Articles

A community-based comprehensive intervention to

reduce cardiovascular risk in hypertension (HOPE 4):
a cluster-randomised controlled trial
Jon-David Schwalm, Tara McCready, Patricio Lopez-Jaramillo, Khalid Yusoff, Amir Attaran, Pablo Lamelas, Paul A Camacho, Fadhlina Majid,
Shrikant I Bangdiwala, Lehana Thabane, Shofiqul Islam, Martin McKee, Salim Yusuf

Summary
Background Hypertension is the leading cause of cardiovascular disease globally. Despite proven benefits, hypertension Published Online
control is poor. We hypothesised that a comprehensive approach to lowering blood pressure and other risk factors, September 2, 2019
http://dx.doi.org/10.1016/
informed by detailed analysis of local barriers, would be superior to usual care in individuals with poorly controlled or
S0140-6736(19)31949-X
newly diagnosed hypertension. We tested whether a model of care involving non-physician health workers (NPHWs),
See Online/Comment
primary care physicians, family, and the provision of effective medications, could substantially reduce cardiovascular http://dx.doi.org/10.1016/
disease risk. S0140-6736(19)31995-6
Population Health Research
Methods HOPE 4 was an open, community-based, cluster-randomised controlled trial involving 1371 individuals with Institute, McMaster University
new or poorly controlled hypertension from 30 communities (defined as townships) in Colombia and Malaysia. and Hamilton Health Sciences,
Hamilton, ON, Canada
16 communities were randomly assigned to control (usual care, n=727), and 14 (n=644) to the intervention. After (J-D Schwalm MD,
community screening, the intervention included treatment of cardiovascular disease risk factors by NPHWs using T McCready PhD, P Lamelas MSc,
tablet computer-based simplified management algorithms and counselling programmes; free antihypertensive and Prof S I Bangdiwala PhD,
S Islam PhD, Prof S Yusuf DPhil);
statin medications recommended by NPHWs but supervised by physicians; and support from a family member or
Research Institute, Fundación
friend (treatment supporter) to improve adherence to medications and healthy behaviours. The primary outcome was Oftalmológica de Santander,
the change in Framingham Risk Score 10-year cardiovascular disease risk estimate at 12 months between intervention Floridablanca, Colombia
and control participants. The HOPE 4 trial is registered at ClinicalTrials.gov, NCT01826019. (Prof P Lopez-Jaramillo PhD,
P A Camacho MD); Masira
Institute, Medical School,
Findings All communities completed 12-month follow-up (data on 97% of living participants, n=1299). The reduction Universidad de Santander,
in Framingham Risk Score for 10-year cardiovascular disease risk was –6·40% (95% CI 8·00 to –4·80) in the control Bucaramanga, Colombia
group and –11·17% (–12·88 to –9·47) in the intervention group, with a difference of change of –4·78% (95% CI (Prof P Lopez-Jaramillo);
Faculty of Medicine, Universiti
–7·11 to –2·44, p<0·0001). There was an absolute 11·45 mm Hg (95% CI –14·94 to –7·97) greater reduction in
Teknologi MARA, Selayang,
systolic blood pressure, and a 0·41 mmol/L (95% CI –0·60 to –0·23) reduction in LDL with the intervention group Selangor, Malaysia
(both p<0·0001). Change in blood pressure control status (<140 mm Hg) was 69% in the intervention group versus (Prof K Yusoff MBBS,
30% in the control group (p<0·0001). There were no safety concerns with the intervention. F Majid BSc); Faculty of Medicine
and Health Sciences, UCSI,
Kuala Lumpur, Malaysia
Interpretation A comprehensive model of care led by NPHWs, involving primary care physicians and family that was (Prof K Yusoff); Faculty of Law
informed by local context, substantially improved blood pressure control and cardiovascular disease risk. This strategy (Prof A Attaran DPhil),
is effective, pragmatic, and has the potential to substantially reduce cardiovascular disease compared with current and Faculty of Medicine
(Prof A Attaran), University of
strategies that are typically physician based. Ottawa, Ottawa, ON, Canada;
Medical School, Universidad
Funding Canadian Institutes of Health Research; Grand Challenges Canada; Ontario SPOR Support Unit and the Autónoma de Bucaramanga,
Ontario Ministry of Health and Long-Term Care; Boehringer Ingelheim; Department of Management of Bucaramanga, Colombia
(P A Camacho); Department of
Non-Communicable Diseases, WHO; and Population Health Research Institute. Health Research Methods,
Evidence and Impact, McMaster
Copyright © 2019 Elsevier Ltd. All rights reserved. University Faculty of Health
Sciences, Hamilton, ON, Canada
(Prof S I Bangdiwala,
Introduction countries at all levels of development that many people Prof L Thabane PhD, Prof S Yusuf);
Cardiovascular disease is the most common cause of with hypertension are unaware of it, or are untreated and Department of Health
death globally, and hypertension the most common risk or uncontrolled.4 Consequently, fewer than 20% of Services Research and Policy,
factor for cardiovascular disease.1 In theory, there is no individuals with hypertension have their blood pressure London School of Hygiene &
Tropical Medicine, London, UK
reason why the goal of reducing cardiovascular disease controlled. Similarly, despite compelling evidence of the (Prof M McKee DSc)
mortality by 30% by 2030, agreed to by the governments benefits of statins in those with hypertension, uptake is Correspondence to:
of the world,2 should not be met. Effective, inexpensive extremely low in most parts of the world.5 Dr Jon-David Schwalm,
medicines to reduce risk factors have been available for This problem shares similarities with HIV a decade ago: Population Health Research
many years.3 The challenge we face is how to identify treatment was available, but people in need were often Institute, McMaster University
and Hamilton Health Sciences,
those who might benefit from treatment, and ensure that unable to benefit. The global health community responded Hamilton, ON, L8L 2X2 Canada
they are treated. We now have extensive information from by investing in health system strengthening, seeking to schwalj@mcmaster.ca

www.thelancet.com Published online September 2, 2019 http://dx.doi.org/10.1016/S0140-6736(19)31949-X 1

