Global Health Action
ISSN: 1654-9716 (Print) 1654-9880 (Online) Journal homepage: https://www.tandfonline.com/loi/zgha20
Evaluating health service coverage in Ghana’s
Volta Region using a modified Tanahashi model
Mallory C. Sheff, Ayaga A. Bawah, Patrick O. Asuming, Pearl Kyei, Mawuli
Kushitor, James F. Phillips & S. Patrick Kachur
To cite this article: Mallory C. Sheff, Ayaga A. Bawah, Patrick O. Asuming, Pearl Kyei, Mawuli
Kushitor, James F. Phillips & S. Patrick Kachur (2020) Evaluating health service coverage in
Ghana’s Volta Region using a modified Tanahashi model, Global Health Action, 13:1, 1732664,
DOI: 10.1080/16549716.2020.1732664
To link to this article:  https://doi.org/10.1080/16549716.2020.1732664
© 2020 The Author(s). Published by Informa
UK Limited, trading as Taylor & Francis
Group.
Published online: 16 Mar 2020.
Submit your article to this journal 
Article views: 1062
View related articles 
View Crossmark data
Full Terms & Conditions of access and use can be found at
https://www.tandfonline.com/action/journalInformation?journalCode=zgha20
GLOBAL HEALTH ACTION
2020, VOL. 13, 1732664
https://doi.org/10.1080/16549716.2020.1732664
ORIGINAL ARTICLE
Evaluating health service coverage in Ghana’s Volta Region using a modified
Tanahashi model
Mallory C. Sheff a, Ayaga A. Bawah b, Patrick O. Asuming c, Pearl Kyei b, Mawuli Kushitor b,
James F. Phillips a and S. Patrick Kachur a
aHeilbrunn Department of Population and Family Health, Mailman School of Public Health, Columbia University, New York, NY, USA;
bRegional Institute for Population Studies, University of Ghana, Accra, Ghana; cDepartment of Finance, University of Ghana Business
School, Accra, Ghana
ABSTRACT ARTICLE HISTORY
Background: The United Nations 2030 Sustainable Development Goals have reaffirmed the Received 13 September 2019
international community’s commitment to maternal, newborn, and child health, with further Accepted 28 January 2020
investments in achieving quality essential service coverage and financial protection for all.
Objective: Using a modified version of the 1978 Tanahashi model as an analytical framework RESPONSIBLE EDITOR
for measuring and assessing health service coverage, this paper aims to examine the system Jennifer Stewart Williams,Umeå University, Sweden
of care at the community level in Ghana’s Volta Region to highlight the continued reforms
needed to achieve Universal Health Coverage. KEYWORDS
Methods: The Tanahashi model evaluates health system coverage through five key measures Health systems; systems
that reflect different stages along the service provision continuum: availability of services; evaluation; health policy;
accessibility; initial contact with the health system; continued utilization; and quality cover- universal health coverage;
age. Data from cross-sectional household and health facility surveys were used in this study. primary health care
Immunization and antenatal care services were selected as tracer interventions to serve as
proxies to assess systems bottlenecks.
Results: Financial access and quality coverage were identified as the biggest bottlenecks for
both tracer indicators. Financial accessibility, measured by enrollment in Ghana’s National
Health Insurance Scheme was poor with 16.94% presenting valid membership cards.
Childhood immunization was high but dropped modestly from 93.8% at initial contact to
76.7% quality coverage. For antenatal care, estimates ranged from 65.9% at initial visit to
25.1% quality coverage.
Conclusion: Results highlight the difficulty in achieving high levels of quality service cover-
age and the large variations that exist within services provided at the primary care level.
While vertical investments have been prioritized to benefit specific health services,
a comprehensive systems approach to primary health care needs to be further strengthened
to reach Ghana’s Universal Health Coverage objectives.
Background health systems barriers to service delivery and access to
quality care at regional, district, and sub-district levels,
The United Nations 2030 Sustainable Development
despite the fact that service delivery in LMICs is often
Goals have reaffirmed the international community’s
decentralized and focused at lower levels of care.
commitment to maternal, newborn, and child health,
Understanding peripheral level health systems rather
with additional investments in achieving quality essential
than those at global or national scale is notably important
service coverage and financial protection for all.
in sub-Saharan Africa where regional variations within
Universal Health Coverage (UHC), promoting the right
the same country can be extensive [5].
to health for vulnerable, poor, and remote populations,
Like many countries, Ghana’s approach to UHC
has therefore become embedded as a global priority.
