Available via license: CC BY 4.0
Content may be subject to copyright.
1
TesemaAG, etal. BMJ Global Health 2022;7:e009025. doi:10.1136/bmjgh-2022-009025
Exploring complementary and competitive
relations between non- communicable
disease services and other health extension
programme services in Ethiopia: a
multilevel analysis
Azeb Gebresilassie Tesema,1,2 David Peiris ,1 Rohina Joshi,3,4
Seye Abimbola ,1,5 Fasil Walelign Fentaye,6 Alula M Teklu,6 Yohannes Kinfu7,8
Original research
To cite: TesemaAG,
PeirisD, JoshiR, etal.
Exploring complementary
and competitive relations
between non- communicable
disease services and other
health extension programme
services in Ethiopia: a multilevel
analysis. BMJ Global Health
2022;7:e009025. doi:10.1136/
bmjgh-2022-009025
Handling editor Sanni Yaya
►Additional supplemental
material is published online only.
To view, please visit the journal
online (http:// dx. doi. org/ 10.
1136/ bmjgh- 2022- 009025).
Received 9 March 2022
Accepted 24 May 2022
For numbered afliations see
end of article.
Correspondence to
Mrs Azeb Gebresilassie Tesema;
azeb18@ gmail. com
© Author(s) (or their
employer(s)) 2022. Re- use
permitted under CC BY.
Published by BMJ.
ABSTRACT
Background Ethiopia has recently revitalised its health
extension programme (HEP) to address the rising burden
of non- communicable diseases (NCDs). We examined the
effects of existing essential HEP services on the uptake of
NCD preventive services.
Methods We applied a mixed- effect non- linear model
with a logit link function to identify factors associated
with a community resident’s probability of receiving NCD
prevention services through the HEP. The data were drawn
from the Ethiopian HEP assessment Survey conducted
in all regions. The analysis included 9680 community
residents, 261 health extension workers (HEWs), 153
health posts, 119 health centres, 55 districts and 9 regions,
which we combined hierarchically into a single database.
Results In the 12 months before the survey, 22% of the
sample population reported receiving NCD preventive
service at least once. The probability of receiving NCD
prevention service increased by up to 25% (OR=1.25,
CI 1.01 to 1.53) if health centres routinely gathered NCD
data from health posts and by up to 48% (OR=.48, CI 1.24
to 1.78) if they provided general (ie, non- NCD specic)
training to HEWs. NCD preventive service uptake also
increased if the HEW held level IV qualication (OR=1.32,
CI 1.06 to 1.65) and lived in the community (OR=1.24,
CI 1.03 to 1.49). Conversely, if facilities delayed general
performance reviews of HEWs by a month, uptake of NCD
prevention services decreased by 6% (OR=0.94, CI 0.91
to 0.97). We observed that better HIV/AIDS programme
performance was associated with a lower uptake of NCD
preventive services (OR=0.15, CI 0.03 to 0.85).
Conclusion Despite efforts to improve NCD services
through the HEP, the coverage remains limited. A strong
HEP is good for the uptake of NCD preventive services.
However, integration requires a careful balance, so that the
success already recorded for some existing programmes
is not lost.
BACKGROUND
Non- communicable diseases (NCDs) account
for one- third of all deaths in Ethiopia.1–3
Rapid urbanisation, epidemiological transi-
tion, population ageing and lifestyle changes
are anticipated to compound the burden of
WHAT IS ALREADY KNOWN ON THIS TOPIC
⇒Ethiopia’s primary healthcare focused health exten-
sion programme (HEP) has been a guiding frame-
work for its health sector development for the past
two decades, with non- communicable disease
(NCD) preventive service being added in 2016.
⇒Previous studies have only examined the impacts of
the HEP such as maternal and child health, HIV/AIDS
and tuberculosis.
WHAT THIS STUDY ADDS
⇒Despite ongoing efforts to improve NCD preventive
services through the HEP, coverage of NCD services
is still limited, more so for underserved communities.
⇒NCD services, is inuenced by improved health in-
formation system, health worker training and timely
performance appraisal practices.
⇒HIV/AIDS service delivery appears to compete with
NCD preventive service provision.
HOW THIS STUDY MIGHT AFFECT RESEARCH,
PRACTICE OR POLICY
⇒A strong HEP is good for the uptake of NCD pre-
ventive services, but integration with existing pro-
grammes may not be straightforward. It requires a
carefully balanced approach, so that the success
already recorded for some existing programmes is
not lost.
⇒Furthermore, strengthening overall training, infor-
mation management system and building trust be-
tween the community and the HEWs are all essential
for a successful HEP, just as they are essential for
individual services within the HEP and for integrating
NCD preventive services.
⇒A range of complementary effects point to the need
for an overall strong HEP capable of delivering exist-
ing services as a prerequisite for successful integra-
tion of new services.
2TesemaAG, etal. BMJ Global Health 2022;7:e009025. doi:10.1136/bmjgh-2022-009025
BMJ Global Health
and unmet need for NCD services in the country even
further. This poses a significant threat to Ethiopia’s
healthcare system, which, thus far, has been focused on
tackling HIV, malaria, tuberculosis (TB) and maternal
and child health (MCH). The mismatch has been more
significant for communities living far from health facili-
ties.4
Rising community expectations for better and inclusive
services and the growing NCD burden in Ethiopia led to
a comprehensive National NCD Prevention and Control
Strategic Action Plan. The plan of action focused on
behavioural risk factors for NCDs, with a clear strategy
to address them through the country’s health extension
programme (HEP).5 Implemented in 2003, the HEP
is a community- based strategy geared towards deliv-
ering essential health promotion, disease prevention
and selected curative health services at the community
and health postlevel. The programme was revitalised
in 2016/2017, introducing mental health and NCDs as
additional packages, primarily focusing on cardiovas-
cular disease, diabetes, cancer and chronic respiratory
diseases.6 7
Through the optimised HEP, the scope of work for
Health Extension Workers (HEWs) has been expanded
over and above their pre- existing role and included NCD
prevention and promotion activities. HEWs—10th- grade
school graduates who have received a 1- year training
in primary healthcare (PHC) before being deployed as
salaried civil servants in health posts in their respective
villages—were initially responsible for implementing 16
preventive, promotive and selected curative packages
focused on four programmatic areas: family health,
disease prevention and control, hygiene and environ-
mental sanitation and health education and communi-
cation.6 8 9
Under the optimised HEP, HEWs role for NCDs
included: awareness creation of NCD risk factors,
promoting of NCD prevention activities through health
education, undertaking community screening for early
detection of NCDs, referring patients with more complex
health needs to health centres and collecting and inter-
preting NCD data in the community.7 10–12 A preservice
training programme had been launched to support the
HEWs new mandate and the newly introduced packages.