 Articles
Research in context
Evidence before this study
Hypertension is the leading cause of cardiovascular disease, with
the majority of the burden in low-income and middle-income
countries. The HOPE 4 programme was initiated in 2013,
to address the burden of cardiovascular risk and hypertension.
Initial phases of the programme included systematic reviews
and mixed methods analysis to inform the design of the HOPE 4
intervention components to be tested in a cluster randomised
controlled trial. We systematically searched electronic databases
including MEDLINE, Embase, Global Health, LILACS, Africa-Wide
Information, IMSEAR, IMEMR, and WPRIM from inception until
May 8, 2013, to identify barriers to appropriate hypertension
control at the patient, health-care provider, and health-system
level. No limits were applied with respect to language.
Controlled vocabulary, keywords (MeSH terms) and free-text
terms were identified for each domain of our health systems
framework which also focused on “hypertension’’, ‘‘barriers’’,
and ‘‘obstacles’’. No limits to study design were imposed.
The barriers identified in the systematic review, coupled with
the findings of qualitative health-system appraisals in Colombia
create effective mechanisms to deliver treat­ment to those
who could benefit at low cost.6 So far, however, this
approach has not been matched by those seeking to
reduce the preventable burden of cardiovascular disease.
By drawing on the literature on health systems
research, we can identify what such a response would
look like. First, it would address all of the building blocks
of health systems, and in particular the need for an
See Online for video appropriate workforce, access to affordable diagnostics
and medicines, and evidence-based guidelines. Second,
it should be tailored to the national context, recognising
that health systems are embedded within wider systems
of governance, beliefs, and norms. The health outcomes
prevention and evaluation 4 (HOPE 4) project is, to
our knowledge, the first attempt to do this.7 Working in
two middle-income countries, Colombia and Malaysia,
we have developed, implemented, and evaluated, in a
cluster-randomised trial, a comprehensive model of care
that takes full account of the local context. This innovative
package has been designed to address the constrained
resources affecting most countries based on systematic
reviews of barriers to effective management of hyper­
tension.8,9 Thus, we used non-physician health workers
(NPHWs) supported by physicians and, crucially, family
and community members; a simple combination of anti­
hypertensive drugs and a statin; and simplified guide­
lines, delivered through a tablet computer (henceforth
called tablets).
Many of these components have been separately
evaluated in previous research, although with mixed
results.10,11 Our intervention is innovative in how, by taking
a systems approach, it is more than the sum of its parts.
Importantly, our intervention is informed by an initial
2 www.thelancet.com Published online September 2, 2019 http://dx.doi.org/10.1016/S0140-6736(19)31949-X
and Malaysia led to the development of an intervention
comprising of modification of health behaviours and initiation
of free combination antihypertensive drugs plus statins,
by non-physician health workers guided by a tablet
computer-based decision support system, supervised by
physicians, and involving an individual’s family or friends to
promote adherence.
Added value of this study
The HOPE 4 study shows that a comprehensive model of care
that is informed by strategies to overcome country-specific
barriers, resulted in a substantial reduction in cardiovascular
disease risk and improved blood pressure control.
Implications of all the available evidence
Adoption of the HOPE 4 strategy could substantially enhance
reduction in cardiovascular disease risk in those with
hypertension, and in doing so help achieve the UN’s General
Assembly target that calls for a one-third reduction in
premature cardiovascular disease mortality by 2030.
detailed health system assessment and barrier analysis in
each country,12,13 which used a combination of quantitative
and qualitative research to identify the challenges that
needed to be overcome in designing the intervention.
In this study, we tested the effectiveness of this compre­
hensive intervention in reducing cardiovascular disease
risk among people with hypertension in two middle-
income countries.
A video abstract is available online.
Methods
Study design and clusters
The HOPE 4 study was a parallel-arm, cluster-random­
ised controlled trial done in 30 urban and rural
communities in Colombia (Fundación Oftalmológica de
Santander) and Malaysia (Universiti Teknologi MARA).
We have previously reported details for the study
rationale, trial design, and methods.7 We have also
reported the results of the qualitative research that
identified barriers and potential solutions to improving
hypertension control, which were the basis for the
development of the multifaceted intervention in the
HOPE 4 trial.8,9,12,13 After screening and enrolment in
each community was completed, community clusters
were randomly assigned to usual care or to participate
in an active cardiovascular disease risk detection and
management programme facilitated by NPHWs for
12 months. The selection of study communities was
informed by country-specific leads to ensure that they
were geographically separated. This separation was done
to reduce contamination bias between clusters (ie,
participants from the control arm communities seeking
health care in the intervention arm communities). The

communities were rural and urban townships identified
by the local governments.
Study population
Screening to identify eligible participants in potential
communities involved a combination of door-to-door
household screening and the use of community outreach
centres or local events in public spaces. Eligible par­
ticipants were aged at least 50 years, with at least one of
the following criteria: average systolic blood pressure
(SBP) of at least 160 mm Hg recorded at one visit; average
SBP of 140–159 mm Hg recorded at one visit and
participant reported a diagnosis of hypertension or was
taking antihypertensive medication; average SBP of at
least 130 mm Hg recorded at one visit and participant
reported a diagnosis of diabetes or was taking medications
for diabetes; did not meet earlier criteria, but had an
average SBP of 140–159 mm Hg recorded on two separate
visits, at least one day apart.
Participants were ineligible if they refused consent,
were concurrently participating in any other study or heart
health programme that would interfere with partici­
pation in HOPE 4, had severe comorbid conditions with
life expectancy of less than 1 year, or had other serious
conditions or factors likely to interfere with study
participation or with their ability to complete the trial.
Before study initiation, the institutional review board or
independent ethics committee of participating institu­
tions approved the HOPE 4 protocol. Reporting in this
publication is consistent with the CONSORT statement.14
Randomisation and masking
A computer-generated, central randomisation system
(located at the Population Health Research Institute,
Hamilton, ON, Canada), allocated communities (1:1) to
either the intervention group or control group. Ran­
domisation was first stratified by country and then by
rural or urban location of the community. Each
community was randomly assigned after the entire
community was screened, and written consent was
obtained to avoid biases in the selection or volunteering
of individuals within the intervention and control
communities. Participants, NPHWs, primary care physi­
cians, and local investigators were not masked to the
study group allocation. Furthermore, lipids were analysed
centrally within each country, without knowledge of
whether the samples were from participants in the
control or intervention communities. Blood pres­sure was
measured as per the WHO STEPS protocol (ie, three
separate readings, in which the last two were averaged) to
ensure consistency of recordings, and an automated
Omron monitor was used at each visit.15 These procedures
avoid potential biases with respect to blood pressure
and lipid measurements. Variability was minimised by
standardised NPHW curriculum training. Furthermore,
outcome assessments in the control and intervention
groups were at similar timepoints.
www.thelancet.com Published online September 2, 2019 http://dx.doi.org/10.1016/S0140-6736(19)31949-X 3
Articles
Procedures
The HOPE 4 study involved a multifaceted intervention
consisting of three core elements provided as a package:
community screening, detection, treatment, and control
of cardiovascular disease risk factors by NPHWs (in
collaboration with local physicians) guided by tablet-based
simplified management algorithms, decision support,
and counselling programmes; provision of free locally
available combination antihypertensive medications and
a statin recommended by NPHWs, but supervised by local
physicians; and support from a participant-nominated
treatment supporter (friend or family member) to improve
adherence to medications and health behaviours.7–9,12,13,16
The NPHWs did the initial screening, recruitment,
and follow-up in both the intervention and control
communities. No structured intervention was provided
in the control group, but they did receive existing
local cardiovascular disease health literature and were
recommended to see their local health-care provider as
usual. The NPHWs in both the intervention and control
communities were a combination of newly hired and
retrained community health workers and research staff.
The NPHWs were paid for by the HOPE 4 study
with partial financial support from local governments.
The NPHWs were expected to have a minimum of a
secondary school diploma or equivalent. The 1-week
training curriculum, which was contextualised to each
country, has been previously published.17
Study medications included a generic single-pill
combination of two antihypertensives (at half or full
doses). Combinations included an angiotensin receptor
blocker or angiotensin converting enzyme inhibitor
coupled with a diuretic or calcium channel blocker
(appendix p 2). A separate generic cholesterol-lowering See Online for appendix
agent (atorvastatin at 20 mg or rosuvastatin at 10 mg) was
also provided (appendix p 2).7 These medications were
also available in local pharmacies for about US$5–8 each
in Malaysia and $14–38 each in Colombia, per 1-month
supply. Both countries offer subsidies for medications
through the public health-care system. Following NPHW
recommendation, and medication prescription pro­
vided by the primary care physicians, medications were
dispensed either from pharmacies, by a professionally
trained pharmacist, or the physician’s office. NPHWs
were responsible for delivering or arranging for
participants to obtain the medications through home or
local clinic visits, on the basis of participant preference
and local context. Participants already taking cardio­
vascular medication were reviewed by the NPHW and
supervising physician for transition (if appropriate) to
the medications provided by the study. Study-associated
medications were recommended, but not mandatory, as
part of the study protocol. These medications were
provided at no cost to the participant if they chose to
use them. Participating NPHWs and physicians were
provided with simplified medication management
algorithms guiding their use by the study (appendix p 3).7
 Articles

However, initiation or up-titration of antihypertensives

4904 individuals from 30 participating provided by the study was optional, and the ultimate
communities (clusters)* screened
for eligibility decisions about treatment in complex or uncertain cases
were left to the treating physician.
The NPHW visited participants in their home, or
3004 excluded
2938 did not meet study criteria
at a local clinic. At 6 months and 12 months after
66 undetermined† randomisation, the NPHW repeated study assessments
in all participants in both the intervention and control
communities and recorded the components of the
1900 eligible individuals identified
Framingham Risk Score (FRS).18 Fasting blood was
collected in participants at baseline and 12 months and
524 excluded analysed for total cholesterol, HDL, LDL, triglycerides,
509 refused consent
15 pre-randomisation deaths and glucose at a single core laboratory in each country
that was affiliated with the coordinating sites. Par­
ticipants in communities randomly assigned to the
1376 participants enrolled intervention group had additional visits at baseline
(one with the NPHW and one with the local physician),
5 pre-randomisation deaths and at 1 month and 3 months after randomisation.
All participants were contacted (via telephone) at
10 months or 11 months after randomisation to confirm
1371 participants from 30 communities
randomly assigned availability for the 12-month visit. Data were collected
using study tablets, which had interactive data
validation measures.

