lth
y a
i b,
ili
t 155, 2000 Antwerp, Belgium
s, Rue E
d Institute of Statistical, Social and Economic Research,
e Noguchi Memorial Institute for Medical Research, Uni
offices for ID cards in remote communities in Senegal.
er the CC BY license
rg/licenses/by/3.0/).
added to the global population between now and 2050 (UN, 2009).
Despite the demographic transition being more advanced in
developed countries, between 1950 and 2000, 66% of the global
increase in people over 60 occurred in low- and middle-income
countries (LMIC); by 2050 it is projected that 80% of all older
derin, 2012; Beard
demographic shift
systems and many
ing adequate, age-
of living for older
ageing has so far
received little attention from both policy makers and researchers.
However, in spite of the low relative share of older people in the
total SSA populations (below 10%); the subcontinent still hosts a
significant aged population, which is expected to grow at a steady
pace. With life expectancy of 16 years for 60 year olds, getting old is
no longer an exception in Africa.
Ageing in Africa raises particular concerns because of its strong
association with increased vulnerability. Several risk factors are
associated with this heightened vulnerability (Crooks, 2009;* Corresponding author.
Contents lists availab
Social Science
se
Social Science & Medicine 119 (2014) 36e44E-mail address: divya.parmar@outlook.com (D. Parmar).© 2014 The Authors. Published by Elsevier Ltd. This is an open access article und
(http://creativecommons.o
1. Introduction
In their ‘manifesto for the world we want’, The Lancet (2012)
identified globally ageing population as a critical issue that must
be addressed to help create sustainable improvements in health. By
2016 it is estimated that there will be more people older than 65
years than children under five, and 1.5 billion people over 60will be
people (i.e. 1.6 billion) will be living in LMIC (Abo
et al., 2011, pp. 4). This unprecedented and rapid
will have far-reaching consequences for health
LMIC already face immense challenges in provid
appropriate healthcare and a decent standard
people.
In Sub-Saharan Africa (SSA) the issue offeatures such as eliminating the registration fee for older people in NHIS and creating administrationa r t i c l e i n f o
Article history:
Received 29 September 2013
Received in revised form
26 June 2014
Accepted 13 August 2014
Available online 13 August 2014
Keywords:
Social health protection
Elders
Ghana
Senegal
Social exclusion
Enrolment
Universal health coverage
Ageinghttp://dx.doi.org/10.1016/j.socscimed.2014.08.011
0277-9536/© 2014 The Authors. Published by ElsevieX Leon Gontran Damas, Fann Residance, Dakar BP25 233, Senegal
University of Ghana, P.O. Box LG74 Legon, Accra, Ghana
versity of Ghana, P. O. Box LG 581 Legon, Accra, Ghana
a b s t r a c t
Although the population of older people in Africa is increasing, and older people are becoming
increasingly vulnerable due to urbanisation, breakdown of family structures and rising healthcare costs,
most African countries have no social health protection for older people. Two exceptions include Sen-
egal's Plan Sesame, a user fees exemption for older people and Ghana's National Health Insurance
Scheme (NHIS) where older people are exempt from paying premiums. Evidence on whether older
people are aware of and enrolling in these schemes is however lacking. We aim to fill this gap. Besides
exploring economic indicators, we also investigate whether social exclusion determines enrolment of
older people. This is the first study that tries to explore the social, political, economic and cultural (SPEC)
dimensions of social exclusion in the context of social health protection programs for older people. Data
were collected by two cross-sectional household surveys conducted in Ghana and Senegal in 2012. We
develop SPEC indices and conduct logistic regressions to study the determinants of enrolment. Our re-
sults indicate that older people vulnerable to social exclusion in all SPEC dimensions are less likely to
enrol in Plan Sesame and those that are vulnerable in the political dimension are less likely to enrol in
NHIS. Efforts should be taken to specifically enrol older people in rural areas, ethnic minorities, women
and those isolated due to a lack of social support. Consideration should also be paid to modify schemeLSE Health, London School of Economics and
b Institute of Tropical Medicine, Nationalestraa
c Centre de Recherche sur les Politiques Sociale 
a Political Science, Houghton Street, London WC2A 2AE, UKEnrolment of older people in social hea
West Africa e Does social exclusion pla
Divya Parmar a, *, Gemma Williams a, Fahdi Dkhim
Felix Ankomah Asante d, Daniel Kojo Arhinful e, Ph
journal homepage: www.elr Ltd. This is an open access articleprotection programs in
part?
Alfred Ndiaye c,
pa Mladovsky a
le at ScienceDirect
& Medicine
vier .com/locate/socscimedunder the CC BY license (http://creativecommons.org/licenses/by/3.0/).
ce &Issahaku and Neysmith, 2013). First, older people in SSA usually
retire in rural areas, characterized by poor infrastructures and acute
problems of basic service provision. Second, many scholars point
out the feminisation of the SSA aged populatione ‘a female society’
according to Apt (2009). This makes Africa's older women twice as
vulnerable, first due to the biological process of ageing, and, second,
due to gender-related discrimination. Third, majority of older
people are illiterate (67% in Africa (UN, 2009)), which is associated
with poor access to public resources. Furthermore, most of the
Africa's older people, especially women, have no formal employ-
ment records and thus no access to formal social security ar-
rangements like pensions. It is estimated that only 17% of older
people in SSA receive an old-age pension (International Labour
Organization, 2014).
Historically the extended family structure in Africa has miti-
gated the effect of these combined risk factors. However, evidence
suggests the situation is changing. Traditional respect and caring
structures are facing substantial social challenges, hence refuting
the widespread African myth of the “inexhaustible capacity of the
extended family to withstand crisis” (Gysels et al., 2011).
Access to appropriate healthcare remains a major concern for
the majority of the ageing population in SSA. Facts speak for
themselves: not only do older people spend more per-capita on
healthcare than others in LMIC, consequently bearing a heavy
burden linked to user fees policies, they also face higher levels of
unmet need for healthcare, with a greater proportion of older
people reporting forgone treatments for illness than younger
groups (McIntyre, 2004; Saeed et al., 2012). The gap between needs
and access is expected to grow further in the short term, especially
due to the escalating epidemic of non-communicable diseases
(NCDs) among the ageing population (Alam et al., 2010; George-
Carey et al., 2012; Holmes and Joseph, 2011).
Recognising the increased vulnerability of older people in rela-
tion to illness and healthcare expenditures, two West African
countries, Ghana and Senegal, have implemented Social Health
Protection (SHP) programs that specifically target older people.