includes several key policy initiatives that prioritize
Indeed, providing effective, affordable, and quality pri- the needs of underserved groups [6]. Ghana’s
mary health-care services can improve maternal and Community-based Health Planning and Services
child health and survival in low-and-middle-income (CHPS) policy, launched and adopted in 1999, aims
countries (LMICs) [1–3], yet this impact is not always to reduce barriers to geographical access to health care
realized. Ongoing initiatives focusing on implementing by providing primary health care (PHC) services at the
vertical programs have constrained official efforts to community level [7]. By posting a resident nurse in
evaluate overall health system coverage and strength, communities with an additional mandate to provide
and understand barriers at various levels of the system outreach services, Ghana aimed to ensure that health
[4]. Few studies provide a comprehensive overview of care reached the most vulnerable and remote
CONTACT Mallory C. Sheff mallory.sheff@columbia.edu Department of Population and Family Health, Columbia University Mailman School of
Public Health, 60 Haven Avenue, B2, New York, NY 10032, USA
© 2020 The Author(s). Published by Informa UK Limited, trading as Taylor & Francis Group.
This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/4.0/), which permits
unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
2 M. C. SHEFF ET AL.
populations. CHPS remains a core component of the Methods
government’s strategy and has contributed to remark-
Study setting
able reductions in maternal and child mortality,
antenatal care (ANC) coverage, and the rising perfor- While the CHPS+ project is implemented in both the
mance of immunization, particularly in underserved Northern and Volta regions of Ghana, this paper
rural areas [8,9]. The CHPS initiative aims to leverage focuses on data from the Volta Region (VR) (now the
the strength of PHC to achieve UHC, with a notable Oti & Volta Regions). In 2010, VR had a census-
emphasis on addressing availability and accessibility of enumerated population of 2.1 million, 38% of which
services [10]. This expanded geographic access is was under the age of 15 [15]. The majority (71%) of the
enhanced by progressive policies intended to address population is Christian. The estimated under-five mor-
financial barriers such as the National Health tality rate is 61 per 1000 live births, although mortality
Insurance Scheme (NHIS) and user fee exemptions risks are known to vary by district [16]. The Volta
for services that reach pregnant women, infants, and Region is one of the poorer performing regions in key
young children [11]. health indicators such as immunization coverage [17].
Project background The analytical framework: a modified Tanahashi
model
In 2016, a five-year project entitled A National
Program for Strengthening the Implementation of The Tanahashi model, first developed in a seminal
the Community-based Health Planning and Services paper in 1978, was modified by UNICEF, the World
(CHPS) Initiative (CHPS+) was launched and imple- Health Organization, and the World Bank in 2002 for
mented by the Ghana Health Service in collaboration use in the Marginal Budgeting for Bottlenecks tool to
with the Regional Institute for Population Studies at estimate the potential impact, resource needs, costs,
the University of Ghana, the University for and budgeting requirements to strengthen national
Development Studies, and the University of Health health systems [14]. This adaptation, applied in ana-
and Allied Sciences, with technical support from lyses in more than 50 countries, is a rare framework
Columbia University’s Mailman School of Public that moves attention beyond access to health services
Health. The CHPS+ project aims to develop and brings quality to the forefront, thus highlighting
a sustainable capacity to implement, monitor, and the effectiveness of health systems interventions and
evaluate health systems strength to improve national potential opportunities to optimize it [18].
capabilities to scale-up community-based primary This paper applies a modified version of the 1978
health-care quality and impact and achieve national Tanahashi model to evaluate health system coverage
UHC objectives [12]. This project aims to foster at the community level through its five key measures
a culture of health service excellence and systems (Table 1) that reflect different stages along the service
thinking previously tested by the Ghana Essential provision continuum [19,20]. These five stages por-
Health Intervention Programme in the country’s tray the complex interaction between the health sys-
Upper East Region [10,12,13]. CHPS+ is being tem and the population in ways that can highlight
scaled up from the Upper East to the Northern and
Volta Regions, where demonstration districts known Table 1. Tanahashi model coverage determinants.
as System Learning Districts provide models of Coverage determinants Definition
excellence in health system delivery and implemen- 1 a. Availability of Refers to the availability of health system
tation at the community level by strengthening the health commodities inputs. These include, for example,medicines and other necessary supplies.
country’s community health service delivery 1 b. Availability of Refers to the availability of trained
platform. human resources professionals at CHPS facilities that can
provided needed care.
2 a. Geographic Physical access to service delivery points.
accessibility
2 b. Financial Financial support or health insurance
Objective accessibility coverage to pay for medical care.
3. Initial contact Refers to the first contact or use of the
Using a modified version of the 1978 Tanahashi model health services or interventions.4. Continued utilization Refers to the repeated contact with the
as an analytical framework for measuring and assessing health system to receive necessary care
health service coverage, this paper aims to examine the 5. Quality coverage Represents the quality of a healthintervention defined by the full course of
system of care at the community level in CHPS+ Volta contact with the health system to receive
Region implementation districts to highlight the con- effective care and the minimum inputs
and processes to achieve defined health
tinued investments and implementation reforms effects.
needed for Ghana to achieve UHC [14]. Adapted from Henriksson et al. [22].