727 participants from 16 communities 644 participants from 14 communities Outcomes

allocated to control* allocated to intervention*
The primary outcome was the mean difference in FRS
10-year cardiovascular disease risk estimate change
35 excluded because 12-month 37 excluded because 12-month from baseline to 12 months, between the intervention
data on risk factors not data on risk factors not and control groups.18 The FRS was chosen as the
available available
15 lost to follow-up 13 lost to follow-up and primary outcome because it is a validated risk score that
and refusals refusals considers several cardiovascular disease risk factors
15 post-randomisation 12 post-randomisation
deaths deaths
addressed by our intervention.19,20 Secondary outcomes
5 withdrawals 12 withdrawals included differences in the changes in SBP between
the inter­vention and control groups at 6 months and
12 months; differences in the changes in LDL, HDL,
692 participants with primary outcome 607 participants with primary outcome
at 12 months at 12 months total cholesterol, triglycerides, and glucose concen­
trations at 12 months; proportion of participants
Figure 1: Trial profile with well controlled blood pressure at 6 months and
*Of the 30 participating communities, 15 were from Malaysia and 15 were from Colombia. For both countries, 12 months (SBP <140 mm Hg); and change in smoking
seven communities were randomly assigned to intervention and eight communities were randomly assigned to status at 6 months and 12 months. Tertiary outcomes
control. †These participants withdrew their consent form before the second screening. Eligibility could therefore
not be determined.
included change in INTERHEART Risk Score21 at
6 months and 12 months and Cholesterol Modifiable
Risk Score at 12 months; change in lifestyle modification
Control (n=727) Intervention (n=644)
(ie, exercise and diet, based on components of the
Age, years 65·8 (9·7) 65·1 (9·1) INTERHEART Risk Score); proportion of par­ticipants
Sex receiving two or more antihypertensive drugs and a
Female 395 (54%) 372 (58%) statin at 6 months and 12 months; and medication
Male 332 (46%) 272 (42%) adherence at 6 months and 12 months (adherence was
Education assessed using the Morisky Medication Adherence Scale
None, primary, or unknown 499 (69%) 446 (69%) 8 in participants actively taking antihypertensive drugs
Secondary or high school 201 (28%) 170 (26%) in both groups; appendix p 4).22 Medication adherence
Trade, college, or university 27 (4%) 28 (4%) was reinforced by NPHWs and treatment supporters as
Past medical history (self-reported) part of the study protocol.7 Safety outcomes were any
Current smoker 68 (9%) 50 (8%) serious adverse events reported in the intervention
Diabetes 270 (37%) 206 (32%) group and clinical events (death, myocardial infarction,
(Table 1 continues on next page) stroke, and cardio­ vascular-related hospitalisations) at
6 months and 12 months in both groups.

4 www.thelancet.com Published online September 2, 2019 http://dx.doi.org/10.1016/S0140-6736(19)31949-X

 Articles

Statistical analysis
Control (n=727) Intervention (n=644)
The study planned to include 1200 participants (40 par­
ticipants from each of the 30 communities; 600 in the (Continued from previous page)
intervention group and 600 in the control group). Based History of known hypertension 523 (72%) 485 (75%)
on findings from PURE study participants who met History of known hypertension and taking 441 (61%) 444 (69%)
antihypertensive medication
criteria for HOPE 4,5 we assumed a within-community
Atrial fibrillation 5 (1%) 12 (2%)
intraclass correlation coefficient of 0·05 for all outcomes,
Stroke 15 (2%) 15 (2%)
a baseline FRS of 24·7% (SD 15%), a mean SBP of
Myocardial infarction 30 (4%) 25 (4%)
160 mm Hg (SD 19), and a mean LDL of 3·58 mmol/L
Angina 19 (3%) 14 (2%)
(SD 1·09 mmol/L). Power was 80% to detect at least
a 3·0% absolute difference in the FRS, a 2·9 mm Hg Congestive heart failure 29 (4%) 29 (5%)

difference in blood pressure, and a 0·32 mmol/L dif­ Liver disease 2 (<1%) 8 (1%)

ference in LDL between the two randomised groups. Kidney disease 12 (2%) 18 (3%)
The analysis of both primary and secondary outcomes Baseline medication use
accounted for the clustering effect and used a mixed One antihypertensive 185 (25%) 205 (32%)
effects model considering community as a random Two antihypertensives 155 (21%) 154 (24%)
intercept effect. An intention-to-treat analysis was done.23 Three antihypertensives 82 (11%) 77 (12%)
Data at 12 months were available from 97% of participants Four or more antihypertensives 28 (4%) 22 (3%)
who were confirmed to be living. We used multiple Statin 195 (27%) 186 (29%)
imputation based on information recorded at baseline to Criteria for enrolment into the study
take into account the 3% missing data for the primary SBP ≥160 mm Hg at one visit 195 (27%) 175 (27%)
outcome as a sensitivity analysis. Two-sided tests at the SBP 140–159 mm Hg in one visit and previous diagnosis 315 (43%) 320 (50%)
0·05 level were considered to be statistically significant. of hypertension or use of antihypertensive medication
Final results are expressed as mean (95% CI) in each SBP ≥130 mm Hg in one visit and a diagnosis of diabetes 125 (17%) 73 (11%)
or use of medications for diabetes
group, and two-sided p values.
Did not meet above criteria but SBP of 140–159 mm Hg 92 (13%) 76 (12%)
The consistency of treatment effects on the primary recorded on two separate visits, at least one day apart
outcome were explored in predefined subgroups, Physical measurements
including sex, education, country, urban or rural com­ SBP, mm Hg 151·7 (15·6) 152·1 (15·4)
munity, and age younger or older than 65 years.7 We
DBP, mm Hg 85·3 (11·9) 84·7 (12·0)
tested whether the effects varied by subgroups using
Waist, cm 93·6 (11·6) 93·2 (10·8)
tests of interactions between subgroup variables and
Hip, cm 100·2 (10·5) 100·2 (10·4)
treatment group. All analyses were done using SAS,
Waist to hip ratio 0·94 (0·1) 0·93 (0·1)
version 9.2.
Body-mass index, kg/m² 27·7 (5·1) 28·3 (5·3)
Laboratory measures, mmol/L
Role of the funding source
Total cholesterol 5·4 (1·2) 5·4 (1·2)
The authors are solely responsible for the design and
LDL 3·4 (1·1) 3·3 (1·1)
conduct of this study, all study analyses, the drafting and
HDL 1·1 (0·4) 1·2 (0·4)
editing of the Article, and its final contents. The external
Triglycerides 2·0 (1·4) 2·0 (1·4)
funders had no role in the design, conduct, analyses, or
Glucose 6·8 (3·4) 6·5 (3·0)
interpretation of the results. SY, J-DS, SI, TM, PL, and
SIB had full access to the study data. SY and J-DS were Framingham Risk Score 10-year risk estimate

jointly responsible for the decision to submit the Article. 10-year risk estimate 35·5% (22·3) 32·6% (21·4)
<30% 371 (51%) 379 (59%)
Results 30–40% 121 (17%) 98 (15%)
Between 2014 and 2017, 4904 participants were screened ≥40% 235 (32%) 167 (26%)
from 30 communities in Colombia and Malaysia Data are mean (SD) or n (%). SBP=systolic blood pressure. DBP=diastolic blood pressure.
(15 each). All screening in Colombia was door to door,
whereas screening of most participants (n=2383, 88%) Table 1: Baseline characteristics of participants
in Malaysia was done at non-medical community events.
1900 people were invited to participate in the study, and analyses; thus, 1371 participants were included in the
1376 (72%) provided written informed consent. The final analysis. 16 communities (n=727) were randomly
proportion of eligible participants versus those enrolled assigned to the control group and 14 (n=644) were
was higher in Colombia (n=616, 86%) than Malaysia randomly assigned to the intervention group. Similar
(n=760, 60%). However, the results were clear, with a (low) numbers of deaths, withdrawals, and losses to
similar magnitude and statistical sig­nificance in both follow-up were noted in both groups. Both groups
countries. Five participants died before random­isation were followed for 12 months and all 30 clusters
of their community, and were not included in the completed follow-up. At study completion, 1299 (97% of

www.thelancet.com Published online September 2, 2019 http://dx.doi.org/10.1016/S0140-6736(19)31949-X 5

