Achieving universal electrification of rural healthcare facilities in sub-Saharan Africa with decentralized renewable energy technologies

Abstract

A potential response to the COVID-19 pandemic in sub-Saharan Africa (SSA) with long-term benefits is to provide electricity for medical equipment in rural health centers and communities. This study identifies a large gap in the electrification of healthcare facilities in SSA, and it shows that decentralized photovoltaic systems can offer a clean, reliable, quick, and cost-effective solution. The cost of providing renewable electricity to each health facility by a stand-alone PV system is analyzed for a given location (incorporating operational costs). The upfront investment cost for providing electricity with PV to >50,000 facilities (mostly primary health posts) currently without electricity is estimated at EUR 484 million. Analysis of the accessibility and population distribution shows that 281 million people could reduce their travel time to healthcare facilities (by an average of 50 min) if all facilities were electrified.

Full text

Article
Achieving universal electrification of rural
healthcare facilities in sub-Saharan Africa with
decentralized renewable energy technologies
Effective strategies for financing the electrification of healthcare remain a
challenge in sub-Saharan Africa. In this study, Moner-Girona et al. identify a large
gap in the electrification of healthcare facilities, and they show that decentralized
photovoltaic systems can offer a clean, reliable, quick, and cost-effective solution.
These findings provide a bottom-up geographic information system (GIS)
framework for policy makers, researchers, consultants, and other stakeholders
bridging two elements of the sustainable development goals: ‘‘energy for all’’
(SDG7) and ‘‘healthcare for all’’ (SDG3).
Magda Moner-Girona, Georgia
Kakoulaki, Giacomo Falchetta,
Daniel J. Weiss, Nigel Taylor
magda.moner@ec.europa.eu
Highlights
In rural sub-Saharan Africa, over
50,000 healthcare facilities lack
electricity supply
PV and battery systems offer a
clean, reliable, and cost-effective
solution
The estimated investment needed
would be just under EUR 500
million
281 million people could reduce
journey time to electrified
facilities by 50 min average
Moner-Girona et al., Joule 5, 2687–2714
October 20, 2021 ª2021 The Author(s).
Published by Elsevier Inc.
https://doi.org/10.1016/j.joule.2021.09.010
ll
OPEN ACCESS

Article
Achieving universal electrification of rural
healthcare facilities in sub-Saharan Africa
with decentralized renewable energy technologies
Magda Moner-Girona,
1,6,
*Georgia Kakoulaki,
1
Giacomo Falchetta,
2,3
Daniel J. Weiss,
4,5
and Nigel Taylor
1
SUMMARY
A potential response to the COVID-19 pandemic in sub-Saharan Af-
rica (SSA) with long-term benefits is to provide electricity for medi-
cal equipment in rural health centers and communities. This study
identifies a large gap in the electrification of healthcare facilities in
SSA, and it shows that decentralized photovoltaic systems can offer
a clean, reliable, quick, and cost-effective solution. The cost of
providing renewable electricity to each health facility by a stand-
alone PV system is analyzed for a given location (incorporating oper-
ational costs). The upfront investment cost for providing electricity
with PV to >50,000 facilities (mostly primary health posts) currently
without electricity is estimated at EUR 484 million. Analysis of the
accessibility and population distribution shows that 281 million peo-
ple could reduce their travel time to healthcare facilities (by an
average of 50 min) if all facilities were electrified.
INTRODUCTION
The United Nations’ (UN) sustainable development goals (SDGs)were adopted by all its
member states in 2015 as a universal action call to end poverty, protect the planet, and
ensure that allpeople enjoy peace and prosperity by 2030.For sub-Saharan Africa (SSA),
there is a direct link between SDG 7 ‘‘Ensure access to affordable, reliable, sustainable,
and modern energy’’ and SDG3 ‘‘Good health and wellbeing for all
1
’’ (Figure 1). Conse-
quently, international organizations such as the World Health Organization (WHO),
Global Fund, the United Nations Development Programme (UNDP), the World Bank,
USAID, and the EU have prioritized specific programs to support and ensure reliable
electricity access to underserved communities in rural areas.
2,3
Electricity is a crucial enabler for the provision of healthcare, education, and other
services, which in turn can aid communities in achieving socio-economic growth. Ac-
cording to the most recent statistics (2019), there remain 570 million people without
access to the most basic electricity services in SSA.
4
Renewable energy has become
the least-cost-effective option for generating electricity in most regions due to insti-
tutional, logistical, transport, and last-mile costs in SSA that run well above the
global averages.
5
As a result of these limitations, SSA is not on target to meet
SDG7 by 2030.
3,6
The lack of energy access extends to healthcare facilities, as one
in four facilities lacks a source of electricity, and three out of four facilities lack reli-
able power.
7–10
This situation is most pronounced in rural areas, and it varies consid-
erably between countries (Figure 2).ThelackofelectrificationinlargepartsofSSA
leaves many healthcare facilities with inadequate power for both basic and emer-
gency services.
15,16
Electricity is essential for the majority of emergency care
Context & scale
Our results shed new light on the
potential of decentralized energy
systems to offer a reliable, quick,
and cost-effective way to increase
access to electricity for rural
healthcare facilities in sub-
Saharan Africa. This study
identified more than 55,000 such
facilities without electricity access
andanalyzedthecostsand
benefits of powering each of these
with solar photovoltaic and
battery storage systems.
Our results can provide a basis for
planning electrification programs
for health facilities in rural sub-
Saharan Africa and can be useful
to policy makers, researchers,
consultants, and other
stakeholders involved in
electrification planning and
healthcare improvement. The
level of granularity, covering
community, national, and regional
levels, is particularly relevant to
prioritizing the allocation of
limited governmental funding,
highlighting where electrification
is most needed and likely to have
the greatest impact on health
services.
Joule 5, 2687–2714, October 20, 2021 ª2021 The Author(s). Published by Elsevier Inc.
This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/4.0/). 2687
ll
OPEN ACCESS

activities, including lighting, laboratory tests, and X-rays, as well as the ventilators
that are often critically important for respiratory support for COVID-19 patients.
Moreover, an estimated 70% of medical devices in the least-developed countries
regularly fail or are unavailable, with poor power quality being a major contributing
factor.
15
The COVID-19 pandemic further increased the pressure on the healthcare system in
SSA, and it highlighted the importance of energy access for delivering reliable
healthcare services. The situation is most challenging for rural communities, which
may also have few primary hospitals, shortages of medical staff, poor health literacy,
no access to clean water, and poor transportation infrastructure. For such commu-
nities, a fast and modular energy solution is urgently needed now, more than
ever.
17
One medium-term approach for boosting resilience in the face of the
COVID-19 pandemic in SSA has been to strengthen the critical services by providing
power to medical equipment in rural health centers and communities with the help of
decentralized renewable energy systems.
18
Decentralized energy resources (DER)
systems typically use renewable energy sources, including small hydro, biomass,
biogas, solar, wind, and geothermal power.
The cost of solar photovoltaic (PV) systems has decreased rapidly over the last
decade, and due to the expansion of global solar panel manufacturing capacity,
Figure 1. Interlinkages of the sustainable development goal 7 target (SDG.7) ‘‘ensure access to
affordable, reliable, sustain able, and modern energy for all’’ with the other sustainable
development goal targets
Source: adapted from https://knowsdgs.jrc.ec.europa.eu/interlinkages/tools?
visualization=chord&edges=0
1
European Commission, Joint Research Centre
(JRC), Ispra, Italy
2
Fondazione Eni Enrico Mattei (FEEM), Milan,
Italy
3
Faculty of Economics and Management, Free
University of Bozen-Bolzano, Bolzano, Italy
4
Curtin University, Bentley, WA, Australia
5
Telethon Kids Institute, Nedlands, WA, Australia
6
Lead contact
*Correspondence: magda.moner@ec.europa.eu
https://doi.org/10.1016/j.joule.2021.09.010
ll
OPEN ACCESS
2688 Joule 5, 2687–2714, October 20, 2021
Article

solar power costs are now expected to decrease at a faster rate than anticipated
before the COVID-19 outbreak.
19,20
The costs of lithium-ion battery systems are
also falling rapidly.
21
Installed solar PV power capacity in Africa increased from
273.5 MW
p
in 2010 to 11 GW
p
in 2020,
22
with a corresponding growth in technical
capacity and confidence in the technology. Such factors make solar power an attrac-
tive option in resource-constrained settings and can make it an affordable option in
many communities.
23,24
This study focuses on modular solar PV systems (including
battery energy storage) as a fast, cost-effective, clean, and reliable solution to supply
the power needed in health centers serving Africa’s rural communities.
As shown in Figure 2A, different countries have adapted different strategies for
supporting the electrification of healthcare; this heterogeneity is reflected in the
percentage of healthcare facilities with electricity access at the country level, which
shows large differences across Africa, ranging from almost 100% of healthcare fa-
cilities with electricity access in Gambia to 42% in Uganda. Since most sub-Saharan
African governments give priority to electrification of social infrastructure (schools
and health centers), the majority of countries have higher rates of electrification for
health centers than in residential buildings. However, these numbers do not cap-
ture other important factors, such as the quality and reliability of the electricity
supply. For instance, an average Nigerian household experiences daily power out-
ages for around 19 h.
25
Electricity access in healthcare facilities in rural SSA has not
been thoroughly explored beyond a few country-specific studies with partial field
data. Figure 2B distinguishes on-grid and off-grid electricity access for three coun-
tries where sufficiently granular data are available, showing the large discrepancies
in access possibilities.
6
In Kenya, 77% of health centers rely on the public national
grid for their primary electricity needs. Conversely, in Niger, 51% of health centers
use off-grid solutions to cover their primary electricity demand.
6
Adair-Rohani
et al.
7,26
found that the proportion of facilities relying only on diesel generators
ranged from an average of 33% in Gambia to only 1% in both Uganda and Zambia.
Excluding Gambia, 4% of all facilities in SSA, on an average, relied only on diesel
generators for electricity.
Figure 2. Percentage of healthcare facilities with electricity access in SSA countries
(A) Percentage of healthcare facilities with electricity access in selected countries (% of total facilities).
(B) Percentage of healthcare facilities (% of total facilities) with on-grid and off-grid electricity access from available country-specific studies (partial field
data, 2007–2012). Source: data compilation from IEA et al.,
6
Adair-Rohani et al.,
7
Pittalis et al.,
11
World Bank Group,
12
UNESCO
13
and World Bank.
14
ll
OPEN ACCESS
Joule 5, 2687–2714, October 20, 2021 2689
Article