This created career transition opportunities and allowed
several HEWs to transition from level III to level IV qual-
ifications while also gaining NCD content as part of the
curriculum.7
Although research on community health workers’
(CHWs) role in NCD prevention is still evolving, the
evidence to date suggests that CHWs can be effective
in NCD prevention in low- income and middle- income
countries.13–18 Several studies have demonstrated an
association between CHW programme performance and
health system factors, such as training and availability of
NCD guidelines.13–18 Furthermore, evidence suggests that
community- level attributes, such as sociodemographic
characteristics and perception towards health workers
and programmes on offer, contribute to the success of
CHW- based NCD programmes.17 19
However, despite growing evidence on CHW
programmes’ role in strengthening health systems and
improving key health outcomes, including NCDs,14–18
existing studies in Ethiopia and elsewhere have several
limitations. First, none of the studies from Ethiopia
focused on NCD programme performance delivered at
the HEP level. Second, although new packages are often
added over existing services, no other research has empir-
ically investigated the complementarity and competition
between NCD service provision and other services. Simi-
larly, not enough is known, if any, health system inputs—
such as supervision and performance appraisal—HEW
attributes and community perceptions towards the HEP
or HEWs have the same effect on NCDs as in other
programmes.19 20 This study uses nationally representa-
tive data to investigate the utilisation of NCD preventive
services through the HEP in Ethiopia. Our specific aims
are to examine (1) the effect of health system inputs
and processes, (2) the association between HEWs’ and
community residents’ characteristics and NCD uptake
and (3) the complementarity and competition between
NCD and other essential preventive services.
METHODS
Study setting
Ethiopia’s HEP, primarily implemented by HEWs, is
part of the country’s PHC strategy to improve health
service coverage and health outcomes. The PHC oper-
ates under a three- tier system: primary, secondary and
tertiary care. The primary care component includes
primary hospitals, health centres and health posts
(the lowest service delivery point at the village level).
The PHC unit comprises five satellite health posts
and a referral health centre. In each health post,
an average of two HEWs serve around 3000–5000
people in their catchment area. In 2019, there were
3790 health centres,21 17 587 health posts and 39 878
HEWs across the country. HEWs are posted within a
health post and receive technical support and super-
vision from the nearby health centre.4 7 A designated
HEW manages health posts, and all HEW working
within the health posts are responsible for delivering
all essential health service packages, including NCD
services.
At the time of data collection, Ethiopia is divided
into nine regions and two city administrations (ie,
Addis Ababa and Dire Dawa). Each region, in turn,
is subdivided into zones, and zones are divided
further into districts, locally known as woredas.22
Woredas are the primary administrative unit in Ethi-
opia’s decentralised system and have a council and
separate operational and sectoral offices, including
one for health, known as a woreda health office.22
A group of kebeles—each of which has around 5000
inhabitants—constitute a single woreda. Kebele is
TesemaAG, etal. BMJ Global Health 2022;7:e009025. doi:10.1136/bmjgh-2022-009025 3
BMJ Global Health
the country’s lowest administrative unit and provides
administrative support to the health post. A detailed
description of Ethiopia’s health system has been
reported elsewhere.23
Study design
Data from the Ethiopian HEP assessment Survey
(EHAS), conducted from 1 October 2018 to 31
October 2019, were used for this study.4 The EHAS
was commissioned by the Ministry of Health (Ethi-
opia) and carried out by the Monitoring, Evalua-
tion, Research and Quality Improvement consul-
tancy group. The national assessment covered all
regions and agro- ecological zones in the country and
employed qualitative and quantitative methods. Only
the quantitative component was used for this study.
The quantitative survey was carried out following a
hierarchical multistage sampling design to identify the
study Woredas, HPS, HCs and community members in
all the nine regions in the country, excluding the two
city administrations. In the first stage, the Woredas
in each region were classified into pastoralist and
agrarian communities based on their primary means
of livelihood. Subsequently, six kebeles were chosen
randomly from each woreda, and the facility assess-
ment surveys were conducted on all health centres
and health posts located in the selected kebeles. The
household survey was conducted only on three kebeles
randomly chosen out of the original six kebeles. The
total sample included 62 woredas, 179 rural health
centres, 343 health posts, 584 HEWs and 12 868
respondents living in 7122 rural households.4 The
woreda level data were supplied by the woreda health
office manager or their representative, while health
centre directors provided the data on health centres.
The head of the health post, a HEW, was responsible
for providing the data related to the health post.
Additionally, each HEW in the selected area provided
data on their age, gender, qualification, year of service,
other HEW attributes and perceptions or views on the
HEP and existing support system. Finally, service utilisa-
tion patterns, including NCDs, were obtained from male
and female community respondents from the selected
areas. The survey also recorded participant characteris-
tics and views towards the HEP and the HEWs working in
their communities.