 Articles
participants confirmed to be alive; n=692 control, n=607
intervention) completed 12-month follow-up visits
(figure 1).
Baseline characteristics of the study participants
were generally similar between the randomised groups
(table 1). However, there were more smokers (n=68,
9·4% vs n=50, 7·8%) and individuals with diabetes
(n=270, 37% vs n=206, 32%) in the control group.
The intervention and control communities were well
balanced, with respect to mean age, population size, and
access to primary and secondary medical facilities within
each community (table 2). The majority of participants
(n=1008, 73·5%) had a history of hypertension and were
taking anti­ hypertensive medications, but their blood
pressure was not controlled. The remaining patients
had newly diagnosed hypertension (table 1). The mean
duration of fasting before obtaining blood samples was
similar in both groups (control 9·7 h, SD 3·0); inter­
vention 9·8 h (SD 2·9)
NPHWs were found to be consistently accurate in their
ability to identify cardiovascular risk (patient identified by
NPHWs as having poorly controlled blood pressure and
medication was indicated) and recommend appropriate
therapies (antihypertensives and statin as per the study
algorithm) when compared with the assessment by local
primary care physicians. Of 3177 assessments in the
intervention group, there was a high proportion of
agreement between the NPHWs and physicians on the
participants, cardiovascular risk (n=3129; 99%) and
identification of contraindications to antihypertensives
and statin (n=3123; 98%). Of 3116 assessments, the
NPHWs and physicians agreed on the need to initiate
antihypertensive and statin medications based on the
Control
(n=16)
Distance from the coordinating 68·4 (72·0)
centre, km
Population of communities 56·9 (90·0)
included, thousands
Number of clinics 6·9 (8·4)
Public clinics 2·1 (1·2)
Private clinics 9·4 (9·0)
Number of hospitals 1·5 (1·2)
Public hospitals 1·0 (0·0)
Private hospitals 1·0 (1·6)
Tertiary care centres 0·3 (0·5)
Specialist centres 0·1 (0·5)
Sex
Female 50·7% (3·2)
Male 49·3% (4·6)
Age ≥50 years in the population 25·9% (3·9)
Data are mean (SD). *Data from the Bureau of Statistics in Malaysia and the
Department of Statistics in Colombia.
Table 2: Characteristics of intervention and control communities*
6 www.thelancet.com Published online September 2, 2019 http://dx.doi.org/10.1016/S0140-6736(19)31949-X
simplified management algorithms on 2886 occasions
(93%).
There was high fidelity of the intervention imple­
mentation. Participants attended 2970 (94%) of the total
planned NPHW visits (n=3160) in the intervention
group. Additionally, a greater proportion of participants
in the intervention group were taking two or more
antihypertensive medications than the control group at
6 months (n=451 [87%] vs n=317 [65%]; p<0·0001) and
12 months (n=463 [84%] vs n=322 [65%]; p<0·0001).
A significantly higher proportion of statin use was
reported in participants in the intervention group
than the control group at 6 months (n=506 [86%] vs
n=245 [36%]; p<0·0001) and 12 months (n=511 [84%] vs
n=262 [38%]; p<0·0001). In the intervention group,
346 (54%) participants and 314 (49%) participants were
taking the free study-associated combination anti­hyper­
tensive and statin medication, respectively, at the first
follow-up visit. 156 (45%) participants taking the study-
associated antihypertensives were on the high-dose
combination at 1 year. 44 (7%) participants in the
intervention group were taking the three-drug
antihypertensive combina­tion at any visit. Adherence to
prescribed antihyper­tensives in the intervention group
was higher than the control group at 6 months
(n=310 [56%] vs n=172 [35%]; p<0·0001) and 12 months
(n=351 [61%] vs n=200 [40%]; p<0·0001). Treatment sup­
porters were most commonly a spouse (n=228, 48%),
offspring (n=159, 34%), other family (n=51, 11%) or
another individual (n=34, 7%) and were present at 971
(74%) visits by the NPHWs in the intervention group.
The reduction in FRS 10-year cardiovascular disease
risk was –6·40% (95% CI 8·00 to –4·80) in the control
group and –11·17% (–12·88 to –9·47) in the intervention
group, with a difference of change of –4·78% (95% CI
Intervention –7·11 to –2·44; p<0·0001; table 3). In the intervention
(n=14) group, there was a 34·2% relative reduction in the FRS
68·3 (78·0) estimate compared with baseline. When considering an
intention-to-treat analysis, using multiple imputation
58·3 (127·0)
to account for missing data at 12 months, the results
were unchanged (FRS at 12 months –6·29 vs –11·34;
6·6 (8·2)
p<0·0001). When body-mass index was used in place of
1·9 (1·2)
missing HDL and total cholesterol values (7% of
9·4 (8·0)
participants) at 12 months, the results were also similar.
1·5 (1·4)
The effects on the primary outcome were consistent in
1·1 (0·4)
all predefined subgroups (with statistically significant
1·1 (1·3)
differences within both countries), with no evidence of
0·5 (0·7)
heterogeneity of the effects of intervention (figure 2).
0·1 (0·3)
Exploratory post-hoc subgroup analysis considering
baseline blood pressure, pre-existing hypertension,
49·4% (2·4) diabetes, and smoking status showed consistent effects
50·6% (2·3) on the primary outcome (appendix p 5).
23·3% (4·1) Most of the secondary outcomes showed reductions
consistent with the primary outcome. There was a greater
absolute reduction in SBP by 11·45 mm Hg (95% CI
–14·94 to –7·97; figure 3), by 0·45 mmol/L in total
cholesterol, and by 0·41 mmol/L (95% CI –0·60 to –0·23)
 Articles

Baseline Change at 12 months from baseline* Difference in change Test between

between intervention intervention and
and control control† (p value)
Control (n=727) Intervention (n=644) Control (n=692) Intervention (n=607)
Mean Framingham Risk Score 10-year 35·47% (22·34) 32·63% (21·37) –6·40% (–8·00 to –4·80) –11·17% (–12·88 to –9·47) –4·78% (–7·11 to –2·44) <0·0001
risk estimate
Cholesterol modifiable risk score 12·39 (5·51) 11·85 (5·17) –1·64 (–2·26 to –1·02) –4·75 (–5·41 to –4·09) –3·11 (–4·01 to –2·20) <0·0001
Total cholesterol, mmol/L 5·37 (1·20) 5·35 (1·21) –0·23 (–0·35 to –0·12) –0·68 (–0·81 to –0·56) –0·45 (–0·62 to –0·28) <0·0001
LDL, mmol/L 3·38 (1·11) 3·34 (1·13) –0·19 (–0·32 to –0·06) –0·60 (–0·74 to –0·47) –0·41 (–0·60 to –0·23) <0·0001
Triglycerides, mmol/L 2·04 (1·42) 2·00 (1·41) 0·12 (0·01 to 0·23) 0·08 (–0·04 to 0·20) –0·04 (–0·20 to 0·13) 0·6413
HDL, mmol/L 1·14 (0·40) 1·19 (0·41) 0·07 (0·03 to 0·11) 0·04 (–0·01 to 0·08) –0·03 (–0·09 to 0·03) 0·3168
Glucose, mmol/L 6·77 (3·39) 6·51 (2·99) 0·21 (0·01 to 0·41) –0·04 (–0·25 to 0·17) –0·25 (–0·54 to 0·04) 0·0949

Data are mean (SD) or mean (95% CI). *The means and 95% CIs are for the within-person differences. †Test statistic and corresponding p value refers to the test of difference between intervention and control
groups at 12 months’ follow-up using generalised linear mixed effects models adjusting for baseline value, and including community as a random intercept effect to take into account the clustered
randomisation design.