These programs aim to reduce the financial barriers faced by older
people in accessing healthcare services.
1.1. The Ghanaian National Health Insurance Scheme (NHIS)
The Ghanaian NHIS, launched in 2003, is a national health
financing systemwith decentralised operations. Each district has its
own insurance fund, which is financed by central-level funds and
premiums. All formal sector employees and their dependents are
automatically enrolled, and their premiums are collected at the
central level via payroll deductions. Self-employed individuals and
informal sector workers need to enrol directly in NHIS. NHIS is
largely funded by value-added tax (61% of the total NHIS revenue in
2009). Investment income (17%), and the Social Security and Na-
tional Insurance Trust pensioners (SSNIT is a national pension
scheme which is mandatory for formal sector employees but
voluntary for self-employed) premiums (16%) constitute a small
proportion of the NHIS budget. NHIS covers almost 95% of the
disease burden in Ghana. Services include outpatient, inpatient and
emergency care, deliveries, dental care, and essential drugs. In-
dividuals need to register with the NHIS once in their lifetime and
then renew their membership annually. Renewal can be done at the
District Mutual Health Insurance Scheme (DHMIS) office or by an
agent of the scheme. Children under 18, pregnant women, in-
digents (i.e. the poor and destitute), and all people over 70 are
exempted from paying premiums, although they still need to reg-
ister and renew their membership annually. Exempt groups
including people over 70 pay a small registration fee at the time of
D. Parmar et al. / Social Scienrenewal. In 2011, 8.2 million people (33% of the population) wereactive members (registered and had renewed their membership
that year) of NHIS and 4.9% of the active members were over 70s
(National Health Insurance Authority, 2011).
1.2. Senegal's Plan Sesame
Unlike NHIS, Plan Sesame directly and exclusively targets older
people. Launched in 2006 during the presidential address to the
nation, Plan Sesame aims to provide free access to public healthcare
services to all citizens over 60 e an estimated 5.9% of the total
population (ANSD, 2012). Costs of consultations, diagnostics,
essential drugs, and hospitalisations are covered by the scheme.
Although there is no specific registration process, older people who
want to benefit from this exemption are required to present a na-
tional ID card at the point of service. The national ID card is
mandatory for all citizens aged over 15 years. It can be obtained in
person at police stations for a fee of approximately $2 and by
presenting a birth certificate or an old ID card (the card is valid for
ten years). According to our study 89% of older people have the new
national ID card. Plan Sesame is largely funded by taxation. Some
funds are also received from the Institut de Pr
evoyance Retraite du
S
en
egal (IPRES) and the Fonds National de Retraite (FNR), the na-
tional contingency/pension fund for formal employees in the pri-
vate sector. Plan Sesame has suffered from under funding and short
staffing, and no communication plan is implemented to promote
the scheme (Mbaye et al., 2013; Leye et al., 2013).
Although the demand for SHP has recently gained momentum
in LMIC, evidence onwhether SHP schemes have been successful in
providing equitable healthcare to older people where they have
access to healthcare on the basis of need, irrespective of their in-
come, age, residency, or sociocultural factors, is limited. In Ghana,
NHIS still struggles to overcome inequities in enrolment (Jehu-
Appiah et al., 2011; Sarpong et al., 2010) and evidence on
whether NHIS has benefitted older people is rare (Lloyd-Sherlock,
2000). Enrolment in itself does not guarantee access to health
services, there are likely to be barriers to accessing healthcare even
for people who are insured (Biritwum et al., 2013). However, these
schemes have been designed in such a way that enrolment is a
prerequisite for accessing free care at the point of use for the
populations covered, and as such enrolment constitutes an
important first step (and potential barrier) to accessing care. It is
therefore crucial to study this step to identify individuals at risk of
inequities in enrolment.
This study tests the hypothesis that socially excluded older
people are less likely to enrol in NHIS and Plan Sesame. This hy-
pothesis is supported by wider literature, which points to social
exclusion in the healthcare sector (Marmot et al., 2008) and in the
social sector more broadly (Popay et al., 2008). To the best of our
knowledge, this is the first study that investigates the role of social
exclusion in SHP uptake among the ageing population in West
Africa.
2. Methodology
2.1. The SPEC (social, political, economic and cultural) framework
We used the SEKN (Social Exclusion Knowledge Network)
definition of social exclusion e ‘dynamic, multidimensional pro-
cesses driven by unequal power relationships interacting across
four main dimensions e Social, Political, Economic and Cultural’:
SPEC (Popay et al., 2008). This definitionwas operationalized in two
steps. First, we conducted an extensive literature review to develop
a SPEC framework based on the SEKN framework of social exclu-
sion, and identified domains related to resources and participation
Medicine 119 (2014) 36e44 37that are important for understanding social exclusion. Resources
are means such as wealth, assets, education or values that can be
used to meet human needs, and participation refers to the power
and ability people have to utilise the resources available to them
(Popay et al., 2008). Within these domains we identified measur-
able indicators or variables of social exclusion, which can be
considered as risk factors of social exclusion. Having a risk factor
does not indicate that the person is socially excluded; rather it in-
dicates that the person is vulnerable to social exclusion. Second,
local expert teams in Ghana and Senegal reviewed the SPEC
framework and identified domains and variables they thought to be
relevant and important in their contexts. Although all four SPEC
dimensions are interdependent and interact to affect social exclu-
sion, the social and cultural dimensions in particular were felt to be
so intricately interlinked that they were combined into one
dimension e sociocultural. The domains and variables, and hy-
pothesis on how they are linked to social exclusion and access to
social protection programs, are described in the Supplementary file
1.
Domains and variables (as shown in Table 1) identified by local
teams, common to both the settings and relevant for understanding
social exclusion among older people were used to construct an
index for each dimension of social exclusion. All dimensions were
given equal weightage (i.e. 100%). Each domain, within the di-
mensions, was also given equal weightage and they add up to 100%.
Therefore, each index has amax value of 100% and higher the index,
greater the vulnerability of the older persons to be social excluded
in that dimension.
e Central, Eastern, Ashanti, Brong-Ahafo and Northern. From each
D. Parmar et al. / Social Science &38Table 1
SPEC dimensions and variables.