Irrespective of this evidence, there is a lack of systematic, recently updated data on
which facilities are with or without access to electricity and hence a limited awareness
of the electrification status. Consequently, it can be difficult to prioritize the provi-
sion of reliable electricity when planning and implementing energy investments in
the health sector in rural areas. This study aims to tackle this information gap by
combining satellite data, national statistical data, and several open-source datasets,
as further described in experimental procedures.
Several studies have examined the potential of decentralized renewable systems to
power healthcare in developing countries. For instance, Dholakia
27
discusses the
potential and critical barriers to the wider uptake of solar power for electrifying
healthcare in developing countries. His review of literature studies shows how power
provision enhances healthcare services. The review argues that it is crucial that
health policies recognize energy as a critical component of the overall infrastructure.
Olatomiwa et al.
28
illustrated the potential of standalone hybrid renewable energy
systems for basic healthcare services in rural areas through an optimization analysis
in Nigeria, highlighting the untapped potential and the noteworthy reliability gains,
even compared with an unreliable national grid. Franco et al.
29
carried out a review
of sustainable energy access and technologies for healthcare facilities in the Global
South. They highlighted that the optimal solution for medium-to-large rural health-
care facilities is a hybrid system coupling a renewable energy source with efficient
batteries and a diesel generator to minimize the cost of coping with the intermit-
tency of renewable energy sources. Orosz et al.
30
examined technical and economic
options for electricity, heating, and cooling in health and education applications in
rural SSA. They noted the significant benefits and cost savings of solutions based on
photovoltaics hybridized with liquefied petroleum gas/propane and micro-concen-
trating solar power tri-generation, compared with conventional diesel and LPG/pro-
pane-based heating and cooling. The World Resource Institute’s Energy Access
Explorer provides information on health center locations, electricity demand, and
renewable energy potential for Kenya, Uganda, and Tanzania.
31
The HOMER Power-
ing Health Tool
32
(USAID, ESMAP, WeCareSolar) is a free online model to create
initial designs of electric power systems for healthcare facilities that have no other
power supply (i.e., diesel generator) or have grid electricity available for a predict-
able period of hours each day.
Our study builds on such country-specific research and is the first of its kind to carry
out a continental-level assessment of the electrification access status for healthcare
facilities (the African electricity access health facility geodatabase: Data and compi-
lation procedure and population clusters and healthcare facility catchment areas),
including estimates of the travel time to each of the facilities with and without elec-
tricity (population clusters and healthcare facility catchment areas). Our analysis also
estimates (1) the energy requirements for powering the health centers and the
optimal size of the PV and battery system (estimation of health facility energy de-
mand) and (2) the associated costs (assessment of electrification costs). The results
are presented at continental, regional, and national levels.
Until recently, the main electricity supply options considered in rural areas of SSA were
the extension of the national grid or the use of standalone diesel generators. This anal-
ysis explores the potential of electrifying healthcare facilities in rural areas using PV mini-
grids (including batteries). A multi-criteria algorithm is used to identify healthcare facil-
ities with a high probability of having no electricity access (referred to as no electricity
access [NEA] healthcare facilities). To do so, we collected, harmonized, and aggregated
a range of open-source datasets (see experimental procedures).
ll
OPEN ACCESS
2690 Joule 5, 2687–2714, October 20, 2021
Article

Figure 3 shows heatmaps of healthcare facilities according to their electricity access
status. The total number of mapped facilities is 122,899 for the whole African conti-
nent, categorized as (1) those with no access to electricity (NEA) (in total, 56,801) and
(2) those in locations with detected electricity access (WEA) (in total, 66,098). Of the
NEA healthcare facilities, 96% are primary health posts (offering very basic services),
and a small number are primary (3%) and secondary (1%) hospitals (see estimation of
health facility energy demand and Table 2). The heatmap in Figure 3A highlights
areas with a higher concentration of facilities without electricity access (yellow); an
example of a dense concentration is in Nigeria, with more than 13,000 NEA facilities.
RESULTS
Following the methodology described in experimental procedures, the study iden-
tified 56,801 health centers in SSA that fall in the NEA category. For each NEA facil-
ity, we assessed the costs of providing electricity with PV decentralized systems and
identified the population served under several travel time assumptions. Table 1 sum-
marizes the key data aggregated at the country level, while the following sections
summarize the results at various geographical levels.
Estimation of population served
The bivariate map (Figure 4) provides a geospatial overview of the most critical level
3 administrative units in terms of lack of electricity access for the population and for
health facilities. The percentage of NEA healthcare facilities ranges from low (left in
the legend) to high (right) and the percentage of the population with electricity ac-
cess from low (bottom in the legend) to high (top in the legend). Gray light colors (A1
Figure 3. Heatmaps of healthcare facilities according to their electricity access status
(A and B) For (A) facilities with detected electricity access (WEA) and (B) facilities with no detected access to electricity (NEA).
ll
OPEN ACCESS
Joule 5, 2687–2714, October 20, 2021 2691
Article

Table 1. Electrification of healthcare fa cilities by PV off-grid systems: total investment costs, average LCOE and share of population
Country
Population Healthcare facilities Costs
Total
Pop. 20
min
walking
to all
facilities
Pop.
20 min
walking
time
to WEA
Pop.
60 min
travel
time
to all
facilities
Pop. 60 min
travel time
to WEA
Total
number
of facilities
Number
of NEA
facilities
Total PV
capacity
Average
LCOE
Total
upfront
costs
Total
NEA
upfront
costs
[million] [%] [%] [%] [%] – – [kWp]
[EUR/
kWh]
[EUR
million]
[EUR
million]
AGO Angola 25.0 53 51 94 92 1,707 861 4,474 0.37 16 6.3
BDI Burundi 11.1 70 50 100 100 775 329 2,876 0.44 10 3.3
BEN Benin 10.9 64 63 100 99 970 67 2,464 0.39 9 0.5
BFA Burkina Faso 18.1 63 58 96 95 1,863 519 4,287 0.36 16 3.7
BWA Botswana 2.3 65 64 99 98 671 224 1,665 0.41 6 1.7
CAF Central
African Rep.
4.9 59 32 97 56 820 715 2,600 0.41 10 6.8
CIV Cote d’Ivoire 22.7 62 58 99 99 2,004 383 6,115 0.39 22 2.9
CMR Cameroon 23.3 81 74 99 96 3,540 1,724 10,504 0.42 38 15.8
COD Dem.
Rep. Congo
77.3 62 38 98 67 14,746 9,790 39,658 0.42 148 100.7
COG Congo 4.6 67 61 95 90 355 167 1,050 0.43 4 1.5
COM Comoros 0.8 62 62 100 100 71 0 83 0.22 0 0
DJI Djibouti 0.9 44 44 91 87 67 30 764 0.31 3 0.2
ERI Eritrea 5.3 60 56 79 75 276 173 646 0.34 2 1.1
ETH Ethiopia 99.2 48 30 95 89 5,399 3,534 12,209 0.40 44 26.3
GAB Gabon 1.7 59 54 99 93 560 397 1,571 0.43 6 3.3
GHA Ghana 27.4 58 54 100 99 2,092 640 5,478 0.39 19 4.8
GIN Guinea 12.6 58 42 99 94 1,648 1,007 3,565 0.37 13 6.9
GMB Gambia 2.0 59 56 100 100 123 26 251 0.34 1 0.2
GNB Guinea-
Bissau
1.8 28 27 81 80 14 2 186 0.37 1 0.1
GNQ Equatorial
Guinea
0.8 42 29 99 99 50 18 292 0.45 1 0.3
KEN Kenya 46.0 54 49 96 90 6,262 1,712 15,148 0.39 54 12.5
LBR Liberia 4.5 63 48 97 91 909 537 2,230 0.42 8 4.4
LSO Lesotho 2.1 34 30 100 100 184 42 820 0.49 3 0.4
MDG Madagascar 24.0 30 15 78 45 4,175 3,723 9,187 0.39 33 28.2
MLI Mali 17.6 40 34 93 78 1,845 1,278 3,752 0.35 14 8.4
MOZ Mozambique 27.8 51 40 92 82 1,643 852 10,702 0.41 38 20.9
MRT Mauritania 4.0 49 47 80 74 712 348 1,146 0.31 4 2.1
MUS Mauritius 1.3 84 84 100 100 176 0 426 0.36 1 0
MWI Malawi 17.3 20 15 100 98 700 353 2,401 0.41 9 2.9
NAM Namibia 2.5 39 38 91 90 393 69 879 0.33 3 0.4
NER Niger 19.9 52 40 89 76 3,018 1,943 4,834 0.31 18 11.2
(Continued on next page)
ll
OPEN ACCESS
2692 Joule 5, 2687–2714, October 20, 2021
Article