In all, this study captured 9680 community residents
nested with 261 HEWs, 153 health posts, 119 health
centres and 55 woredas. The analytic sample included
only those observations with complete data at all levels
of the data collection system. Additional material on
the underlying survey is available elsewhere.4
Conceptual framework
Our analysis focuses on the utilisation of HEWs- based
NCD prevention services. The selection of covariates
affecting service uptake was guided by data availa-
bility and conceptual and empirical evidence on the
role of CHWs in the provision of various community
preventive services in resource- limited settings.24
Accordingly, as shown in figure 1, our core covari-
ates cover four main components: health system
inputs and processes; HEW characteristics and
perceptions towards the HEP; community resident
attributes—such as age, gender, socioeconomic status
(SES) and their perceptions towards HEWs and the
HEP—and cross- programme effects from non- NCD
Figure 1 Factors affecting community health extension worker NCD prevention service in Ethiopia, adapted from Agrawal et
al.24 HEW, health extension workers; NCD, non- communicable disease.
4TesemaAG, etal. BMJ Global Health 2022;7:e009025. doi:10.1136/bmjgh-2022-009025
BMJ Global Health
services delivered through the HEP focussing on TB,
HIV and environmental health services.
Statistical analysis
The EHAS covered six different NCD interventions—
namely, hypertension, heart disease, diabetes, kidney
diseases, cervical and breast cancer—provided by
HEWs through the HEP. Survey respondents were
asked if they used any of the six NCD preventive
services at least once in the year before the survey. The
primary outcome variable was created by converting
those responses into binary outcomes. We coded ‘1’
if residents reported using one or more preventive
services at least once during the past 12 months and
‘0’ otherwise.
We applied a mixed- effects non- linear model with
a logit link function and a random intercept at the
woreda level to estimate the probability of NCD
preventive service utilisation in the country.25 26 The
random- intercept non- linear mixed- effects model
suits the data’s hierarchical structure and the binary
nature of the outcome variable. The random compo-
nent addresses data clustering biases; and the effects
of unobserved heterogeneity or omitted covariates
measured at the woreda level. We fit the random
variance at the woreda level because woredas play a
pivotal role in Ethiopia’s multilayered health system.
They are responsible for the planning, financing and
monitoring health service delivery in their jurisdic-
tion. They manage and coordinate the PHC units
under their administrative control and have a tech-
nical link with the regional health bureaus and the
zonal health department, requiring a high level of
organisational and managerial competence.27–29
In this study, the models were estimated sequentially,
starting with a null model (intercept- only model) that
tested the null hypothesis that there was no between-
cluster variation in NCD service delivery uptake. We
then fitted a separate model, controlling only health
system inputs and process variables. In the third stage,
we added HEW- level attributes and perceived views
on the current support system. Finally, we introduced
variables capturing cross- programme effects, commu-
nity characteristics and perception in the fourth and
final model.
Resident perceptions were measured using 27
Likert- based questions. Out of these, we generated
composite summary indices using principal compo-
nent analysis (PCA). We retained factors with eigen-
values of unity or above, leading to two indices that
broadly corresponded with the community’s views on
HEP and towards HEWs competence to provide avail-
able services. Similarly, we aggregated the 10 Likert
scale questions on HEWs’ perception of HEP and
health system supports posed to HEWs’ into a single
index using PCA. The analyses were conducted in
Stata V.17 and weighted to account for non- response
and design effects.30 31
Patient and public involvement
No patients or public members were directly involved in
the present study’s design as we used secondary data for
the research. There are no plans to involve patients or
the public in disseminating results.
RESULTS
Community characteristics
The mean (SD) age of the sample population was 40.4
(13.81) years, and women were slightly over- represented
(57.4%) in the sample. Twenty- two per cent of the resident
population reported receiving prevention services on
one or more of the six types of NCDs at least once during
the past 12 months. Likewise, 18%, 12%, 16%, 13%, 12%
and 13% reported having received hypertension, heart
disease, diabetes mellitus, kidney disease and cervical
and breast cancer prevention services from HEWs. In
contrast, the uptake of other prevention services such as
for HIV/AIDS and TB was much higher—89% and 84%,
respectively. Receipt of hygiene advice was 47%—lower
than these diseases but two- fold greater than for NCDs.
However, mental health prevention service delivery was
much lower than NCD services, 5.17% (table 1).
HEWs' characteristics
The mean (SD) age of HEWs was 26.3 (4.8) years, and
an average HEW had served about 7.1 (4.7) years. Half
(50 %) of the HEWs had level 4 qualifications, and one
in five (22 %) reported having received short- term NCD
training in the past year. Almost all HEWs (92 %) self-
reported having the skill and being competent in meas-
uring blood pressure (BP).
The majority (76 %) of HEWs were born or grew up
in the same area as they worked in: 28.7% grew up in
the same kebele, and 47.3% were from the same woreda,
while less than a quarter—24.0 %—originated from
outside. The majority (63 %) also currently reside in the
same community that they serve (table 2).
Health system characteristics
Despite 99.9% of health centres saying they provided
supervision visits to health posts, less than half (46.5%)
of HEWs stationed at health posts reported receiving
supportive supervision in the last 6 months (table 3).
About three- fourth (73 %) of the health posts reported
regularly compiling community profiles, while less than
one- third of health centres (31 %) collected and inter-
preted NCD data from health posts. About two- fifth
(36.5%) of health centres reported providing training for
HEP coordinators in the last year. The same proportion,
37%, of health centres also reported providing training
for HEWs in the past 2 years (table 3).
Determinants of utilisation of NCD prevention services
Online supplemental file 1 shows the results of the
multilevel analysis (models 0–3). The null model, repre-
sented by model 0, has no covariates; it provides between-
cluster variance (ie, between woreda). Model 1 captures
TesemaAG, etal. BMJ Global Health 2022;7:e009025. doi:10.1136/bmjgh-2022-009025 5
BMJ Global Health
the effects of health systems inputs on predicting NCD
preventative uptake without controlling for additional
variables. Model 2 simultaneously estimates the associa-
tions between health system- level and HEW- level attrib-
utes and NCD service uptake. The final model, model 3,
portrays the effects of health system- level, health worker-
level and person- level characteristics on NCD utilisation.