Dimension Domain Variables Weightage
Sociocultural
(SC)
Social and
community
participation
Not a member of any association or
club; not participating in religious
events
25%
Social isolation Single or living alone 25%
Social
discrimination
Feel that elders are not respected in
society
25%
Social
dependence
Require help in undertaking
activities of daily living
25%
Political (P) Access to
healthcare
Do not have a health facility close by 20%
Access to
information
Do not have access to a television or
a radio
20%
Political and
civil
participation
Do not exercise their voting rights;
not a member of any political party
20%
Decision
making role
Do not have a decision making role
in an association
20%
Access to safe
housing
Live in an unsafe neighbourhood
(Feel that it is not safe to walk alone
in the night in the neighbourhood)
20%
Economic (E) Relative
poverty
Belong to the poorest quartiles (Q1
& Q2)*
33%
Economic
dependence
Do not have an independent source
of income
33%
Precariousness
of shelter
Not living in a family-owned
household*
33%2.2. Household surveys
We used data from cross-sectional household surveys con-
ducted in 2012 in Senegal and Ghana. Since the target population
for NHIS and Plan Sesame differs, we applied different sampling
strategies in the two countries. In Ghana, the surveywas carried out
in five relatively underdeveloped and predominantly rural regionsNotes: Variables denoted by (*) are estimated at the household level; all other are
individual level variables.region, one district was selected for sampling. The districts were
selected after consulting the Ghana Statistical Service to ensure a
mixof urban and rural areas. In Ghana 52% of the population lives in
urban areas and the remaining 48% in rural areas (CIA, 2014). In
each selected district, 27 nationally representative Enumeration
Areas (EAs) were selected. We visited each EA and created a list of
all households present there. From each EA, 30 households were
randomly sampled for interviews, i.e. a total of 4050 households
overall. We selected 30 households per EA as advised by the
MEASURE Demographic Health Surveys Program (ICF International,
2012).
In Senegal the survey was conducted in Dakar, Diourbel, Matam
and Tambacounda regions. These regions were selected to ensure
that the sample selected would be heterogeneous in terms of
ruraleurban population, poverty and access to healthcare. In
Senegal 42.5% of the population lives in urban areas and 57.5% in
rural areas (CIA, 2014). Except Dakar, the other three regions are
predominantly rural, and all of them have a regional hospital. For
sampling, the list of EAs was obtained from the National Agency of
Statistics and Demography. From this list, EAs were randomly
selected proportional to the number of households, and then
household compounds were randomly selected. If more than one
household was found in the compound, the interviewer randomly
selected the household to be interviewed. If the household did not
have a person over 60, the household was replaced. A total of 2933
households were selected.
The household survey consisted of two parts. Part I was
administered to the household head and collected data on basic
demographics and socio-economic situation of the household and
its members. For Ghana, this part also gathered information on
health status, healthcare utilisation, and awareness and utilisation
of NHIS. Part II included questions on social exclusion and was
administered to those members that were eligible for NHIS and
Plan Sesame in Ghana and Senegal respectively. In Ghana the
household head and his/her spouse were interviewed, while in
Senegal a person over 60, preferably the household head was
interviewed. In case the household head was under 60 or unavai-
lable, another person over 60 from the household was interviewed.
For Senegal, besides questions on social exclusion, Part II also
covered questions on health status, healthcare utilisation and,
awareness and utilisation of Plan Sesame.
Although efforts were made to include similar worded ques-
tions in both surveys, owing to the differences in contexts and
scheme features, many questions were adapted to local settings.
The survey was drafted in English in Ghana and French in Senegal
but interviews were conducted in local languages if the respondent
was not fluent in the official language.
In Ghana, as the survey was administered to study social
exclusion in NHIS and since NHIS is offered to all Ghanaians and not
only to people over 70, the respondents were not necessarily over
70. For Ghana we therefore used data on persons over 70 who had
answered Part II of the questionnaire i.e. 435 older people. In
Senegal all persons over 60 had answered Part II, so we used the
entire sample i.e. 2933. In both countries, the household head was
interviewed if possible. Therefore, older people in our sample are
more likely to be household heads.
For NHIS, although registration needs to be done once, mem-
bership needs to be renewed each year. People over 70 were
considered enrolled if they had a valid card e they were registered
and had renewed their NHIS membership that year. Those that had
renewed their membership but had not received their cards were
considered as not enrolled as they did not possess a valid card to
use NHIS.
In Senegal people over 60 are not required to register for Plan
Medicine 119 (2014) 36e44Sesame. However, they are required to present their national
by a healthcare provider (e.g. doctor) or if they were hospitalised in
the past 12 months. It is expected that older people suffering from
chronic illnesses that require regular healthcare and those having
hospitalisation needs will tend to enrol more.
2.4.1. Sociocultural variables
We created variable, No_association, to reflect whether a person
was a member of any association (e.g. social clubs, sports clubs,
religious associations, women's groups). Variable, Religious_parti-
cipation, was included to reflect the intensity and regularity of
participating in religious events. Variable No_respect indicates
whether older people feel respected. The questionnaire included a
Male 1 if male; 0 otherwise 48.1% 53.4%
Educated 1 if education; 0 otherwise 32.6% 26.8%
0 otherwise
Urban 1 if living in an urban area; 0 otherwise 50.6% 48.8%
HHHmale 1 if household head is male;
0 otherwise
49.9% 73.9%
Chronic 1 if has any chronic illness; 0 otherwise 27.8% 25.1%
Hospitalised 1 if hospitalised in the past 12 months;
0 otherwise
11.5% 7.7%
Sociocultural (SC)
Single 1 if single or living alone; 0 otherwise 56.1% 40.5%
No_respect 1 if feel elders are not respected these
days; 0 otherwise
4.8% 30.7%
No_association 1 if not member of any association or
club; 0 otherwise
64.4% 74.8%
Religious_participation 1 if does not regularly attend religious
events; 0 otherwise
24.8% 46.4%
Needhelp_adl 1 if needs help in activities of daily
living; 0 otherwise
7.6% 25.4%
Political (P)
Lessaccess_info 1 if living in a household without a
radio or TV; 0 otherwise
36.6% 8.9%
Decision_making 1 if not in any formal decision making
position; 0 otherwise
81.6% 85.8%
Political_participation 1 if not part of any political party/group
or did not vote in the previous general
elections; 0 otherwise
4.1% 29.4%
Far 1 if PHC or health post is within 15 min'
walk; 0 otherwise
48.5% 44.3%
NotSafe 1 if feel it is not safe to walk alone in the
night in the neighbourhood;
0 otherwise
8.1% 30.6%
Economic (E)
No_income 1 if does not have any independent
source of income; 0 otherwise
29.0% 41.0%
Rented_house 1 if living in a rented house;
0 otherwise
24.8% 13.8%
Wealth quartiles Q1eQ4, with Q1 referring to the
poorest 25% households and Q4 to the
richest 25% households.