Table 1. Continue d
Country
Population Healthcare facilities Costs
Total
Pop. 20
min
walking
to all
facilities
Pop.
20 min
walking
time
to WEA
Pop.
60 min
travel
time
to all
facilities
Pop. 60 min
travel time
to WEA
Total
number
of facilities
Number
of NEA
facilities
Total PV
capacity
Average
LCOE
Total
upfront
costs
Total
NEA
upfront
costs
[million] [%] [%] [%] [%] – – [kWp]
[EUR/
kWh]
[EUR
million]
[EUR
million]
NGA Nigeria 182.1 80 71 99 98 36,428 13,060 93,701 0.40 339 108.9
RWA Rwanda 11.6 50 32 100 100 607 276 2,146 0.43 8 2.5
SDN Sudan 40.2 27 26 70 69 491 134 2,022 0.30 8 3.2
SEN Senegal 15.1 72 69 99 98 1,538 464 2,947 0.33 11 3.0
SLE Sierra Leone 6.4 69 43 100 87 1,808 1,327 4,084 0.41 14 10.3
SOM Somalia 10.8 60 43 94 71 865 618 1,628 0.30 6 3.8
SSD South Sudan 12.3 58 34 88 67 1,779 1,355 4,188 0.40 15 10.8
STP Sao Tome
Principe
0.2 67 67 97 97 54 249 0.39 1
SWZ Swaziland 1.3 27 26 100 99 139 22 447 0.56 2 0.2
TCD Chad 14.0 64 50 90 72 1,570 1,258 3,785 0.35 14 10.0
TGO Togo 7.3 50 47 99 97 363 141 1,571 0.40 6 1.3
TZA Tanzania 53.4 57 39 96 90 7,547 3,840 17,363 0.39 62 28.5
UGA Uganda 39.1 47 45 99 99 4,404 533 10,000 0.40 36 3.9
ZAF South Africa 54.5 52 51 100 100 4,713 377 17,191 0.49 66 3.8
ZMB Zambia 16.2 31 29 95 86 1,388 893 3,539 0.37 13 7.0
ZWE Zimbabwe 15.6 27 25 87 67 1,437 1,040 4,275 0.43 15 8.7
Total/
average
999.8 54 70 91 88 122,899 56,801 321,399 0.39 1,170 484
The number of population represents the amount of population living in areas within the indicated travel time (60 min optimal travel time or 20 min walking). The total investment costs (for components,
engineering, and soft costs) are calculated aggregating the total cost of decentralized energy options, taking into account the optimized size of the system for each health facility with its specific load con-
sumption and the economy of scales (lower upfront cost for larger systems). The LCOE is calculated as an average of the LCOE values per country, taking only the NEA facilities serviced by PV.
ll
OPEN ACCESS
Joule 5, 2687–2714, October 20, 2021 2693
Article

in the legend) indicate areas with high electricity access rates for both the population
and healthcare facilities. Dark cyan color shades (A3 in the legend) represent areas
with high electrification rates of facilities but low electricity access rates for the pop-
ulation. The darkest mahogany shading (B3, C3) defines areas with low rates of
electricity access for both the population and healthcare facilities, for example, pre-
dominantly in regions in Central Africa. Figure 4 also identifies areas with an asym-
metric relationship between the two measures of electricity access. For instance,
areas with a high level of electricity access to healthcare facilities but where the gen-
eral population still has a low access rate (dark blue).
Unsurprisingly, a review of the literature on healthcare service location found that the
relationship between the proximity to healthcare facilities and health outcomes is
significant.
33
Therefore, to estimate the potential benefits of electrifying healthcare
facilities, the analysis focused on quantifying travel time to the nearest facility, either
currently electrified or not. Two options are considered: travel by the fastest avail-
able transport mode (optimal travel time) and walking time. The latter is particularly
relevant to the rural locations and small primary health posts that make up the ma-
jority of NEA healthcare facilities in this study.
Figure 5Amaps the travel timeto the most accessible healthcare facility at each location
(by any means of transport) and Figure 5B shows t he additional trave l time needed to get
Figure 4. Bivariate representation at administrative level 3 of the percentage of NEA healthc are
facilities and the percentage of population without access to electricity
The percentage of NEA healthcare facilities ranges from light blue hue (low values) to dark (high values),
while for population with electricity access ranges from light red hue (low values) to dark (high values).
ll
OPEN ACCESS
2694 Joule 5, 2687–2714, October 20, 2021
Article

to the most accessible health facility that already has access to electricity. The underlying
accessibility estimation methodology is described in the study conducted by Weiss
et al.
34
Note that this analysis only focuses on facilities that can be reached most rapidly
and ignores complexities such as individuals choosing to go to more distant facilities
because of specific preferences (e.g., public/private hospitals). Also, there is no system-
atic database of private healthcare facilities in SSA. Several hotspot regions are evident
(e.g., the Democratic Republic of the Congo, Madagascar, and Chad), where signifi-
cantly longer journeys are required for people needing access to healthcare facilities
with electricity. Figures 5C and 5D show the same comparison for travel times by foot
(a prevalent travel mode in rural areas). Here, the difference is even larger in some
areas—a significant discrepancy is observed in almost all countries from Eastern and
Central Africa, especially in Somalia, Ethiopia, the Democratic Republic of the Congo,
the Central African Republic, Chad, and Madagascar. Concerning the differences in
healthcare accessibility between countries in Western and Central Africa, the disparity
is largely due to population density and, to a lesser extent, the completeness of the
healthcare facility dataset for each country. For example, Nigeria has a population of
over 200 million, whereas DRC (second largest population in Africa and 2.5 times larger
in area) has about 90 million. Nigeria also has more than 3 times the number of
Figure 5. Travel times to healthcare facilities
(A and B) (A) Optimal travel time (min) to the most accessible healthcare facility (to all facilities,
including NEA) compared with (B) additional time required when traveling to only a facility with
electricity (WEA).
(C and D) (C) Optimal walking time (min) to the most accessible healthcare facility (all including NEA)
compared with (D) the additional time required when walking to a facility with electricity (WEA).
ll
OPEN ACCESS
Joule 5, 2687–2714, October 20, 2021 2695
Article

healthcare facilities. For these reasons, the accessibility map shows that people in West
Africa are closer to healthcare facilities.
In terms of the population affected, Figure 6 shows cumulative curves ofthe share of
population in each country with a given travel time to the most accessible healthcare
facility. These curves illustrate the accessibility inequality between countries for all
health centers (purple lines) compared with the health centers with access to elec-
tricity only (green lines). Figure 6 also highlights the pronounced differences
between walking time distance to any facility (blue lines) and to only electrified
UGA ZAF ZMB ZWE
SLE SOM SSD SWZ TCD TGO TZA
MWI NAM NER NGA RWA SDN SEN
KEN LBR LSO MDG MLI MOZ MRT
ETH GAB GHA GIN GMB GNB GNQ
CMR COD COG COM CPV DJI ERI
AGO BDI BEN BFA BWA CAF CIV
20 60 120 180 20 60 120 180 20 60 120 180 20 60 120 180
20 60 120 180 20 60 120 180 20 60 120 180
0%
25%
50%
75%
100%
0%
25%
50%
75%
100%
0%
25%
50%
75%
100%
0%
25%
50%
75%
100%
0%
25%
50%
75%
100%
0%
25%
50%
75%
100%
0%
25%
50%
75%
100%
Minutes
Share of population
All (optimal) All (walking) Electr. only (optimal) Electr. only (walking)
Optimal travel time to the most accessible facility
Figure 6. Comparison of the accumulative population at optim al travel time (min) to most accessible facility
To all facilities (including NEA, violet line), to the nearest electrified health facility (WEA, green). For walking time to all most accessible facilities
(including NEA, blue), and to the nearest electrified health facility (WEA, yellow)
ll
OPEN ACCESS
2696 Joule 5, 2687–2714, October 20, 2021
Article

healthcare facilities (yellow lines). For instance, a greater percentage of the popula-
tion of Nigeria (NGA) can rapidly access healthcare facilities than in Equatorial
Guinea (GNQ). Disparities between countries are larger for walking travel times. In
the Democratic Republic of the Congo, only about 50% of the population can reach
an electrified health facility in less than 60 min by foot. Note that the 60 min
threshold for traveling on foot generally approximates to the 5 km buffer found to
be meaningful for determining the propensity for care seeking.
35
At the continental level, only 4% of the sub-Saharan African population lives within
>1hoftraveltimefromahealthfacility,but10%livewithin>1htraveltimefroman
electrified health center. Considering travel by foot, 46% of the SSA population lives
within 20 min or more of walking time from a health facility, while 54% from an elec-
trified health center. Limitations on access to electrified facilities is higher in some
countries: for instance, in Madagascar, 55% of the population live more than 1 h
from the nearest electrified facility and 85% with more than 20 min of walking
time.Incontrast,inMalawi,only2%areat60minormore,althoughthevaluefor
>20 min walking time is similar to that of Madagascar (70%). The Central African Re-
public (44% optimal travel time), the Democratic Republic of the Congo (32%),
Figure 7. Distribution maps for NEA healthcare facilities
(A) Upfront cost of NEA healthcare facilities (EUR).
(B) Estimated annual electricity demand per health center (kWh/year)
(C) Optimized PV capacity requirement (kW
p)
.
(D) Optimized battery storage requirement (kWh). Note that the high electricity demand for
healthcare facilities in Sudan and Mozambique reflects the assigned classification; field data would
be needed to fully harmonize this across the continent.
ll
OPEN ACCESS
Joule 5, 2687–2714, October 20, 2021 2697
Article

and South Sudan (26%) also had high percentages of the population living further
than 60 min away from electrified facilities. Figure 6 and Table 1 report detailed
country-level accessibility statistics.
Estimation of the electrification costs for the NEA healthcare facilities
Next, we estimated the optimized PV capacity, the optimized battery capacity, and
the associated costs for each of the healthcare facilities (Figure 7;Table 1), and then
aggregated them to estimate the total capacity and costs at the national level (Fig-
ures 8 and 9). Details of the underlying data, calculations, and assumptions are found
in experimental procedures.Figure 7A shows the estimated PV upfront cost; Fig-
ure 7B shows the estimated annual electricity demand based on the levels of NEA
healthcare facilities. Figures 7C and 7D show the calculated PV system power capac-
ity requirement [kW
p
] and battery storage capacity requirement [kWh], respectively.
These values are location dependent. For primary health posts with the same elec-
tricity demand (1,825 kWh/year), the optimized PV system size ranges from 1.2 to
2.8 kW
p
and battery size from 2.4 to 7 kWh. The two extremes correspond to loca-
tions on the coast of Somalia, which has high solar radiation and lower seasonality,
and on the South African east coast, which needs a larger PV array and battery size
due to lower irradiation and distinct winter/summer seasons. The estimated PV up-
front cost (including hardware, engineering, and soft costs) varies correspondingly
from EUR 4,500 to EUR 11,000.
Figure 8. LCOE (EUR/kWh) for PV systems installed in NEA health centers
LCOE of decentralized PV systems (including storage) for the NEA healthcare facilities in Africa
average at administrative level 3.
ll
OPEN ACCESS
2698 Joule 5, 2687–2714, October 20, 2021
Article