While all the three models with covariates (models
1–3) generally reveal similar patterns, adding all three
sets of covariates together into the same model in model
3 resulted in additional significant variables than those
identified in models 1 and 2. Our final model, model 3,
is presented in a forest plot in figure 2.
The forest plot shows that the odds of receiving NCD
prevention service by a community member increased by
up to 25% (OR=1.25, CI 1.01 to 1.53) if health centres
routinely gathered NCD report from health posts and by
up to 48% (OR=1.48, CI 1.24 to 1.78) if they provided
training to HEWs. These results are similar to those
observed in models 1 and 2. Our findings also showed
that if the HEWs have a level IV qualification and live in
the same community they serve, the odds of NCD utilisa-
tion in the community increase by about 32% (OR=1.32,
CI 1.06 to 1.65) and 25% (OR=1.24, CI 1.03 to 1.49),
respectively. However, long years of service (OR=0.49 CI
0.31 to 0.78) and perceived lack of health system support
by HEWs (OR=0.73 CI (0.63 to 0.86) adversely impact
Table 1 Descriptive statistics of outcome variables,
resident characteristics and perception for NCD prevention
service, Ethiopian health extension programme, 2022
Sample community characteristics and
perception (n=9680)
% or mean
(weighted)
Community NCD prevention coverage* (%) 21.92
Respondents who have heard from HEWs
about (%)
Cervical cancer 12.89
Diabetes mellitus 16.48
Hypertension 18.02
Heart disease 12.33
Breast cancer 13.57
Kidney diseases 13.54
Respondents who heard from HEWs about
tuberculosis
84.42
Respondents who heard from HEWs about
HIV and AIDS
89.14
Respondents who heard from HEWs about
hand hygiene
47.35
Respondents who heard from HEWs about
mental illnesses
5.17
Gender (%), female 57.84
Wealth quintile index (%)
Lowest 14.94
Lower 17.85
Middle 20.56
Higher 24.44
Highest 22.22
Composite index for community perception
►Community trust, acceptance and respect
for HEWs (mean score)
0.89 (3.37)
►Community perception towards HEWs
competence to provide service (mean
score)
0.62 (2.09)
*Percentage of respondents who have heard from HEWs
about any type of NCD prevention service, out of the six, at
least once in the past 12 months.
HEW, health extension worker; NCD, non- communicable
disease.
Table 2 Descriptive statistics of health extension workers'
characteristics and perception towards HEP, Ethiopian
health extension programme, 2022
Health extension workers characteristics
and perception
% or mean
(weighted)
HEWs age (mean, years) 26.3 (4.8)
18–24 years 21.52
25–34 76.42
35–50 2.06
HEWs length of service year (mean, years) 7.1 (4.7)
0 –<5 23.29
5 –<10 19.69
10–16 57.02
Place where the HEW grew up (%)
Outside of woreda 23.94
In the woreda 47.35
In the kebele 28.71
HEWs level of education (%)
Level 1–3 49.89
Level 4 50.11
HEW’s marital status (%)
Currently in union 74.64
Currently not in union 25.36
Area the HEWs' live (%)
In nearby town 37.11
In the kebele 62.89
HEW who participated in NCDs short- term
training (%)
21.86
HEW self- reported level of competence to
measure BP (%)
91.94
HEWs’ perception towards HEP and health
system support, composite index (mean
score, SD)
−0.39 (1.02)
BP, blood pressure; HEW, health extension worker; NCD, non-
communicable disease.
6TesemaAG, etal. BMJ Global Health 2022;7:e009025. doi:10.1136/bmjgh-2022-009025
BMJ Global Health
coverage of NCD prevention uptake. Similarly, delays in
conducting routine HEW performance assessments had
an adverse effect, as were the health centre’s involvement
in providing HEP packages training for HEP coordina-
tors. Each month’s delay in conducting HEWs perfor-
mance reviews lowered community uptake of NCD
prevention services by 6% (OR=0.94, CI 0.91 to 0.97).
There was a negative association between programme
performance on HIV/AIDS and community uptake of
NCD preventive services (OR=0.15, CI 0.03 to 0.85). We
found a strong socioeconomic gradient with NCD preven-
tive service utilisation. Those in the highest, higher
and middle SES categories had a fivefold (OR=5.22,
CI 2.66 to 10.25)), 3.8- fold (OR=3.80, CI 1.91 to 7.56)
and twofold (OR=2.11, CI 1.24 to 3.60)) higher uptake
of NCD services than those in the lowest SES category.
However, other sociodemographic attributes such as age
and gender of residents did not affect the utilisation of
NCD preventive services.
The variance component was statistically significant
across the four models. The addition of health system,
HEW and community- level characteristics have increased
rather than lessened the residual between- cluster vari-
ance. With a between- cluster variance of 1.85 (1.07 to
3.20), p<0.001) shown for the final model, the estimated
intraclass correlation was 0.36, meaning that a third of
the variance in NCD service uptakes was due to the varia-
tions between woredas.
DISCUSSION
For close to two decades, Ethiopia has implemented
a PHC- focused HEP to increase coverage of essential
services and promote health outcomes.32 While studies
have examined the health impacts of the HEP, its effect
on NCD preventive service uptake and the competition
and complementarity with other existing programmes, if
any, remain unexplored.4 33 34
Our results showed that utilisation of NCD prevention
services through the HEP remains limited. Less than a
quarter (22%) of the study population had accessed NCD
services 12 months before the survey, and service uptake
was significantly associated with the SES of residents.
Those in the highest, higher and middle SES categories
had greater access to NCD preventive services than resi-
dents in the lower and lowest SES categories. A previous
study in Ethiopia also showed a similar finding for the
country’s other existing services—mothers from higher
income families were more likely to visit health posts than
lower income families.35 This is concerning, given that
the HEP itself was explicitly designed to improve access
for underserved groups and address existing health
inequalities.7
Our study identified several factors operating at the
health system level that demonstrate the complemen-
tarity and competition between NCD and other essential
preventive services.
