You are on page 1of 11

Health Policy and Planning, 2020, 1–11

doi: 10.1093/heapol/czaa116
Review

Downloaded from https://academic.oup.com/heapol/advance-article/doi/10.1093/heapol/czaa116/5999228 by Karolinska Institutet University Library user on 20 December 2020
Economic evaluation and health systems
strengthening: a review of the literature
Susan Cleary *
Health Economics Unit/Division, School of Public Health and Family Medicine, University of Cape Town, Anzio
Road, Observatory, Cape Town 7925, South Africa
*Corresponding author. Health Economics Unit/Division, School of Public Health and Family Medicine, University of Cape
Town, Anzio Road, Observatory, Cape Town 7925, South Africa. E-mail: susan.cleary@uct.ac.za
Accepted on 2 September 2020

Abstract
Health systems strengthening (HSS) is firmly on the global health and development agenda. While
a growing evidence base seeks to understand the effectiveness of HSS, there is limited evidence
regarding cost and cost-effectiveness. Without such evidence, it is hard to argue that HSS repre-
sents value for money and the level of investment needed cannot be quantified. This paper seeks
to review the literature regarding the economic evaluation of HSS from low- and middle-income
country (LMIC) settings, and to contribute towards the development of methods for the economic
evaluation of HSS. A systematic search for literature was conducted in PubMed, Scopus and the
Health Systems Evidence database. MeSH terms related to economic evaluation were combined
with key words related to the concept of HSS. Of the 204 records retrieved, 52 were retained for full
text review and 33 were included. Of these, 67% were published between January 2015 and June
2019. While many HSS interventions have system wide impacts, most studies (71%) investigated
these impacts using a disease-specific lens (e.g. the impact of quality of care improvements on
uptake of facility deliveries). HSS investments were categorized, with the majority being invest-
ments in platform efficiency (e.g. quality of care), followed by simultaneous investment in platform
efficiency and platform capacity (e.g. quality of care and task shifting). This review identified a
growing body of work seeking to undertake and/or conceptualize the economic evaluation of HSS
in low- and middle-income countries. The majority assess HSS interventions using a disease-
specific or programmatic lens, treating HSS in a similar manner to the economic evaluation of
medicines and diagnostics. While this approach misses potential economies of scope from HSS
investments, it allows for a preliminary understanding of relative value for money. Future research
is needed to complement the emerging evidence base.

Keywords: Economic evaluation, costs, cost-effectiveness, health systems, technical efficiency

Introduction determine the affordability of services, establishing or refining


In 2015, United Nations member states signed the Sustainable in-country processes of Health Technology Assessment to include
Development Goals (SDGs), committing to the achievement of key economic evaluation is becoming a global priority (Sixty-Seventh
health and development targets by 2030 (United Nations, 2015). A World Health Assembly, 2014). Such processes will further contrib-
central pillar of these efforts is the movement towards Universal ute towards the achievement of the UHC intermediate objectives of
Health Coverage (UHC), which entails the progressive realization of efficiency (through the inclusion of cost-effective services within the
equitable access to quality health care, free at the point of use benefits package) and transparency (through an explicit identifica-
(Kutzin, 2013). These goals of equity, quality and financial risk pro- tion of the services that are included).
tection are achievable irrespective of country income level, provided Resource allocation or priority setting for health systems can
that the chosen health benefits package to be provided by the health include decisions around whether to invest in a new technology (e.g.
system is affordable (Glassman et al., 2017). Given the need to medicine, diagnostic or investigation) as well as decisions about

C The Author(s) 2020. Published by Oxford University Press in association with The London School of Hygiene and Tropical Medicine. All rights reserved.
V
For permissions, please e-mail: journals.permissions@oup.com 1
2 Health Policy and Planning, 2020, Vol. 0, No. 0

KEY MESSAGES

Downloaded from https://academic.oup.com/heapol/advance-article/doi/10.1093/heapol/czaa116/5999228 by Karolinska Institutet University Library user on 20 December 2020
• Considerations of value for money are key towards ensuring that services offered within Universal Health Coverage systems achieve
maximum population benefits.
• While a growing evidence base exists regarding the cost and cost-effectiveness of medicines and diagnostics for specific conditions,
less is known about the cost-effectiveness of health systems strengthening (HSS). Similarly, methods for the economic evaluation of
technologies such as medicines and diagnostics are well established, but key gaps exist regarding how to conceptualize and conduct
economic evaluations of HSS.
• This paper offers a review of the literature of economic evaluations of HSS interventions conducted in low- and middle-income coun-
try settings, with a view towards understanding the current scope and methodological practices of the field.
• Of the papers retrieved, the majority assess HSS interventions using a disease-specific or programmatic lens, treating HSS in a similar
manner to the economic evaluation of medicines and diagnostics. While such an approach misses the potential economies of scope
from HSS investments, it allows for a preliminary understanding of relative value for money.