The levelized cost of electricity (LCOE) was computed for each of the healthcare
facilities, with values ranging from 0.25 EUR/kWh in Somalia to 0.62 EUR/kWh in cen-
tral Nigeria. Figure 8 shows the average LCOE values per region. Large geograph-
ical differences can be observed, generally depending on the local climatic condi-
tions
36
and degree of seasonality.
To estimate the total generation capacity per country required to power all the rural
health facilities, the PV capacities of both the WEA and NEA facilities were aggregated
(Figure 9). In a similar way, the overall costs were estimated at the national level (Figure 9,
blue bars) for rural healthcare facilities, reaching a total upfront investment cost for all
SSA of EUR 1,170 million. If only the NEA facilities are considered, the sum is EUR 484
million. The cost of providing universal access to healthcare facilities with electricity in
each country will clearly depend on the percentage of healthcare facilities that already
have access to electricity. For example, in Kenya and Tanzania, the total PV capacity
required to cover all health centers is in the same range (15 MW
p
and 17 MW
p
,respec-
tively), but the costs of electrification for all NEA facilities are much higher in the case of
Tanzania (EUR 28.5 million) than in Kenya (EUR 12.5 million) because in Tanzania, 50% of
health centers do not have access to electricity compared with 26% in Kenya.
Figure 9. Percentage of NEA healthcare facilities per country (orange bars, first column)
Total PV capacity (MW
p
) per country estimated to be installed in the NEA healthcare facilities (purple bars). Total costs (million EUR) of installing PV
systems to NEA healthcare facilities (blue bars). The share of population with more than 60 min travel time to an electrified health facility (dark green
bars) and the percentage of population with more than 20 min walking time to an electrified health center (cyan bars).
ll
OPEN ACCESS
Joule 5, 2687–2714, October 20, 2021 2699
Article

In addition,the avoided greenhousegas (GHG) emissions werecalculated by computing
the emissions of a standalone diesel generator supplying the same electricity demand
over the lifetime of the optimized PV systemper health center (see experimental proced-
ures). Figure 10 shows the estimations of the total avoided GHG emissions per country
when powering all the rural health facilities with renewables. In addition, the total GHG
emissions avoided amounted to 206 kt of CO
2
over a system lifetime of 20 years.
An additional aspect of sustainability is the end-of-life management of PV-battery
mini-systems, as addressed in several studies.
37,38
The volume of PV modules for
NEA electrification is projected to be approximately 5 MW, with a useful life of 25
years. This is less than 0.5% of the 2020 PV market volume in Africa.
22
Analysis of re-
cycling and e-waste policies is beyond the scope of this paper, but it is expected that
theproductsusedinruralelectrificationwillbeabletotakeadvantageofthena-
tional schemes needed to address the overall market.
Table 1 summarizes the total investment costs, average LCOE, and the population
that can reach a potentially new-electrified healthcare facility within 1 h motorized
traveltimeand20minbyfootforeachcountry.
To conclude, the datasets created in this work are being used to provide free online
information via the Clean Energy Access (CEA) Tool (accessible at http://d6-dev-
africap.jrc.it/energy_tool). This has been developed for visualizing and analyzing in-
formation on electricity access in Africa and the overall clean energy outlook. The
tool (Figure 11) has a specific focus on improving the general healthcare in areas
of rural Africa with minimal or no access to electricity. It allows the visualization
and analysis of the EHFDB healthcare facilities singularly or for user-defined areas.
The results can be summarized at the national and/or subnational levels and then
downloaded. The EHFDB database is also available for download from https://
data.jrc.ec.europa.eu/collection/id-0076).
DISCUSSION
Our results shed new light on the potential of decentralized energy systems to offer a
reliable, quick, and cost-effective way to increase access to electricity in rural
Figure 10. Breakdown of avoided GHG emissions (tCO
2eq
) estimated per NEA health center and
aggregated per country in each African region
ll
OPEN ACCESS
2700 Joule 5, 2687–2714, October 20, 2021
Article

healthcare facilities in SSA. This study identified 56,801 NEA health centers in SSA
andanalyzedthecostsandbenefitsofpoweringthesebydecentralizedPVsystems.
We estimated that 281 million people would benefit from reduced travel time to a
health facility with electricity access if all NEA facilities were provided with PV sys-
tems (reaching an average of 50 min). The impact is even more pronounced when
considering only walking as a means of traveling, in which case, 298 million people
would reduce travel time (by an average of 6 h) if all current NEA facilities were pro-
vided with electricity. This highlights a particularly urgent need to power those
healthcare facilities which serve populations with above-average journey times to
health posts and hospitals with electricity. The analysis reported here provides a
means to easily identify such priority areas.
There are no significant resource or technical barriers to using solar photovoltaic-
based systems to provide electricity for rural healthcare facilities. The PV/battery sys-
tem size optimization was computed for 122,899 healthcare facilities, assuming a
level of electricity demand depending on the type of healthcare facility (according
to the WHO electricity demand tiers) and a full coverage of over 95% of days per
year (286 GWh/year in total). Under the chosen assumptions, an annual energy de-
mand of 121 GWh would be needed to cover the needs of the sub-Sahara African
NEA rural health centers. Up to 2020, Africa, the continent with the richest solar re-
sources in the world, had installed 11 GW
p
of solar PV, about 1% of the global to-
tal.
22,39
According to the IEA Africa projections,
5
PV deployment should grow to
almost 15 GW
p
a year and is projected to reach 320 GW
p
in 2040. The results of
this study estimated the total current demand for NEA facilities will be satisfied by
133 MW
p
of solar PV (with 220 MWh of battery storage), which is small relative to
the above projects but would potentially have an enormous societal impact. In addi-
tion, electrifying NEA using PV compared with non-renewable sources would avoid
206 t CO
2
in GHG emissions.
40
Figure 11. Clean energy access tool, including healthcare facilities analysis
Open-source web tool developed by the European Commission-JRC (http://d6-dev-africap.jrc.it/energy_tool).
ll
OPEN ACCESS
Joule 5, 2687–2714, October 20, 2021 2701
Article

The LCOE of the PV systems (including battery storage) installed in the NEA aver-
agesto0.4EUR/kWh,witharangefrom0.26to0.64EUR/kWh(Table 3). These
values are lower than those reported up to now for mini-grid systems in Africa. How-
ever, the cost is decreasing, and the intention is to show what could be affordable
with economies of scale, standardized products, more experienced local suppliers,
and efficient administrative procedures. It is important to note that when the LCOE is
calculated without design, installation, or permitting costs—for instance, for com-
parisons with other generation renewable or fossil technologies—the LCOE values
vary between 0.16 and 0.49 EUR/kWh.
The total upfront investment cost of powering the existing SSA NEAs by PV systems
(including battery storage and soft costs) totals to EUR 484million. This sum is relatively
minor compared with the EUR 15.6 billion financial flows in 2016, supporting clean and
renewable energyin developing countries,
5
as well as the latest IEA figuresfor achieving
universal electricity supply in Africa of around EUR 92 billion a year through 2040. The
funds we envisage to support PV expansion for NEA include those aimed at achieving
SDG3 on good health and well-being, where electrification would be part of an overall
package of measures. This further increases the policy relevance of our bottom-up
assessment. Overall, the analysis presented in this paper and the associated open-
source web tool (http://d6-dev-africap.jrc.it/energy_tool) can support a better integra-
tion of energy and health policy by identifying the areas/countries where the invest-
ments are most needed. Already, some development agencies are mobilizing invest-
ments in the direction suggested by our paper.
32,41,42
Despite the large amount of data collection, computing, and analysis involved in the
development of the study, limitations remain. The data-intensiveness of the analysis
implies growing uncertainty over the reliability of the database, as some sources
such as the existing grid infrastructure and facilities locations and characteristics
might be outdated or incomplete. For example, there are limited data available
on electricity grid lines in a number of sub-Saharan African countries. Another limi-
tation of the approach is the use of night-time lights to estimate electrification as,
for example, it could miss facilities with standalone electricity access closed in the
night or/and not having outdoor lights. The methodology does not address the ex-
isting unreliability of the grid electricity, which is also a key factor to consider in terms
of the quality and continuity of service a health facility can provide. In the analysis, we
rely on the Global Human Settlement Layer (GHSL) gridded population product, but
adoption of different population products might lead to slightly different results, as
highlighted in study conducted by SDSN and TReNDS.
43
Moreover, gridded popu-
lation products are based on statistical downscaling of census data and thus system-
atically exclude invisible populations, as reported in the study conducted by Carr-
Hill et al.
44,45
Although beyond the scope of our paper (which focuses on meeting
the current needs of existing facilities), future research could use the results of
modeling studies projecting future facility needs based on population growth and
the accessibility and/or patient beds target-based in our study
33
to also estimate
the potential future demand arising from adding new facilities. In addition, the
accessibility analysis excludes individual preferences that may push individuals to
seek healthcare at a facility that is not the most accessible one. Finally, the classifi-
cation of facilities for each country might have limitations, as very different national
systems have been manually homogenized into generally valid healthcare tiers.
Validation of the applied model was completed by partially available health facility
data (e.g., Gambia) and visual interpretation of satellite images. However, in specific
countries (such as Zambia and Rwanda), the lack of complete information on medium
ll
OPEN ACCESS
2702 Joule 5, 2687–2714, October 20, 2021
Article