Effects of district-level health system inputs and processes
We found that systematic collection of NCD reports
from health posts facilitates the uptake of NCD services.
These findings are consistent with previous studies where
data reporting and data utilisation played a crucial role
in improving health service delivery at the community
level.24 However, we found that health posts’ engage-
ment in compiling community profiles adversely affects
coverage of NCD services, which could indicate potential
competition between time spent on core activities versus
other commitments. A time–motion study conducted
in Ethiopia showed that recordkeeping, reporting and
managing family folders consume more than a tenth of
available time,36 which was considered substantial and
believed to have had adverse effects on new HEP pack-
ages.
Health system inputs, such as health information
management systems and programmatic processes
like HEW training, supportive supervision, and perfor-
mance appraisal, are essential for developing problem-
solving skills and fostering quality of care practices within
service delivery settings.17 24 Our result showed a strong
link between HEW performance evaluation practices
and NCD prevention services, with delays in conducting
Table 3 Descriptive statistics of health system inputs
for NCD prevention service, Ethiopian health extension
programme, 2022
Sample characteristics (n=9680)
% or mean
(weighted)
Health system inputs and process
Health post compile community (kebele)
information prole (%)
73.38
Health centre’s HEP coordinator received
training on HEP packages (%)
36.49
HEWs’ involvement in community activities
other than HEP (%)
82.16
Health post had received supportive
supervision from woreda health ofce in the
last 6 months (%)
46.49
Health centres provide supervision to health
post (%)
99.89
Health post with blood pressure (BP)
measurement service (%)
62.20
Most recent performance assessment made
for HEW by (Month/s ago) (mean, SD)
1.96 (2.15)
Community members involved in the
performance assessment of HEWs
No 84.24
Yes 15.76
Health centres provide training for HEWs in
the last 2 year (%)
37.37
Health centres collect NCD reports from
health posts (%)
31.09
HEW, health extension worker; NCD, non- communicable disease.
TesemaAG, etal. BMJ Global Health 2022;7:e009025. doi:10.1136/bmjgh-2022-009025 7
BMJ Global Health
assessments having a significant adverse effect on NCD
uptake. While a qualitative study in Ethiopia suggests that
PHC services receiving greater attention are more likely
to be implemented,23 our findings demonstrate comple-
mentary effects, as none of the health system factors—
such as overall HEW training, supportive supervision,
and performance appraisal—captured in our study was
explicitly aimed at NCDs.
However, the finding on training, which showed atten-
dance by HEP coordinators as having a negative impact
on NCD uptake, is contrary to conventional beliefs. In
general, attending training programmes is expected to
improve know- how, enhance health workers’ perfor-
mance and a programme coordinator’s capacity to deliver
services.17 24 While our data do not explicitly capture
the type, quality or duration of training received by
programme coordinators, the contents of most training
opportunities in Ethiopia are known to focus on HIV,
MCH and malaria due to the high priorities accorded to
those services.7 These may explain the unexpected rela-
tionship between HEP coordinator training and NCD
service utilisation observed in the current study.
In the current study, the involvement of woreda admin-
istrators in HEWs supervision also did not show any
significant association with improving NCD prevention
service coverage. While some studies found results like
ours,37 the absence of association in the current research
is contrary to evidence that supportive supervision is crit-
ical for programme success and effective NCD service
delivery by CHWs.14 17 24 As evidence suggests, supervision
needs to be linked to performance goals, targeted to a
specific group and focused on particular knowledge and
skillset to have the desired effect.17 It also requires a well-
trained staff who have subject matter know- how and are
able to guide and support HEWs. The absence of a strong
relationship in the current study may be attributed to the
lack of quality supervision and insufficient attention to
NCD services during supervisory visits. Previous research
Figure 2 Forest plot: health system, health worker and community characteristics effects on receiving NCD prevention
service, Ethiopia: 2022. HEW, health extension worker; NCD, non- communicable disease; SES, socioeconomic status; TB,
tuberculosis.
8TesemaAG, etal. BMJ Global Health 2022;7:e009025. doi:10.1136/bmjgh-2022-009025
BMJ Global Health
on Ethiopia’s NCD programme at the PHC level indi-
cated that supervisors had minimal knowledge of NCDs
while the supervision system itself lacked NCD focus.23
Such a lack of a deeper understanding of NCD knowl-
edge among supervisors is likely to lead to a greater
emphasis on supporting programmes, they were familiar
with and trained on; hence the insignificant association
with NCD observed in the present study.
Effect of HEW characteristics and perception towards lack of
health systems support
The role of HEW attributes in improving access and
coverage of essential health services has been docu-
mented in Ethiopia, especially for MCH, HIV, TB and
malaria programmes.4 33 34 38 39 Even though evidence
on CHWs and NCDs is still evolving, the few studies else-
where highlight the potential role of CHW in delivering
NCD interventions.13–18 In line with these studies, we
found that NCD uptake was positively and significantly
associated with HEW qualification and living arrange-
ments in the communities they served. Uptake of NCD
preventive services tended to be higher if the HEW had
level IV qualifications and lived in the same community.14
This may be because, in Ethiopia, HEWs with level IV
qualifications, as opposed to level III or below, receive
NCD- focused content as part of their additional year- long
training on the HEP.