whether to invest in ‘health system strengthening’ (e.g. information outcomes, approach to decision-making and advantages or disad-
or managerial system). While the former will often have a relatively vantages. Cost analysis is a form of partial economic evaluation that
narrow scope or disease-specific focus, the latter will often impact can include costs from across the perspectives. Similarly, CCA and
across multiple disease-specific areas (Vassall et al., 2019). Both the CBA can include the full spectrum of costs, in addition to outcomes.
methodology and the evidence base of cost and cost-effectiveness to In the former, outcomes are listed separately, while in the latter,
date has been more focused on the former, with less attention paid these are aggregated and monetized. Both however have disadvan-
to the cost-effectiveness of health systems strengthening (HSS). As tages, with the main disadvantage for CCA being the lack of theoret-
argued by Murray et al. (1994, p. 664) ‘cost-effectiveness analysis of ical basis and lack of decision-making guidance, while the main
health interventions, which are more often than not disease specific, disadvantage for CBA is the bias and lack of precision associated
tends to neglect the role of the health system in delivering these with monetizing outcomes (Helmchen and Lo Sasso, 2010;
interventions. There are no explicit analyses of cost-effectiveness of Wiseman, 2016). In between these two extremes are the CEA and
improving the physical or human infrastructure of the health system, CUA. Neither of these accommodates non-health costs nor out-
which provide for direct comparisons between investing in the deliv- comes; and both seek to maximize a given outcome within a given
ery system and purchasing more specific interventions delivered by budget constraint and/or equity constraint if relevant (Cookson
the health system’. Indeed, if health systems are weak, it is possible et al., 2017). For CEA, given that the outcomes are in natural units,
that investments towards HSS would generate more value than comparability is normally limited to the specific disease in question
investments in additional medicines and diagnostics within disease- and to evaluations using the same outcome (e.g. CEAs using life
specific areas (Mills et al., 2006). Generating evidence on the cost- years gained can be compared across a range of HIV-treatment
effectiveness of HSS and the level of investment needed to achieve interventions). The budget constraint in question is then the health
performance outcomes is therefore a priority (Vassall et al., 2019). care budget dedicated to that particular treatment programme (e.g.
This paper seeks to review economic evaluations of HSS inter- the HIV-treatment budget) and issues of priority setting are
ventions from low- and middle-income country (LMIC) settings to addressed within the HIV programme (Cleary et al., 2010).
describe the scope and key methodological aspects of this literature However, in practice, determining the size of the HIV-treatment
and thereby to contribute towards future methodological budget is itself an allocative decision and CEA is unable to guide this
developments. determination. Because of this, CUA is an increasingly popular form
of economic evaluation. Because health care outcomes are estimated
Conceptual underpinnings using multi-attribute approaches such as quality adjusted life years
To gain insight into the methodology of HSS economic evaluations, (QALYs) or disability adjusted life years (DALYs), comparisons can
it is necessary to first describe ‘standard’ approaches. Following be made across diseases and interventions, and outcomes can be
Drummond et al. (2015) an economic evaluation considers two or maximized across the health care budget constraint in general (again
more ways of achieving a given outcome. There are five types of par- subject to an equity or other constraint if relevant). This idea of
tial or full economic evaluation: cost analysis, cost-consequence ana- choosing the set of interventions or programmes that maximize
lysis (CCA), cost-effectiveness analysis (CEA), cost-utility analysis health within the budget (or other relevant) constraint can be termed
(CUA) and cost–benefit analysis (CBA). In addition, costs can be ‘allocative efficiency’ (Stinnett and Paltiel, 1996).
estimated using different ‘perspectives’. A patient perspective Guidelines for economic evaluation recommend that costs
includes out of pocket payments (e.g. user fees), costs of accessing should include any changes in resources that accompany a decision
care (e.g. transport costs) and may include other opportunity costs to implement a particular programme, intervention or course of ac-
of time spent receiving care (e.g. waiting times within facilities) that tion, as appropriate to the perspective of the costing (Husereau
are incurred by a patient and potentially also by a patient’s carer(s). et al., 2013). For example, within a provider’s perspective, costs
A health care provider or health systems perspective includes the would typically include capital costs (e.g. health facility infrastruc-
costs borne by the health sector. A societal perspective is broadest ture, furniture, equipment and vehicles) as well as recurrent costs
and encompasses both patient and provider perspectives. Table 1 (e.g. staff, medicines, laboratory investigations, supplies and over-
outlines each type of analysis, indicating the types of costs included, heads). If relevant, analyses can include ‘above service delivery’ costs
Health Policy and Planning, 2020, Vol. 0, No. 0 3

Table 1 Types of economic evaluation managerial salaries to calculate the mean cost per visit for manager-
ial staff time. Amongst others, this approach assumes that interven-
CCA
tions within a particular setting function independently of each

Downloaded from https://academic.oup.com/heapol/advance-article/doi/10.1093/heapol/czaa116/5999228 by Karolinska Institutet University Library user on 20 December 2020
Costs Can include health care and non-health care costs
Outcomes Multiple health outcomes and non-health outcomes other (Hauck et al., 2019); however, there are multiple ways in
specified in natural units, e.g. cases detected which interventions are interdependent, with the potential for
AND life years gained variations in economies of scope depending on how services are
Decision-making Not defined. Decision maker needs to define own organized within and/or across the platform (Hauck et al., 2019).
rule(s) rules What this means is that—for two or more interventions—costs, ben-
Advantages A broad scope of costs and outcomes can be efits or both costs and benefits are different under joint vs separate
included, including non-health costs and production, and this mutual dependency is argued to support the im-
outcomes
portance of a systems perspective in evaluation (Hauck et al., 2019).
Disadvantages Lack of clarity about usefulness in decision-making
CEA
Following this logic, Hauck et al., have created a typology of three
Costs Only includes health care costs types of HSS investments. The first category (HSS1) concerns invest-
Outcomes Singular health outcomes specified in natural units, ments in the technical efficiency of the particular facility or platform
e.g. cases detected OR life years gained which would impact on the production of all services delivered
Decision-making Maximize outcomes within the budget constraint. within that setting. Examples include health information systems or
rule(s) Can include an equity constraint. Given non- laboratory infrastructure (Morton et al., 2016). The second category
comparability of outcomes, only addresses tech- (HSS2) concerns investments into the capacity of the facility or plat-
nical efficiency
form, including the reorientation of existing resources, enabling a
Advantages Natural outcomes are generally easier to measure
key resource constraint to be relaxed. A common example is scarce
Disadvantages Omits non-health costs and outcomes. Cannot
address allocative efficiency human resources limiting service expansion, with investments such
CUA as task shifting being a potential solution to this capacity constraint.
Costs Only includes health care costs The third category (HSS3) concerns investments in new platform
Outcomes Multi-attribute health outcome measures such as areas such as new mobile services which again potentially alters the
QALYs or DALYs relative cost-effectiveness of existing interventions.
Decision-making Maximize outcomes within the budget constraint. Figure 1 presents these ideas within a conceptual framework,
rule(s) Can include an equity constraint. Given compar- summarizing ‘standard’ economic evaluation together with HSS1–3.
ability of outcomes, can address allocative
In the remaining sections, this conceptualization will be used as one
efficiency
way of categorizing and describing the empirical literature.
Advantages Comparability of diverse programmes is enhanced
through use of multi-attribute health outcome
Disadvantages Omits non-health costs and outcomes
CBA
Methods
Costs Can include health care and non-health care costs Search strategy
Outcomes Health outcomes and non-health outcomes (such as A systematic search for relevant literature was conducted in
process utility, e.g. reassurance value associated
PubMed, Scopus and the ‘Health Systems Evidence’ database,
with negative test result) combined in monetary
including any publications before July 2019. With the help of a spe-
units
Decision-making If benefits (in monetary terms) exceed costs, social cialized librarian, the MeSH heading term ‘Costs and Cost Analysis’
rule(s) welfare is improved through implementing was identified as the preferred term to retrieve all concepts related
intervention to economic evaluation. This approach was complemented by
Advantages Broad scope of outcomes can be measured in mon- searches for the key words ‘cost(s)’ or ‘costing’ within the heading
etary terms and abstract of the paper. Given that all economic evaluations
Disadvantages Issues of bias and precision in monetizing outcomes would include costs, these two strategies were deemed to be ad-
equate. For the concept of health system strengthening, full text
Adapted from Drummond et al. (2015) and Helmchen and Lo Sasso (2010).
searches were conducted within articles using the key words ‘health
system(s) strengthening’ or ‘health system(s) development’.
which include procurement and supply chain activities and demand One of the key challenges in selecting papers is to find a working
generation activities (e.g. mass media campaigns) amongst others definition for HSS. While a number of definitions can be proposed,
(Clift et al., 2016). Within the third edition of Disease Control in their seminal work the World Health Organization (2007) catego-
Priorities (DCP-3), the authors categorize these capital and recurrent rized health systems into six building blocks and sub-components
costs into direct costs at the point of care, including personnel, including service delivery; health workforce; information systems;
drugs, equipment, etc.; costs of facility-level ancillary services, medical products, vaccines and technologies; financing and govern-
including rent, building maintenance, etc. and programme costs, ance, as outlined in Table 2. HSS is then defined as ‘improving these
that support services but frequently are incurred separately from . . . building blocks and managing their interactions in ways that
service delivery such as administration, management, logistics, etc. achieve more equitable and sustained improvements across health
They use the term ‘service delivery costs’ for the first, and group services and health outcomes’ (World Health Organization, 2007, p.
the second and third together under the term ‘health system costs’ 4). However, by this definition, it would be difficult to exclude any
(Watkins et al., 2017). economic evaluation given that they would all evaluate interventions
The ‘standard’ approach to the inclusion of health system costs that have the potential to improve health services and health out-
is to use an allocation factor to estimate the share of these costs used comes. Therefore, following others (Zeng et al., 2018) the service
by the intervention under evaluation (Drummond et al., 2015). An delivery building block is restricted to interventions associated with
example would be the use of annual clinic visits to allocate annual enhancing demand for care, public private partnerships around
4 Health Policy and Planning, 2020, Vol. 0, No. 0