voltage and low voltage grid lines leads to larger numbers of NEA centers than the
publicly available statistics (39% instead of 20% and 30% instead of 18%, respec-
tively). Several institutions (including the WHO) are currently performing a global
assessment of electricity in healthcare facilities with the aim of closing the existing
knowledge and lack of information,
2
and the outcomes can support the validation
of the methodology used here and increase the accuracy of the analysis.
Nonetheless, our methodology is a first approximation to quantify and analyze the cur-
rent situation for all of SSA and can providea basis for further studies that can take advan-
tage of complementary field data collection to create a more comprehensive database
of energy access levels for all social infrastructure, not just health facilities. To address
this need, data collection should aim to offer a multidimensional picture of the availabil-
ity and reliability of existing electricity services, either off-grid or on-grid. This study also
highlights the importance of the aggregation of multiple data sources, both from na-
tional or regional surveys and field data collection that promotes better understanding
of the subject. Representing information with a geospatial dimension can further help
outline the multifaceted picture of energy access in health centers in Africa. Also, the re-
sults provide an opportunity for future studies toaddress factors such as a moredetailed
representation of the health center load profiles, including seasonal variability, future in-
creases in electricity demand with economic and demographic growth, and the use of
PV systems for grid-connected health centers to ensure power reliability.
Our results can be considered as a promising aspect for planning the electrification
of health facilities in rural SSA and are potentially beneficial for policy makers, re-
searchers, consultants, and other stakeholders involved in electrification planning
and healthcare improvement. The level of granularity, covering community, na-
tional, and regional levels, is particularly relevant to the prioritization in the alloca-
tion of limited governmental funding, highlighting regions where electrification is
most needed and likely to have the greatest impact on health services for rural pop-
ulations. Effective strategies for financing electrification of healthcare are critical and
remain a challenge in SSA.
46
Therefore, when planning at the national level, it is of
critical importance to take into account how programs are designed and what prior-
ities are applied by national and/or local authorities. Establishing evidence-based
and multisectoral strategies tailored to each country-specific context remains imper-
ative. Given the complex and multifaceted nature of sustainable energy access,
composite indicators can help attract investment in decentralized electricity gener-
ation. An example of this is the PV Decentralized Energy Investments (PV-DEI) in-
dex,
47,48
which covers the environmental, social, political, and financial aspects
with over 50 individual indicators. High scores in the social dimension imply that
the impacts of investing in decentralized PV are likely to significantly improve various
social outcomes. The methodology introduced in this study could be extended to
assess how solar energy for health facilities and other social infrastructure can be
developed as part of an integrated energy hub for off-grid rural communities.
EXPERIMENTAL PROCEDURES
Resource availability
Lead contact
Further information and requests for resources should be directed to and will be ful-
filled by the lead contact, Magda Moner-Girona (magda.moner@ec.europa.eu).
Materials availability
All unique materials generated in this study are available from the lead contact
without restriction, Magda Moner-Girona (magda.moner@ec.europa.eu)
ll
OPEN ACCESS
Joule 5, 2687–2714, October 20, 2021 2703
Article

Data and code availability
The electricity grid and the electricity access health facility database (EHFDB) gener-
ated during this study are available at: https://data.jrc.ec.europa.eu/collection/
id-0076
Visualization (temporal server with beta version) http://d6-dev-africap.jrc.it/
energy_tool
The methodology subsections were developed to identify the NEA facilities, the
population served by each of these facilities (defined as population clusters or catch-
ment areas), and the associated electricity demand. Figure 12 presents a workflow
chart summarizing the various stages and data inputs and outputs in the generation
and validation of the data.
The African electricity access health facility geodatabase: Data and
compilation procedure
Information on electricity access for healthcare facilities is scarce, not collected
systematically, and limited to specific projects or regional aggregations. As a result,
we developed an electricity access health facility database (EHFDB) in Africa.
Figure 12. Simplified methodology framework showing the data sources and the various stages in the generation and validation of the data: inputs,
processing, and outputs
ll
OPEN ACCESS
2704 Joule 5, 2687–2714, October 20, 2021
Article

Figure 13. Zoom-in maps for an area with healthcare facilities
(A) Zoom-in map of the healthcare facilities classified as (A) with electricity access (WEA) in blue
crosses and no electricity access (NEA) in orange crosses. The nightlights’ intensity background
layer (buffer from yellow [low intensity] to red [highest intensity] )and the electricity grid 5 km buffer
(in shaded brown).
(B) Zoom-in area with background layer, the optimal travel time (by foot in minutes) to healthcare facilities
with access to electricity: the shortest walking time (<15 min) is represented in green to the longest time in
violet. The map includesclusters of population (light orange polygons), WEA and NEA healthcare facilities.
The maps were generated using the following data, which was collected and processed by the authors:
ll
OPEN ACCESS
Joule 5, 2687–2714, October 20, 2021 2705
Article

Tthe EHFDB database can be downloaded from https://data.jrc.ec.europa.eu/
collection/id-0076 for this study using various source data. We accounted for the
lack of available data related to the status of electricity access of African health cen-
ters by extracting information from the available satellite remote sensing archives
and the spatial extent of the existing electricity grid.
49–53
For this study, a new elec-
tricity grid layer was compiled using multiple sources regarding the existing trans-
mission and distribution network. These include the Open Street Map, the World
Bank datasets, Arderne et al.,
54
the Economic Community of West African States
Observatory for Renewable Energy and Energy Efficiency,
55
as well as rural electri-
fication agencies and EU delegations in Africa (Burkina Faso,
56
Kenya,
57
Tanzania
58,59
).
A buffer zone of 5 km was created around the geo-located electrification infrastruc-
ture to identify areas likely to be able to connect to the electricity grid (https://data.
jrc.ec.europa.eu/dataset/624c6e71-3b9c-4f48-8c67-645911798d41). In parallel, we
utilized night-time lights data captured by a sensor aboard the NASA-National
Oceanic and Atmospheric Administration NPP satellite published in 2019 with a
450 m
2
resolution. The nightlight’s intensity layer was used to define areas with
the presence of night-time lights.
50–53,60–62
By combining the buffered electricity
grid and the nightlight layer mask, we created a proxy layer of electrification
coverage (Figure 13). Night light with an intensity of less than 0.5 was treated as
noise and excluded from our analysis.
The resulting EHFDB incorporates (1) the geographic locations of health centers
acquired from the healthcare facilities spatial database published by Maina J.
et al.,
45,59
combined with open-source data, such as OpenStreetMap (OSM;
https://www.openstreetmap.org/)andGoogleMaps(https://www.google.com/
maps/), (2) a geostatistical probabilistic layer on electricity access for healthcare fa-
cilities,
61,62
and (3) the estimated power requirements, the optimized PV and battery
size to meet these requirements, and the costs of the systems to be installed, calcu-
lated as described in the next sections.
Population clusters and healthcare facility catchment areas
Since an accurate and high-resolution geographical distribution of the population is
essential to determine the population without access to electricity, our spatial anal-
ysis used the integrated continental dataset of population distribution (published in
2019) provided by the GHSL framework.
64–67
The GHSL builds on past research and
relies on processing 40 years of Landsat imagery for mapping the global built-up
areas from 1975 to 2015.
64,65
The population grid datasets (GHS-POP) were derived
from the GHSL building density and population census data and were originally
developed to allocate census population data in built-up areas.
Population clusters were generated using the latest GHSL continental population layer
at a 250 m resolution.
67
The ArcGIS ‘‘regions creation’’ functionality was used as a first
step to identify the population clusters instead of people per single pixel (cell). This al-
lowed us to use the connectivity option with 8 neighboring cells, whereby adjacent cells
could become a part of the same cluster. Once the first clusters were defined, a second
Figure 13. Continued
GHS population grid; GHS-POP 16 data, produced and made publicly available by the European
Commission – JRC (https://ghsl.jrc.ec.europa.eu/data.php); and night-time lights Version 4 DMSP-
OLS 17, produced and made publicly available by NOAA’s National Geophysical Data Centre
(VIIRS DNB) (https://ngdc.noaa.gov/eog/dmsp/downloadV4composites.html).
ll
OPEN ACCESS
2706 Joule 5, 2687–2714, October 20, 2021
Article

step was used to connect the clusters within a 250 m distance or one pixel. Populations
were allocated to each unique cluster by aggregating the population for each pixel con-
tained in the delineated cluster. A geospatial mask layer combining the proxy layer (elec-
tricity grid with its 5 km buffer and the night light mask where lights are present) was used
to delineate population clusters with a probability of having access or those without ac-
cess to electricity. For practical purposes, this provides a way to estimate electricity ac-
cess and, therefore, to estimate the population in areas without electricity access (see
Figures 13Aand4). Figure 13B shows an example of an area with the defined population
clusters, where the NEA healthcare facilities are represented as orange crosses and the
WEA facilities as blue crosses.
Healthcare facility catchment areas
The number of potential beneficiaries for each health facility is obtained from estimates
of the population within a certain travel time. Continent-wide travel time maps were
generated for both on-foot travel and motorized travel using an established accessibility
estimation approach.
34
This approach is predicated on applying a least-cost algorithm
to a set of geo-located points in combination with a friction surface containing estimates
of the time it takes to traverse each pixel within a global grid with a spatial resolution of
30-arc-s (approximately 1 km at the equator).
33,34,68–71
The resulting travel time maps
were then used to cleave the population surface into pixels, and thus the population,
beyond a set of minute thresholds from the nearest facility. We also generated facility
catchments, defined as all pixels thathave the shortest travel time to a given healthcare
facility. We next calculated the zonal sum for each facility catchment from the population
layers. Lastly, we intersected the facility points with the catchment layer to account for
instances when multiplefacilities were located within the same 30 arc-s pixel, as defined
by the resolution of the travel time analysis. In doing so, each facility was associated with
a potential patient population, even if that population was shared among several co-
located facilities. The result of this analysis consists of the original table of facilities,
amended with additional columns containing the population of the associated facility
catchment. Note, however, that attributing populations to facilities in this manner ig-
nores individual decisions to seek care from more distant facilities or none at all.
72
Estimation of health facility energy demand
The WHO, along with the World Bank, have developed a multi-tier measurement of
electricity supply in primary and secondary healthcare facilities,
6,9,73
along with
typical functional profiles divided in several tiers
74
. In this analysis, we link the elec-
tricity demand of these tiers to 4 types of healthcare facilities (Table 2) (at the primary
level, these vary between countries in their definitions, specialization, population
served, services provided, infrastructure, and staffing. Health centers, medical cen-
ters, polyclinics, health posts, dispensaries, clinics, health huts, health units, etc. may
have similar functions but may equally represent different levels of service provision
between countries). The appropriate healthcare electricity demand tiers have been
defined following the WHO levels of health services of primary, first referral, second
referral, and tertiary referral
8
levels. Given that there is no universal standardized
definition of health facility types, making cross-country comparisons is difficult,
particularly at the primary level, as these vary between countries in their defini-
tions.
63
Each health facility was checked, and a manual labeling was applied to clas-
sify them in the 4 categories, using database queries based on the provided names
of each facility or the type of the facility if available. This was achieved by extracting
unique facility-type names, assigning the tier value, and parsing the corresponding
tier to each facility electricity demand in the database. The electricity demand for a
given health service level was assumed to remain the same across all countries. The
total electricity consumption for the healthcare facilities is calculated by:
ll
OPEN ACCESS
Joule 5, 2687–2714, October 20, 2021 2707
Article