However, unlike the case for family planning, reproduc-
tive health and HIV/AIDS services, community uptake
of NCD preventive services was not affected by other
demographic attributes of HEWs, such as age, gender
or marital status. On the other hand, their service year
and perceived lack of support from the health system
had an adverse impact on NCD programme uptake. This
demonstrates yet another complementary effect—in
that HEWs’ perception was not specific to NCDs but cut
across other services. Although work experience gener-
ally improves performance, the negative association
observed for service years in the present study may be
attributed to frustration with the system. The absence of
career advancement opportunities also limits long- term
commitment among health workers and encourages
frequent job changes.14 In Bangladesh and China, where
CHWs received the required health system support, such
as training and necessary equipment, they were confi-
dent in delivering essential NCD- related services. In
contrast, in Viet Nam and Nepal, where CHWs were not
adequately trained and did not have the necessary equip-
ment, they were less confident and lacked the skills to
deliver NCD- related services.14
In addition, we observed that HEWs’ participation in
short- term NCD training did not improve community
NCD service uptake. The negative association between
NCD service coverage and short- term training remained
even after we controlled for potential interaction with
HEWs’ level of qualification. The negative effect was, in
fact, more substantial for those with level III qualifica-
tions or below, indicating that skill upgrading through
short- term training alone may not solve the problem.7
This may be so because short- term training for CHWs
often focus on a limited set of skills (ie, BP measurement
for hypertension) rather than comprehensive training
that covers several common NCDs (ie, hypertension,
diabetes and cancer).14 The finding may also further
suggest another complementary effect—the overall low
level of competence of HEWs may itself have limited their
ability to benefit from short- term NCD training while at
the same time adversely affecting the quality- of- service
provision at the community level.7
Effect of community characteristics and perception towards
HEP and HEWs
Several studies have shown that community- level NCD
service provision depends on the broader SES of the
community, their perception of health programmes in
the community and CHWs’ capacity to deliver mandated
services.14 20 Our findings suggest that such complementa-
rity as the uptake of NCD preventive services was higher if
residents already hold favourable views towards the HEP,
trust HEWs and their competency in providing services.
This finding is similar to previous studies in Ethiopia for
other existing HEP services, which demonstrated a posi-
tive association between programme success and commu-
nity perception towards HEWs’ role and the HEP more
generally.19 20 These studies showed that maternal health
outcomes tend to be positive if HEWs have a good rela-
tionship with the local community and are trusted with
their skill set.19 20 Community trust and acceptance of
health workers and positive views on programmes under
implementation create a conducive work environment
for CHWs and contribute to their retention, motivation,
performance, accountability and receiving support from
the community.
Effect of other programmes on NCD service delivery
There is growing evidence that integrating NCD care
with other existing services is feasible and an opportunity
to achieve Universal Health Coverage.40 41 Should inte-
gration prove to be effective, it can increase coverage of
NCD services and potentially improve health outcomes
for those with NCDs. However, our study showed no
association (neither complementary nor competitive)
between TB and hand hygiene programmes and NCD
coverage. In contrast, HIV/AIDS programme perfor-
mance had statistically significant adverse (ie, competi-
tive) effects. The negative association between HIV/AIDs
programme performance and NCD coverage supports
the view that integration with new programmes could,
competitively, spread resources too thinly in the face of
a dwindling funding opportunity for HIV programmes.42
This, in turn, may jeopardise the success of HIV services
and may result in worse health outcomes for people
living with HIV and little benefit for those living with
NCDs.
TesemaAG, etal. BMJ Global Health 2022;7:e009025. doi:10.1136/bmjgh-2022-009025 9
BMJ Global Health
Strengths and limitations of the study
Finally, a few caveats around the research are in order.
The data used in our study were collected in 2019, but
much has changed since in Ethiopia. The ongoing civil
war in the country is likely to have shifted the attention
and focus of the government, reoriented policy and
funding priorities and disrupted health services previ-
ously available to the population. Moreover, like the rest
of the world, Ethiopia is in the middle of the COVID- 19
pandemic, which also impacted priorities and service
coverage across the health sector. Findings from the
present study and application of results for designing
future programmes, thus, need to take these events and
developments into account.
Another limitation of our analysis is the cross- sectional
nature of the data, as it only shows association and does
not necessarily control for lag effects. For example, the
opposite associations between NCD specific short- term
training for HEW and NCD prevention service uptake
could be mere reflections of lack of adequate observa-
tional time and teething effects related to recruitment
and targeting. Second, the competing relationship
between NCD preventive services and HIV/AIDS is likely
to evolve. A complete picture of such dynamic relation-
ships requires much more than a cross- sectional evalu-
ation design can provide. In addition, the inclusion of
other existing services, such as MCH, in the analysis
could have provided additional insight into the synergy
between NCD and other HEP services. Our data had
only a limited number of comparable information on
existing HEP preventive services. Hence, future evalua-
tions should adopt longitudinal designs to address these
shortcomings.
Despite the limitations, the study also has several
strengths. It is the most extensive programme evalua-
tion of HEP in Ethiopia and probably one of the largest
CHWs evaluations in Africa. It is also the first- ever study
globally to empirically explore the complementarity and
competition between NCD and other essential preventive
services.
CONCLUSION
Results on a range of complementary effects point to
the need for an overall strong HEP capable of delivering
existing services as a prerequisite for successfully inte-
grating new (in this case, NCD) services.43 For example,
strengthening overall training and supervision and
building trust between the community and the HEWs
are essential for a successful HEP, just as they are for
individual services within the HEP and integrating new
ones such as NCD preventive services. However, integra-
tion efforts also require a carefully balanced approach,
so that the success already recorded for some existing
programmes is not lost.
Author afliations
1The George Institute for Global Health, Faculty of Medicine, University of New
South Wales, Sydney, New South Wales, Australia
2School of Public Health, College of Health Sciences, Mekelle University, Mekelle,
Ethiopia
3School of Population Health, Faculty of Medicine, University of New South Wales,
Sydney, New South Wales, Australia
4The George Institute for Global Health, New Delhi, India
5School of Public Health, Faculty of Medicine, University of Sydney, Sydney, New
South Wales, Australia
6Monitoring, Evaluation, Research, and Quality Improvement (MERQ), Ethiopia
ofce, MERQ Consultancy PLC, Addis Ababa, Ethiopia
7Department of Public Health, Faculty of Health, University of Canberra, Canberra,
ACT, Australia
8Department of Health Science Metrics, University of Washington, Seattle,
Washington, USA
Twitter Azeb Gebresilassie Tesema @Azebdej, David Peiris @davidpeiris, Rohina
Joshi @RohinaJoshi and Seye Abimbola @seyeabimbola
Acknowledgements AGT would like to acknowledge the nancial support
received from UNSW through the Scientia PhD Scholarship program. All authors
would also like to thank MERQ consultancy group and Ministry of Health, Ethiopia
for providing the national assessment Survey data, which was funded by the Bill &
Melinda Gates Foundation grant.