GLOBAL HEALTH ACTION 3
gaps in service delivery [21,22]. A particularly useful Afadzato South, Nkwanta North, Krachi East, and
element of this approach is its capability to distin- Krachi Nchumuru) over the April to
guish between potential and actual coverage, high- October 2017 period. Sampling was powered to
lighting gaps between available supply and demand detect a 15% reduction in under-five mortality at
or utilization. the end of the project, with 80% power at 5% level
This analysis will utilize CHPS+ cross-sectional base- of significance in each study region, and to draw
line evaluation data to assess health system strength a representative sample of women of reproductive
[14,19]. Given the challenge in accessing quality data age (15–49 years) in the CHPS+ project catchment
for a comprehensive picture of the primary health-care area, collecting information on both the health of
system at the community level, ‘tracer interventions’ women and of their children. The survey used a two-
will allow for the most relevant and local data to repre- stage stratified cluster sampling approach: the first-
sent the overall system [14]. For this analysis, immuni- stage sampled enumeration areas (EAs), and
zation and antenatal care data, which comprise essential the second-stage sampled households. EAs were stra-
components of maternal and child health and founda- tified by type of location (rural or urban) and by the
tional elements of the CHPS initiative, will be used as size of the EA (measured by the estimated number of
tracer interventions to represent the Volta Regional households in the EA). EAs were also stratified by
Health System’s capacity to deliver essential primary size into three groups (small, medium, or large), and
health-care services [23]. We further assess sub- subsequently sampled from each stratum using prob-
national service delivery by incorporating data on ability proportional to population size.
financial access to services in the model via individuals’ Immunization and antenatal care services were
enrollment in Ghana’s National Health Insurance selected as tracer interventions to serve as proxies to
Scheme. assess systems bottlenecks in the study region. The
The five distinct stages of the Tanahashi model dataset for both services was limited to the analysis of
(Figure 1) lend themselves to a bottom-up stepwise women and their living children aged 12 to 23 months
assessment of health services. In order to understand at the time of the survey. To reduce maternal respon-
shortcomings in potential coverage on the supply side dent recall bias, data for immunization were
(availability and accessibility) and those that affect restricted to the interviewer’s visual assessment of
actual coverage on the demand side (initial contact, each child’s immunization card.
continued utilization, and quality coverage), we A total of 11,201 women were interviewed in the
examine the differences between each level. Volta Region, 1515 of which had living children aged
12–23 months. Table 2 presents the socio-
demographic characteristics of the women and their
Data sources children included in the analysis.
Household survey
The CHPS+ project baseline household survey was Health facility assessment
conducted in seven of the 25 districts of the VR The CHPS+ health facility assessment was conducted in
(Central Tongu, Northern Tongu, Akatsi North, July 2018 and collected information on all facilities
Figure 1. Modified Tanahashi model.
4 M. C. SHEFF ET AL.
Table 2. Characteristics of women and their children of peripheral point of provision of primary health-care
12–23 months. services at the community level, especially in remote
Variable (mothers) Volta region locations. We furthermore include the primary health-
Age 15–24 416 (27.46%) care personnel that should always be available at both
25–34 703 (46.40%)
35–44 361 (23.83%) CHPS zones and health centers; these include
45–49 35 (2.31%) Community Health Officers (CHOs) or Community
Education No Education 595 (39.27%)
Primary 357 (23.56%) Health Nurses (CHNs), and Enrolled Nurses (EN).
Middle School/JHS/JSS 473 (31.22%) The health facility survey assessed facility readiness by
Secondary + 90 (5.94%)
Wealth index/SES Poorest 245 (16.17%) interviewing a health worker present at the time of inter-
status Poorer 293 (19.34%) view. This provided information regarding human
Middle 290 (19.14%)
Richer 326 (21.52%) resources and staffing, routine care, services provided at
Richest 361 (23.83%) the facility, and availability of equipment and
Religion parity Other 111 (7.33%) commodities.
Christian 1169 (77.16%)
Muslim 69 (4.55%)
Traditional 166 (10.96%)
1–3 872 (57.56%) Indicators
4–6 461 (30.43%)
7+ 182 (12.01%) The above data sources provide an overview of
Marital status Single 147 (9.70%)
Married 1076 (71.02%) the community health system in our study region.