Downloaded from https://academic.oup.com/heapol/advance-article/doi/10.1093/heapol/czaa116/5999228 by Karolinska Institutet University Library user on 20 December 2020
Figure 1 Conceptual framework of economic evaluation and investments in HSS. Adapted from Hauck et al., (2019)

Table 2 Categories of HSS by health system building blocks


To guide analysis, a data extraction template was created in
Service delivery Microsoft Excel. Data extracted directly from papers included a descrip-
Enhancing demand for care tion of the intervention and the study setting. In addition, based on the
Public–private partnerships around service provision details provided by the author(s), a judgement was made regarding the
Quality and safety of care health systems building block(s) covered within the evaluation (as per
Health workforce Table 2), and the type of HSS investment(s) (as per Figure 1). A distinc-
Supply and distribution
tion was also drawn between disease- or programme-specific HSS (tar-
Management and performance systems
geting system support within specific disease or programme areas) or
Pre-service and in-service training
Medical products, vaccines and technologies
system wide HSS where no explicit disease or programme link is identi-
Public–private partnerships around medical products, vaccines fied, following Warren et al. (2013). Key methodological data extracted
and technologies included type(s) of economic evaluation, costing perspective and out-
Supply chain management come measure(s) if applicable. These data were extracted as per the
Quality and safety of medicines stated method of the author(s) or the article and further coded to enable
Information systems comparison. On type of economic evaluation, studies of the ‘costs of
Health information systems scaling up’ were termed budget impact analysis; while CEA studies that
Management information systems included detailed intermediate outcomes with modelled QALYs or
Financing
DALYs were coded as CEA and CUA.
Payment mechanisms: provider
These data elements are summarized below.
Payment mechanisms: beneficiary
Priority setting or resource allocation General
Governance • Country, countries or region of focus
Oversight, accountability, regulations and incentives (e.g. audits,
• Description of HSS intervention
targets, core standards)
• Health systems building block(s) targeted by the intervention
Strategic planning
Consumer and stakeholder involvement (i.e. community (as per Table 2)
participation) • Type of HSS investment (as per Figure 1)
• System wide or disease-/programme-specific evaluation.
Adapted from Adam et al. (2012) and World Health Organization (2007).

Methods
service provision, and evaluations of quality and safety of care,
• Type(s) of economic evaluation
while economic evaluations of medicines, diagnostics or similar
• Costing perspective(s)
health technologies within a disease-specific area (e.g. GeneXpert
• Outcome measure(s).
for TB diagnosis or antiretrovirals for HIV treatment) are excluded
(see Table 2). In contrast, interventions falling within the medical
products, vaccines and technologies building block are included, Results
given the focus of this building block on public–private partnerships,
The initial search generated 153 hits from PubMed, 51 from
supply chain management and processes associated with securing
Scopus and 60 from the Health Systems Evidence database. One
the quality and safety of medicines.
additional paper was identified as part of a broader DCP volume
In summation, included papers:
that had been retrieved (Levin and Chisholm, 2016) and 2 were
• Conducted a full or partial economic evaluation of an HSS inter- included through prior knowledge of the literature (Lagarde
vention in a LMIC setting (as defined by the World Bank in 2019). et al., 2012; Borghi et al., 2015). After the removal of dupli-
• Were written in the English language. cates, 256 records were screened, of which 204 were excluded
Health Policy and Planning, 2020, Vol. 0, No. 0 5