Contributors Conceptualisation: AGT, YK, DP, RJ and SA. Investigation: AGT,
YK, AMT and FWF. Data curation: AGT, AMT and FWF. Formal analysis: AGT.
Visualisation: AGT. Writing—original draft preparation: AGT and YK. Writing—review
and editing: AGT, YK, DP, RJ, SA, AMT and FWF. Supervision: YK, DP, RJ and SA.
AGT, YK, DP, RJ and SA provided critical intellectual input during the analysis and
interpretation.
Funding The UNSW Scientia Scholarship program supports AGT. SA was
supported by the Australian National Health and Medical Research Council
(NHMRC) through an Overseas Early Career Fellowship (APP1139631). RJ is
supported by the Australian National Heart Foundation (APP 102059) and a UNSW
Scientia Fellowship. DP is supported by NHMRC career Development Fellowship,
Level 2 and Australia National Heart Foundation Future Leader Fellow. AT and FWF
are employed by MERQ Consultancy PLC.
Competing interests AGT, DP, RJ and YK declare no competing interests. SA
is Editor in Chief of BMJ Global Health. AT and FWF are employees of MERQ
Consultancy PLC, however, this does not alter our adherence to BMJ Global Health
policies on sharing data and materials.
Patient and public involvement Patients and/or the public were not involved in
the design, or conduct, or reporting, or dissemination plans of this research.
Patient consent for publication Not applicable.
Ethics approval The EHAS has been approved by the National Research and
Research Ethics Review Board in Ethiopia under the Ethiopian Public Health Institute
(EPHI- IRB- 151- 2018), Ethiopia. We obtained ethics approval from the University of
New South Wales (UNSW) Human Research Ethics Committee (HC210066), Sydney,
Australia to conduct secondary data analysis of the EHAS. Participants gave informed
consent to participate in the study before taking part.
Provenance and peer review Not commissioned; externally peer reviewed.
Data availability statement Data may be obtained from a third party and are not
publicly available. All relevant data contributing to the ndings are within the paper
and in Supplementary table 1. As secondary data users, we are restricted by data
sharing policy and ethical clearance to share additional data. All request for the
original data should be directed to the data custodian, the MERQ Consultancy.
Supplemental material This content has been supplied by the author(s). It
has not been vetted by BMJ Publishing Group Limited (BMJ) and may not have
been peer- reviewed. Any opinions or recommendations discussed are solely
those of the author(s) and are not endorsed by BMJ. BMJ disclaims all liability
and responsibility arising from any reliance placed on the content. Where the
content includes any translated material, BMJ does not warrant the accuracy and
reliability of the translations (including but not limited to local regulations, clinical
guidelines, terminology, drug names and drug dosages), and is not responsible
for any error and/or omissions arising from translation and adaptation or
otherwise.
Open access This is an open access article distributed in accordance with the
Creative Commons Attribution 4.0 Unported (CC BY 4.0) license, which permits
others to copy, redistribute, remix, transform and build upon this work for any
purpose, provided the original work is properly cited, a link to the licence is given,
10 TesemaAG, etal. BMJ Global Health 2022;7:e009025. doi:10.1136/bmjgh-2022-009025
BMJ Global Health
and indication of whether changes were made. See:https://creativecommons.org/
licenses/by/4.0/.
Author note The reexivity statement for this paper is linked as an online
supplemental le 2.
ORCID iDs
DavidPeiris http://orcid.org/0000-0002-6898-3870
SeyeAbimbola http://orcid.org/0000-0003-1294-3850
REFERENCES
1 Assebe LF, Belete WN, Alemayehu S, etal. Economic evaluation
of health extension program packages in Ethiopia. PLoS One
2021;16:e0246207.
2 Global Burden of Diseases (GBD) COMPARE. Analyze updated data
about the world’s health levels and trends from 1990 to 2019 in this
interactive tool using estimates from the Global Burden of Disease
(GBD) study, 2019. Available: https://vizhub.healthdata.org/gbd-
compare/
3 Bollyky TJ, Templin T, Cohen M, etal. Lower- Income countries that
face the most rapid shift in noncommunicable disease burden are
also the least prepared. Health Aff 2017;36:1866–75.
4 Ministry of Health Ethiopia. The National assessment of the health
extension program. abridged report. Addis Ababa: Ministry of Health
Ethiopia, 2019.
5 Federal Democratic Republic of Ethiopia MoH. National Strategic
Action Plan (NSAP) for Prevention & Control of Non- Communicable
Diseases in Ethiopia. 2014 – 2016.
6 Federal Democratic Republic of Ethiopia MoH. Health sector
transformation plan, 2015/16 - 2019/20.
7 Ethiopia Federal Ministry Health. Realizing Universal Health
Coverage Through Primary Health Care. A Roadmap for Optimizing
the Ethiopian Health Extension Program 2020 - 2035. Addis Ababa:
Ethiopia, July 2020.
8 Wang H, Tesfaye R, Ramana GNV. Ethiopia health extension
program an institutionalized community approach for universal
health coverage. A world bank study 2016.
9 Tilahun H, Fekadu B, Abdisa H, etal. Ethiopia's health extension
workers use of work time on duty: time and motion study. Health
Policy Plan 2017;32:320–8.
10 Federal Democratic Republic of Ethiopia MoHHeaPHSD. Realizing
Ethiopia’s Long- term Primary Health Care (PHC) Vision Through the
2nd Generation Health Extension Program (HEP).
11 Federal Democratic Republic of Ethiopia MoH. Second generation
health extension program major non- communicable diseases
prevention and control package 2015.
12 Federal Democratic Republic of Ethiopia MoH. The Ethiopia
Noncommunicable Diseases and Injuries (NCDI) Commission Report
Summary. In: Addressing the impact of non- communicable diseases
and injuries in Ethiopia, 2018.