Table 3: Outcomes at 12 months

N Mean (95% CI) pinteraction

Country
Colombia 690 –4·25 (–7·45 to –1·05)
0·74
Malaysia 545 –5·22 (–8·88 to –1·56)
Location
Urban 659 –4·74 (–8·26 to –1·22)
0·92
Rural 576 –4·83 (–7·99 to –1·66)
Sex
Female 688 –4·66 (–6·86 to –2·45)
0·72
Male 547 –5·08 (–8·25 to –1·91)
Age
Women ≤65 years or men ≤55 years 478 –3·80 (–6·30 to –1·30)
0·31
Women >65 years or men >55 years 757 –4·96 (–7·96 to –1·95)
Education
Primary or less 850 –4·52 (–7·23 to –1·81)
Secondary or higher 337 –5·05 (–8·58 to –1·52) 0·16
Trade, college, or university 48 –1·45 (–7·95 to 5·06)
Overall 1235 –4·78 (–7·11 to –2·44)

–10 –8 –6 –4 –2 0
Difference in mean (95% CI)

Figure 2: Differences in Framingham Risk Score for 10-year cardiovascular disease risk between intervention and control groups

in LDL at 12 months in the intervention compared with
the control group (p<0·0001; table 3). Furthermore, the
change in blood pressure control (<140 mm Hg) status
was more than double in the intervention group (69% vs
31%; p<0·0001) at 12 months. There was a trend to benefit
for most health behaviours in the intervention group,
resulting in a substantial reduction in the overall
INTERHEART Risk Score at 6 months (–1·9 [SD 5·6] vs
–4·5 [SD 5·4]; p=0·0008) and 12 months (–2·2 [SD 5·8] vs
4·8 [SD 5·4]; p<0·0001; table 4). In particular, there
were improvements in reports of physical activity and
dietary modifications (appendix p 7). However, there were
no significant differences in glucose concentrations,
HDL, smoking cessation rates, and weight between the
two groups at 12 months (appendix p 7).
18 serious adverse events were reported in intervention
participants who were taking study-associated medi­
cations (644 participants). In all reported cases, the
study-associated medication was identified to be unre­
lated to the event and the medication was continued
in 17 of 18 (94%) events. There were no significant
differences in deaths (n=12 [2·0%] vs n=15 [2·1%]),
myocardial infarction (n=7 [1·2%] vs n=12 [1·7%]), stroke
(n=5 [0·8%] vs n=6 [0·9%]), their composite (n=23 [3·8%]

www.thelancet.com Published online September 2, 2019 http://dx.doi.org/10.1016/S0140-6736(19)31949-X 7

 Articles

vs n=32 (4·6%)], or cardiovascular hospitalisations

A
4·0 Control
(n=35 [5·8%] vs n=27 [3·9%]) between the intervention
Intervention and control groups, respectively.
3·8

3·6 Discussion
LDL cholesterol (mmol/L)

3·4
A comprehensive, contextually appropriate model of care
resulted in a substantial reduction in cardiovascular
3·2
disease risk in two middle-income countries primarily
3·0 through improvements in blood pressure, LDL, and some
2·8 health behaviours. Other studies using NPHWs have
shown a modest increased uptake of antihypertensives
2·6
without trans­lating into a reduction in blood pressure.24,25
0 We showed a large reduction in blood pressure and cardio­
0 6 12
vascular disease risk, something significant for a
Visit (months)
Control Intervention Control Intervention
community-based intervention. We believe that a major
(N=636) (N=585) (N=636) (N=585) reason for its success was that our intervention package
Mean (95% CI) 3·38 3·34 3·21 2·75 explicitly targeted barriers identified through extensive
(3·30–3·46) (3·25–3·43) (3·13–3·28) (2·67–2·84)
health system appraisals, a step that is frequently omitted
B in developing complex interventions. These appraisals
160 Control pointed to a need for a comprehensive strategy, bringing
Intervention together elements tackling the main health systems
155
building blocks, in contrast with previous studies that have
Systolic blood pressure (mm Hg)

150 addressed only one or a few barriers. The strategy evaluated

145 here was an intervention led by NPHWs in conjunction
with primary care physicians and family, coupled with the
140
provision of free antihypertensives and a statin, as well as
135 strategies to improve adherence to medication and health
130 behaviours. It is a comprehensive and collaborative model
of care. The approach of door-to-door and community
125
event screening identified and led to the management of
0 individuals with previously undiagnosed hypertension
0 6 12
Visit (months) and uncontrolled hyper­ tension. To our knowledge, our
Control Intervention Control Intervention Control Intervention study is the only cluster-randomised controlled trial
(N=674) (N=602) (N=665) (N=587) (N=674) (N=602) of a comprehensive programme that addresses major
Mean (95% CI) 151·75 152·05 142·40 131·47 141·99 130·60 contextual barriers to cardiovascular risk detection,
(150·62– (150·88– (140·8– (130·1– (140·48– (129·35–
152·88) 153·25) 144·0) 132·9) 143·49) 131·83) treatment, and control in middle-income countries.
A recent systematic review of task sharing for the
Figure 3: Change in LDL and systolic blood pressure and over time in the intervention and control groups management of blood pressure in low-income and

Baseline Change at 6 months from baseline* Test between Change at 12 months from baseline* Test between
intervention intervention
and control and control at
at 6 months 12 months†
(p value) (p value)
Control Intervention Control Intervention Control Intervention
(n=727) (n=692) (n=674) (n=589) (n=692) (n=607)
SBP, mm Hg 151·8 (15·6) 152·1 (15·4) –9·0 (–11·4 to –6·7) –20·1 (–22·7 to –17·6) <0·0001 –9·7 (–12·1 to –7·3) –21·1 (–23·7 to –18·6) <0·0001
DBP, mm Hg 85·3 (11·9) 84·7 (12·0) –2·6 (–4·2 to –1·0) –6·1 ( –7·8 to –4·4) 0·0026 –2·9 (–4·4 to –1·4) –6·9 (–8·5 to –5·3) 0·0004
Controlled SBP <140 mm Hg 125 (17·2%) 74 (11·5%) 30·1 (25·6 to 34·7) 63·4 (59·2 to 67·6) <0·0001 30·4 (25·8 to 34·9) 68·9 (64·9 to 72·9) <0·0001
Controlled SBP <140 mm Hg 115 (15·8%) 68 (10·6%) 29·4 (24·9 to 33·9) 59·2 (54·9 to 63·5) <0·0001 28·4 (24·0 to 32·9) 64·8 (60·7 to 68·9) <0·0001
and DBP <90 mm Hg
INTERHEART Risk Score21 14·8 (6·1) 13·8 (5·3) –1·6 (–2·7 to –0·5) –4·4 (–5·6 to –3·2) 0·0008 –1·9 (–2·9 to –0·9) –4·9 (–5·9 to –3·8) <0·0001

Data are mean (SD) or n (%). SBP=systolic blood pressure. DBP=diastolic blood pressure. *The means and 95% CIs are for the within-person differences. †Test statistic and corresponding p value refers to the test
of difference between intervention and control groups at 12 month’s follow-up using generalised linear mixed effects models adjusting for baseline value, and including community as a random intercept effect
to take into account the clustered randomisation design.