Q1 25.3% 25.1%
Q2 25.7% 24.6%
Q3 23.9% 25.1%
Q4 25.1% 25.1%
a Majority religion is Christianity in Ghana and Islam in Senegal.
b Majority ethnicity is Akan, Ga and Ewe in Ghana and Wolof in Senegal.
ce &biometric identity card (or IPRES or FNR cards) at health facilities to
get exemptions. Hence, in Senegal people over 60 require a valid
card and they should be aware of Plan Sesame to avail exemptions.
We therefore considered persons as enrolled, if they had a valid
card and also knew about Plan Sesame.
In this study we wanted to compare older people who are
eligible for free healthcare, however they were defined differently
as per the scheme eligibility. For NHIS, people over 70 are eligible
for exemptions. While Plan Sesame exemptions are offered to
people over 60. We therefore included persons who are benefitting
from SHP as over 70 years in Ghana and over 60 years in Senegal.
2.3. Regression models
Our basic regression model for determinants of enrolment can
be defined as:
logitðpÞ ¼ logðp=1 pÞ ¼ b0 þ Xi$bi1 þ SEi$bi2 (1)
Enrolled is a binary outcome variable indicating enrolment sta-
tus as 0/1, and p is the probability that the individual has enrolled in
the scheme i.e. p ¼ prob(Enrolled ¼ 1). SEi is a set of SPEC variables
(as described in Table 1), Xi is a set of remaining variables that may
influence enrolment, and bs are the model parameters.
We estimated three logistic regression models for each country.
First, we estimated a simple regression model (Model A) with only
Xi variables, we then ran the regression with all variables e Xi and
SEi variables in Model B, and in the third model (Model C) we
included Xi variables and the SPEC indices. Bivariate analyses were
also conducted as a preliminary step to the multivariate models.
These analyses are presented in Supplementary file 2.
Since more than one older person could have answered Part II of
the questionnaire in Ghana (for example if both the household head
and his/her spouse were over 70), the standard errors for Ghana
were adjusted for clustering at the household level. All models
were estimated using Stata 12.
2.4. Variables
The variables used in this study are described in Table 2.
The dependent variable, Enrolled, is a binary variable indicating
the enrolment status i.e. having a valid NHIS card in Ghana. In
Senegal there is no Plan Sesame specific card. Older people are
required to show their national biometric ID, IPRES or FNR cards at
health facilities to access free care. Hence, a person was considered
enrolled if he/she had a valid card and had heard of Plan Sesame.
Core variables identified in Table 2 were included in all re-
gressions. Since majority of persons were illiterate, Education was
constructed as a binary variable and persons who had any formal
education were coded as 1. Whether persons belong to majority or
minority religious and ethnic groups could determine if they
experience exclusion and discrimination. To capture this, two var-
iables, Majority_religion and Majority_ethnicity, were included. In
Ghana majority religion was Christianity while in Senegal Islam.
Akan, Ga and Ewe were regarded as majority ethnicities in Ghana
and Wolof in Senegal. Previous studies (Jehu-Appiah et al., 2011;
Chankova et al., 2008) have found that households in urban areas
and female-headed households tend to have higher levels of
enrolment and healthcare utilisation; we therefore included vari-
ables, Urban and HHHmale. Presence of adverse selection (when
high-risk individuals who have a higher probability of using
healthcare services enrol more than lower-risk or healthier in-
dividuals) in voluntary schemes has been noted before (Parmar
et al., 2012). This is captured by Chronic and Hospitalised. Re-
D. Parmar et al. / Social Scienspondents were asked if they have any chronic illness as diagnosedHead 1 if elder is the household head;
0 otherwise
89.9% 68.1%
Majority_religiona 1 if belongs to the majority religion;
0 otherwise
66.7% 96.5%
Majority_ethnicityb 1 if belongs to the majority ethnicity; 72.6% 47.6%Table 2
Definition and descriptive statistics of variables, by country.
Variables Definition Percentage
Ghana Senegal
Enrolled 1 if enrolled; 0 otherwise 71.7% 47.7%
Core variables
Medicine 119 (2014) 36e44 39statement (‘Older people are respected these days’) and responses
Committee for Research in Health, Senegal [674/MSAS/DS/DER].
illness (28% in Ghana and 25% in Senegal) and 11% in Ghana vs. 8% in
Older people who had a chronic illness were significantly more
likely to enrol in NHIS and Plan Sesame. Being hospitalised in the
ce &Senegal reported being hospitalised in the past 12 months. Access
to health facilities was poor as majority of them lived more than
15 min' walk from any PHC or health post. For those that do not
possess a car (90% of households in Senegal and 99% in Ghana as per
our sample), reaching health facilities could be a challenge. A large
proportion of older people are either single or live alone, most are
not members of any association, but many regularly attend reli-3. Results
3.1. Descriptive statistics
Table 2 shows the percentage or means of variables that were
included in the regressions in the two countries. More older people
were enrolled in NHIS in Ghana than in Plan Sesame in Senegal
(72% vs. 48%). Majority of older people were uneducated in both
countries, although more were educated in Ghana than in Senegal,
despite the fact that persons, as included in this study, were rela-
tively older in Ghana than in Senegal. We have a higher proportion
of older people who were household heads in both settings.
Around 1/4th of the older people reported having a chronic2.4.3. Economic variables
Relative wealth was estimated at the household level by using
principle components analysis (PCA). Household ownership of
durable goods (e.g. TV, radio, car), housing conditions (material of
roof, number of rooms) and livestockwere used. Based on their PCA
scores, household were divided into quartiles with Q1 representing
the poorest 25% and Q4 the richest 25% households. Since
composition of household assets differ in rural and urban areas,
quartiles were constructed separately for rural and urban house-
holds. As a result, quartiles for the whole sample (rural and urban
combined) may not exactly consist of 25% of the persons. Economic
dependence is reflected by variable No_income, which captures
whether a person has any source of independent income e.g. from
pensions, salary or rent.
Ethical approval for this research was obtained from the
Noguchi Memorial Institute for Medical Research Institutional Re-
view Board, Ghana [069/11-12] and from the National Ethicswere captured by a Likert scale (strongly agree to strongly
disagree). Persons who require help with activities of daily living
may have difficulties in participating in social and cultural events
and Needhelp_adl captured this. Persons living alone or who are
single may feel isolated, and may be vulnerable to socially exclu-
sion. We included variable, Single, to investigate this.