Downloaded from https://academic.oup.com/heapol/advance-article/doi/10.1093/heapol/czaa116/5999228 by Karolinska Institutet University Library user on 20 December 2020
Figure 2 PRISMA flow diagram

following title and abstract review. Of the 52 retained for full- Methodological approaches
text review, 18 were excluded (systematic review or conceptual Figure 3 summarizes the type of economic evaluation, costing per-
paper ¼ 7; no full text ¼ 1; no economic evaluation ¼ 9; no HSS spective and outcome measures used across the studies. As is shown,
intervention ¼ 1). A total of 33 papers were retained and cost analysis was the most common study type (25% of 33 studies)
included in the qualitative synthesis. Details are contained in followed by CEA (24%) and BIA (15%). The remaining studies
Figure 2. A database including all papers is available in supple- were classified as a combination of different types of economic
mentary material (Online Table 1). evaluation including CEA, CUA and BIA. As mentioned, studies
Of the 33 included papers, the majority (56%) was from African were classified as CEA and CUA if they provided outcomes in natural
settings, including Benin (n ¼ 1), Ethiopia (n ¼ 2), Kenya (n ¼ 4), units as well as QALYs or DALYs. In terms of perspective, the major-
Malawi (n ¼ 2), Nigeria (n ¼ 1), Rwanda (n ¼ 1) and South Africa ity of studies took a provider’s perspective. Close to half of the
(n ¼ 2). In addition, there were 6 papers from Asia (Cambodia ¼ 1, included studies did not state their perspective and were classified as
China ¼ 3 and Myanmar ¼ 1) and 1 paper from South America likely provider (36%) or likely societal (6%) or unclear if it appeared
(Chile). Finally, some papers took regional perspectives (sub- that different perspectives were being used across different analyses,
Saharan-Africa and South Asia, n ¼ 1) or focused on income level without a clear identification of what was being used (n ¼ 1). No stud-
(low- and lower middle-income settings, n ¼ 2; low- and middle- ies took a patient only perspective (although patient costs are included
income settings, n ¼ 4). In terms of the scope of the literature within the societal perspective). In terms of outcomes, 40% of studies
reviewed, of the 33 included studies, 27 (82%) took a programmatic did not include outcomes (of which 25% are cost analyses and 15%
or disease-specific lens to the evaluation and 6 (18%) were system- are BIA) while 33% presented QALYs or DALYs (frequently in add-
wide. The majority of papers (n ¼ 22) was published in the period ition to intermediate or natural outcome measures). The remaining
2015–19. 27% presented only natural or intermediate outcomes.
6 Health Policy and Planning, 2020, Vol. 0, No. 0

Downloaded from https://academic.oup.com/heapol/advance-article/doi/10.1093/heapol/czaa116/5999228 by Karolinska Institutet University Library user on 20 December 2020
Figure 3 Type of economic evaluation, costing perspective and outcome measure

HSS within system wide evaluations inspections, licencing, etc.) which have been considered in standard
Of the 33 included papers, 6 studies could be described as system- EE as ‘above service delivery’ costs, and frequently excluded (Clift
wide, although as illustrated below, such a breadth of analysis et al., 2016).
makes it hard for the authors to provide sufficient details of what The remaining two system wide papers focused on the health
costs and outcomes are being estimated. Of these, 1 paper estimated workforce building block and HSS2 (i.e. relaxing a health system
a benefit package for Malawi (Ochalek et al., 2018) and 1 estimated capacity constraint). One considered upstream and downstream
benefit packages for low- and lower middle-income settings nurse retention strategies for rural parts of South Africa (Lagarde
(Watkins et al., 2017). In Ochalek et al. (2018), the case for inves- et al., 2012); this paper was unique in that it defined a HS perform-
ting in HSS is made through an argument around poor coverage or ance outcome measure in the cost-effectiveness findings (rural-nurse
low implementation levels for cost-effective disease-specific inter- year) as opposed to a health-related outcome. The other performed
ventions. In effect the authors argue that these low implementation a cost analysis of health extension workers (a form of community
levels indicate the quantum of investment in HSS that would allow health worker, CHW) for Ethiopia (Canavan et al., 2017). None of
for increased availability or coverage of disease-specific interven- these papers sought to unpack the impact of HSS on health out-
tions. In other words, they do not specifically evaluate the costs or comes or on the uptake or utilization of disease-specific services.
cost-effectiveness of the various types of HSS that could be imple-
mented in Malawi but instead provide an indication of the scale of HSS within disease-specific or programmatic
the investment that could be made while remaining cost-effective. evaluations
Similarly, Watkins et al. (2017) examine coverage levels of interven- The majority of papers (27) considered various forms and combina-
tions recommended within their benefits package and argue that this tions of HSS investments, across a range of building blocks, using a
would be a starting point for countries to think through the key disease-specific or programmatic lens. This lens could be relatively
health systems bottle necks that prevent increased coverage. While wide, including clusters of conditions or services [e.g. Mental,
these authors include ‘health systems’ costs as part of the unit costs Neurological and Substance Use Disorders (MNS); childhood ill-
of disease-specific interventions, they also do not directly estimate nesses; surgery; vaccines; Maternal and Child Health services] or
the resource needs or provide details of HSS. more narrow (e.g. Malaria treatment; facility-based births; diabetes
In contrast, 2 of the system wide papers focused only on assess- care; specific contraceptives).
ing costs, with 1 offering a costing of the health strategic plan in In terms of type of HSS investment included in the disease-
Cambodia (Cantelmo et al., 2018) and 1 estimating the costs of specific category, 10 papers included a focus on HSS1 (i.e. interven-
meeting the SDGs in LMICs (Stenberg et al., 2017). The approach tions seeking to change the technical efficiency of the health system),
to costing builds on existing disease specific and HSS strategic goals, 1 focused on HSS2 (i.e. relaxing a health system capacity con-
and then assesses the costs needed to achieve these goals. These cost- straint), 3 focused on HSS3 (i.e. developing a new health system
ing papers used the OneHealth Tool to generate their estimates, platform) and 9 focused on HSS1&2 (i.e. relaxing a health system
which includes disease-specific and health systems modules. The capacity constraint plus implementing measures to ensure that the
authors argue that their approach includes inputs such as infrastruc- new resources are used in a technically efficient manner). The
ture and supply chain (as per standard EE approaches) but in add- remaining 4 papers sought to include a broader range of HSS; of
ition includes the costs of strengthening HS performance (audits, which 2 papers focused on surgical services and 2 focused on MNS.
Health Policy and Planning, 2020, Vol. 0, No. 0 7