Widowed 16 (1.06%) We use indicators from these CHPS+ data sources
Divorced 9 (0.59%)
Separated 70 (4.62%) to evaluate the five stages of the modified Tanahashi
Living together 197 (13.00%) model as per Table 4 below:
Occupation No occupation 221 (14.59%) Data were analyzed using Stata 13; the construc-
Student 13 (0.86%)
Farming 795 (52.48%) tion of the Tanahashi model graphical representation
Trading/Selling 328 (21.65%) was done in Microsoft Excel.
Hairdressing/Dressmaking 107 (7.06%)
Housewife 10 (0.66%)
Other 41 (2.71%)
TOTAL 1515* Results
Variable (children) Volta region
The below information provides step by step results
Child gender Male 813 (52.76%)
Female 728 (47.24%) for indicators tailored to each stage of the modified
Place of delivery Home 810 (52.56%) Tanahashi model for the two tracer interventions:
Govt Hospital 362 (23.49%)
Govt Health Center 277 (17.98%) childhood immunization and antenatal care. Figure 2
CHPS Compound 35 (2.27%) provides results according to the model.
Other 57 (3.70%)
Post-natal care at the 686 (44.52%)
facility
Post-natal care at 567 (36.79%) Tracer intervention: immunization
home
TOTAL 1541* 1a. Availability of human resources: The presence
*Our data set included information on 1541 children 12–23 months of of a CHO/CHN at a CHPS compound or
age (including 28 pairs of twins) belonging to 1515 individual women. health center in our study area is 92.41%.
Enrolled nurses are present in 57.93% of
delivering health services in the 7 CHPS+ survey dis- these community-based primary health-care
tricts of the VR. The survey included CHPS zones with centers. The overall presence of one of these
and without functioning service posts, commonly health cadres is 94.48%.
known as ‘CHPS compounds’, Sub-district Health 1b. Availability of essential health commodities and
Centers, private clinics, and District Hospitals. For this equipment: Of the 145 facilities surveyed, 141
analysis, we are limiting data to Ghana Health Service (97.24%) provide immunization services.
designated CHPS service catchment areas that are either Using vitamin A and a vaccine refrigerator as
equipped with or lacking functioning compounds and a proxy for availability of all immunization
to government health centers (HC) (Table 3). Where program commodities, overall availability is
services are operational, they function as the most 51.72%.
2a. Financial accessibility: Of the women with chil-
dren aged 12–23 months, only 16.94% had
Table 3. Number and facility type in the Volta Region. a valid NHIS card that was presented to and
7 CHPS+ Districts in VR seen by the CHPS+ project fieldworkers.
Facility type 2b.Geographic accessibility: Geo-located data were
CHPS zone without compound 42 used to determine geographic accessibility and
CHPS zone with compound 72
Health Center 31 calculate the distance between each surveyed
TOTAL 145 household and the nearest CHPS zones.
GLOBAL HEALTH ACTION 5
Table 4. Measures for immunization and ANC as indicators of health system strength.
Stage/Process component Indicator Definition Source
1a. Availability (commodities ● Proportion of CHPS zones and health centers (HC) with the necessary immunization services and Health facility
& equipment) supplies at the time of survey survey
● Proportion of CHPS facilities and HC with the necessary supplies to provide ANC, including manual
BP apparatus, Sulfadoxine/pyrimethamine, iron, tetanus-toxoid vaccine, and folic acid
1b. Availability (human ● Proportion of CHPS zones and HC with trained health personnel (CHO/CHN/EN) at the time of Health facility
resources) survey survey
2a. Accessibility (geographic) ● Proportion of the population within 5 km of a CHPS zone or HC Household
survey
2b. Accessibility ● Proportion of the population with a valid National Health Insurance card Household
(financial) survey
3. Initial Contact ● Proportion of children who received the BCG vaccine Household
● Proportion of women who have gone to a CHPS zone or HC for their first ANC visit survey
4. Continued Utilization ● Proportion of children who received all three doses of the DPT vaccine Household
● Proportion of women who have been to 4+ of their ANC visits at a CHPS zone or HC survey
5. Quality/Effective coverage ● Proportion of children who have received all vaccines mandated by Ghana’s Expanded Household
Programme on Immunization (EPI) by 24 months Survey
● Proportion of women who have received the minimum package of ANC services
Figure 2. Tanahashi model results for immunization and ANC coverage.
Overall, 64.767% of women live within 4 km of live within 4 km of either a CHPS zone or
a CHPS zone compound, as per national CHPS a health center.
policy, and 29.68% of women live within 4 km 3. Initial contact: Overall, BCG coverage was high
of a health center. Overall, 80.85% of women at 93.80%.