2.4.2. Political variables
Some groups may not have access to necessary resources that
are needed to participate in society e either due to the unavail-
ability of resources or due to barriers that prevent these groups
from accessing these resources. Variables that capture access to
health facilities and safe housing and the ease of accessing infor-
mationwere included. For Ghana, clinics run bymissions and NGOs
could have NHIS accreditation and be entitled to provide free care
to enrolled persons. For this reason these two types of clinics were
also considered while defining access to health facilities by vari-
able, Far. Variable, Political_participation, reflects participation in
democratic processes. It includes two activities e participated in
recent elections and being member of a political party.
D. Parmar et al. / Social Scien40gious events, with attendance being higher in Ghana.last 12 months increased the odds of enrolling by four folds in NHIS
and almost two folds in Plan Sesame.
With regard to sociocultural variables, those who were not
members of sociocultural associations were less likely to enrol in
Plan Sesame. Surprisingly, those who felt that older people are not
generally respected in society were more likely to enrol in NHIS.
This could be because a very small sample of them (4.8% as
mentioned in Table 2) actually felt that older people are not
respected.
Older people vulnerable to political exclusion as reflected from
lower political and civic participation, living in relatively unsafe
neighbours and having limited access to information channels, had
lower odds of enrolling in Plan Sesame. For NHIS, lower political
and civic participation (significant only at 10%) and living far from
health facilities were found to lower the odds of enrolling.
The results show that older people belonging to richer house-
holds (Q3 and Q4) are 1.6 and 2.3 times more likely to enrol in Plan
Sesame and 2.6 and 4 times in NHIS as compared to the poorest 25%
households. This shows that both schemes have a pro-rich bias,
with NHIS being more inequitable.
As mirrored in the odds ratio in Model B for both countries,
thosewhowere vulnerable in all dimensions of social exclusion had
lower odds of enrolling in Plan Sesame, as seen from the SPEC
indices in Model C. For NHIS, only the political dimension was
found to significantly lower the odds of enrolling. However, it
should be noted that the indices, although found to be highly sig-
nificant, are very close to 1.
4. Disscussion
The implementation of Plan Sesame and NHIS exemptions for
people over 70 represent significant efforts to remove financial
barriers to healthcare access for older people. There is currently
limited evidence to show how successful these schemes have been
at enrolling the ageing population and whether inequities in
enrolment exist. Our study has addressed these issues by analysing
data from household surveys in Ghana and Senegal to ascertain
enrolment rates for targeted persons, to analyse determinants of
enrolment, and to explore whether social exclusion is restricting
access to these SHP schemes.
Our results show evidence of persisting inequities in enrolment
for older people in NHIS and Plan Sesame caused by a combination
of economic, political and sociocultural dimensions. The impact of
each of these dimensions is discussed in turn in the remainder of3.2. Determinants of enrolment
Results from the logistic regressions models, with enrolment
status as the dependent variable, are presented in Table 3.
With regard to the core variables, men, household heads, having
some formal education and living in urban areas increased the odds
of enrolling in Plan Sesame by almost two folds. This remained
relatively consistent for all models, even after other covariates and
SPEC indices were introduced. For NHIS these variables were not
found to be significant. Belonging to majority ethnicity (i.e. Wolof)
increased the odds of enrolling in Plan Sesame while belonging to
majority religious group (i.e. Christianity) increased the odds for
enrolling in NHIS, although found to be significant only in Model A
for NHIS.
Strong evidence of adverse selection was found for both
schemes, with NHIS being associatedwith higher adverse selection.
Medicine 119 (2014) 36e44this section.
Table 3
Determinants of enrolment, 2012.
Variables Ghana e National Health Insurance Scheme Senegal e Plan Sesame
Model A Model B Model C Model A Model B Model C
OR SE CI OR SE CI OR SE CI OR SE CI OR SE CI OR SE CI
Male 0.696 (0.511) 0.165e2.938 0.617 (0.405) 0.170e2.235 0.698 (0.573) 0.139e3.492 2.207 (0.307)*** 1.680e2.900 2.141 (0.345)*** 1.561e2.937 2.114 (0.306)*** 1.592e2.808
Educated 1.048 (0.301) 0.597e1.839 0.708 (0.222) 0.384e1.308 0.762 (0.223) 0.429e1.353 2.038 (0.202)*** 1.679e2.474 1.596 (0.169)*** 1.297e1.963 1.766 (0.181)*** 1.445e2.157
Head 1.250 (0.544) 0.533e2.936 1.636 (0.696) 0.710e3.767 1.309 (0.609) 0.526e3.259 1.735 (0.207)*** 1.373e2.191 1.482 (0.197)*** 1.142e1.923 1.378 (0.172)** 1.079e1.758
Majority_religion 2.478 (0.913)** 1.204e5.102 1.439 (0.582) 0.651e3.180 1.852 (0.702) 0.882e3.892 0.997 (0.225) 0.640e1.551 1.086 (0.257) 0.683e1.727 1.003 (0.231) 0.638e1.576
Majority_ethnicity 0.563 (0.228) 0.254e1.246 0.862 (0.412) 0.338e2.198 0.833 (0.355) 0.362e1.919 1.408 (0.115)*** 1.198e1.653 1.212 (0.109)** 1.017e1.445 1.394 (0.119)*** 1.180e1.647
Urban 1.420 (0.329) 0.901e2.236 1.414 (0.402) 0.810e2.469 1.392 (0.355) 0.844e2.296 1.810 (0.169)*** 1.507e2.173 1.953 (0.217)*** 1.571e2.429 1.963 (0.191)*** 1.621e2.376
HHHmale 0.911 (0.645) 0.228e3.646 0.873 (0.597) 0.228e3.336 0.614 (0.488) 0.129e2.917 1.011 (0.146) 0.761e1.341 0.909 (0.140) 0.672e1.230 0.890 (0.133) 0.664e1.193
Chronic 2.141 (0.650)** 1.181e3.882 2.044 (0.686)** 1.058e3.946 2.276 (0.725)*** 1.219e4.249 1.686 (0.164)*** 1.393e2.040 1.689 (0.175)*** 1.378e2.069 1.731 (0.173)*** 1.424e2.105
Hospitalised 4.161 (2.262)*** 1.434e12.077 3.841 (2.214)** 1.241e11.888 4.430 (2.482)*** 1.477e13.284 1.596 (0.243)*** 1.184e2.151 1.856 (0.298)*** 1.354e2.544 1.788 (0.281)*** 1.314e2.432
Sociocultural Single 1.118 (0.393) 0.561e2.226 0.753 (0.089)** 0.598e0.949
No_respect 0.147 (0.076)*** 0.053e0.403 0.963 (0.089) 0.804e1.153
No_association 0.785 (0.293) 0.378e1.633 0.494 (0.068)*** 0.377e0.647
Religious_participation 0.744 (0.224) 0.412e1.344 1.027 (0.104) 0.842e1.252
Needhelp_adl 0.971 (0.478) 0.370e2.549 0.885 (0.090) 0.725e1.079
Political Lessaccess_info 0.956 (0.349) 0.467e1.956 0.671 (0.110)** 0.486e0.926
Decision_making 0.701 (0.347) 0.265e1.851 1.003 (0.173) 0.716e1.406
Political_participation 0.414 (0.215)* 0.149e1.148 0.522 (0.053)*** 0.428e0.638
Far 0.559 (0.145)** 0.336e0.930 0.916 (0.081) 0.771e1.089
NotSafe 1.681 (0.886) 0.598e4.725 0.797 (0.079)** 0.657e0.967
Economic No_income 1.257 (0.366) 0.710e2.225 0.846 (0.079)* 0.705e1.016
Rented_house 0.553 (0.170)* 0.302e1.009 0.950 (0.126) 0.733e1.233
Wealth: Q2 1.287 (0.440) 0.659e2.514 1.237 (0.154)* 0.970e1.578
Wealth: Q3 2.575 (1.120)** 1.098e6.041 1.557 (0.200)*** 1.211e2.003
Wealth: Q4 3.861 (1.771)*** 1.572e9.486 2.286 (0.310)*** 1.753e2.982
SC Index 0.993 (0.008) 0.978e1.009 0.991 (0.002)*** 0.987e0.995
P Index 0.978 (0.008)*** 0.963e0.993 0.983 (0.002)*** 0.978e0.987
E Index 0.992 (0.005) 0.982e1.002 0.990 (0.002)*** 0.987e0.994
Observations 435 434 435 2917 2902 2917
Dependent variable: Binary choice variable for enrolment.