The papers on surgical services were based on a combination of lit- correct treatment for malaria in Kenya (Goodman, 2006); super-
erature review and cost (Verguet et al., 2015) or cost-effectiveness vised CHWs for diagnosis and treatment of malaria in Myanmar
(Mock et al., 2015) modelling across all LMICs in the former or all (Kyaw et al., 2016) and child health days (campaigns) to increase

Downloaded from https://academic.oup.com/heapol/advance-article/doi/10.1093/heapol/czaa116/5999228 by Karolinska Institutet University Library user on 20 December 2020
low- and lower middle-income countries (LLMICs) in the latter. For coverage of Vitamin A and other key child health priorities such as
Mock et al, which is part of DCP-3, the innovation is a focus on the deworming in Ethiopia (Fiedler and Chuko, 2008).
cost-effectiveness of the delivery platform, seeking to recommend The final three papers in this sub-group were all focused on
that surgeries are conducted at the most cost-effective level of care HSS3 and investments in new platform capacities. Of these, two
where delivery is feasible. In addition, the authors focus on quantify- focused on supply chain management within the medical products,
ing infrastructure needs, given that first level hospitals with operat- vaccines and technologies building block (1 on scaling up the vac-
ing rooms will need to be built to increase access to these surgical cine supply chain across LMICs (Portnoy et al., 2015) and 1 on inte-
procedures. Similarly, in their cost analysis, Verguet et al. estimate grating nutritional commodities into the existing supply chain in
the number of operating rooms that will need to be constructed to Kenya (Eby et al., 2019)]. The final paper in this set was focused on
achieve scale up targets. The two MNS papers included one DCP-3
a remote monitoring system for home-based non-invasive positive
chapter (Levin and Chisholm, 2016) focused on LMICs, and an add-
pressure ventilation of children with upper airways obstruction in
itional analysis focused on specific countries within SSA and South
China (Zhou et al., 2012).
Asia (Chisholm et al., 2017) and specific MNS disorders (psychosis,
These were all detailed empirical analyses seeking to test out in-
depression and epilepsy). The latter paper focused on the develop-
novative approaches to improve quality and demand for services.
ment and application of a new module within the OneHealth Tool
These detailed empirical analyses are valuable for providing ideas of
to cost these conditions. As mentioned, this tool has a health systems
what might work in different contexts and within different disease
application, although the authors caution that the data and know-
areas.
ledge requirements to use the health system components—such as
logistics and governance—are seldom available. Similar to the DCP
chapter on essential surgery, the MNS chapter has a focus on the de- HSS and type of building block
livery platform and/or delivery agent as being of key concern from It was possible to classify by type of building block (noting that
an efficiency and coverage (or equity) perspective (Levin and many papers included >1 building block), using the categorizations
Chisholm, 2016). from Table 2, as illustrated in Figure 3 for 25 of the included papers.
The 10 papers focusing on HSS1 (i.e. seeking to influence the The most common building blocks were service delivery (particular-
technical efficiency of the health system) included one on pay-for- ly quality of care and demand for care) and health workforce (with
performance in Tanzania (Borghi et al., 2015); various quality im- close to equal numbers for supply, management and training of
provement initiatives including for IMCI in Kenya (Barasa et al., health workers). The financing building block was third most com-
2012) and in Rwanda (Manzi et al., 2018); various voucher schemes mon, with provider payment mechanisms being more frequently
encouraging ANC and facility-based births in Myanmar (Kingkaew evaluated than beneficiary payment mechanisms. Few papers
et al., 2016) and Uganda (Alfonso et al., 2015) and encouraging focused on medicines (e.g. issues of safety or supply chain) or on in-
pregnant women to purchase ITNs in Tanzania (Mulligan et al., formation systems. There were no papers focused on governance.
2008); a ‘detailing’ intervention consisting of external supervision While some papers focused on community mobilization, this was
and training to encourage the uptake of intra-uterine devices in expressed in terms of encouraging the demand for services and was
Kenya (Wesson et al., 2008); capitation as a provider payment not expressed in terms of community participation to encourage ac-
mechanism for diabetes care in South Africa (Volmink et al., 2014); countability or responsiveness. These interventions were therefore
an IT-based surveillance mechanism to detect disease outbreak in classified as part of the service delivery building block and not as
China (Ding et al., 2015) and a pharmacist intervention focusing on governance.
safe and rational drug use within an intensive care unit in China Figure 4cross-tabulates the type of HSS investment against build-
(Jiang et al., 2012). ing block subcomponents. Of the 25 empirical papers considered
Nine papers focused on interventions that were a combination of
here, 40% could be categorized as HSS1, 12% as HSS2, 12% as
HSS1 and HSS2. Largely, these interventions introduced a new form
HSS3 and 36% were a combination of HSS1 and HSS2. Within the
of health worker or infrastructure to relax a health system capacity
reviewed literature, HSS1 was more often focused on the service de-
constraint (HSS2) in addition to various measures to encourage the
livery and health workforce building blocks. Examples include inter-
efficient utilization of the new resource (technical efficiency—
ventions to enhance quality of care and voucher schemes to enhance
HSS1). Interventions within this subset tended to be multicompo-
demand for care within the service delivery building block while
nent, including facility rehabilitation, training, supervision and
training was a common HSS1 investment under the health work-
performance-based financing in Benin (Paul et al., 2017); a subsi-
force building block. HSS2 is an investment that is designed to relax
dized insurance system including quality of care interventions in
Nigeria (Ruelas et al., 2012); provision of training, medical supplies a platform capacity constraint; here the most commonly evaluated
and equipment together with home visits by CHWs to antenatal and intervention was increasing supply of health workers through task
postnatal mothers in Uganda (Ekirapa-Kiracho et al., 2017); com- shifting. HSS3 refers to an investment into a new element within the
munity mobilization plus a CHW package of care for mothers and platform. Here, the few examples within the reviewed literature
newborns in Malawi (Greco et al., 2017); a voucher scheme, com- included two supply chain evaluations (one for vaccines and one for
munity mobilization, training, supportive supervision and provision nutrition commodities) and one evaluation of a remote monitoring
of supplies to encourage facility-based births in Uganda (Mayora system for home-based ventilation of children and infants with
et al., 2014); CHW home visits during pregnancy and postpartum upper airway obstruction. Finally, of the 36% of papers that consid-
plus provision of equipment in Tanzania (Manzi et al., 2017); train- ered a combination of HSS 1&2 (i.e. technical efficiency and relax-
ing of shopkeepers including provision of job aids and charts plus ing a key capacity constraint), the most common intervention was to
follow-on support by CHWs and community mobilization to enable encourage demand for care (e.g. through vouchers and community
8 Health Policy and Planning, 2020, Vol. 0, No. 0