Conshealth;n=½Cidin(Equation 1)
where:
n: identified health facility
i: health facility category
c
i
: daily electricity consumption for category i[kWh/day]
d
i
: number of operational days per category i[days]
Assessment of electrification costs
The cost of providing renewable electricity to each health facility by a standalone PV
and battery system is analyzed in two ways: (1) the capital expenditures (CAPEX)
based on the system specification for a given location and (2) the corresponding
LCOE, a measure of the average net present cost of electricity generation for a
generating plant over its lifetime.
Costs breakdown
Based on the field installation costs of PV systems in SSA,
40
we grouped the initial
investment costs into three main factor groups for the hardware components and
an additional one for the engineering and soft costs (see Table 3). The component
costs and their shares are extracted using the bottom-up gathering methodology
used by Moner-Girona et al.
40
In this study, the engineering and soft costs were calculated using the cost shares of
each factor group as a percentage of the total cost,
40
as the financial costs are highly
dependent on the country and local conditions, in particular, for accessing remote or
difficult to reach locations. The share of total capital costs used was 20% on an
average for the PV array and mounting structure, 20% for BOS, 27% for storage
and monitoring, and 33% for the engineering and soft costs.
Levelized cost of electricity
The LCOE was estimated for each facility by a location-specific analysis of the energy
output and reliability of the PV system, as described by Huld et al.
36
The PV array and
Table 2. Electricity demand of health facility per level of services category
Health
services level
Energy
demand
[kWh/year] Electricity tier Description
Tertiary
referral
hospital
91,250 tier 5 –
full access
Coverage for >120 beds. High-energy
requirements. May contain sophisticated
diagnostic devices requiring additional
power and perform surgical procedures.
Secondary
hospital
14,600 tier 4–
advanced
access
Coverage for 60–120 beds. Moderate
energy requirements. May accommodate
sophisticated diagnostic medical equipment.
First hospital 7,300 tier 3 –
intermediate
access
Coverage for 30–60 beds. Low/moderate
energy requirements.
Primary
health post
1,825 tier 2 –
basic access
No beds other than for emergencies/
maternity care. Typically located in a
remote setting with limited services and a
small staff. Typically operates weekdays.
Low energy requirements.
Source: Moner-Girona et al.,
57
Maina et al.,
63
USAID,
75
and AFREA
76
ll
OPEN ACCESS
2708 Joule 5, 2687–2714, October 20, 2021
Article

battery storage sizes are optimized for each location to ensure the least-cost solution
with a power outage on less than 5% of days in a year. This criterion reflects the fact
that the grid reliability in many countries is low, and over one-third of households
have a connection that works half or less of the time.
77
In view of this, the 5% daily
outage frequency should provide a quality of service, at least as good as that of
the national grid, while keeping the system cost at a competitive level. The PV per-
formance is calculated from location-specific hourly solar radiation values derived
from satellite data, supplemented with surface temperature and wind speed data
from climate reanalysis.
78
The daily electricity demand load profile is as defined in
the study conducted by Huld et al.,
36
scaled to match the postulated annual energy
demand of a given facility (see Figure 14). The PV module and battery performance
algorithms incorporate measured data on PV module and battery performance using
Li-ion batteries.
36
The LCOE was computed for each health facility according to Equation 2,consid-
ering solar resource,
78
the electricity demand for the relevant health facility type (Ta-
ble 2;Figure 14), together with initial investment costs, CAPEX, replacements, and
operation, and maintenance for an operational lifetime of 20 years for the PV array
and 10 years for the battery.
57
The system capital costs are based on field specific
data presented in the study conducted by Moner-Girona et al.
40
The CAPEX
0
values
incorporate the economy of scale effects, with a module price equal to 0.83 EUR/W
p
and higher prices for systems smaller than 1 kW
p
(0.99 EUR/W
p
)andlowerpricesfor
systems larger than 100 kW
p
(0.63 EUR/W
p
). The Li-ion battery prices were 350 EUR/
kWhnominalwhenthebatterysizewassmallerthan50kWhand280EUR/kWhwhen
it was larger than 50 kWh. In the case of social energy infrastructure and based on
cost-benefit analysis, international institutions
79,80
recommend the use of low social
discount rates between 3% and 5%, and the latter value is used for all locations.
Table 3. Summary of cost factor groups for off-grid PV systems
# Factor group Components
1 PV array PV modules
PV mounting structure
2 Balance of system (BOS) PV cabling
PV earthing
charge controller
DC protections board
inverter
AC protections
AC cabling
3 Storage and monitoring battery bank and rack
DC battery protections
DC battery cabling
control and battery room
monitoring board and Software
4 Engineering and soft costs installation,
civil works and miscellaneous materials
system design and project management
training
permitting fees, taxes, and financing
transport
Other equipment (for operation
and maintenance)
spare parts and storage
Source: adapted from Moner-Girona et al.
40
ll
OPEN ACCESS
Joule 5, 2687–2714, October 20, 2021 2709
Article

LCOEn=
CAPEX0+PT
t=1ðRt+OtÞð1+rnÞt
PT
t=1ðESnÞð1+rnÞt(Equation 2)
Where:
n: identified health facility
LCOEn: levelized cost of electricity in facility n[EUR/kWh]
CAPEX0: initial PV system investment cost at t=0[EUR]
t:timeinyearst= 0 is the installation year
T: economic lifetime of the PV system (years)
O
t
: operation and maintenance cost (2%) in year t (EUR)
Rt: replacement cost in year t (EUR)
E
Sn
: average annual electricity production from the given system depending on solar
radiation and electricity demand in health center n(kWh)
r
n
:discountrate
Import taxes for photovoltaic modules are non-existing or low in most African coun-
tries.
81,82
This is due to the fact that strategies to support the deployment of renewable
energies in Africa are increasing (IRENA, 2020).
1
Moreover, as the realization of the elec-
trification of health infrastructure is part of a national strategy in collaboration with inter-
national organizations, the analysis expects measures to reduce investment risk, in the
form of grants (not debt), exemption from VAT (as many countries already have), and
exemption from import duties. This is due to the fact that policies tend to favor projects
with positive externalities for their overall societal benefits. Figure 15 shows the effect of
the variation in VAT and import duties on the LCOE for 16 countries. The LCOE
computed for all health facilities, taking into account the particular VAT and import
Figure 14. Daily electricity demand profile for healthcare facilities (high daytime consumption)
given as hourly % values of the total daily power consumption
ll
OPEN ACCESS
2710 Joule 5, 2687–2714, October 20, 2021
Article

duties per country, shows a higher degree of dispersion (spread), higher skewness in the
lower values (blue box), and also higher variability outside the upper and lower quartiles
(whiskers). Burundi (high import duty on batteries) and Mozambique (import duties on
PV modules and batteries) show the largest differences, with an increase in the LCOE
of 0.1 EUR/kWh. 70% of the studied countries show an increase in the LCOE of less
than 0.07 EUR/kWh, and Mali and Zambia show no significant difference (due to exemp-
tion of VAT and import duties).
Adoption of diesel/petrol generators
The current methodology considers 100% renewable energy systems (due to the
mentioned sustainability focus, specifically to reach the SDG7, affordable, reliable, sus-
tainable, and modern energy for all). Despite the fact that the model focuses on sustain-
able rural electrification, we used location-specific estimates of the relative electricity
costs (USD/kWh) of solar photovoltaic versus diesel. The methodology used for this com-
parison is explained step by step in previous publications,
23,24,83
where the generation
cost of electricity that relies on solar photovoltaics is compared with that of diesel. Fig-
ure 16 shows the results of the analysis when comparing the differences between the
diesel and PV production costs: the minimum, first quartile, median, third quartile,
and maximum of all NEA health centers per country. It is observed that for almost all
countries, the long-term costs (accounted for 20 years) of diesel are higher than the
PV costs, with an average range between 5 and 50 USD cents.
Estimating avoided emissions
The estimation of the carbon mitigation potential of using fully renewable mini-grids
is based on the avoided GHG emissions in CO2
eq
.
40
The avoided GHG emissions
Figure 15. Comparison of levelized cost of electricity computed for health facilities taking into
account the particular VAT and import duties per country (in blue) and LCOE calculated without
import duties and with a homogeneous 15% VAT (in green).
The LCOE values with in-country specific values (blue box) show a higher degree of dispersion
(spread box) and higher skewness in the lower values and also higher variability outside the upper
and lower quartiles (whiskers).
ll
OPEN ACCESS
Joule 5, 2687–2714, October 20, 2021 2711
Article

were calculated by computing the emissions of a standalone diesel generator sup-
plying the same electricity demand over the lifetime of the PV plus battery storage
systems. The annual GHG emissions were calculated for 20 years lifetime multiplying
the computed emission factor of 1.7 tCO2/MWh for the electricity demand per each
health center.
ACKNOWLEDGMENTS
The authors would like to thank DG INTPA F.1 – the European Commission for
providing valuable insights and supporting the development of the interactive
tool for this study. We would like to acknowledge Pere Roca Ristol (Joint Research
Centre – the European Commission) for the development of the PV-DEI web tool
and Fernando Fahl and Marco Pittalis for their valuable insights in the development
of the tool. The authors would like to thank Luc Severi (SE4ALL) for providing valu-
able insights to this study and the Power Africa team at the U.S. Agency for Interna-
tional Development. Disclaimer: The views expressed are purely those of the authors
and may not, under any circumstances, be regarded as stating an official position of
the European Commission.
AUTHOR CONTRIBUTIONS
Conceptualization, M.M.-G.; methodology, M.M.-G. and G.K.; data curation, G.K.
and D.J.W.; investigation, M.M.-G., G.K., G.F., D.J.W., and N.T.; visualization,
G.K. and G.F.; writing – review & editing, M.M.-G., G.K., G.F., D.J.W., and N.T.
DECLARATION OF INTERESTS
The authors declare no competing interests.
Received: May 1, 2021
Revised: August 3, 2021
Accepted: September 23, 2021
Published: October 20, 2021
Figure 16. Location-specific estimates of the relative electricity production costs (¢USD/kWh) of solar photovoltaic versus diesel per country
Minimum, first quartile, median, third quartile, and maximum of all NEA health centers per country.
ll
OPEN ACCESS
2712 Joule 5, 2687–2714, October 20, 2021
Article