Acronyms: Odds Ratio (OR); Standard Errors (SE); Confidence Interval (CI); Sociocultural (SC); Political (P); Economic (E).
Robust SE in parenthesis.
***p < 0.01, **p < 0.05, *p < 0.1.
ce &4.1. Economic exclusion and enrolment
Pronounced economic inequalities are evident in both schemes,
with older people in the richest quartiles being more likely to enrol
than those in the poorest quartile. In Ghana, older people who live
in family-owned houses are more likely to enrol in NHIS than those
living in rented houses (nearly 25%), while older people with their
own source of income in Senegal (nearly 59%) were more likely to
enrol in Plan Sesame than individuals with no income. These
findings are consistent with existing literature that finds substan-
tial evidence of low enrolment for the most economically vulner-
able individuals in SHP schemes that are specifically targeted
towards the poor in LMIC (Jehu-Appiah et al., 2011; Sarpong et al.,
2010). Although both schemes have taken steps to reduce financial
barriers to access, it is clear that richer persons are more likely to
activate their entitlements to access Plan Sesame or to receive ex-
emptions from NHIS. In Ghana, potential financial barriers to
enrolment for the poor still exist. Despite being exempt from
paying the full premium, older people are required to pay a small
registration to join. Removal of this registration fee may increase
enrolment among the poor as the existence of even a small fee has
previously been shown to prevent the poor from enrolling (Wang
et al., 2005).
4.2. Political exclusion and enrolment
Our results show that the political dimension is an important
determinant of enrolment in NHIS and Plan Sesame. Older people
that take an active role in civic society by voting were shown to
have higher odds of enrolment. For Plan Sesame this is partly
explained by individuals who intend to vote being required to have
the same ID card to obtain a voter registration ID, as they need to
access Plan Sesame; an individual that intends to vote therefore has
an additional incentive to acquire an ID card. Voting may also be
significant as individuals who take an interest in politics are more
likely to have heard of these financing schemes and to be influ-
enced by the opinions of political and local leaders. Additionally,
peoplewho participate in the democratic systemmay have a higher
level of trust in the democratic system and political structures and
may be more likely to participate in government initiatives such as
Plan Sesame and NHIS (Ottone, 2007). Geographical inequities are
shown to exist for individuals living in remote areas, with in-
dividuals living close to a health facility in Ghana having signifi-
cantly increased odds of enrolling in NHIS and urban individuals
more likely to enrol in Plan Sesame. Links between lack of physical
access to healthcare and lower enrolment has beenwell established
(Schneider and Diop, 2004). This apparent geographical segrega-
tion and exclusion suggests that further investment in health
centres is needed in Ghana. In Senegal, distance to a health facility
is not significantly related to enrolment, suggesting that low
enrolment among rural individuals may instead be due to barriers
to accessing administrative offices to apply for an ID card. Efforts
should therefore be taken to create ID card centres in remote areas
of Senegal.
4.3. Sociocultural exclusion and enrolment
Our results also show the importance of social support networks
and solidarity in influencing enrolment patterns in Plan Sesame.
SHP schemes are founded on the principle of social solidarity and
we initially hypothesized that greater feelings of solidarity would
increase odds of enrolment (Barrientos and Lloyd-Sherlock, 2000).
In our regression models respect afforded to older people were
included as proxy measures for solidarity felt with fellow citizens.
D. Parmar et al. / Social Scien42Having a positive perception on the respect afforded to olderpeople in Ghana increased the odds of being enrolled in NHIS,
confirming our initial hypothesis. Being a member of an association
significantly increased the odds of being covered by Plan Sesame,
while being single, decreased odds of enrolment. These variables
signify the existence of a social support network and indicate that
an individual is not isolated from the community. Social support
networks are important in ensuring that individuals are able to
actively participate in society and benefit from opportunities that
are afforded to them as networks with many social connections are
more likely to introduce new ideas and opportunities to their
members (Putman, 1993). Being a member of an association is
therefore likely to enhance the probability of having heard of Plan
Sesame and an individual's ability to register for an ID card.
The SEKN and Commission on Social Determinants of Health
have both affirmed that health inequalities and access to SHP are
affected by hierarchical systems of social stratification based on
gender, class, education, age, ethnicity, and religion (Marmot et al.,
2008; Popay et al., 2008). Unsurprisingly, a number of these so-
ciocultural variables are shown to be significant determinants of
enrolment in Plan Sesame and NHIS. In Senegal, being male,
educated, household head or belonging to the majority ethnicity
increases the odds of being covered. In Ghana, although these
variables are not significant, belonging to the majority religion
significantly increases the odds of enrolment. The significance of
these variables may indicate that discriminatory processes are
present and religious minorities in Ghana and ethnic minorities
and women in Senegal are being excluded from SHP. Efforts should
be made to target older women and minorities to ensure they are
fully aware of their entitlements. As highlighted in previous liter-
ature, education is unsurprisingly significant as higher education
leads to a greater understanding of the scheme and therefore a
higher tendency to enrol (Chankova et al., 2008; Jehu-Appiah et al.,
2011).