Table 4: Secondary and tertiary outcomes at 6 and 12 months

8 www.thelancet.com Published online September 2, 2019 http://dx.doi.org/10.1016/S0140-6736(19)31949-X


middle-income countries reported reductions in SBP
with a mean of 4·8 mm Hg (95% CI 3·6–6·1).26 The
benefits varied depending on the type of NPHWs
involved in the intervention.26 The best results were
achieved with pharmacists (–8·1 mm Hg) as compared
with com­munity health workers (–3·7 mm Hg) as the
NPHWs and when NPHWs were guided by a protocol or
decision support.26 However, effects were limited by the
NPHWs’ inability to prescribe effective medications or
lack of availability of medications. This review noted that
there is a need for further research assessing community-
based, collaborative models of care. The HOPE 4 inter­
vention was informed by extensive previous work on
barrier assessments and health system appraisals done
in both Colombia and Malaysia to target barriers to
cardiovascular risk reduction (appendix p 6).7–9,12,13
The HOPE 4 study advances our understanding of how
NPHWs can be more effective in reducing cardiovascular
disease risk in several ways. First, with proper training
and supervision, even individuals with minimal training
can learn to successfully use a tablet-based decision
support system to diagnose and counsel individuals
with respect to their cardiovascular disease risk.17 This
is evidenced by the accuracy of the NPHWs’ risk
assessments and treatment recommendations when
compared with the primary care physician. Furthermore,
the effect of the NPHW’s counselling of the participant
and treatment supporter is shown by the improve­ment in
several health behaviours (captured in the INTERHEART
Risk Score) and the significant reductions in blood
pressure and LDL in the intervention group.
Second, our study was an international cluster-ran­
domised controlled trial that showed success in urban
and rural settings in two different countries and
continents, with widely varying health systems and
cultural backgrounds, which suggests our findings are
widely applicable. Given that country-specific barrier
assessments informed the design of HOPE 4, such
approaches would need to be considered to tailor the
intervention in different settings. Furthermore, the
intervention was effective in all subgroups, including
men and women, and those of different ages, with
differing education levels, in both countries, and in rural
and urban communities. It should be noted that there
was a non-significant trend for the intervention to be less
effective in participants with higher education; however,
most HOPE 4 participants (69%, n=945) had low educa­
tion; such populations have higher risk of cardiovascular
disease.27
Third, our study used a community-based intervention;
participants were recruited from their homes, community
outreach centres, or events based within public spaces.
Unlike studies based in medical clinics, which only
include those who seek medical help, our approach
overcomes the traditionally low detection of hypertension
in the community. Previous efforts of systematic screening
alone have had little additional effect in reducing blood
www.thelancet.com Published online September 2, 2019 http://dx.doi.org/10.1016/S0140-6736(19)31949-X 9
Articles
pressure or risk factors because they were not followed up
with pharmacological treatments.28,29 HOPE 4 overcomes
this by coupling community screening with a multifaceted
intervention, including the prescription and provision of
antihypertensive medications and statins, resulting in
substantial reductions in cardiovascular disease risk.
Furthermore, home visits by NPHWs helped to address
common barriers to attending medical clinics.12,13 One-
third of study participants reported that they had to travel
by taxi or use public transportation to attend physician
visits in the past. This time and expense, along with lost
wages, was identified as a barrier in the initial health
system appraisals. The community-based component
of the intervention probably contributed to improved
adherence to medical follow-up, a necessity for optimal
cardiovascular risk management.
Fourth, the HOPE 4 intervention involved task sharing
in a collaborative model involving NPHWs, physicians,
and family members, as recommended by WHO.30,31
This strategy addresses the health system barriers we
identified relating to physician shortages and policies
that limit the NPHWs’ ability to prescribe medications.
By training NPHWs and supporting them with mobile-
health decision support to screen, diagnose cardiovascular
disease risk, and recommend effective medications, the
management of cardiovascular disease risk by local
physicians is made easier and more efficient.7 The
HOPE 4 approach to initiating combination antihyper­
tensive drugs requiring only one or two blood pressure
measurements (with simultaneous reductions in lipids
and other risk factors) resulted in a substantial reduction
in 10-year cardiovascular disease risk as compared with
the control group, and the change in blood pressure
control was more than double in the intervention group.
Furthermore, this simplified approach limits the need
for multiple medical visits, resulting in additional savings
beyond the cost-effective strategy of using NPHWs in
low-resource settings.11 These findings and experiences
might also be applicable to high-income countries, as
the detection, treatment, and control of hypertension
and other cardiovascular disease risk factors is also
suboptimal in these settings.4
Fifth, the HOPE 4 study overcame significant health
system barriers relating to the cost and availability of
antihypertensive and statin medications in low-income
and middle-income countries by providing free locally
sourced medications to eligible participants in the
intervention group. It is well established that blood
pressure control is better achieved using combinations
of low doses of antihypertensive agents, an approach
endorsed by several guidelines.32–35 Furthermore, single-
pill combination antihypertensives have been recently
added to the WHO essential medicine list.36 Given
evidence supporting the use of statins in individuals with
hypertension regardless of lipid concentrations, this
medication was also recommended and used in eligible
participants.37–41 Higher proportions of combination
 Articles
antihypertensive (89% vs 55%) and statin medications
(89% vs 32%) were seen at 12 months in the intervention
group, showing the success of the intervention imple­
mentation. Although free medication is an important
component of the intervention and might explain the
large effects on blood pressure reduction, it probably
does not account for the full effect of the HOPE 4 strategy.
Other studies have shown a benefit to task sharing with
NPHWs despite medications being provided for free in
both the intervention and control groups.42
Finally, the HOPE 4 intervention actively involved the
participant’s family and friends as treatment supporters
in their management of cardiovascular disease risk.
The treatment supporters were encouraged to help the
participants adhere to their medications and healthy
behaviours between NPHW and physician visits. Such a
strategy has been shown to be effective at improving
medication adherence and mortality in patients with
HIV.16 Three-quarters of our intervention participants
had their treatment supporter present during inter­
vention visits, suggesting that they were actively engaged
in the participant’s management. Medication adherence
to antihypertensive drugs and statins was found to be
better in the intervention group. Further supporting the
role of treatment supporters and counselling from the
NPHWs, there was a trend toward improvements in
several reported healthy behaviours (appendix p 7);
however, there was no significant change in smoking
rates, weight, or HDL at 12 months.
The HOPE 4 study was not designed to assess the
safety of specific antihypertensive medications or statins
used in the study; however, there was no excess in
adverse events reported in the intervention group.
All study-provided medications were purchased from
approved, commercially available supplies within each
country, and are commonly used in clinical practice.
NPHWs were found to accurately assess cardiovascular
disease risk, medication contraindications, and treatment
recommendations when compared with physicians. This
supports the need to consider changes in regulations that
would allow trained NPHWs to prescribe a limited
number of commonly used drugs, such as safe anti­
hypertensive drugs and statins.10
Our study has some potential limitations. First,
participants in the control group were screened for
cardiovascular disease risk at baseline, had their blood
pressure assessed, and were provided with information
about local cardiovascular disease health programmes
and recommendations to see a health-care provider as
appropriate. Therefore, our screening was an intervention
in itself, because some participants in the control group
might have modified their behaviour after learning their
blood pressure was elevated, which could have led to
more people in the control communities receiving
treatment than would be the case in usual care. This
modification might have reduced the real differences
in primary and secondary outcomes between the
10 www.thelancet.com Published online September 2, 2019 http://dx.doi.org/10.1016/S0140-6736(19)31949-X
intervention and control groups, meaning that the
differences in cardiovascular disease risk reductions
between the intervention and control groups we observed
are likely to be an underestimate of the full effects of
our intervention strategy. Second, contamination was a
potential risk, but it was minimised by the cluster design
because only the communities randomly assigned to the
intervention had access to NPHW counselling (guided by
the intervention tablet application only), management
algorithms (inter­ vention tablet application only), and
free study medi­cations. In addition, communities were
chosen with adequate geographical separation from each
other. This approach was successful because we saw
marked differences in reduction of blood pressure,
lipids, and several other risk factors. Third, blinding was
not feasible in HOPE 4, which tested two strategies;
however, several steps were implemented to minimise or
avoid potential biases. Masking outcome assessors or
using a separate set of assessors to record risk factors
during follow-up was not feasible in a large study
involving several communities. Substantial additional
resources would have been needed to hire, train, and
oversee a second national team in each country, to do
duplicate visits at 6 months and 12 months, and to
develop separate streams of data collection processes and
tablet software. Therefore our approach of ensuring that
the blood pressure recordings were from automated
devices and lipids were measured centrally without
knowledge of whether they were from intervention or
control communities minimised biases in the most
practical manner possible for the two key outcomes of
the trial. The extensive analyses undertaken to confirm
that there was no bias in the blood pressure recorded in
HOPE 4 are reported in the appendix (p 8). Fourth,
surrogate outcomes were used in this study. There is
good evidence to support improved clinical outcomes
with the reductions in blood pressure and LDL shown in
this trial. To show a benefit in clinical events, a much
larger and longer trial would be required. It was not the
goal of our study to examine the direct effect of our
intervention on clinical events but instead to develop and
test more effective implementation strategies to reduce
risk factors in those with hyper­tension. The ASCOT-LLA
trial,37 the analysis of those with elevated blood pressure
in the HOPE 3 trial,38 and the results of the PolyIran
trial43 indicate that substantial reductions in cardio­
vascular disease will occur from reducing both blood
pressure and lipids in individuals with elevated blood
pressure. Fifth, given the stratified randomisation by
country, there were an uneven number of clusters in the
intervention and control groups, leading to an imbalance
of some characteristics. However, sensitivity analysis,
adjusting for differences in baseline characteristics
including country, age, sex, education, baseline use of
antihyper­tensives, hypertension, diabetes, and smoking
status did not significantly change the primary outcome.
Finally, the single-pill combination antihypertensive and