Downloaded from https://academic.oup.com/heapol/advance-article/doi/10.1093/heapol/czaa116/5999228 by Karolinska Institutet University Library user on 20 December 2020
Figure 4 Types of HS building blocks included in economic evaluations

Figure 5 Types of building blocks and types of HSS investments

mobilization) and quality of care while task shifting services to in- (Adam et al., 2012), such a definition would lead to the exclusion of
crease the supply of health workers (Figure 5). the majority of the economic evaluation literature. For this reason,
disease-specific or programmatic HSS evaluations were included. A
smaller set of papers (8/33) focused on specifying benefit packages
Discussion and included HSS within overall estimates. Only 1 paper was from
This review of the literature has identified 33 economic evaluation South America, while most papers were from African settings.
of HSS investments from LMIC settings. Most papers (25/33) under- In terms of methods, the programme- or disease-specific HSS lit-
took economic evaluations of HSS within disease-specific or pro- erature included in this review followed similar approaches as would
grammatic areas, such as Malaria prevention or facility-based be taken within standard economic evaluations of medicines or diag-
births. While some consider HSS to be horizontal or system wide nostics. These papers provided detailed empirical data on costs as
Health Policy and Planning, 2020, Vol. 0, No. 0 9

well as natural outcome measures. In addition, a small number of of the papers retrieved by a recent review on this topic (Turcott-
these papers demonstrated the feasibility of providing modelled Tremblay et al., 2016). While this shortcoming had potentially been
multi-dimensional units (such as DALYs) to enable the fuller as- mitigated through the inclusion of studies that are indexed in the

Downloaded from https://academic.oup.com/heapol/advance-article/doi/10.1093/heapol/czaa116/5999228 by Karolinska Institutet University Library user on 20 December 2020
sessment of allocative efficiency. This is an encouraging approach, Health Systems Evidence database, it is likely that studies have been
enabling a consideration of the relative opportunity cost and value missed. Thus while this paper does not attempt to provide a system-
for money of HSS in comparison to disease-specific investments. atic review, these limitations nevertheless underscore the prelimin-
This review did not retrieve any studies evaluating economies of ary nature of this synthesis.
scope from single vs joint production, or estimating the cost-
effectiveness of HSS interventions across the health system more
broadly, e.g. the implementation of a health information system. Conclusion
It therefore remains to be seen whether the conceptualizations of In conclusion, this review identified a small body of economic evalu-
Hauck and colleagues (2019) would be feasible within empirical
ations of HSS investments from LMIC settings. The detailed empir-
research.
ical papers included in this review demonstrated the feasibility of
In contrast to these detailed empirical papers, eight studies
conducting economic evaluations of HSS using standard
focused on estimating benefits packages more broadly, whether for
approaches. This is complemented by the broader benefits package
specific settings (e.g. LLMICs) or for specific programmatic areas
papers which provide a mechanism for estimating the overall level
(e.g. surgery or MNS across LMICs). These papers are unable to
of investment in HSS that could be cost-effective, or that would be
provide details of the relative value for money of HSS in a granular
needed to achieve key goals. However, the overall evidence base
or nuanced way, however they are able to provide an indication of
what level of overall investment in HSS might be cost-effective or remains small, and further research should be encouraged to support
might be needed to achieve key strategic goals. In other words, these ongoing progress towards UHC.
papers will not give insight into the cost-effectiveness of implement-
ing a health information system, e.g. they can provide an indication
Funding
of how much in total should be invested in HSS.
Standard approaches towards economic evaluation have typical- Susan Cleary is a member of the Consortium for Resilient and Responsive
ly recommended the inclusion of the costs associated with infra- Health Systems (RESYST). This article is an output from a project funded by
the UK Aid from the UK Department for International Development (DFID)
structure, facility management and facility information systems,
for the benefit of developing countries. However, the views expressed and in-
etc., frequently categorizing these as overhead costs, capital costs
formation contained in this article are not necessarily those of or endorsed by
and above service delivery costs (Drummond et al., 2015; Clift
DFID, which can accept no responsibility for such views or information or for
et al., 2016). This review identified a growing practice of naming
any reliance placed on them.
these costs ‘health systems’ costs, which is useful for enabling the
economic evaluation literature to speak to a broader health policy
and systems audience. In addition, this approach helps to highlight Acknowledgements
that the majority of costs is health systems costs. This is an import-
The author gratefully acknowledges the inputs from two anonymous
ant insight for donors that have pursued disease-specific outcomes
reviewers.
without funding the full range of health system inputs, with negative
implications for crowding out the services that are not donor prior- Conflict of interest statement. None declared.
ities, but which depend on the same health system platforms for Ethical approval. No ethical approval was required for this study.
delivery.
This paper has a number of limitations. First, given the focus on
scoping the literature and describing current methods, no attempt References
was made to limit the inclusion of studies to those meeting particu- Adam T, Hsu J, de Savigny D et al. 2012. Evaluating health systems strength-
lar quality standards. Second, this paper has a single author, mean- ening interventions in low-income and middle-income countries: are we ask-
ing that only one person carried out the data extraction and ing the right questions? Health Policy and Planning 27: iv9–19.
analysis. Part of this analysis included making a judgement about Alfonso YN, Bishai D, Bua J et al. 2015. Cost-effectiveness analysis of a vou-
the type of economic evaluation and type of costing perspective, in cher scheme combined with obstetrical quality improvements: quasi experi-
the instances where this was not clearly identified within the papers. mental results from Uganda’. Health Policy and Planning 30: 88–99.
Third, it is relatively difficult to distinguish between disease-specific Barasa EW, Ayieko P, Cleary S et al. 2012. A multifaceted intervention to im-
interventions that are about HSS vs those that are about medicines prove the quality of care of children in district hospitals in Kenya: a
and diagnostics. Realistically, there is a great deal of overlap given cost-effectiveness analysis. PLoS Medicine 9: e1001238.
Borghi J, Little R, Binyaruka P et al. 2015. In Tanzania, the many costs of
that evaluations of the cost-effectiveness of medicines and diagnos-
pay-for-performance leave open to debate whether the strategy is cost-effec-
tics will also include health systems costs. Fourth, 3 papers were not
tive. Health Affairs 34: 406–14.
found in the electronic search, of which 1 was part of a broader
Canavan ME, Linnander E, Ahmed S et al. 2017. Unit costing of health exten-
DCP volume that had been retrieved (Levin and Chisholm, 2016)
sion worker activities in Ethiopia: a model for managers at the district and
and 2 were included through prior knowledge of the literature
health facility level. International Journal of Health Policy and
(Lagarde et al., 2012; Borghi et al., 2015). The latter two papers did Management 7: 394–401.
not use the term ‘health system(s) strengthening OR development’ Cantelmo CB, Takeuchi M, Stenberg K et al. 2018. Estimating health plan
within the body of the article. It is therefore likely that there are costs with the OneHealth tool, Cambodia. Bulletin of the World Health
other economic evaluations of HSS in LMICs that have not been Organization 96: 462–70.
retrieved for similar reasons. Indeed, while this paper did find a few Chisholm D, Heslin M, Docrat S et al. 2017. Scaling-up services for psychosis,
articles on performance-based financing, the search did not find all depression and epilepsy in sub-Saharan Africa and South Asia: development
10 Health Policy and Planning, 2020, Vol. 0, No. 0

