AJ, et al. Genetically deter- vascular chronic diseases in 0¢5 million adults in China: a
mined blood pressure, antihypertensive drug classes, and prospective cohort study. Lancet Glob Health 2018;6:
risk of stroke subtypes. Neurology 2020;95:e353-e361. e641-e649.
30. O’Donnell MJ, Chin SL, Rangarajan S, et al. Global and 34. Sorganvi V, Kulkarni M, Kadeli D, et al. Risk factors for
regional effects of potentially modifiable risk factors asso- stroke: a case control study. Int J Curr Res Rev 2014;6:46.
ciated with acute stroke in 32 countries (INTERSTROKE): 35. Hinkin TR. A review of scale development practices in
a case-control study. Lancet 2016;388:761-775. the study of organizations. J Manag 1995;21:967-988.
31. Krishnamurthi R, Feigin VL. Global stroke burden and 36. Dennis CL, Faux S. Development and psychometric test-
prevention strategies. Oxford Textb Neurol Neuropsy- ing of the Breastfeeding Self-Efficacy Scale. Res Nursing
chiatr Epidemiol 2020:183. Health 1999;22:399-409.
32. Lewington S, Clarke R, Qizilbash N, et al. Age-specific 37. Kline P. The Handbook of Psychological Testing. Psy-
relevance of usual blood pressure to vascular mortality: a chology Press; 2000.
meta-analysis of individual data for one million adults in 38. Schinka JA, Velicer WF, Weiner IB. Handbook of Psychol-
61 prospective studies. The Lancet 2002;360:1903-1913. ogy: Research Methods in Psychology, 2. John Wiley &
33. Lacey B, Lewington S, Clarke R, et al. Age-specific associ- Sons Inc; 2013.
ation between blood pressure and vascular and non-