statin medications were provided free to participants in
the intervention group and were paid for as part of the
HOPE 4 study. Although some countries provide these
medications at low or no cost, this is not the case in
several other countries. We hope that the combined
effect of our results, in the context of the ASCOT-LLA,37
HOPE 3,38 and PolyIran trial43 results, along with the
inclusion of combination antihypertensive drugs and
statins in the Essential Medicines List will persuade
governments to provide such medicines at low or no cost
to those with hypertension. This would overcome a key
barrier (cost and availability of medicines) to improved
control of hypertension and reduction of cardiovascular
disease.
A comprehensive model of care led by trained NPHWs
with physicians and family members resulted in a sub­
stantial reduction in estimated cardiovascular dis­ ease
risk, primarily through improvements in blood pressure,
LDL, medication adherence, and some health behaviours.
In addition, this study led to the detection of hypertension
and consequent initiation of treatment in a significant
number of individuals who otherwise were unlikely to
have been diagnosed or managed in the near future. The
HOPE 4 strategy could help to attain the UN General
Assembly Action Plan for a one-third reduction in
premature mortality from cardiovascular disease.44 On
the basis of the substantial reductions in cardiovascular
disease risk and the improvements in blood pressure
control shown (as well as the additional benefits of
community screening), this target is achievable if health
system strategies, such as the one evaluated here, are
adapted to local contexts and then adopted.
Contributors
SY designed and supervised the study and data analysis, interpreted
the data, and reviewed and commented on all drafts of the Article.
J-DS helped to design and supervise the study, analysed data, led the
writing of the analysis plan, and had the primary responsibility for
writing the Article. TM coordinated the study implementation globally,
and reviewed and commented on all drafts of the Article. SI participated
in writing the analysis plan, did the analysis, and reviewed and
commented on all drafts of the Article. MM designed the health systems
barriers assessment, helped design the intervention package, reviewed
and commented on the data analysis and drafts of the Article. AA, PL, LT
and SIB reviewed and commented on the data analysis and drafts of the
Article. All other authors implemented the study in their respective
countries and provided comments on drafts of the Article.
Declaration of interests
J-DS, TM, PL, SIB, SI, and SY report that their institution received grants
from the Canadian Institutes of Health Research, Ontario Ministry of
Health and Long-Term Care, Boehringer Ingelheim, and the Department
of Management of Non-Communicable Disease, WHO, for the conduct of
the study. PL-J and PAC report that their institution received unrestricted
grants from Grand Challenges Canada, and the Santander Departmental
Secretary of Health, during the conduct of the study. KY and FM report
that their institution received a grant from Grand Challenges Canada
during the conduct of the study. MM reports that his institution received a
grant from the Canadian Institutes of Health Research. AA reports grants
from Canadian Institutes of Health Research during the conduct of the
study. LT declares no competing interests.
Data sharing
The Population Health Research Institute has a formal data sharing
policy. Data will be disclosed only upon request and approval of the
www.thelancet.com Published online September 2, 2019 http://dx.doi.org/10.1016/S0140-6736(19)31949-X 11
Articles
proposed use of the data by a review committee created by leaders of the
study. This review will serve to ensure that patient privacy and rights,
and data and research integrity, can be maintained. Review criteria will
include demonstrated competence in data security and analysis and data
will be shared to achieve the objectives in the approved protocol only.
Individual participant data and a data dictionary will be made available,
subject to requirements or restrictions from research ethics board or
institutional review boards, existing contracts or agreements,
and conditions set forth in participant consent forms. Data provided will
be limited to data which underlies the results in the main publication.
after de-identification. The protocol and statistical analysis plan for
analysis of the primary results will be shared. Data can be disclosed from
2 years after the main paper is published, plus 6 additional months for
every year of study conduct, to a maximum of 7 years. Data will be made
available through secure data transfer methods overseen by the
Population Health Research Institute (PHRI), or by having analyses
performed by the PHRI Department of Statistics, subject to capacity.
Each proposal must identify and provide funding to defray the costs of
data preparation, storage, transfer, and analysis for the organisation
incurring these costs.
Acknowledgments
We thank our funders and partners at WHO for their collaboration on
the NPHW training curriculum at the onset of this project.
We acknowledge support from the following physicians in Malaysia and
Colombia: Ng Kien Keat, Khairul Shafiq Ibrahim, Adibah Hanim Ismail,
Norhaslira Abdul Rahim, Noor Hasliza Hassan, Ruziaton binti Hasim,
Nor Anizah bt Muzaid, Sabariah Idris, Johanna Otero,
Jose Patricio López-López, Sandra Rueda-Quijano, and
Luz Maria Gomez-Peña. This work was supported by the Canadian
Institutes of Health Research (CIHR) and Grand Challenges Canada
(GCC), as part of the Global Alliance for Chronic Disease program
(CIHR grant number 120389; GCC grant numbers 0069–04 and
0070–04); CIHR’s Strategy for Patient Oriented Research (SPOR),
through the Ontario SPOR Support Unit, as well as the Ontario Ministry
of Health and Long-Term Care; an unrestricted grant from Boehringer
Ingelheim; the Department of Management of non-communicable
diseases, WHO; the Santander Departmental Secretary of Health,
Bucaramanga, Colombia; and the Population Health Research Institute.
This research programme does not represent the official view of the
Global Alliance for Chronic Diseases, WHO, or the governments of
Canada, Malaysia, or Colombia. No official support or endorsement of
this article by the Global Alliance for Chronic Diseases hypertension
programme, WHO, or federal governments of participating countries is
intended or should be inferred.
References
1 Roth GA, Johnson C, Abajobir A, et al. Global, regional,
and national burden of cardiovascular diseases for 10 causes,
1990 to 2015. J Am Coll Cardiol 2017; 70: 1–25.
2 NCD Countdown 2030 Collaborators. NCD Countdown 2030:
worldwide trends in non-communicable disease mortality and
progress towards Sustainable Development Goal target 3.4.
Lancet 2018; 392: 1072–88.
3 Hoffmann R, Plug I, McKee M, et al. Innovations in medical care
and mortality trends from four circulatory diseases between 1970
and 2005. Eur J Public Health 2013; 23: 852–57.
4 Chow CK, Teo KK, Rangarajan S, et al. Prevalence, awareness,
treatment, and control of hypertension in rural and urban
communities in high-, middle-, and low-income countries.
JAMA 2013; 310: 959–68.
5 Yusuf S, Islam S, Chow CK, et al. Use of secondary prevention
drugs for cardiovascular disease in the community in
high-income, middle-income, and low-income countries
(the PURE study): a prospective epidemiological survey. Lancet
2011; 378: 1231–43.
6 Balabanova D, McKee M, Mills A, Walt G, Haines A. What can
global health instutitions do to help strengthen health systems in
low income countries? Health Res Policy Syst 2010; 8: 22.