6 M. C. SHEFF ET AL.
4. Continued utilization: Of the 1306 children whose 4. Birth preparedness and complication planning
immunization card was seen, 1191 (91.19%) (birth and emergency plan, breastfeeding
received all three doses of a diphtheria, pertussis, counseling)
and tetanus (DPT)-containing vaccine. Sensitivity The proportion of women provided with the full
analysis comparing full DPT immunization to full minimum package of ANC services at either
polio immunization as an indicator for continued a CHPS zone or health center is 23.59%.
utilization showed that both measures were statis-
tically comparable with a high level of significance
(p < 0.001). Discussion
5. Quality/effective coverage: For our immunization
tracer indicator, we define effective coverage as Assessing health system strength for UHC
children having received all of the basic vaccina- The modified Tanahashi model provides an overview
tions as required by Ghana’s Expanded of the current state of the primary health-care system
Programme on Immunization (EPI), which in seven districts of Ghana’s Volta Region. Through
includes one dose of BCG at birth, three doses the application of immunization rates and antenatal
of the oral polio vaccine (excluding the dose visits as tracer indicators, the final models illustrate
given at birth), three doses of a DPT- gaps in the system that remain obstacles to ensuring
containing vaccine and hepatitis B vaccine at 6, access, coverage, and quality health services in the
10, and 14 weeks, and one dose of the measles VR. Results provide a deeper understanding of the
vaccine. The proportion of children with full provision of, and demand for, health services, as well
immunization is 76.72%. as Ghana’s position in achieving its Universal Health
Coverage ambitions.
Our data show that financial accessibility, as mea-
sured by NHIS enrollment, is poor – only 16.94% of
Tracer intervention: ANC services
women surveyed had valid NHIS cards. However,
1a. Availability of human resources is same as above. this did not appear to constrain the initiation of
1b. Availability of essential health commodities and childhood immunization or antenatal care, with
equipment: Of the 145 CHPS and health cen- 93.8% and 65.7% respondents making the first visits
ters included in the analysis, 86.62% had the for each service, respectively. These well-established
necessary commodities to provide a minimum interventions may be so highly valued by commu-
package of ANC services, including a manual nities that women and families were willing to pay
blood pressure apparatus, an adult scale, out-of-pocket [24,25] or relied on user fee exemp-
Sulfadoxine/pyrimethamine, iron tablets, teta- tions to lower the potential financial access barrier.
nus-toxoid vaccine, and folic acid. Results subsequently showed a substantial drop-off in
2a. Financial accessibility is the same as above. complete quality coverage for both tracer interven-
2b. Geographic accessibility is the same as above. tions. For childhood immunizations, this was modest,
3. Initial contact: 65.74% of women visited either dropping from 93.8% at initial contact to 76.7% qual-
a CHPS zone or HC as the location for their first ity coverage based on a composite indicator across
ANC visit. multiple antigens. But for ANC coverage, estimates
4. Continued utilization: 69.79% of women went to dropped from 65.7% at initial visit to 23.6% for
either a CHPS compound or a HC for at least 4 quality coverage.
of their ANC visits. Vertical policy and financial investments in ANC
5. Quality/effective coverage: For our ANC tracer and immunization have been significant over the
indicator, effective coverage means the provision years, and have enhanced the overall strength of the
of the minimum basic package of antenatal care, system. Ghana’s 1978 EPI operational policy aims to
including: promote immunization delivery to all children under
1. Identification of pre-existing health condi- five through routine provision of essential vaccines at
tions (weight and nutrition status, anemia, appropriate months of childhood age, and through
hypertension) national immunization days, and campaigns to combat
2. Early detection of complications arising during specific diseases such as poliomyelitis and measles [26].
pregnancy (pre-eclampsia, gestational diabetes) The Ghana Health Service implements this strategy
3. Health promotion and disease prevention through their Reaching Every District approach that
(tetanus toxoid vaccine, prevention and utilizes the country’s decentralized health system, spe-
treatment of malaria, nutrition counseling, cifically leveraging CHPS posts and the assignment of
micronutrient supplements [folic acid and personnel to hard-to-reach areas. Investment in immu-
iron tablets], family planning counseling) nization has likewise been very successful, with global
GLOBAL HEALTH ACTION 7
support from international actors including GAVI – be able to improve these outcomes through modest
The Vaccine Alliance, committing $US 224, 538,541 efforts that support this integration by leveraging
between 2001 and 2016 to strengthen logistics, training, evidence-based planning and data-driven decision-
and service delivery mechanisms [27]. making as key guiding components of health systems
Moreover, investments in Ghana’s ANC program strengthening.