13 Rachlis B, Naanyu V, Wachira J, etal. Community perceptions of
community health workers (CHWs) and their roles in management
for HIV, tuberculosis and hypertension in Western Kenya. PLoS One
2016;11:e0149412- e.
14 Abdullah AS, Rawal LB, Choudhury SR. Use of community health
workers to manage and prevent non- communicable diseases. In:
Asia WHOROfS- E. New Delhi, 2019.
15 Rawal L, Jubayer S, Choudhury SR, etal. Community health
workers for non- communicable diseases prevention and control in
Bangladesh: a qualitative study. Glob Health Res Policy 2021;6:1.
16 Musoke D, Atusingwize E, Ikhile D, etal. Community health workers'
involvement in the prevention and control of non- communicable
diseases in Wakiso district, Uganda. Global Health 2021;17:7.
17 WHO. Who guideline on health policy and system support to
optimize community health worker programmes. Sected highlights
2018.
18 Jeet G, Thakur JS, Prinja S, etal. Community health workers for
non- communicable diseases prevention and control in developing
countries: evidence and implications. PLoS One 2017;12:e0180640.
19 Getachew T, Abebe SM, Yitayal M, etal. Health extension workers'
perceived health system context and health post preparedness to
provide services: a cross- sectional study in four Ethiopian regions.
BMJ Open 2021;11:e048517.
20 Kok MC, Kea AZ, Datiko DG, etal. A qualitative assessment of
health extension workers' relationships with the community and
health sector in Ethiopia: opportunities for enhancing maternal
health performance. Hum Resour Health 2015;13:80.
21 Master Facility Registry. (MFR) [Internet]. Ministry of Health,
Ethioipia, 2022. Available: https://mfrv2.moh.gov.et/#/dashboard
22 United Stated Agency for International Development (USAID). Health
Facility Governance in the Ethiopian Health System. In: Bethesda
MD, ed. HealthSystems20/20. 20814 | USA, 2012.
23 Tesema AG, Abimbola S, Mulugeta A, etal. Health system capacity
and readiness for delivery of integrated non- communicable disease
services in primary health care: a qualitative analysis of the Ethiopian
experience. PLOS Glob Public Health 2021;1:e0000026.
24 Agarwal S, Sripad P, Johnson C, etal. A conceptual framework for
measuring community health workforce performance within primary
health care systems. Hum Resour Health 2019;17:86.
25 Goldstein H. Nonlinear multilevel models, with an application to
discrete response data. Biometrika 1991;78:45–51.
26 Rabe- Hesketh S, Skrondal A. Multilevel and longitudinal modeling
using Stata. 3rd ed. Stata Press books from StataCorp LP, 2012.
27 World Health Organization. Primary health care systems
(PRIMASYS): case study from Ethiopia. Geneva, 2017.
28 Wang H, Tesfaye R, Ramana GN. Ethiopia health extension program:
an institutionalized community approach for universal health
coverage. World Bank Publications, 2016.
29 Bergen N, Ruckert A, Kulkarni MA, etal. Subnational health
management and the advancement of health equity: a case study of
Ethiopia. Glob Health Res Policy 2019;4:12.
30 StataCorp. Stata: release 17. statistical software. College Station,
TX: StataCorp LLC, 2021. https://www.stata.com/manuals/u.pdf
31 Goldstein H, Rasbash J. Improved approximations for multilevel
models with binary responses.. J R Stat Soc Ser A Stat Soc
1996;159:505–13.
32 Assefa Y, Gelaw YA, Hill PS, etal. Community health extension
program of Ethiopia, 2003- 2018: successes and challenges toward
universal coverage for primary healthcare services. Global Health
2019;15:24.
33 Karim AM, Admassu K, Schellenberg J, etal. Effect of Ethiopia's
health extension program on maternal and newborn health care
practices in 101 rural districts: a dose- response study. PLoS One
2013;8:e65160.
34 Jackson R, Hailemariam A. The role of health extension workers in
linking pregnant women with health facilities for delivery in rural and
Pastoralist areas of Ethiopia. Ethiop J Health Sci 2016;26:471–8.
35 Yitayal M, Berhane Y, Worku A, etal. Health extension program
factors, frequency of household visits and being model households,
improved utilization of basic health services in Ethiopia. BMC Health
Serv Res 2014;14:156.
36 Mangham- Jefferies L, Mathewos B, Russell J, etal. How do health
extension workers in Ethiopia allocate their time? Hum Resour
Health 2014;12:61.
37 Getachew T, Abebe SM, Yitayal M, etal. Association between a
complex community intervention and quality of health extension
workers' performance to correctly classify common childhood
illnesses in four regions of Ethiopia. PLoS One 2021;16:e0247474.
38 Medhanyie A, Spigt M, Kie Y, etal. The role of health extension
workers in improving utilization of maternal health services in rural
areas in Ethiopia: a cross sectional study. BMC Health Serv Res
2012;12:352.
39 Admassie A, Abebaw D, Woldemichael AD. Impact evaluation of
the Ethiopian health services extension programme. J Dev Effect
2009;1:430–49.
40 Tapela NM, Tshisimogo G, Shatera BP, etal. Integrating
noncommunicable disease services into primary health care, Botswana.
Bull World Health Organ 2019;97:142–53.
41 Foo CD, Shrestha P, Wang L, etal. Integrating tuberculosis and
noncommunicable diseases care in low- and middle- income countries
(LMICs): a systematic review. PLoS Med 2022;19:e1003899.
42 Adeyemi O, Lyons M, Njim T, etal. Integration of non- communicable
disease and HIV/AIDS management: a review of healthcare policies and
plans in East Africa. BMJ Glob Health 2021;6:e004669.
43 Topp SM, Abimbola S, Joshi R, etal. How to assess and
prepare health systems in low- and middle- income countries for
integration of services- a systematic review. Health Policy Plan
2018;33:298–312.