4.4. SPEC indices of social exclusion
A similar pattern to the above results is seenwhen lookingmore
closely at the SPEC indices of social exclusion. In our third regres-
sion model (Model C), sociocultural, political and economic indices
are all significantly related to enrolment in Plan Sesame, indicating
that individuals at risk of social exclusion are less likely to enrol.
Results from Ghana suggest that individuals at risk of social
exclusion in the political domain are less likely to enrol in NHIS,
although the economic and sociocultural indices are not significant.
These results confirm our initial hypothesis and indicate that access
to SHP schemes can be restricted by social exclusion. Efforts to
reform both schemes should be undertaken to ensure that socially
excluded individuals are aware of their entitlements and to ensure
that all older people enjoy the same rights and ability to access
these schemes.
More research on NHIS and Plan Sesame is needed to explore
the underlying causes or pathways that explain the results pre-
sented in this study. A number of non-financial indicators, such as
access to information and administrative processes, may be pre-
venting older people from enrolling. Richer and educated in-
dividuals are likely to have better access to media, giving them
greater exposure to scheme-related information and education
campaigns which enhance their awareness and understanding of
SHP (Schneider and Diop, 2004). Indeed, results from our regres-
sion analysis indicate that older people with access to a TV or radio
were more likely to enrol. Although general awareness of NHIS is
high in Ghana, we are not aware of any study that has looked at the
awareness of exemptions among older people.
In addition to lack of awareness, complex administrative pro-
Medicine 119 (2014) 36e44cedures may be discouraging vulnerable older people from
ce &enrolling. In Senegal older people are required to obtain an ID card
from officials by either presenting a birth certificate or taking three
Senegalesewitnesses to confirm their Senegalese nationality. These
requirementsmay be difficult for some persons to fulfil, particularly
those that are socially isolated, in poor health or living in remote
communities.
4.5. Limitations
Our results are based on an analysis of quantitative data on
social exclusion and thus capture social exclusion as a set of binary/
dichotomous variables. Although a quantitative assessment of the
multiple indicators of disadvantage presents important insights
into the analysis of social exclusion, further research is needed to
fully capture the complexity and dynamics of social exclusion
processes. In addition, the survey in Ghana was part of a larger
study exploring enrolment of all individuals in NHIS and was not
specifically targeted towards the ageing population. In comparison,
our Senegalese survey was targeted towards older people, ensuring
the sample size in Ghana was far smaller (435) than in Senegal
(2933). This may have been responsible for some variables to be
insignificant in the regression models for Ghana. Practical consid-
erations ensured that interviews were primarily conducted with
older persons who were household heads. We therefore miss the
experience of those who are not household heads, who may be
more likely to experience social exclusion and barriers to accessing
SHP and healthcare. Furthermore, due to the differing design of the
schemes, our analysis of NHIS was conducted on persons over 70
while our analysis of Plan Sesame was for over 60s. This should be
taken into consideration while comparing the results between
these two schemes. It was also not possible to explore all possible
variables influencing enrolment. More research is needed to
explore whether prior experience of enrolling in SHP schemes and
utilizing healthcare services, and perceptions on quality of
healthcare can further explain the differences in enrolment rates
among the socially excluded groups in these countries. Since NHIS
has been operational for nine years and Plan Sesame for almost six
years, there is a need to understand program-level barriers faced by
older people and how these barriers can be removed so that the
most vulnerable are not left out. Finally, we did not study the
impact of either scheme on utilization of healthcare or health
outcomes as this was beyond the scope of the study. Previous
studies have shown that although SHP schemes often increase
utilization and reduce the risk of catastrophic health payments;
these benefits are often unequally distributed amongst vulnerable
groups, particularly those in the lowest quintile (Lu et al., 2012;
Lloyd-Sherlock et al., 2012). It is therefore important for future
research to determine if these SHP schemes benefit the most
vulnerable older people.
5. Conclusion
Our study makes a valuable contribution to the evaluation of
SHP for older people in LMIC. Plan Sesame and exemptions for older
people in NHIS are both significant policies that have taken steps to
address the inequities experienced by older people in relation to
healthcare access. However, results from our study indicate that
older people at risk of social exclusion are currently disadvantaged
in enrolment and neither scheme has yet achieved the goal of eq-
uity in access for older people. Although these schemes aim to
reduce financial barriers to enrolment, economically vulnerable
persons still suffer from inequities in enrolment and efforts should
be taken to identify the very poorest to ensure they are aware of
and enrol in SHP schemes. Simply targeting the removal or
D. Parmar et al. / Social Scienreduction of financial barriers may not be enough. Enhanced effortsshould also be made to reach older populations in remote areas,
those who belong to ethnic minorities, women, and those isolated
due to a lack of social support. Consideration should also be paid to
modifying scheme features such as eliminating the registration fee
for older people in NHIS and creating administration offices for ID
cards in remote communities in Senegal. Recognising and taking
steps to address factors hindering enrolment of older people at risk
of social exclusion will ultimately improve the prospect of
achieving equity and universal coverage in older populations.
Acknowledgements
The research leading to these results has received funding from
the European Commission's Seventh Framework Programme FP7/
2007 under grant agreement no. 261440. The funder was not
involved in the research and preparation of the article, including
study design; collection, analysis and interpretation of data;
writing of the article; nor in the decision to submit it for
publication.
This study is part of a larger project, Health Inc. The authors
would like to thank members of the Health Inc. consortium who
have contributed to the development of the social exclusion
framework and research methodology. We would also like to thank
our respondents who gave their valuable time and the anonymous
reviewers for their feedback.
Appendix A. Supplementary data
Supplementary data related to this article can be found at http://
dx.doi.org/10.1016/j.socscimed.2014.08.011.
References
Aboderin, I., 2012. Global poverty, inequalities and ageing in Sub-Saharan Africa: a
focus for policy and scholarship. J. Popul. Ageing 5 (2), 87e90.
Alam, N., Chowdhury, H.R., Bhuiyan, M.A., Streatfield, P.K., 2010. Causes of death of
adults and elderly and healthcare-seeking before death in rural Bangladesh.