have demonstrated its importance as a component of
primary health care, and a key service provided at the
Strengths and limitations
community level through CHPS. In 2005, Ghana’s
Ministry of Health introduced free maternal health- Since data used for this study were collected from one of
care delivery nationwide, of which antenatal care was Ghana’s 10 regions, results are not generalizable to the
a significant component. The policy aimed to reduce entire country. Moreover, the Volta Region study dis-
financial barriers to seeking maternal services [25]. In tricts were selected because they met the criteria for the
2007, this policy was formally integrated into the CHPS+ intervention and control areas, and not because
NHIS, which also expanded free ANC coverage they were representative of the VR. The cross-sectional
from public health facilities only to all public and study design that is employed does not permit temporal
accredited private and faith-based health-care provi- analyses of the causes and consequences of relationships
ders [28]. This was found to significantly increase the discerned for health-seeking behavior. In particular,
percentage of children whose mothers obtained at financial access is a known determinant of care seeking
least four ANC visits [29]. that cannot be investigated.
Despite policy integration to support free maternal Further, analyses are based on tools developed and
and child health services to promote universal access to data collected prior to the implementation of the
health care, the proportion of women who had verified CHPS+ project and the development of the current
valid insurance cards showing activemembership in the study. While estimates derived for each stage of the
NHIS was significantly lower than other reported stu- model were based on the best available project data,
dies [30]. Additional data from our study area show that the CHPS+ household and health facility surveys
86% of women responded yes to having ever registered were not specifically designed for the Tanahashi
as a member of NHIS, and approximately 36% reported model. If data collection was designed for estimating
having valid cards that were not valid upon inspection. the model, more specific information could have
This likely signifies that respondents are either misin- been compiled, especially regarding vaccine availabil-
formed about their insurance status or that barriers to ity by antigen. The range of data sources required to
regularly renewing memberships – including difficulty develop this modified Tanahashi model may also not
in affording renewal payment, poor satisfaction with be available in settings that do not have comparable
the services, timing of premium payment – are challen- data resources.
ging to overcome [25]. Service delivery challenges such The quality of immunization services could also
as inadequate registration materials may further delay have been more strictly defined as the proportion of
renewal and exacerbate low membership [25]. children who have received all doses of the vaccine at
Given the combined level of financial and pro- the exact time and interval specified in Ghana’s EPI
grammatic investment in both immunization and operational policy guidelines. However, authors
ANC, these two health-care components should remained consistent with the approach used to esti-
demonstrate near-perfect coverage in both the supply mate immunization coverage in standard cross-
and demand domains of the Tanahashi model, and sectional surveys such as the Demographic and
therefore be benchmarks for guiding the way towards Health Survey. Authors therefore looked at overall
achieving Universal Health Coverage. Yet, despite the immunization completion rather than timely immu-
fact that EPI has been integrated into CHPS, much of nization by antigen to measure the final component
its funding and management have remained highly of the model.
vertical. In addition, it has been possible to introduce The strength of this analysis lies in the general
new vaccines into the program without drastically overview it provides of the health system, guiding
altering the delivery method for further systems inte- implementers and policymakers to look into compo-
gration. These factors may contribute to the relatively nents that require additional attention, funding, and
higher level of quality coverage in our model. In programming to further strengthen comprehensive
contrast, ANC services require that peripheral health service delivery.
workers and subnational health systems integrate
multiple funding, supply, equipment, training, and
Conclusion
personnel streams to achieve comprehensive quality
care. These complexities may account for the much The modified Tanahashi models presented above
larger erosion of ANC coverage after initial contact. highlight the difficulty in achieving high levels of
Our findings suggest that district-level managers may quality service coverage in the Volta Region, and
8 M. C. SHEFF ET AL.
the large variations that exist within services provided Tanahashi model provides a deeper understanding of the
at the primary care level – even in the same facilities provision of, and demand for, health services, as well as
and at the hands of the same health workers and Ghana’s position in achieving Universal Health Coverage
ambitions.
managers. In order to provide quality health-care
services and reduce health disparities, policymakers
must have a comprehensive grasp of the factors that Data availability statement
constrain or enable the distribution of care, especially
The data that support the findings of this study are available
in reaching its most vulnerable population. This from the corresponding author, MCS, upon reasonable
model can therefore be used as a tool to identify request.
specific systems bottlenecks to better understand the
health system, leveraging results to target financial
and programmatic investments towards increasing ORCID
access and achieving high quality of care. Results Mallory C. Sheff http://orcid.org/0000-0002-2787-4045
further strengthen the need to balance vertical global Ayaga A. Bawah http://orcid.org/0000-0003-4864-4182
and national health investments with programming Patrick O. Asuming http://orcid.org/0000-0002-5010-
and implementation efforts focused on strengthening 3484
Pearl Kyei http://orcid.org/0000-0001-6233-5891
the overall health system in order to achieve UHC. Mawuli Kushitor http://orcid.org/0000-0001-8668-7162
James F. Phillips http://orcid.org/0000-0002-6720-7204
S. Patrick Kachur http://orcid.org/0000-0002-8586-6019
Acknowledgments
The authors gratefully acknowledge the contribution of the
Regional Institute for Development Studies team members References
to the development and execution of the project interven- [1] Rowe AK, de Savigny D, Lanata CF, Victora CG. How
tion and research program. can we achieve and maintain high-quality perfor-
mance of health workers in low-resource settings?