7 Schwalm JR, McCready T, Lamelas P, et al. Rationale and design of
a cluster randomized trial of a multifaceted intervention in people
with hypertension: The Heart Outcomes Prevention and
Evaluation 4 (HOPE-4) study. Am Heart J 2018; 203: 57–66.
 Articles
8 Maimaris W, Paty J, Perel P, et al. The influence of health systems
on hypertension awareness, treatment, and control: a systematic
literature review. PLoS Med 2013; 10: e1001490.
9 Khatib R, Schwalm JD, Yusuf S, et al. Patient and healthcare
provider barriers to hypertension awareness, treatment and follow
up: a systematic review and meta-analysis of qualitative and
quantitative studies. PLoS One 2014; 9: e84238.
10 Joshi R, Alim M, Kengne AP, et al. Task shifting for
non-communicable disease management in low and middle income
countries—a systematic review. PLoS One 2014; 9: e103754.
11 Gaziano T, Abrahams-Gessel S, Surka S, et al. Cardiovascular
disease screening by community health workers can be cost-effective
in low-resource countries. Health Aff 2015; 34: 1538–45.
12 Legido-Quigley H, Lopez PAC, Balabanova D, et al. Patients’
knowledge, attitudes, behaviour and health care experiences on the
prevention, detection, management and control of hypertension in
Colombia: a qualitative study. PLoS One 2015; 10: e0122112.
13 Risso-Gill I, Balabanova D, Majid F, et al. Understanding the
modifiable health systems barriers to hypertension management in
Malaysia: a multi-method health systems appraisal approach.
BMC Health Serv Res 2015; 15: 254.
14 Campbell MK, Piaggio G, Elbourne DR, Altman DG,
CONSORT Group. Consort 2010 statement: extension to cluster
randomised trials. BMJ 2012; 345: e5661.
15 WHO. WHO STEPS surveillance manual: the WHO STEPwise
approach to chronic disease risk factor surveillance. Geneva:
World Health Organization, 2008.
16 Nachega JB, Chaisson RE, Goliath R, et al. Randomized controlled
trial of trained patient-nominated treatment supporters providing
partial directly observed antiretroviral therapy. AIDS 2010; 24: 1273–80.
17 Khan M, Lamelas P, Musa H, et al. Development, testing,
and implementation of a training curriculum for nonphysician
health workers to reduce cardiovascular disease. Glob Heart 2018;
13: 93–100.
18 D’Agostino RB Sr, Vasan RS, Pencina MJ, et al. General cardiovascular
risk profile for use in primary care: the Framingham Heart Study.
Circulation 2008; 117: 743–53.
19 Artigao-Rodenas LM, Carbayo-Herencia JA, Divison-Garrote JA, et al.
Framingham risk score for prediction of cardiovascular diseases:
a population-based study from southern Europe. PLoS One 2013;
8: e73529.
20 Chia YC, Gray SY, Ching SM, et al. Validation of the Framingham
general cardiovascular risk score in a multiethnic Asian
population: a retrospective cohort study. BMJ Open 2015;
5: e007324.
21 Yusuf S, Hawken S, Ounpuu S, et al. Effect of potentially modifiable
risk factors associated with myocardial infarction in 52 countries
(the INTERHEART study): case-control study. Lancet 2004;
364: 937–52.
22 Morisky DE, Ang A, Krousel-Wood M, Ward HJ. Predictive validity
of a medication aderence measure in an outpatient setting.
J Clin Hypertens 2008; 10: 348–54.
23 Schafer JL. Analysis of incomplete multivariate data. Boca Raton:
CRC Press, 1997.
24 Peiris D, Praveen D, Mogulluru K, et al. SMARThealth India:
a stepped-wedge, cluster randomised controlled trial of a
community health worker managed mobile health intervention for
people assessed at high cardiovascular disease risk in rural India.
PLoS One 2019; 14: e0213708.
25 Tian M, Ajay VS, Dunzhu D, et al. A cluster-randomized,
controlled trial of a simplified multifaceted management program
for individuals at high cardiovascular risk (simcard trial) in rural
Tibet, China, and Haryana, India. Circulation 2015; 132: 815–24.
26 Anand TN, Joseph LM, Geetha AV, Prabhakaran D, Jeemon P.
Task sharing with non-physician health-care workers for
management of blood pressure in low-income and middle-income
countries: a systematic review and meta-analysis. Lancet Glob Health
2019; 7: e761–67.
12 www.thelancet.com Published online September 2, 2019 http://dx.doi.org/10.1016/S0140-6736(19)31949-X
27 Rosengren A, Smyth A, Rangarajan S, et al. Socioeconomic status
and risk of cardiovascular disease in 20 low-income, middle-income,
and high-income countries: the prospective urban rural
epidemiologic (PURE) study. Lancet Glob Health 2019; 7: e748–60.
28 Krogsbøll LT, Jørgensen KJ, Grønhøj Larsen C, Gøtzsche PC.
General health checks in adults for reducing morbidity and
mortality from disease: Cochrane systematic review and
meta-analysis. BMJ 2012; 345: e7191.
29 Jorgensen T, Jacobsen RK, Toft U, et al. Effect of screening and
lifestyle counselling on incidence of ischaemic heart disease in
general population: Inter99 randomised trial. BMJ 2014; 348: g3617.
30 Abegunde DO, Shengelia B, Luyten A, et al. Can non-physician
health-care workers assess and manage cardiovascular risk in
primary care? Bull World Health Organ 2007; 85: 432–40.
31 WHO. WHO CVD-risk management package for low- and
medium-resource settings. Geneva: World Health Organization,
2002.
32 López-Jaramillo P, Barbosa E, Molina DI, et al. Latin American
consensus on management of hypertension in the patient with
diabetes and the metabolic syndrome. J Hypertens 2019; 37: 1126–47.
33 Mancia G, De Baker G, Dominiczak A, et al. 2007 Guidelines for the
management of arterial hypertension: The Task Force for the
Management of Arterial Hypertension of the European Society of
Hypertension (ESH) and of the European Society of Cardiology
(ESC). Eur Heart J 2007; 28: 1462–536.
34 Chobanian AV, Bakris GL, Black HR, et al. Seventh report of the
Joint National Committee on Prevention, Detection, Evaluation,
and Treatment of High Blood Pressure. J Hypertens 2003; 42: 1206–52.
35 Malaysian Society of Hypertension. Working Group on Hypertension.
Clinical Practice Guidelines—Management of Hypertension, 5th edn.
Malaysia: Malaysian Society of Hypertension, 2018.
36 WHO. World Health Organization model list of essential medicines,
21st list, 2019. Geneva: World Health Organization, 2019.
37 Sever PS, Dahlöf B, Poulter NR, et al. Prevention of coronary and
stroke events with atorvastatin in hypertensive patients who have
average or lower-than-average cholesterol concentrations,
in the Anglo-Scandinavian Cardiac Outcomes Trial—Lipid
Lowering Arm (ASCOT-LLA): a multicentre randomised controlled
trial. Lancet 2003; 361: 1149–58.
38 Yusuf S, Lonn E, Pais P, et al. Blood-pressure and cholesterol
lowering in persons without cardiovascular disease. N Engl J Med
2016; 374: 2032–43.
39 Lonn EM, Bosch J, López-Jaramillo P, et al. Blood-pressure lowering
in intermediate-risk persons without cardiovascular disease.
N Engl J Med 2016; 374: 2009–20.
40 Yusuf S, Bosch J, Dagenais G, et al. Cholesterol lowering in
intermediate-risk persons without cardiovascular disease.
N Engl J Med 2016; 374: 2021–31.
41 Grundy SM, Stone NJ, Bailey AL, et al. 2018 AHA/ACC/AACVPR/
AAPA/ABC/ACPM/ADA/AGS/APhA/ASPC/NLA/PCNA Guideline
on the Management of Blood Cholesterol. Circulation 2019;
139: e1082–143.
42 Ogedegbe G, Plange-Rhule J, Gyamfi J, et al. Health insurance
coverage with or without a nurse-led task shifting strategy for
hypertension control: a pragmatic cluster randomized trial in
Ghana. PLoS Med 2018; 15: e1002561.
43 Roshandel G, Khoshnia M, Poustchi H, et al. Effectiveness of
Polypill for primary and secondary prevention of cardiovascular
diseases: a pragmatic cluster randomized controlled trial.
SSRN 2019; published online June 24. https://ssrn.com/
abstract=3408063 (preprint).
44 Mamudu HM, Yang JS, Novotny TE. UN resolution on the
prevention and control of non-communicable diseases:
an opportunity for global action. Glob Public Health 2011; 6: 347–53.