REFERENCES
1. Ritchie, R., and Mispy, O.-O. (2018). SDG
tracker. https://sdg-tracker.org/energy.
2. WHO. (2020). Call for expression of interest:
development of the ‘global assessment of
electricity in healthcare facilities’ report.
https://www.who.int/docs/default-source/air-
pollution-documents/eoi-report-electricity-in-
healthcare-facilities.pdf?sfvrsn=d7e5a965_2.
3. Puig, D., Moner-Girona, M., Kammen, D.M.,
Mulugetta, Y., Marzouk, A., Jarrett, M., Hailu,
Y., and Naki
cenovi
c, N. (2021). An action
agenda for Africa’s electricity sector. Science
373, 616–619.
4. World Bank (2019). Sustainable energy for All
(SE4ALL ) database from the SE4ALL global
tracking framework. https://data.worldbank.
org/.
5. International Energy Agency (2019). Africa
Energy Outlook (IEA). https://www.iea.org/
reports/africa-energy-outlook-2019.
6. IEA, IRENA, UNSD, World Bank, and WHO.
(2020). Tracking SDG 7: The Energy Progress
Report (World Bank Publications).
7. Adair-Rohani, H., Zukor, K., Bonjour, S.,
Wilburn, S., Kuesel, A.C., Hebert, R., and
Fletcher, E.R. (2013). Limited electricity access
in health facilities of sub-Saharan Africa: A
systematic review of data on electricity access,
sources, and reliability. Glob. Health Sci. Pract.
1, 249–261.
8. Mikul, B., Nicolina, A., Ruchi, S., Elisa, P., Elaine,
R., F, Susan, W., Heather, A., and Carlos, D.
(2018). Access to Modern Energy Services for
Health Facilities in Resource-Constrained
Settings (World Health Organization).
9. World Health Organization (WHO), and World
Bank. (2014). Access to Modern Energy
Services for Health Facilities in Resource-
Constrained Settings: A Review of Status,
Significance, Challenges and Measurement
Reprinted in 2015 with Changes (World Health
Organization).
10. Ouedraogo, N.S., and Schimanski, C. (2018).
Energy poverty in healthcare facilities: a ‘‘silent
barrier’’ to improved healthcare in sub-Saharan
Africa. J. Public Health Policy 39, 358–371.
11. Pittalis, M., Kyriakarakos, G.G., and Moner-
Girona, M. (2021). Driving clean energy access
for social infrastructure: existing business
models and emerging trends in sub-Saharan
Africa. In Encyclopedia of Sustainability
Science and Technology, R.A. Meyers, ed.
(Springer).
12. The World Bank Group (2019). World Bank
indicators. https://data.worldbank.org/.
13. UNESCO (2019). UNESCO database. http://
data.uis.unesco.org/.
14. World Bank (2020). Multi-Tier framework
database. https://mtfenergyaccess.esmap.
org/.
15. Sustainable energy for All. (2020). Powering
health care. https://www.seforall.org/
powering-healthcare.
16. Cronk, R., and Bartram, J. (2018).
Environmental conditions in health care
facilities in low- and middle-income countries:
coverage and inequalities. Int. J. Hyg. Environ.
Health 221, 409–422.
17. IRENA (2020). The Post-COVID Recovery: an
Agenda for Resilience (Development and
Equality (IRENA)).
18. Kumar, A. (2021). Decentralize the energy
system so COVID does not slow clean energy.
Nature 589, 198.
19. ReportLinker. (2020). United States solar
energy market - growth, trends, and forecasts.
(2020 - 2025). https://www.reportlinker.com/
p05954264/United-States-Solar-Energy-
Market-Growth-Trends-and-Forecasts.html?
utm_source=GNW.
20. Cohen, A. (2020). Solar power prices sink in the
age of covid. Forbes. https://www.forbes.com/
sites/arielcohen/2020/09/08/solar-power-
prices-sink-in-the-age-of-covid/?
sh=2a79ca9c28e4.
21. Ziegler, M.S., and Trancik, J.E. (2021). Re-
examining rates of lithium-ion battery
technology improvement and cost decline.
Energy Environ. Sci. 14, 1635–1651.
22. Ja
¨ger-Waldau, A. (2020). Snapshot of
photovoltaics–February 2020. Energies 13, 930.
23. Szabo
´, S., Pinedo Pascua, I., Puig, D., Moner-
Girona, M., Negre, M., Huld, T., Mulugetta, Y.,
Kougias, I., Szabo
´, L., and Kammen, D. (2021).
Mapping of affordability levels for
photovoltaic-based electricity generation in
the solar belt of sub-Saharan Africa, East Asia
and South Asia. Sci. Rep. 11, 3226.
24. Szabo
´, S., Bo
´dis, K., Huld, T., and Moner-
Girona, M. (2011). Energy solutions in rural
Africa: mapping electrification costs of
distributed solar and diesel generation versus
grid extension. Environ. Res. Lett. 6, 034002.
25. Oseni, M.O. (2017). Self-generation and
households’ willingness to pay for reliable
electricity service in Nigeria. Energy J. 38,
165–194.
26. Porcaro, J., Mehta, S., Shupler, M., Kissel, S.,
Pfeiffer, M., Dora, C.F.C., and Adair-Rohani, H.
(2017). Modern Energy Access and Health,
State of Electricity Access Report (The World
Bank). http://hdl.handle.net/10986/26648.
27. Dholakia, H.H. (2018). Solar powered
healthcare in developing countries. Nat.
Energy 3, 705–707.
28. Olatomiwa, L., Blanchard, R., Mekhilef, S., and
Akinyele, D. (2018). Hybrid renewable energy
supply for rural healthcare facilities: an
approach to quality healthcare delivery.
Sustain. Energy Technol. Assess. 30, 121–138.
29. Franco, A., Shaker, M., Kalubi, D., and
Hostettler, S. (2017). A review of sustainable
energy access and technologies for healthcare
facilities in the Global South. Sustain. Energy
Technol. Assess. 22, 92–105.
30. Orosz, M.S., Quoilin, S., and Hemond, H.
(2013). Technologies for heating, cooling and
powering rural health facilities in sub-Saharan
Africa. Proc. Inst. Mech. Eng. A: J. Power
Energy 227, 717–726.
31. World Resources Institute (WRI) (2021). Energy
access explorer. https://www.
energyaccessexplorer.org/.
32. HOMER (2020). Homer powering health tool.
(Health Communications). http://tools.
poweringhealth.org/.
33. Falchetta, G., Hammad, A.T., and Shayegh, S.
(2020). Planning universal accessibility to public
health care in sub-Saharan Africa. Proc. Natl.
Acad. Sci. USA 117, 31760–31769.
34. Weiss, D.J., Nelson, A., Vargas-Ruiz, C.A.,
Gligori
c, K., Bavadekar, S., Gabrilovich, E.,
Bertozzi-Villa, A., Rozier, J., Gibson, H.S.,
Shekel, T., et al. (2020). Global maps of travel
time to healthcare facilities. Nat. Med. 26,
1835–1838.
35. Golub, G., Herman-Roloff, A., Hoffman, S.,
Jaoko, W., and Bailey, R.C. (2016). The
relationship Between distance and post-
operative visit attendance following medical
male circumcision in nyanza Province, Kenya.
AIDS Behav. 20, 2529–2537.
36. Huld, T., Moner-Girona, M., and Kriston, A.
(2017). Geospatial analysis of photovoltaic
mini-grid system performance. Energies 10,
218.
37. Factor and Clean Energy Solutions Center.
(2016). PV/battery waste management in the
context of rural electrification. https://www.
wearefactor.com/docs/PV-
Battery_Waste_Management.pdf.
38. Manhart, A., Hilbert, I., and Magalini, F. (2018).
End-of-Life Management of Batteries in the
Off-Grid Solar Sector, D. Hinchliffe and E.
Gunsilius, eds. (Deutsche Gesellschaft fu
¨r
Internationale Zusammenarbeit (GIZ) GmbH).
39. IRENA (2020). The Renewable Energy
Transition in Africa (KfW Development Bank,
Deutsche Gesellschaft fu
¨r Internationale
Zusammenarbeit (GIZ), IRENA). https://www.
irena.org/-/media/Files/IRENA/Agency/
Publication/2021/March/Renewable_
Energy_Transition_Africa_2021.pdf.
40. Moner-Girona, M., Solano-Peralta, M.,
Lazopoulou, M., Ackom, E.K., Vallve, X., and
Szabo
´, S. (2018). Electrification of Sub-Saharan
Africa through PV/hybrid mini-grids: reducing
the gap between current business models and
on-site experience. Renew. Sustain. Energy
Rev. 91, 1148–1161.
41. Power Africa. (2020). (USAID)/power Africa
announces $2.6 million in healthcare
electrification grants to solar energy
companies in nine countries in sub-Saharan
Africa. https://powerafrica.medium.com/
usaid-power-africa-announces-2-6-
c003fa75f004.
42. IRENA. (2020). African Union and IRENA to
advance renewables in response to COVID-19.
https://www.irena.org/newsroom/
pressreleases/2020/Apr/African-Union-and-
IRENA-to-Advance-Renewables-in-Response-
to-Covid19.
43. SDSN, and TReNDS. (2020). Leaving No One
off the Map: A Guide for Gridded Population
Data for Sustainable Development. https://
ll
OPEN ACCESS
Joule 5, 2687–2714, October 20, 2021 2713
Article