Author contributions Lancet. 2005 Jan;366(9490):1026–35. DOI: 10.1016/
S0140-6736(05)67028-6
MCS and SPK developed the idea for the manuscript. AAB, [2] Travis P, Bennett S, Haines PA, Pang T, Bhutta PZ,
POA, and PK led the data collection and data cleaning. Hyder AA, et al. Overcoming health-systems con-
MCS led the majority of the data analysis with contribu- straints to achieve the Millennium Development
tions and support from MK. MCS developed the draft of Goals. Lancet. 2004;364(9437):900–6. DOI: 10.1016/
the manuscript with support and contributions from SPK S0140-6736(04)16987-0
and JFP. [3] Haynes A, Sanders D, Lehmann U, Rowe AK, Lawn
JE, Walker DG, et al. Achieving child survival goals:
potential contribution of community health workers.
Disclosure statement Lancet. 2007;369(9579):2121–2131. DOI: 10.1016/
S0140-6736(07)60325-0
No potential conflict of interest was reported by the [4] Sanders D, Haines A. Implementation Research Is
authors. Needed to Achieve International Health Goals. PLoS
Med. 2006 Jun;3(6):e186. https://doi.org/10.1371/jour
nal.pmed.0030186
Ethics and consent [5] Speybroeck N, Paraje G, Prasad A, Goovaerts P,
Ebener S, Evans DB. Inequality in Human Resources
Ethical clearance to conduct this study was obtained from for Health: Measurement Issues. Geogr Anal. 2012
Columbia University’s Institutional Review Board (IRB), Apr;44(2):151–61. DOI: 10.1111/j.1538-4632.2012.
protocol number AAAR0315 and the IRB at the Ghana 00842.x
Health Service. [6] Neelsen S, O’Donnell O. Progressive universalism?
The impact of targeted coverage on health care access
and expenditures in Peru. Heal Econ (United
Funding information Kingdom). 2017;26(12):e179–203. DOI: 10.1002/
The CHPS+ project is implemented with support provided hec.3492
by the Doris Duke Charitable Foundation, Inc. (Grant # [7] Ministry of Health. National community health plan-
2016107). ning and services. 2014 Nov. p. 1–19.
[8] Phillips JF, Bawah A a, Binka FN. Accelerating repro-
ductive and child health programme impact with
Paper context community-based services: the Navrongo experiment
in Ghana. Bull World Health Organ. 2006 Dec;84
The provision of quality primary health-care services can (12):949–55. DOI: 10.2471/blt.06.030064
improve maternal and child health in lower and middle [9] Baatiema L, Skovda M, Rifkin S, Campbell C, Von
income countries, yet this impact is not always realized. Schoen-Angerer T. Assessing participation in a com-
Previous initiatives focusing on vertical programs have munity-based health planning and services pro-
constrained efforts to evaluate health system coverage and gramme in Ghana. BMC Health Serv Res.2013;13
understand barriers to care. Applying the modified (233). https://doi.org/10.1186/1472-6963-13-233
GLOBAL HEALTH ACTION 9
[10] Awoonor-Williams JK, Sory EK, Nyonator FK, Phillips District Empowerment for Scale-up (CODES): a com-
JF,Wang C, SchmittML. Lessons learned from scaling up plex district-level management intervention to
a community-based health program in the Upper East improve child survival in Uganda: study protocol for
Region of northern Ghana. GlobHeal Sci Pract. 2013Mar a randomized controlled trial. Trials. 2016;17
21;1(1):117–33. DOI: 10.9745/GHSP-D-12-00012 (135):1241–4. https://doi.org/10.1186/s13063-016-
[11] Brugiavini A, Pace N. Extending health insurance in 1241-4
Ghana: effects of the National Health Insurance [22] Henriksson DK, Fredriksson M, Waiswa P, Selling K,
Scheme on maternity care. Health Econ Rev. 2016 Peterson SS. Bottleneck analysis at district level to illus-
Dec;6(1):7. DOI: 10.1186/s13561-016-0083-9 trate gaps within the district health system in Uganda.