and application of a mental health systems planning tool (OneHealth). management of childhood illness care in rural Rwanda’. PLoS One 13:
Epidemiology and Psychiatric Sciences 26: 234–44. e0194187.
Cleary S, Mooney G, McIntyre D. 2010. Equity and efficiency in Manzi F, Daviaud E, Schellenberg J et al. 2017. Improving Newborn Survival

Downloaded from https://academic.oup.com/heapol/advance-article/doi/10.1093/heapol/czaa116/5999228 by Karolinska Institutet University Library user on 20 December 2020
HIV-treatment in South Africa: the contribution of mathematical program- in Southern Tanzania (INSIST) trial; community-based maternal and new-
ming to priority setting. Health Economics 19: 1166–80. born care economic analysis. Health Policy and Planning 32: i33–41.
Clift J, Arias D, Chaitkin M et al. 2016. Landscape Study of the Cost, Impact, Mayora C, Ekirapa-Kiracho E, Bishai D et al. 2014. Incremental cost of
and Efficiency of above Service Delivery Activities in HIV and Other Global increasing access to maternal health care services: perspectives from a de-
Health Programs. Results for Development Institute. http://www.r4d.org/ mand and supply side intervention in Eastern Uganda. Cost Effectiveness
wp-content/uploads/R4D_ServDeliveryLandscaping_Final-0825_web.pdf and Resource Allocation 12: 14–9.
Cookson R, Mirelman A, Grif S et al. 2017. Using cost-effectiveness analysis Mills A, Rasheed F, Tollman S. 2006. Strengthening health systems. In:
to address health equity concerns. 20: 206–212. doi: Jamison D, Breman J, Measham A (eds). Disease Control Priorities in
10.1016/j.jval.2016.11.027 Developing Countries, 2nd edn. Washington, DC/New York: The
Ding Y, Sauerborn R, Xu B et al. 2015. A cost-effectiveness analysis of three International Bank for Reconstruction and Development/The World
components of a syndromic surveillance system for the early warning of epi- Bank/Oxford University Press.
demics in rural China Health policies, systems and management in low and Mock C, Donkor P, Gawande A et al. 2015. Essential surgery: key messages of
middle-income countries’. BMC Public Health 15: 1–9. this volume. In: Debas H et al. (eds). Essential Surgery: Key Messages of
Drummond MF et al. 2015. Methods for the Economic Evaluation of Health This Volume. In Essential Surgery: Disease Control Priorities, 3rd edn.
Care Programmes, 4th edn. Oxford: Oxford University Press. Washington, DC, 1–19. doi: 10.1596/978-1-4648-0346-8_ch19
Eby E, Daniel T, Agutu O et al. 2019. Integration of the UNICEF nutrition Morton A, Thomas R, Smith PC. 2016. Decision rules for allocation of finan-
supply chain: a cost analysis in Kenya’. Health Policy and Planning 34: ces to health systems strengthening. Journal of Health Economics 49:
188–96. 97–108.
Ekirapa-Kiracho E, Barger D, Mayora C et al. 2017. Uganda Newborn Study Mulligan JA, Yukich J, Hanson K. 2008. Costs and effects of the Tanzanian
(UNEST) trial: community-based maternal and newborn care economic national voucher scheme for insecticide-treated nets. Malaria Journal 7:
analysis’. Health Policy and Planning 32: i42–52. 1–10.
Fiedler JL, Chuko T. 2008. The cost of Child Health Days: a case study of Murray CJ, Kreuser J, Whang W. 1994. Cost-effectiveness analysis and policy
Ethiopia’s Enhanced Outreach Strategy (EOS). Health Policy and Planning choices: investing in health systems. Bulletin of the World Health
23: 222–33. Organization 72: 663–74.
Glassman A, Giedion U, Smith PC. 2017. What’s In, What’s Out? Designing Ochalek J, Revill P, Manthalu G et al. 2018. Supporting the development of a
Benefits for Universal Health Coverage. Washington, DC: Center for health benefits package in Malawi. BMJ Global Health 3: e000607.
Global Development, 1–398. Paul E, Lamine Dramé M, Kashala J-P et al. 2017. Performance-based financ-
Goodman CA. 2006. The cost-effectiveness of improving malaria home man- ing to strengthen the health system in Benin: challenging the mainstream ap-
agement: shopkeeper training in rural Kenya’. Health Policy and Planning proach. International Journal of Health Policy and Management 7: 35–47.
21: 275–88. Portnoy A, Ozawa S, Grewal S et al. 2015. Costs of vaccine programs across
Greco G, Daviaud E, Owen H et al. 2017. Malawi three district evaluation: 94 low- and middle-income countries’. Vaccine 33: A99–108.
community-based maternal and newborn care economic analysis. Health Ruelas E, Gomez-Dantes O, Leatherman S et al. 2012. Strengthening the qual-
Policy and Planning 32: i64–74. ity agenda in health care in low- and middle-income countries: questions to
Hauck K, Morton A, Chalkidou K et al. 2019. How can we evaluate the consider. International Journal for Quality in Health Care 24: 553–7.
cost-effectiveness of health system strengthening? A typology and illustra- Sixty-Seventh World Health Assembly. 2014. Health Intervention and
tions. Social Science & Medicine 220: 141–9. Technology Assessment in Support of Universal Health Coverage.
Helmchen LA, Lo Sasso AT. 2010. Outcomes in economic evaluations. Health WHA67.23(24 May 2014), 7–9. http://apps.who.int/medicinedocs/docu