J. Health Popul. Nutr. 28 (5), 520e528.
ANSD & ICF International, 2012. Enqu^ete D
emographie et de Sant
e a Indicateurs
Multiples (EDS-MICS). ANSD et ICF International, Calverton, Maryland, USA.
Apt, N.A., 2009. The extra burden of ageing women in a poor economic environ-
ment. In: Population Ageing: Towards Improvement of the Quality of Life,
pp. 1e12. Brussels.
Barrientos, A., Lloyd-Sherlock, P., 2000. Reforming health insurance in Argentina
and Chile. Health Policy Plan. 15, 417e423.
Beard, J., Biggs, S., Bloom, D., Fried, L., Hogan, P., Kalache, A., et al., 2011. Global
Population Ageing: Peril or Promise. World Economic Forum, Geneva.
Biritwum, R., Mensah, G., Yawson, A., Minicuci, N., 2013. Study on Global AGEing
and Adult Health (SAGE) Wave 1: The Ghana National Report. World Health
Organization, Geneva.
Chankova, S., Sulzbach, S., Diop, F., 2008. Impact of mutual health organizations:
evidence from West Africa. Health Policy Plan. 23, 264e276.
CIA. The World Factbook. https://www.cia.gov/library/publications/the-world-
factbook/index.html (accessed 02.06.14).
Crooks, D., 2009. Development and Testing of the Elderly Social Vulnerability Index
(ESVI): A Composite Indicator to Measure Social Vulnerability in the Jamaican
Elderly Population. FIU Electronic Theses and Dissertations Paper 186.
George-Carey, R., Adeloye, D., Chan, K.Y., Paul, A., Kolci
c, I., Campbell, H., Rudan, I.,
2012. An estimate of the prevalence of dementia in Africa: a systematic analysis.
J. Global Health 2 (2), 20401.
Gysels, M., Pell, C., Straus, L., Pool, R., 2011. End of life care in Sub-Saharan Africa: a
systematic review of the qualitative literature. BMC Palliat. Care 10, 6.
Holmes, W.R., Joseph, J., 2011. Social participation and healthy ageing: a neglected,
significant protective factor for chronic non communicable conditions. Global.
Health 7 (1), 43.
ICF International, 2012. Demographic and Health Survey Sampling and Household
Listing Manual. MEASURE DHS. ICF International, Calverton, Maryland, USA.
International Labour Organization, 2014. World Social Protection Report. ILO,
Geneva, p. 4.
Issahaku, P.A., Neysmith, S., 2013. Policy implications of population ageing in West
Africa. Int. J. Sociol. Soc. Policy 33 (3/4), 17.
Jehu-Appiah, C., Aryeetey, G., Spaan, E., De Hoop, T., Agyepong, I., Baltussen, R., 2011.
Equity aspects of the National Health Insurance Scheme in Ghana: who is
Medicine 119 (2014) 36e44 43enrolling, who is not and why? Soc. Sci. Med. 72, 157e165.
Lancet, 2012. A manifesto for the world we want. Lancet 380 (9857), 1881.
Leye, M.M.M., Diongue, M., Faye, A., Coum
e, M., Faye, A., Tall, A.B., Tal-Dia, A., 2013.
Analysis of free health care for the elderly in the context of the “Plan S
esame” in
Senegal. Sant
e Publique 25.
Lloyd-Sherlock, P., 2000. Population ageing in developed and developing regions:
implications for health policy. Soc. Sci. Med. 51 (6), 887e895.
Lloyd-Sherlock, P., Minicuci, N., Beard, J., Chatterji, S., 2012. Social protection and
preventing illness in developing countries: establishing the health effects of
pensions and health insurance. Int. Soc. Secur. Rev. 65 (4), 51e68.
Lu, C., Chin, B., Lewandowski, J.L., Basinga, P., Hirschhorn, L.R., Hill, K., et al., 2012.
Towards universal health coverage: an evaluation of Rwanda Mutuelles in its
first eight years. PLoS One 7, e39282.
Marmot, M., Friel, S., Bell, R., Houweling, T.A., Taylor, S., 2008. Closing the gap in a
generation: health equity through action on the social determinants of health.
Lancet 372, 1661e1669.
Mbaye, E.M., Ridde, V., K^a, O., 2013. “Good intentions are not enough”: analysis of a
health policy for the elderly in Senegal. Sant
e Publique 25, 107e112.
McIntyre, D., 2004. Health policy and older people in Africa. In: Lloyd-Sherlock, P.
(Ed.), Living Longer: Ageing, Development and Social Protection. Zed Books,
London.
National Health Insurance Authority, 2011. National Health Insurance Scheme:
Annual Report 2011. National Health Insurance Authority, Accra.
Ottone, E., 2007. Social Cohesion: Inclusion and a Sense of Belonging in Latin
America and the Caribbean. Santiago de Chile.
Parmar, D., Souares, A., De Allegri, M., Savadogo, G., Sauerborn, R., 2012. Adverse
selection in a community-based health insurance scheme in rural Africa: im-
plications for introducing targeted subsidies. BMC Health Serv. Res. 12, 181.
Popay, J.E.S., Hern
andez, M., Johnston, H., Mathieson, J., Rispel, L., 2008. Under-
standing and Tackling Social Exclusion. Final Report to the WHO Commission
on Social Determinants of Health. Social Exclusion Knowledge Network,
Lancaster.
Putman, R., 1993. The prosperous community: social capital and public life. Am.
Prospect 13, 35e42.
Saeed, B.I.I., Oduro, S.D., Ebenezer, A.M.F.E., Zhao, X., 2012. Determinants of
healthcare utilization among the ageing population in Ghana. Int. J. Bus. Soc. Sci.
3 (24), 66e77.
Sarpong, N., Loag, W., Fobil, J., Meyer, C.G., Adu-Sarkodie, Y., May, J., et al., 2010.
National health insurance coverage and socio-economic status in a rural district
of Ghana. Trop. Med. Int. Health 15, 191e197.
Schneider, P., Diop, F., 2004. Community-based Health Insurance in Rwanda. Health
Financing for Poor PeopledResource Mobilization and Risk Sharing. World
Bank, Washington DC, pp. 251e274.
United Nations, 2009. World Population Ageing. Working Paper ESA/P/WP/212.
Department of Economic and Social Affairs, Population Division, New York.
Wang, H., Yip, W., Zhang, L., Wang, L., Hsiao, W., 2005. Community-based health
insurance in poor rural China: the distribution of net benefits. Health Policy
Plan. 20, 366e374.
D. Parmar et al. / Social Science & Medicine 119 (2014) 36e4444