[12] Phillips JF, Awoonor-Williams JK, BawahAA. TheGhana Glob Health Action. 2017;10:1. DOI: 10.1080/
Essential Health Intervention Programme (GEHIP): A 16549716.2017.1327256
plausibility trial of health system strengthening in four [23] Bos E, Batson A. Using immunization coverage rates
rural districts of northern Ghana. Final report. 2016. for monitoring health sector performance:
[13] Phillips JF, Koku Awoonor-Williams J, Bawah AA, Measurement and Interpretation Issues (English).
Nimako BA, Kanlisi NS, Sheff MC, et al. What do you 2000. HNP discussion paper series. Washington, DC:
do with success? The science of scaling up a health World Bank. http://documents.worldbank.org/
systems strengthening intervention in Ghana. BMC c u r a t e d / e n / 6 0 7 7 2 1 4 6 8 7 6 3 7 8 3 0 7 0 /
Health Serv Res. 2018;18:484. DOI: 10.1186/s12913- Usingimmunization-coverage-rates-for-monitoring-
018-3250-3 health-sector-performance-Measurement-and-
[14] O’Connell T, Sharkey A. Reaching universal health Interpretation-Issues
coverage: using a modified Tanahashi model sub- [24] Witter S, Garshong B, Ridde V. An exploratory study
nationally to attain equitable and effective coverage. of the policy process and early implementation of the
New York, NY: UNICEF; 2013. free NHIS coverage for pregnant women in Ghana. Int
[15] Ghana Statistical Service. 2010 Population and Housing J Equity Health. 2013 Feb;12(1):16. DOI: 10.1186/
Census Report: Summary Report of Final Results. Accra, 1475-9276-12-16
Ghana. 2014. Available at: https://www.statsghana.gov. [25] Arthur E. Wealth and antenatal care use: implications
gh/gssmain/storage/img/marqueeupdater/Census2010_ for maternal health care utilisation in Ghana. Health
Summary_report_of_final_results.pdf Econ Rev. 2012 Dec;2(1):14. DOI: 10.1186/2191-1991-
[16] Ghana Statistical Service - GSS, Ghana Health Service 2-14
- GHS, and ICF International. 2015. Ghana [26] Government of Ghana. Comprehensive Multi-Year
Demographic and Health Survey 2014. Rockville, Plan for Immunizations (2015-2019). 2017; 2017
Maryland, USA: GSS, GHS, and ICF International. (2):473–82. Available at: https://extranet.who.int/coun
[17] Ghana Health Sevice. The Health System in Ghana, tryplanningcycles/sites/default/files/planning_cycle_
facts and Figures. 2018; Available from: internal-pdf:// repository/ghana/ghana_cmyp_2015-2019.pdf
0240438844/Facts-and-figures-2015_GHS.pdf% [27] Gavi Alliance. Ghana Injection of Hope: Success with
0Ahttp://www.moh.gov.gh/wpcontent/ uploads/2017/ dual vaccine introductions and overall high immuni-
07/Facts-and-figures-2015.pdf sation rates makes Ghana a model for others in Africa.
[18] McCollum R, Taegtmeyer M, Otiso L, Mireku M, 2017; Available from: http://www.gavi.org/delivering/
Muturi N, Martineau T, et al. Healthcare equity ana- ghana/
lysis: Applying the Tanahashi model of health service [28] Dalinjong PA, Laar AS. The national health insurance
coverage to community health systems following scheme: Perceptions and experiences of health care
devolution in Kenya. Int J Equity Health. 2019;18 providers and clients in two districts of Ghana.
(1):1–12. Health Econ Rev. 2012;2(1):1–13. DOI: 10.1186/
[19] Tanahashi T. Health service coverage and its evalua- 2191-1991-2-13
tion. Bulletin of the World Health Organization. 1978; [29] Bonfrer I, Breebaart L, De Poel E Van. The effects of
56 (2), 295-303. https://apps.who.int/iris/handle/ Ghana’s national health insurance scheme on mater-
10665/261736 1978;56(2):295–303. nal and infant health care utilization. PLoS One.
[20] Chopra M, Campbell H, Rudan I. Understanding the 2016;11(11):1–13. https://doi.org/10.1371/journal.
determinants of the complex interplay between cost- pone.0165623
effectiveness and equitable impact in maternal and [30] La Vincente S, Aldaba B, Firth S, Kraft A, Jimenez-
child mortality reduction. J Glob Health. 2012;2 Soto E, Clark A. Supporting local planning and bud-
(1):010406. DOI: 10.7189/jogh.02.010406 geting for maternal, neonatal and child health in the
[21] Waiswa P, O’Connell T, Bagenda D, Mullachery P, Philippines. Heal Res Policy Syst. 2013;11(1):1–6.
Mpanga F, Kiwanuka D, et al. Community and https://doi.org/10.1186/1478-4505-11-3