Economics 19: 1300–17. ments/s21463en/s21463en.pdf
Husereau D, Drummond M, Petrou S et al. 2013. Consolidated Health Stenberg K, Hanssen O, Tan-Torres T et al. 2017. Financing transformative
Economic Evaluation Reporting Standards (CHEERS)—explanation and health systems towards achievement of the health Sustainable Development
elaboration: a report of the ISPOR Health Economic Evaluations Goals: a model for projected resource needs in 67 low-income and
Publication Guidelines Task Force. Value in Health 16: 231–50. middle-income countries. Lancet Global Health 1–13. doi:
Jiang SP et al. 2012. Effectiveness of pharmaceutical care in an intensive care 10.1016/S2214-109X(17)30263-2
unit from China’. Saudi Medical Journal 33: 756–62. Stinnett AA, Paltiel AD. 1996. Mathematical programming for the efficient al-
Kingkaew P, Werayingyong P, Aye SS et al. 2016. An ex-ante economic evalu- location of health care resources. Journal of Health Economics 15: 641–53.
ation of the Maternal and Child Health Voucher Scheme as a Turcotte-Tremblay, Anne Marie, Jessica Spagnolo, Manuela De Allegri, and
decision-making tool in Myanmar. Health Policy and Planning 31: 482–92. Valéry Ridde. 2016. Does performance-based financing increase value for
Kutzin J. 2013. Health financing for universal coverage and health system per- money in low- and middle- income countries? a systematic review. Health
formance: concepts and implications for policy’. Bulletin of the World Economics Review 6 (1). 10.1186/s13561-016-0103–9.
Health Organization 91: 602–11. United Nations. 2015. Transforming our World: The 2030 Agenda for
Kyaw SS, Drake T, Thi A et al. 2016. Malaria community health workers in Sustainable Development. doi: 10.1201/b20466-7
Myanmar: a cost analysis. Malaria Journal 15: 1–7. Vassall A, Bozzani F, Hanson K et al. 2019. Considering health-systems con-
Lagarde M, Blaauw D, Cairns J. 2012. Cost-effectiveness analysis of human straints in economic evaluation in low- and middle-income settings. Oxford
resources policy interventions to address the shortage of nurses in rural Research Encyclopedia of Economics and Finance 1–25. Oxford University
South Africa’. Social Science & Medicine 75: 801–6. Press. doi: 10.1093/acrefore/9780190625979.013.38
Levin C, Chisholm D, 2016. Cost-effectiveness and affordability of interven- Verguet S, Alkire BC, Bickler SW et al. 2015. Timing and cost of scaling up
tions, policies, and platforms for the prevention and treatment of mental, surgical services in low-income and middle-income countries from 2012 to
neurological, and substance use disorders. In: Patel V et al. (eds). Mental, 2030: a modelling study. The Lancet Global Health 3: S28–37.
Neurological, and Substance Use Disorders: Disease Control Priorities, 3rd Volmink HC, Bertram MY, Jina R et al. 2014. Applying a private sector capi-
edn. Washington, DC, 219–236. doi: 10.1017/CBO9781107415324.004 tation model to the management of type 2 diabetes in the South African pub-
Manzi A, Mugunga JC, Iyer HS et al. 2018. Economic evaluation of a mentor- lic sector: a cost-effectiveness analysis. BMC Health Services Research 14:
ship and enhanced supervision program to improve quality of integrated 1–9.
Health Policy and Planning, 2020, Vol. 0, No. 0 11

Warren AE, Wyss K, Shakarishvili G et al. 2013. Global health initiative countries: a systematic review of methodological frameworks. Health
investments and health systems strengthening: a content analysis of global Economics S37–54. doi: 10.1002/hec
fund investments. Globalization and Health 9: 30. World Health Organization. 2007. Everybody’s Business: Strengthening

Downloaded from https://academic.oup.com/heapol/advance-article/doi/10.1093/heapol/czaa116/5999228 by Karolinska Institutet University Library user on 20 December 2020
Watkins D, Jamison D, Mills T et al. 2017. Universal Health Coverage and es- Health Systems to Improve Health Outcomes: WHO’s Framework for
sential packages of care. In: Jamison D et al. (eds). Disease Control Action. Geneva. doi: 10.1371/journal.pone.0013372
Priorities: Improving Health and Reducing Poverty, 3rd edn. Washington, Zeng W, Li G, Ahn H et al. 2018. Cost-effectiveness of health systems
DC: The International Bank for Reconstruction and Development/The strengthening interventions in improving maternal and child health in low-
World Bank, 43–68. and middle-income countries: a systematic review. Health Policy and
Wesson J, Olawo A, Bukusi V et al. 2008. Reaching providers is not enough to Planning 33: 283–97.
increase IUD use: a factorial experiment of “Academic Detailing” in Kenya. Zhou J, Liu D-B, Zhong J-W et al. 2012. Feasibility of a remote monitoring
Journal of Biosocial Science 40: 69–82 system for home-based non-invasive positive pressure ventilation of children
Wiseman V, Mitton C, Doyle-Waters M et al. 2016. Using economic evidence and infants. International Journal of Pediatric Otorhinolaryngology. 76:
to set healthcare priorities in low-income and lower-middle-income 1737–40.

You might also like