www.sdsntrends.org/leaving-no-one-off-the-
map.
44. Carr-Hill, R. (2013). Missing millions and
measuring development progress. World Dev.
46, 30–44.
45. Carr-Hill, R. (2014). Measuring development
progress in Africa: the denominator problem.
Can. J. Dev. Stud. 35, 136–154.
46. Ifeagwu, S.C., Yang, J.C., Parkes-Ratanshi, R.,
and Brayne, C. (2021). Health financing for
universal health coverage in Sub-Saharan
Africa: a systematic review. Glob. Health Res.
Policy 6,8.
47. Moner-Girona, M., Bender, A., Becker, W.,
Bo
´dis, K., Szabo
´, S., Kararach, A.G., and
Anadon, L.D. (2021). A multidimensional high-
resolution assessment approach to boost
decentralised energy investments in Sub-
Saharan Africa. Renew. Sustain. Energy Rev.
148, 111282.
48. Bender, A., Moner-Girona, M., Becker, W.,
Bo
´dis, K., Szabo
´, S., Kararach, A.G., and
Anadon, L.D. (2021). Dataset for
multidimensional assessment to incentivise
decentralised energy investments in Sub-
Saharan Africa. Data Brief 37, 107265.
49. Ouma, P.O., Maina, J., Thuranira, P.N.,
Macharia, P.M., Alegana, V.A., English, M.,
Okiro, E.A., and Snow, R.W. (2018). Access to
emergency hospital care provided by the
public sector in sub-Saharan Africa in 2015: a
geocoded inventory and spatial analysis.
Lancet Glob. Health 6, e342–e350.
50. NASA Earth Observatory (2019). Earth at night.
https://earthobservatory.nasa.gov/.
51. Falchetta, G., Pachauri, S., Parkinson, S., and
Byers, E. (2019). A high-resolution gridded
dataset to assess electrification in sub-Saharan
Africa. Sci. Data 6, 110.
52. Doll, C.N.H., and Pachauri, S. (2010). Estimating
rural populations without access to electricity in
developing countries through night-time light
satellite imagery. Energy Policy 38, 5661–5670.
53. Dugoua, E., Kennedy, R., and Urpelainen, J.
(2018). Satellite data for the social sciences:
measuring rural electrification with night-time
lights. Int. J. Remote Sens. 39, 2690–2701.
54. Arderne, C., Zorn, C., Nicolas, C., and Koks,
E.E. (2020). Predictive mapping of the global
power system using open data. Sci. Data 7,19.
55. ECREEE (2019). ECOWREX: ECOWAS
observatory for renewable energy and energy
efficiency. http://www.ecowrex.org/.
56. Moner-Girona, M., Bo
´dis, K., Huld, T., Kougias,
I., and Szabo
´, S. (2016). Universal access to
electricity in Burkina Faso: scaling-up
renewable energy technologies. Environ. Res.
Lett. 11, 084010.
57. Moner-Girona, M., Bo
´dis, K., Morrissey, J.,
Kougias, I., Hankins, M., Huld, T., and Szabo
´,S.
(2019). Decentralized rural electrification in
Kenya: speeding up universal energy access.
Energy Sustain. Dev. 52, 128–146.
58. Aly, A., Moner-Girona, M., Szabo
´, S., Pedersen,
A.B., and Jensen, S.S. (2019). Barriers to large-
scale solar power in Tanzania. Energy Sustain.
Dev. 48, 43–58.
59. Moner-Girona, M., Ghanadan, R., Solano-
Peralta, M., Kougias, I., Bo
´dis, K., Huld, T., and
Szabo
´, S. (2016). Adaptation of Feed-in Tariff
for remote mini-grids: Tanzania as an
illustrative case. Renew. Sustain. Energy Rev.
53, 306–318.
60. US NOAA National Oceanic and Atmospheric
Administration, and National Geophysical
Data Center and US Air Force Weather Agency.
(2014). Version 4 DMSP-OLS nighttime lights
time series. https://ngdc.noaa.gov/eog/dmsp/
downloadV4composites.html.
61. Andrade-Pacheco, R., Savory, D.J., Midekisa,
A., Gething, P.W., Sturrock, H.J.W., and
Bennett, A. (2019). Household electricity access
in Africa (2000–2013): closing information gaps
with model-based geostatistics. PLoS One 14,
e0214635.
62. Falchetta, G., Pachauri, S., Parkinson, S., and
Byers, E. (2019). A High-Resolution Gridded
Dataset to Assess Electrification in Sub-
Saharan Africa.. https://github.com/giacfalk/
Electrification_SSA_data.
63. Maina, J., Ouma, P.O., Macharia, P.M.,
Alegana, V.A., Mitto, B., Fall, I.S., Noor, A.M.,
Snow, R.W., and Okiro, E.A. (2019). A spatial
database of health facilities managed by the
public health sector in sub Saharan Africa. Sci.
Data 6, 134.
64. Pesaresi, M., Ehrlich, D., Ferri, S., Florczyk, A.,
Freire, S., Haag, F., Halkia, M., Julea, A.M.,
Kemper, T., and Soille, P. (2015). Global human
settlement analysis for disaster risk reduction.
In 36th International Symposium on Remote
Sensing of Environment (International Society
for Photogrammetry and Remote Sensing
(ISPRS)), pp. 837–843.
65. Pesaresi, M., Ehrlich, D., Florczyk, A.J., Freire,
S., Julea, A., Kemper, T., and Syrris, V. (2016).
The global human settlement layer from
landsat imagery. Int. Geosci. Remote Sens.
Symp. IGARSS.
66. Paresi, M., Melchiorri, M., Siragusa, A., and
Kemper, T. (2017). Atlas of the Human Planet.
Mapping Human Presence on Earth with the
Global Human Settlement Layer, EUR 28116
(Publications Office of the European Union).
67. European Commission and Joint Research
Centre (JRC) (2019). Global human settlement
layer. GHS-POP R2019A. https://ghsl.jrc.ec.
europa.eu/ghs_pop2019.php.
68. Weiss, D.J., Nelson, A., Gibson, H.S.,
Temperley, W., Peedell, S., Lieber, A., Hancher,
M., Poyart, E., Belchior, S., Fullman, N., et al.
(2018). A global map of travel time to cities to
assess inequalities in accessibility in 2015.
Nature 553, 333–336.
69. The malaria atlas project. (2018). Accessibility
to cities: Global Malaria Epidemiology. https://
malariaatlas.org/research-project/
accessibility-to-cities/.
70. Wigley, A.S., Tejedor-Garavito, N., Alegana,
V., Carioli, A., Ruktanonchai, C.W., Pezzulo, C.,
Matthews, Z., Tatem, A.J., and Nilsen, K.
(2020). Measuring the availability and
geographical accessibility of maternal health
services across sub-Saharan Africa. BMC Med.
18, 237.
71. Hulland, E. (2020). COVID-19 and health care
inaccessibility in sub-Saharan Africa. Lancet
Healthy Longev. 1, e4–e5.
72. Battle, K.E., Bisanzio, D., Gibson, H.S., Bhatt,
S., Cameron, E., Weiss, D.J., Mappin, B.,
Dalrymple, U., Howes, R.E., Hay, S.I., and
Gething, P.W. (2016). Treatment-seeking
rates in malaria endemic countries. Malar. J.
15,20.
73. Bhatia, M., and Angelou, N. (2015). Beyond
Connections: Energy Access Redefined,
Energy Sector Management Assistance
Program (ESMAP) (World Bank).
74. Practical Action. (2013). Poor People’s Energy
Outlook 2013: Energy for Community Services
(Practical Action Publishing).
75. USAID (2006). Powering health, electrification
options for rural health centers. https://pdf.
usaid.gov/pdf_docs/PNADJ557.pdf.
76. Africa Renewable Energy Access Program
(AFREA) (2010). Photovoltaics for Community
Service Facilities (World Bank).
77. Afrobarometre. (2019). Progress toward
‘reliable energy for all’ stalls across Africa.
https://afrobarometer.org/press/progress-
toward-reliable-energy-all-stalls-across-africa-
afrobarometer-survey-finds.
78. European Commission- Joint Research Centre
(JRC) (2019). Photovoltaic geographical
information system (PVGIS). https://ec.europa.
eu/jrc/en/pvgis.
79. Joint Regulators Group (JRG) (2011). In
Discounting for CBAs Involving Private
Investment, but Public Benefit, B. Basalisco, ed.
(Office of Gas and Electricity Markets
(OFGEM)).
80. ENTSO-E. (2020). Guideline for Cost Benefit
Analysis of Grid Development Projects
(European Network of Transmission System
Operators (ENTSO)). https://
eepublicdownloads.entsoe.eu/clean-
documents/tyndp-documents/Cost%
20Benefit%20Analysis/191023_CBA3_Draft%
20for%20consultation.pdf.
81. Renewable energy association East Africa
(NREAs), USEA, KEREA, and UNREEEA. (2020).
The East African Regional Handbook on Solar
Taxation. https://sun-connect-news.org/
fileadmin/DATEIEN/Dateien/New/EAC-
Regional-Handbook-on-Solar-Taxation-Final-
Copy-2-1.pdf.
82. GET invest Mobilising investment in renewable
energy. (2021). Country briefs insights and data
to help you make informed decisions. https://
www.get-invest.eu/market-information/
country-briefs/.
83. Szabo
´, S., Bo
´dis, K., Huld, T., and Moner-
Girona, M. (2013). Sustainable energy planning:
leapfrogging the energy poverty gap in Africa.
Renew. Sustain. Energy Rev. 28, 500–509.
ll
OPEN ACCESS
2714 Joule 5, 2687–2714, October 20, 2021
Article